i=&or::o'"ie!-'. - Physical Review Link Manager

i=&or::o '"ie!-'. " cai "i —"hays cai Soc "'yÃf ': "le S':I.C tf C 'ol. 3 0 "I 'CS~Si'c 'I 0 a "I;I "IC

'or coa u'i iza:ion anc ayre' "i":ic '9" proc uc:ionAPS Study Group Participants

Bernard R. COOPer, Chairman

West Virginia University, Morgantown, West Virginia 26506

Wayne R. Gruner, ExecutiveDirector

6305 Tulsa Lane, Bethesda, Maryland 20034

Larry Anderson,

Un&'versity of Utah, Salt Lake City, Utah 84112

Robert H. Davis

Florida State University, Tallahassee, Florida 32306

Paul Engelking

University of Oregon, Eugene, Oregon 97403

Jose D. Garcia

University ofArizona, Tucson, Arizona 85721

Edward Gerjuoy

University ofPittsburgh, Pittsburgh, Pennsylvania 15260

Robert I. JaffeeElectric Power Research Institute, Palo Alto, California 94304

Philip G. Kosky

General Electric Research and Development Center, Schenectady, New York 12301

Leonidas Petrakis

Gulf Research and Development Company, Pittsburgh, Pennsylvania 15230

Robert T. PoeUniversity of California, Riverside, California 92521

Richard Pollina

Montana State University, Bozeman, Montana 59717

Reviews of Modern Physics, Vol. 53, No. 4, Part II, October 1981 Copyright 1981 American Physical Society

APS Study Group on Coal Utilization and Synthetic Fuel Production

C. Dwight Prater

Mobil Research and Deuelopment Center, Paulsboro, Negro Jersey 08066

Robert L. Thomas

Wayne State Uniuersity, Detroit, Michigan 48202

Sandor Trajmar

California Institute of Technology, Jet Propulsion Laboratory, Pasadena, California 91103

Ex Officio Panel Members

G. J. Dienes

Brookhauen National Laboratory, Upton, Neio York 11973

Nancy M. O'Fallon

Argonne National Laboratory, Argonne, Illinois 60439

APS Council Review Committee

HarVey BrOOks, Chairman

Harvard Uniuersity, Cambridge, Massachusetts 02138

William A. Fowler

California Institute of Technology, Pasadena, California 91125

Everett Gorin

72 Longmood Driue, San Rafael, California'a 94901

Norman F. Ramsey

Haruard University, Cambridge, Massachusetts 02138

M. D. Schlesinger

4766 Wallingford Street, Pittsburgh, Pennsyluania 15213

The report presented here is the result of a study conducted over a period of approximately fourteen months,

commencing in April 1980, on research planning for coal utilization and synthetic fuel production. This study

is the most recent one of a series on topics related to energy technologies, initiated and organized by the

American Physical Society through its Panel on Public Affairs (POPAj. Financial support for the work of the

study panel has been provided by the U.S. Department of Energy, through the Office of Energy Research and

the Office of the Assistant Secretary for Fossil Energy. The report is directed toward a twofold audience. The

primary intended audience is that of physicists, particularly those who may become active in the research

areas discussed here. Our aim is to show these physicists where useful research can be done, to show the

fundamental scientific interest in the particular work, and to give enough information to get people started.

Our second intended audience is the DOE sponsors and others in the "user community". Here we aim to

suggest areas where contributions from physicists may most profitably be sought and to suggest what

resources are necessary to optimize that contribution.

Rev. Mod. Phys. , Vol. 53, No. 4, Part ll, October 1981


CONTENTS

I. Summary and Principal RecommendationsII. Introduction

III. Research on CoalA. Primer on coal

1. Origin and petrography2. Classification and nomenclature3. Chemical structure4. Physical structure5. References

B. Pore structure of coal1. Introduction2. Research areas

a. Quantitative characterization of coalpore size distribution

b. Study of molecular species resident inand moving through pore structure

3. Recommendations4. References

C. Analytical techniques1. Introduction2. Selected problem areas

a. Elemental compositionb. Chemical structurec. Physical structured. Surface characterizatione. Physical and chemical transformationsSelected techniquesa. Introductionb. Spin resonancec. Electromagnetic spectroscopiesd. Electron probe and electron

spectroscopye. Other techniques

4. Recommendations5. Summary6. References

IV. Research for the TechnologiesA. Primer on coa1 utilization technologies

1. Pyrolysis2. Synthetic fuels3. Gasification: Importance of general

features4. Flow regimes in heterogenous reactors5. Some specific gasifiers6. Liquefaction

a. Some important direct processesb. Indirect processes

7. Direct combustiona. Pulverized coal combustion in boilersb. Fluidized bed combustion

8. Advanced concepts embodying combinedcyclesa. Pressurized fluidized bed combined

cycle

b. Magneto-hydrodynamic conversion"MHD"

c. High temperature fuel cells9. Mining

10. Coal preparation11. References

B. Instrumentation and control1. Introduction

a. General nature of the difficultyb. Plan of presentation

2. Instrumentation and control needs andproblemsa. Level measurementb. In-situ analysis of solid and/or liquid

process streamsc. Gas analysisd. Pressure

3. Multiphase mass flow measurement4. Monitoring of particulates

a. General discussionb. Intensity ratio of forward scattered

laser lightc. Visibility method using split laser beamsd. Particle sizing using laser diffractione. Photon correlation spectrometryf. Imp actorsg. Overview

5. Temperature measurementsa. Thermometry for coal technologyb. Temperature measurements and

modeling6. Potentialities of acoustics

a. Reasons for investigating acoustictechniques

b. Present applications of acoustics in coalutilization

c. Some possible applications of acousticsto coal technology

7. Recommendations —instrumentation andcontrol

8. ReferencesC. Modeling and flow theory

1. Introduction2. Modeling at the particulate 1evel3. Zone models4. Example of modeling at the coal reactor

level5. Multiphase flow modeling6. Need for model test facility7. Theoretical aspects: Summary and

recommendations8. References

D. Materials in coal conversion and utilization1. Introduction2. Environmental fatigue and fracture

a. The hydrogen environmentb. Crack nucleation and growth



3. Erosion4. Corrosion5. Structural ceramics6. Refractories and slags

a. Refractoriesb. Slags

7. Recommendations8. References

E. Size dependent phenomena1. Introduction2. Coal preparation3. Mechanical properties of coal4. Chemical comminution5. Coal-oil mixtures6. Coal-water systems7. Beneficiation jhigh gradient magnetic

separation (HGMS)8. Recommendations

9. ReferencesF. Catalysis

1. Introduction2. Zeolite size and shape selective catalysts3. Mineral matter catalytic effects in direct

liquefaction4. Surface science related to transition metal

heterogeneous catalysts for gas synthesis5. Catalyzed gasification of carbons and coals6. Denitrogenation and desulfurization

catalysts7. Recommendations8. References

Appendix I. CO2 Emissions: Comparisons of the Use ofCoal for Energy Production by VariousRoutes

Appendix II. AcknowledgmentsAppendix III. Glossary

Rev. IVIod. Phys. , Vol. 53, No. 4, Part II, October 1981

Summary and Principal Recommendations

I. SUMMARY AND PRINCIPALRECOMMENDATIONS

This report is meant to provide guidanceto, scientists and planners for maximizing thecontribution of physics research t oimprovement of coal technology, includingsynthetic fuel production. The Study Panelhas identified a set of research issues havingpotential usefulness for the development ofthe technology, as well as of feringopportunities for increasing our understands. ngof fundamental physical processes.

The areas selected include both researchon coal itself and research pertinent ' totechnology development. To gain improvedunderstanding of coal as a material we havefocused on the development of new and improvedanalytical techniques, and we also o f fersuggestions for work to clarify the nature androle of the complex pore networkscharacteristic of the physical structure ofcoals. The research issues pertinent totechnology development include instrumentationand control, modeling and flow theory,materials problems, size-dependent phenomena(size reduction of coal and coal/fluidinteractions), and cataysis. Researchassessments have been developed for theseselected problems, and recommendati onsoffered. For success in attaining the desiredprogress in these areas, and in coalutili zation and synthetic fuel technologygenerally, cooperative ef forts of physicistsworking with other physical-phenomena-orientedresearchers (chemists, materials scientists,engineers) are necessary. It is highlydesirable to consciously foster such andinterdisciplinary e f fort . One contributiontoward this is to organize interdisciplinarysymposia and topical conf er ence s i n r e sear chpertinent to coal utilization, and syntheticfuel production, as is being done under theauspices of the American Physical SocietyCommittee on the Applications of Physics.

Environmentally acceptable utilization ofcoal as an ener gy source is a centralmotivation for the technology, and hence thepertinent research, for coal ut ili zation andsynthetic fuel production. Many of theresearch areas we discuss affect monitoringand controlling production of undesir ablesubstances. Me have not dealt with questionsof environmental impact as such. However, thecoupling between success in the research that

/

i s the topic of this r epor t and therninimi zation of adverse environmental effects,should be borne in mind. Since CO is aninevitable emission in the utilization offossil fuels, and the amoun t pr od~uc ed i sinversely related to the thermal efficiency ofprocesses, we have deemed it appropriate anduseful to provide an analysis comparingvarious coal utilization schemes as sources ofCO~. If our research recommendat ions bearfrui t, that should help to overcome some keyobstacles i n the way o f improved coalutili zation and synthetic fuel pr oduc t iontechnology. Here we present our highestpriority recommendations, both research andinstitutional. More detailed lists o fresearch recommendations appear in thesections treating each selected research area.

A. Principal research recommendations, (not ranked)

~ Me recommend that support for physicsresearch in the area of coal utilization makeprovision for a non-negligible percentage ofhighly novel, even speculative, approaches tooutstanding applied problems. The potentialeconomic payof fs from solutions to theseproblems is so great that support for unusualnontraditional approaches is much mor ejustified in the coal research area than inmost applieD areas.

~ Ne recommend that presently available anddeveloping techniques be applied to thedetermination of chemical structural featuresof coals, including hydroaromatie and aromaticcluster configurations, functional groups, andmol e cu 1 a r we i gh t di str ibutions. Goodcandidates for such techniques are magneticresonance, vibrational, laser, photo electr on,mass, and X-ray spectroscopies.

e Me recommend syst ematie analyses of thephysical properties of coals to elucidate thepore structure of coal particles includingpore shapes and pore size distributions.There should be emphasis on measurements underdynamic conditions, i.e. following changeswith time under realistic operating conditionsof temperature and pressure. This could bedone by use of small angle X-ray and neutronscattering, X-ray tomography, and electronmicroscopy.

o Ne recommend strenuous efforts to developbetter dynamic models of advanced coalprocesses. Such models are necessary to

Rev. IVIod. Phys. , Vol. 53, No. 4, Part II, October 1981


permit reliable scale-up of the processes andto serve as the bases for design of controlsystems . Economi c motivations andconsiderations of safety provide considerableincentive for this research. Very shortresponse times or the need to integratesubsystems with very di f ferent time constantsoften compound the difficulties.

e Me recommend initiation of a carefullyplanned program to evaluate presentlyavailable instruments, and ~here necessary todevelop new instruments, usable both forsystematic collection of data on advanced coalprocesses and for incorporation into controlsystems for these processes. Such dataacquisition is necessary for producing andverifying dynamic models as recommended above.Instruments are needed which can measuremult iphase flow, temper ature, particulatematter, compositions (without sampling) ofdense phase and gaseous material, levels ofgas-solid and 1iqu id -solid inter faces,pressure, and viscosity of the processmaterials in plants. The measurements mustoften be done in real time under adverseconditions. Attention should be gi ven tonovel and advanced (e.g. acoustic and laser)techniques, which have been insuf ficientlyexploited in the past.

~ Me recommend an intensive ef fort ontheoretical work needed to formulate problemsconcerned with multiphase flow behavior, andto develop innovative computer simulations ofsuch behavior. Such research is important tofurther progress in coal utilization, and canlead to improvements in efficiency andsigni ficant decreases in undesirableemissions.

elate recommend an extensive research effortto understand the effects of coal conversionand combustion environments on materials ofconst, ruction (alloys, refractory ceramics,composites, etc. ), including long durationexposure to the process environment. Theability to predict such effects is essentialto ensure reliable and economic operationunder process conditions which may r eachtemper atures as high as 2600C (NHD combustor)and pressures to 2000 psi, which involveatmospheres that are either highly oxidizingor reducing (or that alternate cyclicallybetween these), and which may include soakingin a highly corrosive slagging environment, orexposure to erosive, particle-laden streamflows (as in the direct, coal-fired gasturbine, for example) . Hope ful ly, suchunderstanding will also lead to the

improvement of pertinent materials properties.The recommended research e f fort shouldinvolve. a microscopic approach, using thetechniques of sur face science and theoreticalsemi -empir ical potential modeling, toelucidate the ef feet of the environment(especially hydrogen) on fatigue and fracturephenomena, a fundamental analysis of thephysical mechanisms in solid particle erosionprocesses; and a study of mass and chargetransfer through solid and liquid oxides.

~ He recommend pursuing experimental andtheoretical investigations to delineate thephenomena in si ze reduction processes, such asgrinding and chemical comminution. Currentpractices and technologies are inefficient;and they involve large expenditures of energy,especially to make very small particles.Efforts need to be made to relate themechanical properties and physical (pore)structure to fracture init iation andpropagation in coals.

eMe recommend further work to bringsophisticated spectroscopic methods of solidstate and sur face physics to bear on systemswith compositions, and conditions of pressureand temperature, as closely akin as possibleto those used in practical heterogeneous metalcatalysis for fuel synthesis, i.e. , startingfrom syngas (H2 and CO) mixtures obtained fromcoal gasi f ication. It is part icula r lydesirable to develop spectroscopic techniqueto study intermed iate compounds formed onmetal surfaces in catalytic synthesis and inreacting coal chars.

~ Me recommend increased emphasis onunderstanding and exploiting the properties ofsize and shape selective catalysts. There areat tractive oppor tun itic s for fundamentalscience with great potential impact on thedevelopment of indirect liquefactiontechnology. It is of particular interest tounderstand dif fusion processes in zeoli tesystems of thi s type, and t o make and studysystems with very small part icles oftransition metals imbedded in zeolites.

B. Principal institutional recommendations

eMe recommend the est ab1 i shment o f asample bank of well selected, characterized,and preserved coal samples, available toresearch and analytical groups. This shouldbe preceded by research sufficient to verifythe chosen storage technique(s).

Rev. Mod. Phys. , Vol. 53, No. 4, Part II, October 1981

Summary and Principal Recommendations

This recommendation makes explicit ourstrong agreement wi th the pr imaryrecommendat ion of the recent Basic CoalScience Project Advisory Committee Report tothe Gas Research Institute (GRI), and our wishto encourage the existing interest in the U. S.Department of Energy (DOE) in establishment ofsuch a sample bank. A recent workshop (March27-28, 1981), sponsored by GRI and DOE, washeld to delineate the cur rent sample needs ofthe research and technology communities. Theestablishment of a National Premium CoalSample Bank is under consideration as a resultof this workshop and other meetings ofinterested groups. Such establishment wouldmeet the aims of our recommendation, providedthat the quality and availability of suchsamples is maintained for analyt ical andresearch purposes.

For the maximum effectiveness of theresearch and analyses recommended in ourpresent report to be achieved, standardmaterials need to be available to manyresearch workers. In the past, theeffectiveness and reproducibility of researchin coal science has been adversely affected byinsufficient attention to sample quality. Thecomplex heterogeneous nature of coal from anyone source, together with the diversity ofcoals from dif ferent sources, makes carefulatt ention to sample select ion andcharacteri zation essential i f di f ferentresearchers' results are to be compared andsynthesi zed into a larger understanding. Me

recognize that, there is no consensus at thistime as to the preferred route of stablestorage of coal samples. Our coal sample bankrecommendation therefore includes doingsufficient research to verify the validity ofthe chosen storage technique(s).

The implementation of this recommendationfor a National Premium Coal Sample Bank shouldbe in stages. The initial stage should be touse the best o f the cur rent ly availabletechniques, e.g. coal storage in deioni zedwater at a few degrees Celsius, coal stor agein deioni zed ice, or coal storage in an inertatmosphere of nitrogen or argon. Each ofthese techniques has certain disadvantages innot preserving one or more moieties in thecoal. Therefore our recommendation includes acont, inuing longer term program of coalcharacter i zation, including ef fects o fstorage, to further refine the storage modesand to define new methods of storage for thesamples in the Pr emium Coal Bank.Consideration should also be given to storage

of samples from di f ferent locations in a seamand in different states of aggregation,varying from sizeable pieces to groundsamples. Me recognize that in addition to therecommended storage bank for premium samples,to be available in relatively limitedquantities, the need will continue for other,already exi sting, sample banks that havelarger samples available for pur posesrequiring less rigorous storage control.

~ Me recommend estab li shment of a programto gain fundamental understanding of processesfor coal utilization and synthetic fuelproduction. Thi s would focus on thedevelopment and testing of theoretical modelsand of process control techniques, togetherwith the development of instrumentation foracquisition of the data necessary for thispurpose.

Me recommend that a significant programaimed at, fundamental understanding andimprovement of coal utilization and syntheticfuel production processes be undertaken, withthe focus on modeling of liquefac tion andga s i f ication reactors, combustors, andmultiphase flow systems. This program wouldinclude development of instrumental techniquesfor use at dedicated test facilities tocollect data necessary for development andverification of models, and might have toinvolve construction of new facilitiesspecially designed to achieve the objectivesof this program. The increased understandingof the process kinetics will permit reliablescale-up, and will aid the design of betterprocesses. Benefi ts wi 11 also accrue for thedevelopment of control str ategies andinstrumentation .

In order to'. encourage optimum use ofpeople and test facilities at a variety ofinstitutions (universities, government andnon —profit laboratories, industry);systematize the data collection and retrieval;and plan for the best use of funds forexpanded or additional facilities, -- it isrecommended that o ver al 1 r esponsibility forlong range pl ann ing, coord ination andtechnical oversight be assigned to anappropriate group. This group could be at ana t iona 1 labor atory, un iver s i ty, or non -pro fi torganization, or it might consist of acommittee made up of representatives from suchinstitutions, as well as from industry. Theactivities of the Gas Research Institute andthe Electric Power Research Institute serve asmodels, suggesting an arrangement in which theresearch work would be distributed to existing

Rev. lVlod. Phys. , Vol. 53, No. 4, Part II, October 1981


laboratories.In the area of i nstrument ation

development, the program would no t a im t ocarry instrumentation all the way tocommercialization, but rather to advance thestate of the art in key instrumentation areas,both by coordinating research and by providinga centr al repository of information ond i agnost i c and process controlinstrumentation. The ideal would be acooperatively funded program, with long rangefunding coming partly from government agenciesand partly from industry.

In pl arming and e stab 1 i sh ing thesuggested program, the relationship o f

instrumentation for data acquisition to thedevelopment and validation of process modelsis central. The refinement of such models haseconomi c va lue through the r esu 1 tingrefinement of coal utilization technologies.Under current practice, instrumentationef forts are devoted almost entirely tooperation and control of existing workingsystems, while insufficient effort has beendevoted to systematic collection of data totest reliability of modeling concepts, i.e. toprovide the data necessary to model a priorifor better processes. Setting up the proposedprogram should help to correct thisdeficiency.

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 198'I

I ntroduction

II. INTRODUCTION

All plans giving the U. S.A. 's energyneeds and resources for the transition decadesfrom the twentieth into the twenty-firstcentury project a substantially increasingutili zation of coal. It is ther eforeimportant to deal with the technology problemsof that increased use, including thetrade-offs of time and cost, particularlybear ing in mind that the use of coal must bedone in an environmentally acceptable manner.There is an important and, interesting role forphysicists to play in these pr'oblems. Thereport presented here is the result of a studyconducted over a per iod of approximatelyfourteen months, commencing in April 1980, onresear ch planning for coal utili zation andsynthetic fuel production. Financial supportfor the wor k of the study panel has beenprovided by the U. S. Department of Energy,through the Office of Energy Research (OER)and the Office of the Assistant Seer etary forFossil Energy (ASFE). Our aim is to provideguidance to scientists and planners inmaximizing the positive impact of current andfuture physics research on the development oft echnology for coal energy utilization. Byexample, we will show that for physicists thequestions arising out of the needs of coaltechnology lead to science that is interestingand significant in itself.

A synthetic fuel production program basedon coal requires an investment of hundreds ofbillions of dollars and a probable time periodof decades to meet a relatively modest(10-15%) fraction of the country's liquid fuelneeds. In add i tion, advanced technologiesusing coal for the generation -of electricitycall for further heavy investment of time andmoney. The latter level of investment issignificant, in part, because more thanincremental technology and retro-f it isenvisaged. For example, t he coal-f iredintegrated gasification combined cycle (i.e. ,gas and steam turbines) power plant maysupersede the conventional coal power plant,since the former has comparable or higherefficiency and superior emissions (NO, SOand particulates) contr ol. This developmentholds the promise of relieving the present useof oil and natural gas without the level ofenvironmental damage normally attributed tocoal.

Because of the long lead time t o put thenew synthetic fuel and electricity generating

technologies into place, research must be donenow for impact on the ener gy systems of ten ortwenty year s, or even longer, from now. +

Also, information gained fr om research cancrucially af feet planning for the very lar geinvestments involved in development anddemonstration, as well as the ult imate vastc ap i ta 1 investment involved in productionfacilities. The aggregate investment requiredis such that even a small percentage reductionin costs resulting fr om research findingscould have dramatic economic impact. Much o fthe research discussed in Part IV, Researchfor the Technologies, is also pertinent to the-production of liquid fuels from oil shale.

Our report is or iented primarily towar dphysicists. A number of physicists alreadyare active in coal related research, but thatnumber is small r elative to the needs andchallenges of the problems. Problems such a sthose of mater ial s and instrumentation, verymuch in the bailiwick o f physicists, arecentr al to the economic viability o f c oaltechnologies. Theoretical modeling of thehuge reactors necessary for coal conversion togaseous and liquid fuels is basically anengineer ' s problem, but a large element inthat problem, the linkage between modeling andinstrumentation, has been neglected. Thephysicist ha s un ique qual i f ication tocritically examine and improve the physicalassumptions of a model, and to deviseinstr umentation for acquir ing data to testthat model.

Understanding and control of multiphaseflow is a central issue throughout thetechnology. Hi sto r ic ally, fundamentalunderstanding of this question is aphysicist's game, pnd this seems to be thetime for physicists to reactivate thisinterest. There is a need in the multiphaseflow theory area for a serious look to seewhether advances in computing .technique inrecent year s offer opportunity to includeenough of the essential physics to gain realunderstanding. Also, further data arenecessary on which to base any s uc hcomputational advances.

~For general discussion of research needs forcoal utilization and synthetic fuel productionsee Energy Research Advisory Board, 1980;Fossil Energy Research Morking Group, 19"(9 and1980; Gas Research Institute, 1980; Gor baty etal. , 1979; Grimes e t al . , 1 980; Keller, 19(9;National Academy of Sciences, 1979.


10 APS Study Group on Coal Utilization and Synthetic Fuel Production

In addition to working on problemsaf fecting material s o f construction, manysurface, chemical, and solid-state physicistscan expect to beneficially affect catalysisquestions central to t he economic per for manceof the technologies. Physicists have played asignificant role in gaining under standing ofthe crystal structures of the size and shapeselective zeolite catalysts that are coming todominate refining technology. Yet knowledgeof these materials, and the opportun ities andneeds for new research on them for syntheticfuel technology, i s not widespread amongphysicists. In fact, oppor tunities offered bythe str iking nature of zeolite crystalstructures is a beauti ful example of thegenuinely new science that relates to theneeds of coal technology. Zeolite structurescontain crystallographically defined cavitiesthat ar e per fectly designed to hold tinyparticles of transition metals (~7-10A indiameter) . Moreover, needs of pr act icalcatalysis chemistry have provided techniquefor making samples with such embedded metalparticles. Such systems of fer excitingprospects for exploration of size andsurface-dependent electronic properties; andone has the happy situation, described moreful 1y in Sect ion IVF, that these systems o fferthe possibility of combining desired featuresfr om two competing methods for the synthesisof high octane gasoline from syngas (CO and H

mixture) produced by coal gasification.The objective of much of the technology

we will discuss is conceptually simple. Coalhas major deficiencies as a fuel. Thesea f feet environmental acceptability andconvenience in use. Coal is non-uniform, itcontains impurities; and its solid form isinconvenient for many uses, especially as atransportation fuel. These problems can besolved by liquefact ion or gasi fication, bothof which in essence simply involve increasingthe ratio of hydrogen to carbon, i.e. , the H/Cr atio in coal is much lower than in petroleumand lower yet than that in natural gas(methane). ,Of course, this addition ofhydrogen is not simple. A great complicationis that the starting material, coal, iscomplex and poorly under stood. Thusd evelopmen 0 o f anal yt ical t echn i ques iscentral to the hope for genuinely newtechnology. Trad it ional ly, physic i sts haveplayed a leading role in the development ofnew analyt ical techniques. Of ten thesetechniques have then been refined and appliedby other workers. A good example in the coal

area is the introduction o f Mossbauerspectroscopy as an analytical tool. Use ofthis technique to characterize coal impurities(some of which have an impor tant catalyticfunction) and transformations dur ingprocessing was introduced by physicists only afew years ago. Yet already Mossbauerspectroscopy has become a standard technique.

To design efficient ways for reactingcoal, we require knowledge of its complexphysical structure, as well as of the chemica1structure. This physical structure basicallyconsists of an intricate mesh, with pores ofvarying size. The por e structure controls theaccess and egress of reagents and reactionproducts. Development o f nonintrus i v ephysical probes of this structure, such asnuclear magnetic r esonance (NMR) and smallangle x-ray and neutron scattering (SAXS andSANS) are desirable to delineate the nature ofthe por e str ucture and the pr ocesses by whichreactive species migrate through it. A

par ticul arly promi sing way for physicists tocontribute to the necessary understanding isby designing research progr ams whichcoordinate the measurements of molecularmigration on di f ferent time scales, e.g.recent studies o f water in coal havesuccessfully correlated results from specificheat and NMR measurements.

Besides being addressed to the physicscommunity, our report is intended to providethe DOE sponsors, and other s in the "usercommunity", with some indication of how agroup -- predominantly physicists, butcontaining chemists and chemical engineersrepresenting quite a wide range ofspecializations, per ceives the sc ienti f icaspects of coal utilization after aconscientious reconnaissance —and to suggestareas where contr ibut ions from physicists maybe most profitably sought, as well as tosuggest what resources are necessary tooptimize that contribution.

Nature of the study

The study r eported here is the mostrecent one of a series initiated and organizedby the American Physical Society through itsPanel on Public Affairs (POPA) ~ At present,physicists have participated only to a limitedextent in coal ut il i zation related research.Besides input on speci fic technical questions,it was desired in general to gain the newpoint of view r esulting from an incr easing

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 1981

Introduction

participation by physicists in t;his researchar ea. As part; of gaining a "fresh look",roughly half of the study panel members hadlittle or no previous experience in coal. Theother half of the panel, in quite varyingdegrees, were already fami1 iar wi th the areaof coal science and technology, and providedthe experience base necessary to prevent"reinventing the wheel. " In addition to theknowledge brought in by these people, thestudy panel sought and studied the backgroundknowledge and views of a great number ofconsultants. These consultants are listed inAppendix II. In composing the panel, adiversity of background and interest played astrong role. Mhile physicists are in themajority, the panel also contains chemists andchemical engineer s.

At the outset it was decided to aim atselectivity —to identify and treat a limitednumber of impor tant technical issues in somedepth, rather than to aim at a comprehensivetreatment. Af ter an ext ended per iod ofinformation gathering and discussion, thepanel selected seven priority areas fordetailed assessment o f r esearch needs andopportunities. The ar ea s selected wer e '.

Physical (i.e. Por e) Structur e of Coal,Analytical Techniques, Instr umentation andControl, Modeling and Flow Theory, MaterialsProblems, Size-dependent Phenomena (i.e. sizereduction o f c oa1 and coal/fluidinteractions), and Catalysis. The bulk ofthis report then consists of these researchassessments, which in these areas delineatethe practically important and scientificallysigni ficant opportunities for physicists.

Our principal recommendations pr ecedethis Introduction, as part of the ExecutiveSummar y. They inc 1ud e bot h r ecommendat ionsfor speci f ic pr ior i ty research and forinstitutional arrangements necessary to getthe most out of that research, specifically(l) establishment of a program for developmentand testing of theoretical models and processcontrol techniques in coordination wi thr equi si te wor k on development ofinstrumentation, and (2) provision for a"bank" of well characterized coal samples.Nore detailed lists of recommendations areincluded as part of the r esearch assessmentsfor the individual technical areas.

Organization of the report

The remainder of the report presents ther esearch assessments in the selected areas,

including the backgr ound necessary to'understand both motivation and the context ofwhat is being recommended. The materia'1 isdivided into two large sections (Sections IIIand IV), one dealing with research on coali tsel f, the other deal ing wi th researchpertinent to the technology of coalutilization and synthetic fuel production. On

the assumption that t he coal r elatedbackgr ound of many reader s i s limited, eachsection begins with a primer, i.e. the "coalprimer" in Section III A and the "coaltechnology primer" in Section IV A. There isalso an appendix on the amount of CO producedby various means of coal utili zation orconversion.

The coal section deals with two researchareas: physical (i.e. pore) structure of coalin III 8 and analytical techniques in III C

The motivation for dealing with these topicsis simple. To utilize coal most effectively,one wants to understand as much about coal aspossible.

For example, in the view of advocates ofdir ect liquefaction processes, the aim is topreserve as much as possible of the desirablemolecular structure of the organic part of thecoal, minimizing molecular rearrangements whenit is converted into a liquid. That school ofthought holds that thi s should be inherentlymore energy efficient than indirect processes,where one completely destroys the molecularstructure of coal in creating a "syngas"mixture o f -CO and H&. whether or not thisexpectation of the relative ef fic iency o fdirect liquefaction is ultimately vindicated,the development o f e f ficient dir ectliquefaction pr ocesses requir es knowledge ofthe molecular structure of coal; indeed, itrequires knowledge of whether one should thinkof a molecular struct ur e at all or shouldadopt some other heuristic picture for thechemical structure. Such need for knowledgeof the chemical structure of coal will occuroften in technology development. Clear ly thisrequires analytical techniques. The problemis that coal is a highly variable conglomeratemater ial composed o f several distinguishableorganic subst ances and minerals. Moreover,there is no "standard" coal, nor any obviousway 4o de f ine one. The Analytical Techniquessection addresses itself to the question ofwhat the physicist can do, and what i s neededto introduce new techniques in the face ofthese difficulties.

Me have already mentioned above the keyrole played by the por e struc tur e in

Rev. tVIod. Phys. , Vol. 53, No. 4, Part I I, October 1981


controlling chemical reactions in coal.Di ffusion through t he po r e system and themechanism of adsor ption at the pore sur facesindeed may limit the rates of such reactions.Methods for characterizing the pore structure,and the behavior of molecular species ofinterest within this structure, are discussedin the Pore Structure Section.

The "coal technology primer" of IV A isfollowed by the di scus sion of Instrumentationand Control in Section I V B. Contr ol i sneeessar y for safety and reliable operationunder severe conditions involving extremelyhigh temperatures and pressures and harshphysical and chemical environments, where theneed for short response times often greatlycompounds the di f f icul t i e s. A3. so theinstrumentation must be done in an economicalway. Up until now, instrumentation designefforts have been concentrated on needs ofcontrol for safety and process operations.Another main ob jeeti ve o f instrumentationdesign should be for the data aequi si tionnecessary for realistic modeling of conversionand combustion processes as described inSection IV C. Development of such predictivetheoretical models is important to theeconomic improvement of coal technology,particularly in view of its sensitivity topr ocess details. One must captur e thesedetails, for example the ef fects of mineralcontent, in the compl i ca ted set ofdifferential equations used to formulate themodel for some process, and then find atractable way to solve these equations, againpreserving the details of interest.

Single phase turbulent fl ow andmulti-phase flow ar e pervasive features ofpr ocesses of interest. For example, hightemperature, high velocity slurries of coalpar ticles in fluids, as commonly occur in coalconver sion systems, involve problems oftwo-phase flow. Development of the theory offlow phenomena, also discussed in IV C, isc en tral to improvement o f model ing andconsequent improvement of the processes.

Finding dur able mater ials ofconstruction, able to prov id e des i r edpr operties, and assure adequate lifetime,under the har sh operating conditions of coalutil i zation and synthetic fuel technology, maybe the key to economic viability. To beeconomically successful, a plant must notmerely wor k, it mu s t continue to work,reliably and at reasonable maintenanceexpense. The pertinent materials requirementssuggest research needs and programs that are

discussed in Section IV D.Coal util i za tion typical ly i nvolves

par ticle size reduction. Also interactionswith fluids (whether native to the coal oradded for t r an sport ation, processing, orduring size reduction) depend on phenomena atthe coal/fluid interface and hence are alsolikely to depend significantly on coalparticle sizes. Sever al topics concerned withcoal' preparation and coal/fluid mix'tures aretreated under the category of "size-dependent,phenomena" in Section IV E.

Effective catalysts are impor tant for aneconomically viable synthetic fuel productiontechnology, because of the desire to per formthe necessary r eactions rapidly under "mild"conditions of temperature and pressure. Ofcourse, in current practice, these conditionsare not at all mild, so progress in catalysiscould have very substantial economic impact.Also, catalysis plays a key role in productselectivity, e.g. getting more of the mediumweight hydrocarbons that consti tute gasolineand less of heavier or lighter hydrocarbons.Maintaining the ef fectiveness of catalysts isimportant because catalysts can be veryexpensive ,'and therefore understanding andcombating "po i soning" of c atalysts i simportant. Questions of cat alysi s r elate toresearch interests of physicists in sever always, particularly in the area of electronicstructure at surfaces and interfaces, and insmall metallic particles. Such questions arediscussed in Section IV F.

There is, of course, much interplaybetween r esearch questions in the broadresearch areas discussed in Sections III andIV. Ne air eady have emphasi zed the linkbetween instrumentation development for dataaequi si tion and modeling. Technical problemsin development of analyt ica1 techniques forlearning about the chemical structure of coalare sometimes r elated to those ofinstrumentation for proces s control andunderstanding. A number of questions ofcatalysis depend on the pore structure ofcoal, or involve materials, such as zeolites,themselves having a (much more well defined)pore str uc tur e; and comminution ( sizereduction) of coal and coal/fluid interactionsalso depend upon the physical structure o fcoal. Other examples of interrelationshipsbetween the different areas will be apparentto the reader.

In this r eport the focus i s on researchtopics pertinent to better characterization ofcoal and to improvement o f coal uti1 i zation

Rev. Mod. Phys. , Vol. 53, No. 4, Part I l, October 'l98'1

I ntroduction

processes. Although the Panel was aware ofthe impor tance of environmental ef feets incoal utilization, we have not dealt withquest, ions of environmental impact. Thereforeresearch topics o f envi ronmental ef feet sconcomi tant, to the use o f coal such a satmospher ic effects, global heating, acidrain, etc. have not. been addressed. (That is,we have not dealt wi th impacts of theseeffects. The technical topics we have treatedrather deal with monitoring and controllingsuch emissions. ) Such topics form researchareas of their own and require considerationsdiffering in generic as well as detailedchar aeter from those considered here. (Seediscussion in Of fice of Technology Assessment,1979; Green, 1980; wilson, 1980; Landsber g,1979; National Academy of Sciences, 1 980)However, it is necessary to recognize thecoupling between these research areas an& thesubjects of this report since di fferent coalut il i zation processes produce, in addition todesirable forms of energy and fuel, variousyields of nitrogen oxides, su'lfur ox ides,aerosols, respirable particles, small r esiduesof natural radioactivity, and carbon dioxide.

In par ticular, CO& i s an inevitableemission in the utili zatxon of fossil fuels.The amount of CO is inversely related to thethermal efficiency. This means that if one isconcerned about CO production, one has noalternative to a deep concern forthermodynamic efficiency in every aspect, ofhydrocarbon util i zation. In Appendix I, wepresent a comparison of var ious coalutilization schemes as sources for CO . Ouranalysis leads to some numbers that give afeeling for the amounts of CO that would beproduced by synthetic fuel technology a spresently contemplated.

General remarks

Me have reeogni zed many specificdi f f icul ties needing solut ion, oruncertainties needing clarification, for thefur theranee of coal utilization and syntheticfuel produe tion technology. However, werealize that some of these, for example thosehaving to do wit, h erosion and corrosioneffects on materials of construction, aresear eely new. In fact the Department ofEnergy, both through ASFE and OER, has longmaintained research pr ograms in some of these'areas -- as have several other federalagencies and industry. , Ho~ever it should beemphasi zed that coal and i ts associated

technologies present many of these problemsand di fficulties in extraordinarily sever eform, and that extraordinary effor ts areneeded. It is important for physicists tobring to bear new per spectives for solvingthese problems.

Much invention is needed (e.g. somethingbetter than lock-hoppers for dry materialsfeeding). And though invention often resultsindirectly from research it is not anautomatic consequence —and the dir ect, ion ofinvent, ion ean scarcely be anticipated on thebasis of the direction of research. Severalother Studies and Reports have remarked on thefact that many problems probably will comeinto clear focus, only as a result ofoperating experience with rather large scalefacilities, i.e. large pilot (about 250tons/day) or demonstration-size (thousands oftons/day) facilities

In addition, much basic scienti f ieresearch is needed, includ ing that byphysicists. The perception of this need wasthe initiating impetus to this report, and ouref forts have r eaf firmed this thesis. Verymany of the open questions in coal utilizationand synthetic fuel development are at thelimits of pr esent physical under standing,e.g. , of sur faces and of multi-phase flow.Moreover, the techniques physicists ar eaccustomed to employ —laser s, NNR, neutr onscattering, etc. -- are hopeful tools foransw ring those problems. Me especiallystress that coal researches by physieists canextend basic knowledge in directions whichother physieists will find instructive andwill be happy to f ind published incontemporary physics journals.

To create an atmosphere conducive toinvention and solutions to pr oblems it isimpor tant to expose the whole area of coaltechnology to a br oader technical community,Me believe that the government agencies andprofessional societies should do as much aspossible to provide avenues of publication andrecogni tion and to strengthen communicationamong scientists and engineers in alldisciplines working on coal. The holding ofmeetings and topical con ferenee, as alreadydone under the sponsorship of DOE and APSamong other s, is one example of desir ableaction.References

Energy Research Advi sory Board, 1980,- Research and Development, Needs in the



, Department of Energy, Interim Report ofthe Resear ch and Development Panel,Energy Resear ch Ad vi sory Board,September.

Fossil Energy Research Morking Gr o up, 1979,Assessment of Long-Term Research Needsfor Coal-Gasi f ication Technologies,Fossil Energy Research Morking Group, S.S. Penner — Chairman, MITRE TechnicalReport MTR-79M00160.

Fossil Energy Research Morking Group, 1980,Assessment of Long-Term Research Needsfor Coal-Liquefaction Technologies,Fossil Energy Research Morking Group, S.S. Penner —Chairman.

Gas Resear ch Institute, 1980, Basic CoalSciences Pr oject Advisory CommitteeReport, Gas Research Institute, Chicago,Illinois.

Gor baty, &. L. , F. J. Mright, R. K. Lyon, R.B. Long, R. H. Schlosberg, Z. Baset, R.Liotta, B. S. Silber nagel, D. N. Neskora,1979, Sei ence 206, 1029.

Gr een, A. E. S. , 1980, Coal Burn in g Is s u e s,edited by A. E. S. Green (Univer sityPress of Florida, Gainesville).

Gr imes, M. R. , G. D. Br unton, H. D. Cochr an,Jr. , J. R. Hightower, Jr. , J. S. Johnson,Jr. , I. L. Thomas, J. A. Matson, and F.M. Miffen, 1980, "Assessment of BasicResearch Needs and Priorities in Supportof Fossil Energy" in Basic Research Needs

in Seven Energy Re 1 a ted Technologies,DOE-ER-OO60.

Keller, 0. L. , 1979, ORNL Coal Chemistry1979, ORNL 5629.

Landsberg, H. H. , 1979, Ener gy. The NextTwenty Years; Report by a study groupsponsored by the Ford Foundation andadminstered by Resources for the Future;H. H. Landsberg, Chairman (BallingerPublishing Co. , Cambridge).

National Academy of Sciences, 1979. SomeAspects of Basic Research in the ChemicalSciences, Na tional Academy o f Sciences,1979-

National Academy of Sciences, 1980, Energyin Transition 1985-2010, Final report ofthe Committee on Nuclear and AlternativeEnergy Systems, Nat ional Academy ofSciences (W. H. Freeman and Company, SanFrancisco).

Office of Technology Assessment, 1979, TheDirect Use of Coal, Congress of theUnited States, Of fice of TechnologyAssessment (Super intendent of Documents,U. S. Go vernmen t P'r in ting Of f ice,Ma shington, D. C. , Stock No.052-003-00664-2) .

Milson, C. L. , 1980, Coal -- Br idge to theFuture, Carroll L. Mil son, Pro jec t,Di r e c tor, Repor t, of the Nor ld Coal Study(Ballinger Publishing Co. , Cambridge) .

Rev. Mod. Phys. , Vol. 53, No. 4, Part I t, October 198't

Research on Coal

I II. RESEARCH IN COALIII. A. PRIMER ON COAL

1. Origin and petrographyThe term "coal" i s appl ied to a ver y wide

range of substances. All are carboniferoussolids formed over geological per iods by themetamorphosis of vegetable matter under heatand pressur e. Coal in the words of oneauthority, (Francis, 1954), is:

a compact stratified mass ofmummified plants which have beenmodified chemically in varyingdegree, interspersed with smalleramounts of inorganic matter. "

This having been said, one should be prepared .

from the outset for the fact that coaldisplays on every se al e —— a tomi c,micr oscopic~, and macroseopie -- an intrinsicheterogeneity which reflects both the naturalbiological variability of vegetation and thediver sity of geologic histories resulting ineoalification.

The starting matter is ehemieally andphysically heterogeneous -- containing wood,bark, leaves, . needles, pitch , spores, algae,etc. , and all their diverse chemicalconstituents such as resins, waxes, lignins,hemi-cellulose, cellulose, ete. over pe r i od sr angi n g from 60 to 200 million years,mor cover, the various coals have experiencedan extremely wide range of differing geologichistor ies. Processes o f coal i f ication arecomplex and poorly understood. The transitionfrom biological processes (e.g ~ decompositionmediated by microbes) to purely physicalproeesse s cannot be un ambi guou slyreconstructed. At present, it is not possibleto discuss the structure of coals in terms ofknown trans formations o f known pr ecur sor s.(Flaig, 1966, Stach, 1975)

The conglomerate character of coal isvisually apparent under the petrographer'smicr oscope. (Petrakis, 1980) The carbonaceous(organic) material is composed of severaldistinguishable substances called "macerals"which are disposed in irregularly shapeddomains. Macerals may be classif ied intoabout four pr inc i pal easily d istinguishablevarieties, with numer ous finer distinctionspossible. The domains range in size from a

few microns to the macroscopic. Differentmaeer als are most immediately distinguishablebecause their optical ref lectivities differ,and they are found also to differ chemicallyand in regard to physical characteristicsother than refleetivity. Different maceralsare believed to arise from differentconstituents of the or iginal vegetablemat ter. Coal also contains inclusions o finorganic matter, these are for the most partstandard minerals, not unique to coal, andwell studied in other contexts. They do havespecial significance for the chemical andphysical processing of coal, however.

The serious student of coal must beprepared to come to grips with non-ideal i zedreality. It is not possible to commence thescientific study of coal by "purifying" it.In fact, neither the notion "pure" nor theconcept "perfect", as familiar to solid statephysicists and chemists, is relevant in thecase of coal. The nearest possible thing to aclassical (pure substances) approach is tostudy the properties of segregated macerals,one variety at a time. The need to do thisimparts some importance to the devel. opment ofe f f icient techniques for separating andsorting macerals (Dyrkacs and Horwitz, 1981;Kr oger, 1957; Go 1oushin, 1959) .2. Classification and nomenclature

Di f ficulties and ambiguit i es o fclassification are unavoidable for substancesas varied and complex as the family of coals.Indeed, no system has yet been devised whichsatisfactorily reflects the variations ofatomic arrangement of dif ferent coals or theirbehavior as chemical reactant;s. (Theprerequisite fundamental chemical and physicalknowledge is still lacking. )

The traditional schemes of classificationare according to "rank" and "type. " Each (inseveral versions) is in current use. (Neavel,1978; Berkowitz, 1980) Together, they furnisha scheme of two independent dimensions -- inthe sense that a gi ven rank may f indrepi esentation in coals of several types andvice versa.

Rank

It is useful, in the case of coal, torecogni ze a "microscopic" scale intermediatebetween the atomic and the macroscopic, sincemany of the interesting features fall in therange of a few microns.

Assignment o f r ank provides a grossaverage character ization of a coal indicatingprimarily its heating value as a fuel indirect combust ion; i t i s not concerned withmicro-structure. Lignite, sub-bi tuminous

Rev. Mod. Phys. , Vol. 53, No. 4, Part I l, October 1981


E 00

OQ

~ vH 'M

0

I

N

c9 (Q OOO

2 0

OOOOOOOO~I j IOOOOOO

V)

bQ

~ ~

o ~0E

YQCd)

O OOOOvjmO"o"~I IOOOOOO

OCh Oo

OO00

IOOO O

V/

O

0

E~ M

00oQ

0

O

0

cj

O~ ~

cd0

0~ ~F

bQ

ccj0

00

~ 'M

OQH~ ~

I

E~ENcn +

~ E.E

0 ~~Q

G6

V/

M~DOA

O ~ TH

O

2 4 cn

~ W~ WO

I I

QOao

IoO

c80 0O Ocn v&

~ ~0 ~ 0

o ~ w 0E&E0~ZXZ

0~ W

E

V/ V/ V/

cga

~ oO

~~

Rt

0 & 0

A"Ai A

N VtoooO O

c6 D

0 0 0~ lrt ~ THEEE

MMM

06

M

(DM ~ 'H

~ TH

0S4 ~

cn .&E

(QQ + IN

oPC/J

on

X

0~~

ccrcO

«~~

~ ~. 2.

O O

& o0cn

O

M

~ ~

z

O~ ~ 0

~ W

E~ ~

+~ 0 0

gQ

~ ~y. O

OCO ~~- O."-

~E0bQ

cn 0

j 0O0 g.g cn

0


Research on Coal

TABLE III-A-2. Comparison of Coal Macerals, Thiessen (U. S.) and Stopes-Heerlen (Europe)

Thiessen system Stopes-Heerlen system

anthraxylon and humic matter

spores and pollencuticlesalgal remainsresinous and waxy substances

opaque matter, brown matterfusainfusain and brown matterfusain and brown matter

vitrinite

exiniteor

liptinite

lner tinlte

colinitetelin&te

I sporinitecutinitealginiteresmste

micrinitefusinitesemi fusinitesclerotinite

coal, bituminous coal, and anthracite are themain rank classes. Peat, isconsidered a coalpr ecur sor. Several systems exi st forassigning r ank to coal. The one currentlyused in the United States (ASTM-D-388-66) isshown in Table III-A-1. The par ameter s usedto differentiate between ranks are: 1) "fixedcarbon", (FC) when the value of this isgreater than 69 per cent (fixed carbon is thedry, miner al free residue from a "volatilematter" determination) . Volatile matter (VM)is defined as the per cent by weight, lost byheating a dry sample to 950 C for sevenminutes in the absence of air (thus on a dry,mineral-matter-fr ee basis: FC+VM = 100); 2)when the fixed carbon value is less than 69per cent, differentiation is by calorific(heating) value and/or agglomeration (thecharacteristic of forming a coke button in thevolatile mat, ter test).

Geologically, the progression o f r anksfrom lignite to anthracite may be viewed as anordering according to increasing degree ofmetamorphosis or coalification of the originalvegetable matter. (Parks, 1963; Berkowitz,1980) Hi storically, however, rank wa soriginally devised to differentiate betweencoals to be used in coking and combustion.Even for these pur poses r ank alone does notsuf fice and must be augmented with "grade"information (Neavel, 1978) specifying themineral matter content, sulfur content, andash fusion temperature —qualities which donot vary systematically with rank. Differentcoals, moreover, exhibit di fferent swell ing,so f tening, agglomer ation, and "caking"behaviors which are important in coke makingand other conversion technologies (Beimann etal. , 1963; Thompson, 1981).

Type classification accords r ecognitionto coal's small scale heterogeneity and to itspetrographic characterization -- type beingnow assigned according to the proportions inwhich the var ious m'acer al var ieties arepresent. (An obvious prerequisite is to havecatalogued and described the most commonvar ieties of maceral. ) Typing according tomaceral content dates back to Stopes (1919)and Thiessen (1920, 1938).+ Stopes based hisclassification on observations using reflected1 ight, while Thiessen based his onobservations of thin sections by transmitted

- light. Table III-A-2 gives a comparison oft erms used in t he two systems of maceraldefinition and identi fication.3. Chemical structure

Coal chemistry is faced with formidabletasks if coals are ever to be described at themolecular level, but the potential payoffs maywell be commensurate with the difficulty ofthe task. Presently available synthetic fuelprocesses represent a star t on science basedrefinement and extension of technologies firstconceived decades ago. There is little toindicate that the processes are efficientcompared to others that might be developed. A

bet ter knowledge of the molecular str uctur e ofcoal might suggest completely new reactions,catalysts, or processing conditions.

The nomenclature of coal types is much olderand goes back to earlier type classificationsbased upon visible di f ferences and uponproper ties like texture and mode of fracture.



Despite the difficulty of the problemsome progress has been realized in recentyears as new analyt ical techniques fromphysics and chemistry have been applied tocoal (Karr, 1978, 1979) . Ther e r emain,however, promising scientific tools which haveyet to be applied to coal including some, atleast, which could be extended or modified inways appropriate to the particular problems ofcoal. Section III C of the report is devotedto a mor e extensive assessment of thesequestions.

Studies of segregated maceral varietiesreveal that not only their optical propertiesbut also their element al compositions,densities, and microscopic physical structuresdiffer. See Table III-A-3. One sees here animportant general point, more fully discussedin Section IV-A: even the highest hydrogencontent not;ed (that for exinite) correspondsto an H/C ratio of r oughly 0. 9, ~hereas highgrade liquid transport fuels typically exhibitratios in the r ange 1.9 to 2. 0. In reactionstudies dif ferent macerals respond di fferentlyto hydrogenation, solubilization, pyrolysis,etc. (Thiessen and Francis, 1929; Given, 1975)and even a single maceral variety does notshow the constant, well defined propertiesthat would be char ac teristic of a puresubstance.

Structure at the atomic level isdif ficult to investigate because coal isinsoluble in common solvents, non-crystalline,and nearly opaque. It is easily pyr olzed(decomposed by heat in the absence of oxygen),but coal as such does not exist in the vaporstate. Although coal in some ways resembles apolymer (Lucht and Peppas, 1981) and exhibitssome polymer type behavior, it is not a truepolymer since it has no known repetitivestructural unit. Even a single maceral is a

mixture of macro-mol ecules. Thus coalpresents several obstacles to the deploymento f modern chemi cal and physicalinstrumentat ion in the customary ways. Anddespite a great deal of research and some goodresults already accomplished, knowledge ofcoal structure at the atomic level remainsunsatisfactor y. Section III-C of the reportgives an assessment of cur rent prospects inthis area.

It does seem clear that there is nounique "structure of coal"; each seam and,probably, each sample is different in its

. compo s i t i on, str uc tur e, and proper ti es. Oneexpects that the macr omolecular entities incoal are composed of sub-units some of whichmay themselves be released more or less intactupon mild dissolution or other chemicaltreatment which only affects the weakest bondsor specific types of bonds iri the structure.In this view a credible aspiration for thedescription of coal at the atomic level wouldbe to develop a statistical char acterizationof:

~ molecular size distribution

assizes

and types of sub-units (includingsuch features as their aromaticityand chemical functionalit, ies)

~degree and types of crosslinking betweensub-un i ts

To be useful such a characterization mustspecify distributions and not merely averages,and the cross-linking information must includesome treatment of steric effects and threedimensional structure.

Such descriptions have been offered byFuchs and Sandhoff (1942), Given (1961), and

TABLE III-A-3. Chemical Composition and Physical Properties of Macerals of a High Volatile ABituminous Coal.

Composition, wt. %

Component S N 0Density

g/cm3% Reflectance

in oilRefractive

indexSurface

area(m 2/gram)

ExiniteVitriniteMicriniteFusinite

0.71.00.60.4

1.1

1.60.70.6

5.58.08.043

86.2 6.584.1 5.385.9 4.891.5 3.2

1.211.291.321.48

0.320.911.612.65

1.681.812.011.91

2.52. 1

4 69.8

Rev. Mod. Phys. , Vol. 53, Na. 4, Part I I, Qctcber 1981

Research on Coal

Miser (1978). Figure III-A-1 depicts Miser 'srepresentation, it is intended not to depictan actual coal stueture, but r ather as aheuristic aid to shov the main types ofst r uet ural units, functiona1 groups, andlinkages that might be pr esent in a typicalbit uminous coal . Undoubtedly an act ua1structure is three dimensional and variable.A three dimensional heuristic model veryclosely based on Miser's representation isshovn in Fi gure II I-A-2. To obtain this"ster ically permissible" space-fillingstructure it vas necessary to modify Miser 'sr epresentat, ion only slightly (Spiro, 1981).Of course representations of this type, ifthey vere to be extended to coals of otherr anks, vould have to be constr ueted ofcorrespondingly different sub-units asindicated in Figure III-A-3 (Mender, 1976) ~

Ver y eoneise discussions of coalchemistry and str ueture, in r elation to .coaltechnology may be found in Gorbaty et al, 1979and Gas Resear ch Institute, 1980. A viderrange of topics in coal science is discussedin Lovry, 1963; Elliot t, 1981; van Krevelenand Schuyer, 1957; van Krevelen, 1957; Cooper,1978; and Cooper and Petr akis, 1981.

4. Ph/slcSI ItfUccUF8

In terms of its phy si e al struetur e, coalhas been found to be a heterogenous solidcontaining capillaries, cr acks, and open andclosed pores (with and without, occludedmatter ) in at least, three broad r anges ofcharacteristic dimension. The pores have beencategorized as: macropore ( ~200 L), mesopore(20 —200 A), and micropore ( &20 A) (Grimes,1980) . Some coal samples exhibit specificinter nal surface ar ea as high as severalhundred meter /gram. (Results differ markedlyfr om one coal to another and also dependsigni fieantly on the measur ement techniquechosen. ) The origin or cause of thisporosity, although i t i s o f t echnological asveil as scientific interest, is not yet knovn(Gay, Nandi, and Malker, 1972).

The crack and pore str ueture is of greatpractical importance since it influences theaccess and egress of reagents and reactionproducts in various chemical processes and thedetailed mechanism of combustion of smallpar ticles (Larsen, 1981). (See Section IV B)In fact it appears that for many industrialprocesses these factors and the pr esence

H H

I l

0 SH ~ //

C

HHH.

H O NH2

H C2H5

H~ A~ r~H C~& —0 —~ Sr

c=o S

2H $

H-C-HHHACH)

H—3 —C-

HON

HCH f 0H

H-C-HI

H-C-H

S g ggHI

I

Or y~

H H

H

J„HH

I $ /H H H H

HP H2OH

I

FIG. IIIA-l. Schematic representation of structural groups and connecting bridges in bituminous coal(Wiser, 1978).


S20 APS Study Group on Coal Utilization and Synthetic Fuel Production

4 1 RSI

ie I I 44

, 4

sI~ I I I 4 ls

si

,". ')'.4

4 .. ~I:.. i, 4':::ea~ .;: ' .' '::.:".:::S

4

k . . .: , ,::",",:. S

,R

Ai3lj'jj, :::~.:-':..;8'I~

el I IRII

4—

I I

II I 4)

Ils.144

I 4 4IMI

ill, R I =

..='31!!!" R I ISRR IS 'I RK

44SSS i I BSII

i[ Ii ~ 44&

Fa

I' IG. IIIA-2. Space

R 4 IISSSSlia

jii iw.-,'= 4 I RS.

SMS SRRSR RSI Rg

SR 1M SIRE4 444s 4$%81P

~~~ IN%@!Nil RRI 4IRI mii iiis

l

ISR 4 RRI IIIII.

4 'SM4 RIS e RR 114'

4S:„="BII4SR~RSR~ISS~RR RMSMSM iiiiiiiwse 444IKsseIM~ ~

Sl R'4 I 4MIIi

I!4 4 lllss\ 114 4SS 4

Is s44 44

Ql II ~ 'i If'

''Rl IS 4 M :II~I I SM 4 I ISISIIIRI 4 4 M IRS i ice

filling model of Wiser's representation of bituminous coal (Spiro, I 98 I ).

Ig l I'sss=s4s4s

4

~' SI I I

LSSSSS44. S

Bonkof

cool

YN~ek~rN~~l ~rr I-Q-

Anthraci teSubbI tumlnous

~rQ ~ra~~~PL~I

Lo w-volati Ie

bituminous

Hi gh-vole ti le

bi tuminaus

CHO0

!

CHp

CH~

!

CH-

COOH

Q

HQ~~~ CHQH

!

Q

CHaCCH, QH

~CHp

LigniteRonk

ofcool

OH

FIG. IIIA-3. Representative partial structures of different ranks of coal (Wender, 1976).

Rev. Mod. Phys. , Vol. 53, No. 4, Part I l, October 198'1

Research on Coal

occluded moisture, gases, and minerals may-bemor e important determinant s than the intr insie

. chemistry. Porosity is especially amenable toinvestigation by physics techniques and isdiscussed in Sect, ion III B.

Stability

As taken from the mine, many coals,especially those of lower rank, are unstablein the presence of air . Their qualities asfuel or as input to chemical processes arechanged by slow oxidation and irreversiblewater migration; some even undergo spontaneouscombustion initiated by heat of adsorption andother sur face reactions. These phenomena areo f some consequence for the overall design ofcoal utilization systems, and they must bevery carefully controlled in any at tempt toprovide standard samples for laboratoryresearch. The matter of standard samples isdiscussed in Section III C, and is the subjectof one of our Principal Recommendations asdiscussed in Section I.

References

Beimann, W. , H. S. Auvil, and D. C. Coleman,1963, Chapter ll in Chemistry of CoalUtilization, (Supplementary Volume)edited by H. H. Lowry, John Wiley andSons, New York.

Berkowi tz, N. , 1980, Introduct ion to CoalTechnology, Academic Press.

Cooper, B. R. , 1979, Edit, or, ScientificProblems of Coal Uti1 i zation, TeehniealInformation Center, U. S. Dept. of Energy.(Available as CONF-770509 from NationalTechnical In formation Service,Springfield, VA 22161).

Cooper, B. R. and L. Petrakis, 1991,Editors, Chemistry and Physics of CoalUt, il i zation —1990, Amer ican Inst, it, ute ofPhysics Con ference Proceedings No. 70.

Dyrkae z, Gar y and E. P. Horwi t z, Fuel, inpress.

Elliott, M. A. , 1981, Editor, Chemistr y ofCoal Util i zat ion, Second Supplement,Wiley, New Yor k.

Flaig, W. , 1966, in Coal Science, P. Given,ed. , (Am. Chem. Soc. , Wash. D. C. ) p. 58.

Francis, Nil fr id, 1954, "Coal, Its Formationand Composition", Edward Arnold(Publishers) Ltd. p. l.

Fuehs, W. and A. G. Sandho ff, 1942, Ind.Eng. Chem. , 34, 567.

Gan, H. , S. P. Nandi, and P. L. Walker Jr. ,1 972, Fuel 51, 272.

Gas Research Institute, 1980, Basic CoalScience Pro ject Ad vi sory CommitteeReport, Gas Research Institute, Chi ca go,IL.

Given, P. H. , 1961, Fuel 39, 147, 1960 and40

Given, P. H. , D. C. Cronauer, W. Spackman,H. L. Lovell, A. Davis and B. Biswas,1975, Fuel 54, No. 1, 40-49.

Goloushin, N. S. , 1959, J. Appl. Chem. USSR

32, 1969.Gor baty, M. L. , F. J. Wright, R. K. Lyon, R.

B. Long, R. H. Sehlosberg, Z. Baset, R.Liott, a, B. S. Silbernagel, and D. N.Neskora, Science 206, 1029.

Grimes, W. R. , 1990, "The Physical Structureo f Coal" in Advances in Coal Chemistry,Vol. 1, Chapter 2, Academic Press.

Karr, Clarence Jr . , 1979, 1979, AnalyticalMethods for Coal and Coal Products, 3 v. ,Academic Press.

Kroger, Car 1, 1957, Gluckauf, Vol 93, p.122.

Larsen, J. W. , 1991, pp. 1-27 in Chemi stryand Physics of Coal Uti&ization -1980,edited by B. R. Cooper and L. Petrakis,American Institute of Physics ProceedingsNo. 70, New York.

Lowry, H. H. , 1963, Editor, Chemistry ofCoal Utilization, Supplementar y Volume,Wiley, New York.

Lucht, L. N. , and N. A. Peppas, 1981, pp.29-49 in Chemistry and Physics of CoalUtilization -1980, edited by B. R. Cooperand L. Petrakis, American Institute ofPhysics Proceedings No. 70, New York.

Neavel, R. C. , 1979, pp. 77-79 in Scienti ficProblems of Coal Utilization, edited byB. R. Cooper, U. S. Dept. o f EnergySymposium Series.

Parks, B. C. , 1968, Chapter 1 in Chemistryo f Coal Util i zation, edited by H. H.Lowry, John Wiley apd Sons, New York.

Petrakis, L. and D. W. Grandy, 1990, Journalof Chemical Educ. , 57 $(10, 683.

Spiro, C. L. , 1981, to appear in Fuel.Staeh, E. , M. —Th . Mackowsky, N.

Teichsmuller, G. H. Taylor, D. Chandr a,and R. Teichsmuller, 1975, Textbook ofCoal Petrography, English trans. by D. G.Murchison et al. , (Gebruder Borntraeger,Berlin, Stuttgart).

Stopes, N. D. , 1919, Proc. Roy. Soc.(London) B90, 470-487.

Thiessen, R. , 1920, "Coal Age" 18,


S22 APS Study Group on Goal Utilization and Synthetic Fuel Production

1183-1189, 1223-1228 and 1275-1279.Thiessen, R. and M. Francis, 1929, U. S. Bur.

Nines Tech. Paper 446, 27pp.Thiessen, R. , G. C. Spr unk and H. J.

O'Donnell, 1938, U. S. Bur. Mines I.C.7021, 1939; Fuel 17, 307-315.

Thompson, R. R. , 1 981, pp. 09-65 inChemistry and Physics of Coal Utilization-1980, edited by B. R. Cooper and L.Petrakis, American In sti tute o f PhysicsProceedings No. 70, New York.

van Kr evelen, D. M. and J. Schuyer, 1957,

Coal Science, Elsevier, est erdam.van Kr evelen, D. M. , 1957,

Coal-Typography-Chemistry-Physics-Constitution.

Mender, I. , 1976, Catalytic Synthesis ofChemicals from Coal, Catalysi s ReviewsScience and Engineering, 14(l ): 101, Fig.2.

Miser, M. H. , 1978, pp. 219-236 inScienti fic Probl ems o f Coal Utili zation,edited by B. R. Cooper, U. S. Dept. o fEnergy Symposium Ser ies.

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 'l984

Research on Coal S23

III. B. PORE STRUCTURE OF COAL

1. Introduction

A number of impor tant physical andchemical properties of coal are intimatelyconnected with, and sometimes uniquelyin fluenced by, the pore structur e of coal(Walker, 1980; Grimes, 1980). Among these areeffects relating to adsorption, absorption,diffusion, mass transpor t, catalytic andcombustion processes. Ef fective coalutili zation requires a much improvedunderstanding both in the physical descriptionof the pore structure, and, in the way thatpore structure influences how chemicalreactions occur for coal.

As stated by Grimes, 1980, "...details ofthe pore structure — influence the behavior ofcoal more directly than does virtually anyother property. " This is r eadily appreciatedwith the rea1ization that it is only throughthe surfaces, especially the internal poresurfaces (the integrated value of which canr ange from 10's to over 200 m Ig) that thecoal comes into cont act with the externalenvironment.

Certain coals behave like deformableviscoelastic colloids. Because of this, thephysical character i sties of coal, andtherefore its (slit-shaped) pores, will changeand alter significantly as a function oftemperature, pressure, and the intrusion ofmolecular species'. For example, the wellknown swelling of bituminous coals in thepresence of water is an illustration of this.

Below we give some examples in which thepores in coal play an important role in itsutilization.

Coal Gati f'ication. The reactivity in coalgasi f ication processes depends on the extentof the microporosity in coal. However, at therequired elevated temperature and during theheating process, the microporosity of the coalundergoes drastic changes, with the nature ofthe changes significantly depending on thetype of coal used. For sub-bituminous coaland lignites the elevated temper atur e r esultsin the removal of volatiles in coal andconsequently both an increase in average poresize of the open rnicroporosity and the openingup of pr eviously sealed pores. For mostbituminous coals, on the other hand, much ofthe open porosi t y i s lost when the coal goesthrough a plastic phase upon heating t o

gasification temperatures. Treatment of coal(e.g. , addition of oxygen) during the heatingprocess can significantly alter the porosity,and therefore can increase r eactivity in thecoal gasification process.

Coal Lique f'act ien . In some directliquefaction processes, hydrogen is. added tomolecular fragments formed by ther mal bondbreaking, which occurs at temperatures ofabout 400 C, through the use of a donorsolvent. The microporosity in coal enablesthe donor solvent to come into close contactwith, and donat;e hydrogen to, the free bondsof the molecular fragments. This inbibitionthrough the micr opores (of diameter 20 or 30Aand less) must occur sufficiently rapidly withincreasing temperature so that the pi ocess i scomplete before the thermal plasiticity of thecoal leads to the closing of the micropores.

Coal/Gas Inter face.'One of the prominentcoal/gas interfacial effects is the productionand retention of methane gas in coal seams.The amount retained and its depen. dence ontemperature and pressure will be relevant bothin the safety aspect of coal mining operationsand in the prospects for methane fuel gasextraction in undergr ound mines.

Ceal/Liquid Inter f'ace . In the importantarea of long distance coal transport, a recentdevelopment i s the coal slurry pipeline. Inthe case of coal/water slurries, the coalpar t icles are suspended in water to make them"purnpable", but water inside coal pores willdecrease the ef ficiency of coal combustion.Slurr ies -involving other fluids, such asmethanol, liquid carbon dioxide, or oil, maybe of interest under certain circumstances.Another area where the coil/liquid interfacecomes into play is in wetting and floation, ofimportance in coal preparation and cleaning(Taylor et al. , 1981). (See IVE for a fur therdiscussion of coal/liquid inter f acephenomena. )

2. Research areas

a. Quantitative characterization of coal pore sizedistribution

There has been considerable research inthe charactertization of the internal (i.e.pore } structure in coals. The prevalent

Rev. IVlod. Phys. , Vol. 53, No. 4, Part I I, October 1981


methods used are "Molecular Methods" -- theuse of molecules as probes. This includes thetechniques of unsteady-state di ff us i on o fgases, gas adsorption, and gas and liquiddisplacement (Walker, 1980; Grimes, 1980) .These techniques do not yield unique values ofsurface area and pore vol'umes, but depend onthe molecular probes used, e.g. , pore volumesusing helium and mercur y as displacementfluids are quite dif ferent, presumably becausemercury even at a pressure of 60 kbar s doeynot penetrate pores of less than 200Adiameter. (In fact, even different gases, saycarbon dioxide and nitrogen, give differentpore volumes. ) The origin of this nonuniquemeasurement of internal sur fac."e ar ea arisesfrom a combination of several factors: thesi ze of molecular probe, the influence of poresurfaces on the behavior' and volume of theprobing molecule, and the influence of theprobing molecules on the pore volume it issupposed to measure. In addi tion, mo3. ecularprobes can only measure the volume of poresthat are part o f a network that communicatesto the external sur face. Molecul ar probetechniques generally yield integrated internalpore sur f ace area. By measuring, withdifferent probes, with di f fering abilities topenetrate pores of varying si ze, one candevelop models of the pore size distribution(Walker, 1980; Grimes, 1 980 }.

The influence on the pore structure ofmolecular probes can be avoided by the use o fnonintrusive physical probes, which have otheradvantages as well. Small angle x-rayscatter ing (SAXS ) (Chi. che et al . , 1 965;Kalliat et al. , 1981; Lin et al. , 1978;Schelton and Hendricks, 1978; Schmidt et al. ,1981) is such a technique, and shows promisefor determing the size distribution of thepores. Clearly, the SAXS methods is capableo f probing all pores, including sealed as wellas open (i.e. part of network extending toouter surface) pores. A great advantage ofSA XS i s the abi 1 i ty t o per form suchmeasurements und r di f ferent temperature andpressure conditions, thus making it possibleto folio~ changes in the pore structuredependent on these conditions. Theavailability of powerful synchrotron radiationX-ray sources, such as the forthcomingNational Synchrotron Light-Source Facility atBrookhaven National Laboratory, enhances thepossibilities for success wi th SAXS on coal,because of the gr eat intensity, definedpol ar i za t ion, and wa vel eng th t~anab i l i ty o f theX-r ays.

Small angle neutron scatter ing (SANS)might be used to complement SAXS studies ofcoal pore structure. Since the wavelengths ofthe neutrons ar e longer than those of X-rays,SANS would have an advantage in resolvingbehavior of macropores (those with diametersof hundreds of l}. The great penetratingpower of neutrons, allowing the use of largersamples, might also be an advantg ge. Theability of neutrons to look at all elementsand distinguish between them (includingchanges of magnitude and sign of scatteringlength through the use of isotopes) could beused to examine behavior of molecular speciesresident in the pores.

Determination of the si ze distributionfunction for coal pores by SAXS or SANS wouldbenefit from improvement in the data analysistechnique. This analysis (Chiche et al. ,1965; Lin et al. , 1978; Schmidt et al. , 1981)i s based on theor i e s o f smal 1 angle X-rayscattering from two phase (i.e. in our case,coal and voids) systems. The experimentsmeasure the intensi ty I as a function ofscattering vector K. For spher ical scatters(in our case, spher ical voids), this intensityis related to the size distr ibution functionN(R) through the integral relationship

&&&& = cfN&»&v &R » («R&s~R (III-B-1)

where V(H} is the volume of the scatterer andF(KR) is a shape function. Thus someapproximations must be made to extract N(R)from this integral r elationship. ~ Thetechniques used currently require assuming theshape of the pores, and this is taken asspherical. Thus one surr enders thepossibility of extracting information aboutthe shape of the pores. Since, in fact, poresar e thought to be more slitlike, this isinformation of interest, . (Information aboutthe pore shape can, however, be obtained fromelectron microscopy experiments as long as thepores are reasonably large, i.e. greater thana few tens of Lngstroms in mean diameter. )

Even with the assumption of spherical poreshape, both techniques curr ently used toextract N(H) from knowledge of I(K), throughthe relationship given by Eq. (IIIB—1), haved i f f i cu1 t i e s . One technique involves

~H. R. Child of Oak Ridge National Laboratoryhas pr ovided detailed information on currentanalysis technique to obtain N(R) from smallangle scattering experiments.

Rev. IVlod. Phys. , Vol. 53, No. 4, Part I I, October 198't


numer ical inversion of the I(K) vs K curve toget N(R) vs. R. For spherical pore shape,this involves the Nellin transform equation.The numerical integrations involved requireknowing I(K) to large K values, and ar emeaningful only if I(K) shows Poqod limitingbehavior (Lin et al. , 1978), wl. /K, for largeK. The other technique currently usedinvolves a least mean squares fit of the I(K)vs K data to a sum of coef ficients times shapefunction values,

N21(K) = Z A (R)F (KR) (III-B-2)

b. Study of molecular species resident in and movingthrough pore structure

As already discussed, reactivity (in coalgasification and liquefaction processes and incombustion) depends to a great extent on thetransport of gas and liquid (e.g. , donorsolvent) in the network of' .coal por es. Thisinvolves the interaction of the molecularspecies of interest with the pores (Simons andFinson, 1979). straw and Silbernagel, 1981,have pointed out the value of combined

where the assumption of spher ical shape givesF (KR). The A (R) from the fittings are thentPie values of the size distribution functionN(R) at certain R=R , i.e. A (R) = N(R ). If

n nthe fit is done dir ectly, the N(R ) obtainedndo not give reasonable behavior, because of

the limited number s o f A ' s used and thefinite range of K measured. weighting of thevar ious ter ms in the expansion of Eq. (1118-2)and other constraints are used to "guide" thefitting procedure towards a reasonable result.

Finally, with regard to quan ti tati vechar aeter i zation of the pore si zedistribution, we should again emphasize thedesirability of doing this under dynamicconditions. That is, we want to be able tofollow changes wi th time under realisticconditions of temper atur e and pre ssur e a sex i st in gasi f ication and liquefactionconditions. Recently Naylotte et al. , 1981,have shown that X-ray computed tomography,aided in some circumst ances by xenon actingboth as a probe and contrast agent, can beused to follow changes in the average porestructure of coals and reacting coal charsin-situ.

stud i e s, coordinating result s frommeasurements determining long-time,macroscopic average behavior of a fluidresident in the por e structure with resultsfrom other measurements providing a picture ofmotion on thy molecular scale at short times(10 to 10 seconds) . In their own studyof water in coal (Mraw and Silbernagle, 1981),heat capacity measur ements (Mraw andNaas-O'Rourke, 1979) have been used to givethe time-aver aged behavior, while nuclearmagnet, ic resonance (NMR) (Silbernagel et al. ,1.979; Mhittingham and Silbernagel) has beenused to study dynamic effects.

Near 120K and again at room temperature,the heat capacity of water in coal pores issimilar to that of bulk water; but, atintermediate temperatures where bulk water hasthe first-order freezing transition, there aremarked changes (Mr aw and Silbernagel, 1 981) .For low concentr ations (0.16 g H20/g coal) ofwater, the heat capacit, y (i.e. incrementaldifference from dry coal) gives no evidence ofa phase transition; while at hi gherconcentration (0. 32 or 0. 37 g H 0/g coal) abroad heat capacity maximum is observed, i.e.still no evidence of a latent heat. Mraw andSi1bernagel, 1981, have analyzed thi s behaviorwith a model that suggests that approximatelytwo monolayer s o f sur face adsorbed waterinside coal por es become "non-freezing" (or"already frozen" ).

sideline measur ements o f the shape andwidth of the NMB absorption, and transient NHR

measurements o f the spin-lattice r elaxationtime (T ), probe molecular motions on the10 -1 second time scale. For thewater-in-coal studies, use o f deuteratedsamples, i.e. 020, were useful both todecrease background signal from the organicmater ial, and to provide complementarypictures of molecular motion (because thedeutron inter action with its sur r oundings isof a different, much more local nature, thanthat of a pr oton) . This meant that D 0observations provided an opportunity to studythe molecular dyn ami c s o f water moleculesresident in coal por es wi thou t theinterference from paramagnetic impurities thatoccurred for the pr otons in H20. Such NMR

studies showed a continuous decrease in rateof the molecul ar motion o f water in coal asthe temperature is lowered, with activationener gies in the range of 0.15-0.20 e V,comparable to those in bulk water.

Besides the need for inno va t i v eexper imental techniques, the study of the



behavior of molecular species resident in, ormoving thr ough, coal may also benefit frommodern theoret ical approaches developed inother areas of physics. To whatever degreethe interconnecting pores wi thin coalrepresent a random networ k of transport,channels, st, ati st ical mechanical andprobabilistic approaches may prove to beuseful in the description of the mass transferprocess in pores. The percolation theory(Kir kpatr ick, 1973; Balian, 1979) approach,for example, could be useful here, since ithas proven sueeessful in solid state physicsand other applicat ions. It is also ofinterest that the effective medium theories ofinhomogeneous materials developed within thecontext of condensed matter physics have beensuccessfully applied to the study of theelectr ical (Sen et al. , 1981; Mendelsohn andCohen, 1981) and mechanical ( Johnson and Sen,1981) properties of the porous, sedimentaryrocks of impor tance in oil and gas reservoirs.Indeed it has been sho~n that, the Biot slowwave velocity within a fluid containing aporous medium is the same as the velocity offourth sound in super fluid helium containedwithin the same medium, and the two velocitiesare related to the d.c. conductivity when thefluid is an electrolyte ( Johnson and Sen,1981). Monte Carlo studies offer yet anotherappr oach that may prove to be useful. Thesemacr oscopic appr oache s could reveal valuable,gener al, systematic behavior in coal tr anspor tthat may otherwise be masked in studies at amicroscopic level. As another example ofmodern theoretical appr oaehes that might becarried over to studies of molecular speciesresident, in the coal pore structure, the ideasof de Gennes, 197$, on inter action of confinedpolymer solutions with walls might berelevant.

3. Recommendations

~Me str ongly urge further work onquantitative characterizat, ion ofcoal pore st rue tur e, especiallyincluding the development ofreliable non-intrusive physicalprobes such as SAXS and SANS, andwith emphasis on measurement, underdynamic conditions (i.e. followingchanges wi th time under realisticoper ating condit ions of temper atureand pressure. )~Me- recommend emphasis on studies ofinteraction o f resident molecular

species with the pore surfaces, andof transport of such species throughthe pore structure. Such studiescan provide information o f greatpractical use. Research programscoordinating measurements yieldinglong-time average behavior wi tho the r me a sur ements yi eldingshort-time dynamic behavior mayprove especially valuable.~Greater attention should be paid tothe possible rel evance o fpercolation theory and otherstatistical approaches to the studyof molecular transport pr oeesse sthrough the coal pore structure.

4. References

Balian, R, 1979, Lectures in Ill-CondensedMatter, North Holland.

Chiche, P. , S. Durif, and S. Preger main,196$, Fuel 44,

Cohen, M. H. , 1981, to be published.deGennes, P. G. , 1979, Scaling Concepts in

Polymer Physics, Cornel 1 Un i ver si tyPress; especially see pp. 88-97.

Grimes, M. R. , 1980, "The Physical Structureo f Coal", in Advances in Coal Chemistry,Vol. 1, Chapter 2, Academic Pr ess.

Johnson, D. L. and P. N. Sen, 1981, to bepublished.

Ka1 1 iat, M. , C. Y. Kwak, and P. M. Schm i d t,1991, in New Approaches in Coal Chemistry(B. D. Blaustein, editor ), Amer. Chem.Soe. Symposium, Pi t t sburgh, PA, Nov. ,1980, published by Amer . Chem. Soc. ,Washington (in press).

Kirkpatrick, S. , 1973, Rev. Mod. Phys.$70.

Lin, J.S. , R. M. Hendrick, L. A. Harris, andC. S. Yust, 1978, J. Appl. Cryst. 11,621.

Maylotte, D. H. , R. L. St. Peters, P. G.Kosky, E. J. Lamby and C. Spiro, 1981,"Use o f X-Ray Computerized Tomography inCoal Characterizat ion", presented at theAmerican Chemical Society Meeting,Atlanta, March, 1981.

Mendelsohn, K. and M. H. Cohen, 1981, toappear in Geophysics.

Mr aw, S. C. , and D. F. Naas-0 'Rourke, 1979,Science 205, 901.

Mr aw, S. C. , and B. G. Si lbernagel, 1981,pp. 332-343 in Chemistry and Physics ofCoal Utilization — 1980, AIP Conference


Research on Coal 827

Proceedings No. 70 (edited by B. R.Cooper and L. Petr aki s, New York.

Schelton, J. and R. M. Hendricks, 1978, J.Appl. Cr yst .

Schmidt, P. M. , N. Kalliat, and C. Y. Krak,1981, in Chemistry and Physics of CoalUtil i zation — 1980, AIP Confer enc eProceedings Series No. 70 (edited by B.R. Cooper and L. Petrakis, New York. )

Sen, P. N. , C. Scala, and N. H. Cohen, 1981,Geophysics Q6, 781.

Silbernagel, B. G. , L. B. Ebert, R. H.Schlosberg and R. B. Long, 1981, ACSAdvances in Chemistry 192, 23.

Simons, E. and J. Finson, 1979, Comb. Sci.Tech ~ 19, 217; 19, 227.

Taylor, S. R. , K. J. Miller, and A. J.

Deur brouck, 1981, pp. $44-356 inChemistry and Physics of Coal Utilization-1980, AI P Conference Pr oceedings SeriesNo. 70 (edited by B. R ~ Cooper and L.Petr akis, New York) .

Walker, P. L. , 1980, "Microporosity in CoalIt s Character i zation and Its

Implications for Coal Utili zation",Invited Talk, Royal Society Meeting.

Mhittingham, N. S. and B. G. Silbernagel,1978, "NNR Techniques for Studying IonicDi f fusion", in So 1 id Electrolytes.General Principles, Character ization,Materials, Applications, edited by P.Hagenmuller and M. van Gool, AcademicPress, Nex York.

Rev. Mod. Phys. , Yol. 63, No. 4, Part I I, October 1981


III. C. ANALYTICAL TECHNIQUES

1. Introduction

This section is intended- to pinpoint someselected analytical techniques which could bepr ofitably applied to coal utilization andsynthetic fuel production. A comprehensiveand recent discussion of current analyticaltechniques may be found in the three volumework of Karr (1978., 1979) . He have selectedtechniques which show the most promise forprovidinp, si gni f icant in format ion on keyproblems or which o ffer new approaches tothese pr oblem ar eas. Some are presentlyutilized by the physics community and some areunder extensive development but have not beenapplied to coal. The complexity of coal an'dits derivatives imposes unusual demands onanalytical techniques.

The development of analytical techniqueswould be greatly assisted if properly selectedand preserved reference coals were widelyavailable to analysts and basic coal researchgroups. There have been only a few samplebanks in the United States in recent decadeswher e a scientist or analyst could obtain acharacter i zed sampl e. Pennsylvania StateUniversity in coordination with the Of f ice o fCoal Research (later the Energy Research andDevelopment Administration and presently theDepartment of Energy) has had selected coalsamples available. Other repositories are theNational Bureau of Standards, State GeologicalSur vey Labor armories and several pr ivatecorporations (whose samples have not beenavailable to all interested researchers), e.g.Exxon, General Electric, etc. Analytical andbasic r esearch r esults would be significantlymore valuable if a limited number of carefullyselected, character i zed and stored coals wer estudied so that r esults from different studiescould be directly compared.

2. Selected problem areas

a. E lemental composition

Of the major elements in coal, accurate(t, wo significant figures) and directdetermination of oxygen is most lacking andconsidered to be most important for theunderstanding of chemical structure. A rapidsimultaneous d et ermi nation o f t+e i mpor'tantnitrogen and sulfur heter oatom content, aswell as trace element analysis, would be

important for control of processes andpollution. Neut, ron acti vat ion, var ious X-ray,electron, proton or alpha particle probes, andlaser diagnostic methods appear to have somepromise as analytical techniques. litany of thestandard techniques (Kar r, 1978,1979) forelemental composition (e.g. ultimate analysis)will continue to find useful application.

b. Chemical structure

As indicated earlier, there is no uniquechemical str ucture of coal. Several modelsha v e been pr oposed whi ch appr oach thi squestion from different and useful vantagepoints; e.g. , treating coal as a cross-linkedmac r omo 1 ec u1 ar gel, an amorphoussemiconductor, a conglomer ate or ganicstructural unit, or a collection of micellesembedded in a plastic matrix. Di f ferentmodels make di f ferent assumptions concerningthe carbon "backbone, " and the nature anddegree of linkages. Information concerningthe distribution and local environments of theheteroatoms would be useful for a critical

omparison with model str uctur es. Inaddition, knowledge of the distributions ofsome of the heteroatoms (N and S) mightfacilitate their removal dur ing coalpr ocessing to circumvent problems of catalyticpoi soning, product inst, abilities, andpollution. A recent review of this subjecthas been given by Davidson (1980). A numberof recent studies are also to be found in theGerman and Japanese literature -(Ting, 1979;Hombach et al. , 1979; Hombach, l 980; Maekawaet al. , 1977). Refinements and extensions ofMAR and various double resonance techniquesseem to be promising approaches. Introductionof NHH-active species may prove to be a usefulsupplementary tool for these studies.

One of the structur al parameters ofinterest in coal i s the molecular weightdistr ibution of the cluster s betweencross-links of the coal structure (Lucht andPeppas, 1981). A major difficulty inobtaining this information is the -pr oblem ofbreaking the cross-links and separating thef ragments under conditions mild enough toper mit interpreting the results asrepresentative of the solid. A furtherproblem is that many o f the presentexperimental measurements (e.g. gel permeationchromatogr aphy and vapor phase osomometry)deduce number-average molecular weights, whichtend to give excessive weight to low molecular

Rev. IVlod. Phys. , Vol. 63, No. 4, Part II, October 1981

S30 APS Study G'roup on Goal Utilization and Synthetic Fuel Production

weight fractions. Various techniques of massspectrometry (e.g. plasma desorption and fieldioni zation) should yield distributions ofmolecular we i ghts . In t e r pretat ion o f theexperirnen tal r esul ts al so depends onassumptions about the shapes and structur s ofthe fragments. Development of more directanalytical techni ques would be useful inaddressing this important structural problem.Sma 11 angle neut ron scat ter ing i s apossibility.

c. Physical structure

The pore struct, ure of coals af fectsaccess and egress of reactants and products,sur f ace we tt ing, and bulk mechanicalproperties. The classic adsorptiondesorption measurements (Karr, 1978, 1979)yield surface ar ea estimates that dependstrongly on the nature of the probingrnolecules and on the experimental conditions.Physical techniques, such as low angle x-rayand neutron scatt ering, can also indicatesurface area and pore size distributions.Re a sonable correlation between the t woapproaches has been observed. Furtherrefinement of the physical techniques i sdesirable. (This is discussed in Section III8) It may be useful to introduce materialsinto the pores which would alter thescattering contrast. A more directvisualization of sur face pores and -- byremoval of surface layers, -- interior poresas well, may be achieved by scanning electronmicroscopy.

Different models of coal struetur e implydif ferent mechanical properties of coal. Forexample, the entropy change with extension canbe used to discriminate among various polymermodels.

d. Surface characterization

The techniques for characterizing smallarea, single crystal catalytic sur faces havebeen well developed for studies in catalysis.(Somorjai, 1979; Goodman et al. , 1980) SeeSection IV F for further details. Some of thesame experimental techniques may also provedirectly applicable to coal and coalderivatives. Among the most promising x-rayand electron probe techniques are electronmicroscopy, energy dispersive x-ray analysis,extended X-ray absorption f ine structure(EXAFS) and its recently developed surface

var i ati on, SEXAFS. In addition,photoelectron-, Auger —, and hi ph r esolutionelectron energy loss spectroscopy could beuseful for studying surfaces and surfacelayers. Further discussion of the utilizationof EXAFS and SEXAFS is found in Section IV Fof this report. In the case of coal, it maybe useful to supplement these techniques withthe use of heavy atom labelling, oxidation andreduction, or sit, e-specific reactions. Sincethese techniques are inherently associatedwith a high vacuum environment, it will benecessary to cool the samples at, least toliquid ni trogen temperature, to preventoutgassing and loss of volatiles.

e. Physical and chemical transformations

Many of the t chniques mentioned abovemay be used to study the physical and chemicaltransformations of coal which take place inco a 1 u ti 1 i zation and synthet, ic fuelproduction. The most commonly accepted modelfor coal behavior during pyrolysis is that ofthermolyt. ic bond cleavage with the productionof free radicals (Petrakis and Grandy, 198la).ESH would appear to be a natural technique forinvestigating such phenomena. One dif ficultyin applying this technique, however, is thepresence of free radicals in native coal evenat r oom temperature, whose si gnal masks thesignal from radicals newly created dur ingprocessing. ESR studies to date have shownwide variations in radical for mation atelevated temperatures for di ffer ent maeeralsand -- in the case of vitrinite -- substantialincr eases wi th temperature. Thesemeasurements can be extended to typicalprocess temperatures and pr essur es, and ajudgment concerning free radical, or non-freeradical, involvement in liquefactionmechanisms may be possible (Petrakis andGr andy, 198lb ) .

At elevated temperatures bituminous coalsshow a plastic (mesophase) behavior. Thisbehavior is important in the preparation ofrnetallurgieal eok . Its swelling determinesthe balance between increase in porosity anddecrease in str ength. An understanding ofthe physical and chemical changes in thisphase is also important in other coalprocesses. Experimental techniques which haverecently been util i zed to study thesestructural changes are Transmi ssion ElectronMicroscopy (TEN), modified to include a hot


Research on Coal

stage (Thompson, 1981), and x-r ay computer izedtomography during pyrolysis (Naylotte, 1981).

'awhile the study o f coal i t sel f presentssignificant problems resulting from theambiguity and non-uniformity of physical andchemical structures, one need not be over lyconcerned about this in the case ofcombustion, which is by far the most commonmode of util i zat ion. Coal may be burnedfully, as in a power plant, or partially, asin certain gasifiers. Even liquefaction orgasification of coal is generally aimed atproducing intermediate fuels which ultimatelywi 11 be burned. Thus in every coalutilization str ategy, combustion plays animportant role, sometimes more than once.

Under standing flames and flame phenomenarequires that we consider simultaneouslychemical kinetics, gas dynamics, radiativetransfer, acoustics, dif fusion, evapor ationand nucleation. Understanding combustion,then, is a br oad task involving manyspecialties. All, however, share the need fordevelopment of the best possible analyticalinstrumentation. There is need to combinehigh quality measurement of parameters such astemperatur e, velocity, and concentration ofchemical species with high spatial andpossibly temporal resolution.

Re cently, the clear leader i n

instrumentation for combustion r esearch hasbeen the laser. Many new investigativetechniques have been developed which are basedupon the utilization of laser light sources. 'laser induced fluorescence (Eckbreth et al. ,1978; Cat tolica, 1981), intracavity absorption(Brink, 1980), doppler velocimetry (Dyer andMiotze, 1979), Rayleigh scattering (Graham etal. , 1970), Raman spectroscopy (Eckbreth etal. , 1978; Lapp et al. , 1972), stimulatedRaman spectroscopy (Eckbreth et al , I978;Bloembergen, 1967), coherent anti-Stokes Raman

spectroscopy CARS (Eckbr eth et al. , 1978;Harvey and Nibler, 1978; Klick et al, 1981),Hie scattering (Hyatt, 1980), particulatevaporization (Harde sty, I 980), andoptogalvanic spectroscopy (Vidal, 1980). Moreare possible. For example, the study oftur bul ence may be aided by picosecondtomography (Goulard and Emmerman, 1980).

Because of the potential of lasertechniques for acquiring detailed information,the resurgence of interest in opticalspectroscopy of combustion species seems sureto continue. Even the basic thermodynamicproperties of simple molecules and radicalsparticipating in combustion are still poorly

understood. These simple molecular systemscan be studied in the laboratory byincreasingly sophisticated techniques of lasers pec tr oscopy, e lee tron — andmolecular -beam-spectroscopies.

In addition to labor atory studies offlames, it is possible to study separately the-individual processes which make up the wholecombustion process. Studies o f reactionkinetics are of fundamental importance her e,and various techniques such as shock tube(Li f shitz, in press), flash photolysis(Michael and Lee, 1979), fast flow reactor(Howard, 1979), and molecular beam (Sibener etal. , 1980) experiment, s are all potentiallyuseful for such studies.

Theor etical studies of. molecules andradicals found in combustion environments haveprogressed to the point of being predictive.Examples can be found in the lit erature ofstate-of-the-art, calculations (Harding, 1981)that correctly predict react, ion products, aslat, er confirmed by unambiguous experiments(Buss et al, 1981), in spite of preexisting,contr ary, data. Fur thermore, molecularcalculations can give a picture of reactionpathways in detail unobtainable by directmeasurement s. A reasonable goal of 'suchcalculations .is to unders t and the i nd i v i du a 1molecular processes that, are important inspeci f ic combustion reactions. Presently,methane combustion in air is attr actingtheor etical interest, (Dunning, 1981).

HD r e actors, oper ating at hightemperatures, have their own particular needs.Not, only ar e their materials problems severe,but even the gas phase in the HHD channel i sstill poorly char acterized. Both exper imentaland theoretical techniques need improvement.A par ticular example o f the lat ter, which hashad some success, is the program to establisha basic understanding of the conductivity ofthe gas in the channel. This program hasinvolved theory at a level as basic aselectron scattering of f atoms and molecules(Lane, 1980).

3. Selected techniques

a. Introduction

In choosing each of the following set ofexperimental techniques, we have avoideddiscussing a number of standard analyticaltechniques which are routinely used tochar acterize coal, coal pr oduct s, cat alysts,

Rev. IVlod. Phys. , Vol. 53, No. 4, Part ll, October 1981


and catalyst suppor ts. A note of cautionshould be introduced at the outset to theeffect that coal, being the ill-definedsubstance it is, a number of studies have beencarried out in the past on ".. .poorlyselected, poorly collect ed, poorly prepared,poorly preserved, and poorly character izedcoal samples. . ." (Neavel, 1979). In order toimprove upon this situation, the primaryrecommendation of the recently concluded BasicCoal Sciences Pro ject Report to the GasResearch Institute (Basic Coal SciencesProject, 1980} is the "establishment of afacility to collect, prepare, store,char acteri ze and distribute samples ofr eference coals in strict accordance withspecified protocols. " A comprehensive coalcharacterization program to evaluate coals asprocess feedstocks has been initiated (Neavel,1980) by EXXON Research and EngineeringCompany, Baytown, Texas. Even whenstandardized reference coals become available,however, it must be kept in mind that keystructural parameters may not be wellrepr esented by any particular selection ofreference coals, and indeed coal propertiesmay change with time, even if the samples arestored under carefully controlled conditions.In fact, the optimum storage conditionsthemselves have not yet been established.

A recent development has been theapplication of density gradient centrifugation(Dyrkacs et al. , 1981) for the separation ofcoal macerals. This is a much more selectivemethod than hand sorting, but the resultingsarnpl s are still limit d to mg quantities.There is continuing need to develop new (e, g.ultrasonic) techniques for physical separationof macerals.

One step which might prove important forelucidating key features of coal structurewould be to develop laboratory means forsimulating the natural processes that leadfrom biological precursor materials to coal(Adams, 1970; Teichmuller and Teichmuller,1966; Pan, 1967; Tsukashima, 1967) . Ifrealistic synthetic model substances could bef abr icated on a reasonable time scale,introduction of atomic labels and probes wouldbe facilitated and the appropriate analyticaltechniques could be applied in a systematicfashion.

b. Spin resonance

Magnetic resonance studies have beenamong the most successful in providing

structural information for whole coals,soluble fractions and liquefaction products(Karr 1978, 1979). For example, by measuringvarious resonance parameters the distributionsof molecular and electronic environments ofimpor tant atoms, such as the carbon"backbone, " and the heteroatoms, S, 0, and N

can be probed, and motion of molecules in theporous str uctur e of coal can be detected.More detailed structural information can beacquired from resonance studies of coalderived liquids and can be useful inmonitoring process conditions, stability ofstored liquid products, and the basicmechanisms involved in liquefaction.

unclear Nagnetic Resonance (1NR)—The nuclei which have been studied mosteffectively iq goal and coal derived liquidsare H and C. NNB investigations arecomplicated by the fact thy' three cIf coal'smost abundant nuclei, C, 0, and S havezero nuclear dipole moments. Seve~pl ot)ernucl ' which might be useful, namely N, 0,and S have non-zero nuclear quadrupolemoments, and have low abundance in coal. Evenfor the spin 1/2 nuclei, H and C, stronghomonuclear or heteronuc lear dipolarinteractions, chemical shift anisotropies, andlong spin-lattice relaxation times make itdifficult to obtain structural informationfrom chemical-shift, distributions in solidcoals. Furthermore, solvent extract studiesare limited by the poor solubility of coal,and the lack of knowledge as to how theextract structures are related to those of theparent coals. The multiple pulse techniquesor the selective decoupling techniques enhancethe spectral resolution by suppressing dipolarinteractions. The residual line broadening,which is mainly due to anisotropic chemicalshi f ts, can be f ur ther reduced bysuperposition of magic angle sample spinning(Lowe, 1 959; Bartuska et al, 1977; Naciel etal, 1981). The cross polarization techniquewill be cr ucial for the rare nuclei NiMR

detection (Pines et al, 1973; flaciel et al,1981). It will transfer polarization fromproton spins to the rare nuclei spins during atime which is much shorter than thespin-lattice relaxation times of rare nucleispins, allowing for orders of magnitude fasterdata acquisition. Additional improvement canbe achieved by making use of the multiplexadvantage of pulsed Four ier Transform (FT)



methods, or of cA mult, iple fast passage(correlation NMR). Although combinations ofthese sophisticated techniques have been quitesuccess ful in r eso 1ving interesting spectraldetail, and have yielded quantitat, ive data, insome cases resolution of the spectra or thesensitivity of the techniques needs to be1mprovedo

One of the most basic (hitherto veryelusive) pieces of information relative to thepostulated coal str uc tur al models is thecarbon aromaticity of the coal. Thetraditional view that coal is stronglyaromat, ic has been challenged periodically over

e past ten years, but the most dir ectC-NMR measurements confirm this view and are

in goqd agreement with indirect deductionsfrom H-NMR. Multiple-pulse H NHR with magicangle spinning currently has a resolution justless than that required to resolve directlythe aromatic and aliphatic pr'otons.

The development o f b et ter line-nar rowingtechniques is a matter of hi gh pr ior ity, andif successful, could lead to significantlybetter microscopic information regar dingchemical structures in coal. More thoroughanalysi s and improvements of the presenttechniques, perhaps in con junction with themagic angle spinning, will make importantcontributions in improving the spectralresolution of solid coal. For enhancement of~psitivitg especially for the low abundant

N and C nuclei, the dynamic nuclearpolarization technique has promise which canbe achieved through micro~ave saturation ofthe free radicals which are naturally abundant,in coal samples.

NMR studies of the important heteroatom,oxygen, have been restr icted primarily tohydroxyl groups. These have been determinedby a var iety of chemical modi fications,followed by litMR an~sis for the added group,e.g. using H or Si to classify the OH'sinto specific gr ups followingtr imethylsilylation, or F following fluor inesubstitution. It would also be of interest tostudy other oxygen functional groups, such asether linkages, which are thought to beimportant in the primary liquefaction step.

If line narrowing techniques could bedeveloped, or if iqgtope enriched samplescould be prepared, 0 NMR studies might beqyssible in the solid state. Such studies for

0, or t, he more difficult nuclei, sulfur andnitrogen, could- pr ovide speci f ic and usefulstr uctural information.

Mideline and transient NMR techniques can

provide useful information about the motion ofnuclei and the pore distribution and dipolarinteractions with neighboring nuclei, as wellas quadrupolar interactions with the localinhomogeneous electric fields ~

In fact,molecules labeled with D and F have beenused effectively (Silbernagle et al. , 1980) tostudy the temperature dependences of motionsof molecules introduced into the porestructures of whole coals. (See Section III 8for di scussion of NMR studies on water incoal. ) Mhen combined wi th ESR spin labelstudies, such NHR exper iments can yieldimpor tant information about the coal's porestructure, internal surface adsorption anddiffusion characteristics for the "guest"molecu les, and inter ac t ion s between themolecules and chemical functionalities in thecoal matrix. Similar measurements may also becar ried out on the pores of the importantzeolite catalysts (See Section IV F). Nuclearspin resonance has been used for some time asa simple process monitor for the quantity ofmoisture in coals.

K1ect ron Spin Resonance (KSR ) -- Thistechnique has been used extensively to studythe distribution of geologically stable freeradicals in native coals and the pyrolyticgener ation of additional free r adical. s inwhole coals, coal fractions and coal macerals(Petrakis and Grandy, 1981b; Retcofsky et al. ,1979; Petrakis and Grandy, 1978; Petrakis andGrandy, 198la) . The observed variations ing-value among whole coals correlate reasonablywell with the variations in atomic fraction ofoxygen and sulfur, weighted according to thespin-orbit coupl ing constants o f the s eheter oatoms (Retcofsky et al. , 1979). Thisinformation has been compared with measuredg-values for pur e compounds to suggest.possible model compounds related to coal(Petrakis and Grandy, 1978; Petr akis andGrandy, 1981a).

A striking r esult of ESR studies ofmacerals is the observed di fference of spinconcentrations by 0-5 orders of magnitude(Petrakis and Grandy, 1981a) (fusinitevitrinite & resinite) . Free rad ical,concentrations of all mace' als, mor cover,remain essentially unaf fected by temperatureup to QOOC. At higher temperatures, those ofvitrinites. incr ease rapidly with temper atureand approach typi c al fusinite spinconcentr ations (Petrakis and Grandy, 198la).Changes in g-value with temper ature for the



var ious macerals provide additional indir ectstructural information.

Sin~e it is thought that the initial freeradical formation upon heating may be quicklyf ollowed by the formation of polymericproducts, it, is useful to carry out ESRexperiments as a funct, ion of time. A recentpyr olysis experiment (Grandy and Petrakis, inpress) along these lines, carried out, on asmall sample heated in a microwave cavity,showed a linear increase in spin concentrationfor the first 2 to 5 minutes of theexperiment, fol lowed by a decreasepresumably confirming that polymeri zat ionoccur's ~

ESR studies of the properties of freeradicals as a function o f various solventtreatments appropriate for coal liquefactioni nd i cate that residual free radicalconcent rations in resinite and fusinitemacerals are relatively insensitive to t, hetypes of solvent, wher eas high volatilebituminous vitrinites have r adiealconcentrations which are solvent sensit ive(Petrakis and Grandy, 1981a).

In view of the probable importance of freeradical behavior in determining the rate ofcoal liquefaction, further ESR experiments ona variety of coal components and derivativesshould be carried out. The availability ofbetter separated maceral constituents (Dyrkacset, al. , 1981) could be quite helpful instudying the systematics; and currentlya va i 1 able instrumentation (Gr andy andPetrakis, in press) for ESR experiments at,pressures up to 2000 psi and temperatures to500 C will allow measurements of free radicalsduring coal liquefaction. Further short scale(minutes), time-dependent exper iments wouldalso be useful.

Nue1ear Quadrupole Resonance (NQR)—This technique has found limited applicationi n coal r esearch because of the lowcqnceqgration and/or isotopic abundances of

0, N, and S, and because of the lack ofsufficiently sophisticated equipment, (Marinoet al. , 1980) . It has become possible& j.nrecent years, however, to detect O

quadrupole resonance signals inpolycrystalline solids with the isotope in itsnatural abundance ( 0. 037'g ) using doubleresonance techniques (Smith, 1980).

In their review of potential applicationsof NQR to energy related problems, Marino et

al. (1980) suggest the possibility ofcontinuous measurement of the distribution ofsulfur between organic and inorganic phases incoal, since the nuclear quadrupolar couplinginteraction is very sensitive to the localenvir onment of the probe nucleus. Betterknowledge about the sulfur would help greatlyin understanding the structures in coal, indealing with the poisoning of coal processcatalysts, and in dealing with emissions ofsulfur oxides (SO ) .

Me think the possiblity of selectivelyintr oducing these quadrupolar isotopes in thecoal matrix or its pores as NQR probes of thechemical and physical structures of the coalshould also be explored.

Electron Nuclear Double Resonance(ENDOR)-- Potentially, this technique iscapable of providing direct, information aboutthe free radical intermediates thought to playkey roles in coal liquefaction mechanisms.Most previous st, udies (Schlick et al. , 1978;tliyagawa and Alexander, 1979; Niyagawa andAlexander, 1981; and Nest, and Cannon, 1 981)have resulted in single-line pr oton matrixENDOR resonance s true tur es which requireconsiderable analysis to ext ract, usefulstructural information. Quite differentenvironments for the unpair ed electrons havebeen demonstrated for the macerals vitrain,elarain and fusain from a bituminous coal, butbecause of the unresolved hyperfine structure,no direct conclusions could be drawnconcerning the details o f the structures. Incontrast, to this experience, ENDOR spectra ona vitrain-rich Pitt, sburgh coal, obtained inthe absence of air, showed well-resolvedhyperfine str ucture (Retcofsky et al. , 1979)which disappeared when air was admitted. Thisresult has since been reproduced in anotherlaboratory (Retcofsky, 1980) and suggests thatENDOR techniques could play a leading role inthe future in unraveling the details of thefree radicals in coal and their role inliquefaction processes.

c. E lectromagnetic spectroscopies

The most obvious feature about coal isthat it is a highly absorbing material: "blackas a lump of coal. " Hi gher rank coals ar eshiny, mainly as a result of the refr activeindex change at the sur face. A considerableamount of scatter ing from inhomogeneities


Research. on Coal S35

occurs at all wavelengths, and coals have ar emar kab le degree o f fluorescence, especiallythe lower rank coals that retain a resemblanceto their "humic matter" origins. Theseproperties can for m the basis for theinvestigation of coal and coal transfor mationsby electromagnetic radiation.

Direct imaging or diffraction techniques,depend upon the wavelength of the radiation forthe scale of the features in the material to beprobed. These techniques will naturally useshor t wavelengths to obser ve structure on anatomic scale. Longer wavelength techniques mayalso provide such information if probes, eithernatural or artificial, are used to label sitesin the mat erial. As a result, the ability toidentify natur ally occur ring labels orabsorption centers in coal, or to introduceinto a coal or coal -der ived mat er i al ar ti ficialmolecular sites or atoms that may be used toprovide information about t heir microscopicenvironments, will be extremely impor t ant inanalyzing the str uctur e of coal and how ittransforms under process conditions.

Bef leetanee and adsorption inUltraviolet (UV) ~ visible and nearInfrared (Ii) -- Reflectance is used toidentify and classi fy coal, both in thelaboratory and in industry. Petrographicsorting and typing of macerals is done by, orassisted by, r eflectance measurements {Ting,1978; Davis, 1978). Classic studies ofreflectance, and thus of the refractive indexchange at an air/coal or oil/coal inter face,have been used to gain insight, into suchfundamental properties as the average size ofthe aromatic units in coal. The index ofrefraction of vitrinite is approximately 1.5 to2. 0 in the visible (van Krevelen and Schuyer,1957). There is, however, little under standingof the basic reasons relating to the structureof the material for the functional form of theindex versus wavelength. This situation isonly made worse by lack of data outside of thevisible region.

A firm understanding of the nature of theabsorbing structures in coal would increase theuse fulness of UV, visible, and near IRspectroscopy as analytical techniques in coalchemistry and coal processing. In the visible,the imaginary component o f the index ofrefraction of vitrinite is approximately 0. 02to 0. 5 (van Krevelan and Schuyer, 1957). Ourunderstanding of absorption mechanisms in coalis disappointing, and in some spectral regions

0CL i

0U

U)00 'l0

A (pm)20

FIG. IIIC-1. Schematic of the electromagnetic absorp-tion spectrum of an 85% carbon coal. Dashed lines indi-cate regions of the spectrum where there are no publisheddata: wavelengths shorter than 0.2 pm, the near IR re-gion between 0.7 and 2 pm, and wavelengths longer than15 pm. Many of the observable absorption features havebeen assigned to molecular structures within the coal.Two notable features, the strong UV absorption around0.25 pm and the IR absorption near 6 pm have uncertainassignments, yet they are both very characteristic ofcoals.

there are no published data (see Fig. 1). A

major complication in acquiring such data isthe presence of strong scattering; in thisregard photoacoustic techniques may prove to beuseful. Ther e is a consistent patter n tooptical absorption in coals —an almost steadyiricr ease in absorpt ion f rom the IR to the UV.It is unlikely that the coal coloring resultsf rom a single type of absorbing center, oneexpects a wide variety of absorbing groups inkeeping with the complex diver sity o f coalstructure.

Certain featur es may be the result of arestricted class of absorbers. It is possibleto produce the UV spectrum of pyridine extractsof coal using a mixture o f aroma tiehydrocarbons (de Ruiter and Tschamler, 1958),but this procedure cannot be said to uniquelyidenti fy the absorbing sites in coal a saromatic groups. For example, the UV spectrumshows a feature at 265nm, which may be causedby aromatic r ing structures (Humphreys-Owensand Gilbert, 1958), or, as has beenalternatively suggest;ed, by di-ketones (Friedeland Queiser, 1959).

The distr ibut ion of absor bing centers isnot, known. Thin sections of coal under themicroscope have been shown to be quitecolor ful, and this o f fer s the hope thatlocally, on a microscopic scale, absorptionspectra may show more spectral detail than doesa bulk sample. Spectroscopic studies of coal

Rev. Mod. Phys. , Voi. 53, No. 4, Part II, October 1981


under the microscope appear not to have beendone as yet. Only recently has the increasedtransparency of coal and coke in the near IRbeen shown to be useful in microscopy (Harrisand Yust, 1979). The visual appearance of coalunder the microscope has long been asignificant criterion for classification ofvarious coal types, and the combination ofmicroscopy wit;h various forms of spectroscopymay provide important new information.

Ultimately, we want a consistent, andrelatively comprehensive, picture of the natureof absorption and dispersion of electromagneticradiation in coal. Me need answers to basicquestions regarding the cause for thei n creasing absorption o f coal in theult;raviolet, the nature of the absorbingcenters, the variability of absorption fromcoal to coal, and the role played by theabsorbing centers in the other physical andchemical properties of coal.

Additional in formation mi ght be gained,perhaps, by shifting attention away from thenaturally occurring absorptions in coal, andfocusing attention on various modi fied coals.It may be possible to "stain" coal with variousabsorbing or fluore sc ing compounds whichhighlight particular features. Also, opticalstudies of synthetic coal at various stages ofsynthesis could contribute to our understandingof the optical properties of natural coal.

Vi br a 0 $.oa a l Spect roscopy. IR andRaman -- Molecular vibrations generally haveassociated with &them wave numbers in the regionfrom 4, 000 cm down to about 100 cm, ar e g i on customar i ly t errned the "vibrationalinfrared. " In contrast to other portions ofthe spectrum, absorption of coal in thevibr ational IR exhibits a great deal ofstructure, wi th many r e son ance s correspondingto frequencies of molecular vibrations. (Brown,1955) Certain of these speet;r al f eatur escorrespond to the vibrational motion o fwell-known chemical conf igurations, andestimates of the amount of aromaticity of thecarbon, along with the pr obable nature of thebonding of heteroatoms, have been inferred fromthe IR data. (van Vucht, et al. , 1955;Retcofsky, 1977)

Interpretation of IR data is not alwaysunarnbi guous, however, and i s beset with bothqualitative and quant i tati ve problems. A

typical problem concerns prevalence o f thecarbonyl (C=Q) configuration in the organicpor t;ion of coal. Ordinarily, the analytical

organic chemist gets excellent in formation onthis question )rom the intensity of absorptionin the 1650 cm region, characteristic of C=Ostretch motion. In coal, there 1is a strongdist, inct absorption at 1600 cm (Retcofsky,1977), and a straight, forward interpret, at, ionwould be to assign this feature to a earbonylmotion (Roy, 1957; Depp and Neuwr oth, 1958).However, the known presence of aromatic r ingsin coal complicates the interpretation byi n t r o d uc i n g another poss ibi 1i ty '. some worker shave assigned this same spectral region to C-Cmotions in aromatic rings (van Vucht et al. ,1955; Br own, 1955; and Orchin et al. , l951).Furthermor e, the possible existence of extendedgraphitic structures in coal has led to

'suggestions that this absorption may be theresult, of donor-acceptor behavior (Elofson andSchulz, 1967), or of noncrystalline, graphit, icfeatures of the coal structure (Friedel et al. ,1971). In this way absence of uniqueinterpretations limits the ability of IRspectroscopy to answer questions of compositionand structure directly.

Inherent in vibr ational spectroscopy arealso pr'ob lems o f quanti tat i ve in ten sit ydeterminations. The vibrational frequencies ofa particular group of atoms r emain relativelyconstant, while their IR absorption strengthvaries from molecule to molecule. This makesdif ficult the use of IR as a quantitativeindication of the amount of any particularchemical functionality present in a sample ofunknown or diverse character, such as coal.

Sub jeet to the foregoing limitations, IRcan provide considerable versatility in theanalysis of coal and coal derived products. A

c lear ad vance has been made in IRinstrumentation through the introduction of theFourier Tr ansform IR spectr ometer (FTIR). Thisdevice en joys both a multiplex and athrough-put advantage. It collects data at allfrequencies at all t imes, and util i zes asignificantly larger solid angle than acomparable grating instrument. Furthermore,the instrumental resolution inherent in itsoptics can be made high. These advantages makeFTIH well suited for the rneasur ements of smalldifferences in the spectra of various coals, ora single coal during processing (Solomon,1981).

One method of circumventing some of thespectroscopic ambiguities in structural studiesi s by chemi cal l y modi fying speci ficfunetionaliti es wi thin the coal and observingthe resulting shifts in the spectr um. Fore x amp 1e, chemicals. y at tacking C =0 groups whi1e


Research on Goal S37

leaving aromatic structures intact might be oneway of identifying the nature pf the functionalgroups giving the 1600 cm bands in coal.Some limited work of this type has been done oncoal liquids, (Fujii, 1963) but the inabilityof chemical reagents to penetrate the bulk coali s a problem. Som type of surfacespec troscopy ma y pr o v i d e t he techn i quenecessary to pursue t, his type o f wo r k;phot oacou st i c IR a ppears promising. Thistechnique has been used to observe the aging ofcoal sur faces, exhibiting details notaccessible with other methods (Rockley andDevlin l980) . The possibility of monitoringsur face ef fec ts, a s we 11 as the capabil. ity touse hulk samples rather than finely dividedpowders or thin sections, makes the furtherdevelopment of the photoacoustic IR t,echniquevery attractive.

Raman spectroscopy, like IR spectroscopy,is capable of supplying i. n for matign aboutmolecular vibrations. About one in 10 photonsgf visible light impinging upon a sample

- scatters inelastically. By measur ing t. heenergy loss of the scattered photon, the energyof an accompanying vibrational excitation isdetermined. Because the selection rules for IRand Raman di f fer, the two spectroscopiesprovide complementary information onvibrations. Raman spectroscopy of coal hasbeen limited by the presence of a backgroundfluoreseenee signal, which may obscure theRaman signal (Hakovsky et al. , 1971). Thisproblem might be circumvented by wor king atshorter wavelength (Ziegler and Hudson, 1 981 ).By working at shor ter wavelengt, hs, such as 200nm, it may be possible to shi f t the Ramansignal to wavelengths shorter than those of thefluorescence, thereby r emovi ng theinter ference.

Another advantage may accrue in the UV.At these shor t wavelengths one would be doingresonance Raman spectroscopy, since thewavelength will correspond to that of one ormore electronic absorptions in coal. Byinvestigating the dependence of the Ramansignal upon wavelength, it may be possible toidentify not only the Raman active vibrations,but the electronic ab sorbing gr oupsparticipating in the Raman pr ocess. Thus itmay be possible to cross-correlate featuresthat are observed in elect ronie spectroscopywith those in vibrational spectroscopy. An

example of this is the Raman featurecorresponding to a vibration of about 1 600cm, which may correspond to an aromatic ringmotion. If the assignment of the previously

described (Davis, 1979; Schuyer and vanKrevelan, 1959; de Ruiter and TSchamler, 1958;Humphreys-Owens and Gi lber t, 1958; Friedel andQueiser, 1959; and Speight, 1971) 265 nm UVabsor ption to aromatic structure~ is correct,the Raman signal at 1600 cm should beenhanced as the wavelength of light used in theRaman measurement approaches 265 nm. If, onthe other hand, the inter pretation of the 265nm UV absorption as a di-ketone feature iscor rect, ketone features -in the Raman spectrumshould be enhanced. This example suggeststhat, despite signi f icant instrumentaldif ficulties, Raman spectroscopy should pr oveuseful in the analysis of coal.

Laser Pluxe SpeetreseopySigni ficant r adiation loading can seriouslydamage a sample of coal. This is especiallypossible with laser techniques. In most casesone wishes to minimi ze damage by using lowpower levels and efficient data collection, but,in some cases this damage can itself be useful.A laser can induce very rapid but controlledheating o f a sample, and this may be useful instudies of desor ption or pyrolysis (Hanson andVander borgh, 1980). At very high powerloading, a plume of vaporized mater ial isemitted from a surface that is almostinstantaneously heated by a short, focusedlaser pulse. The ability of a laser tovaporize material in this manner gives rise toseveral laser plume spectroscopies. A pulsecan vapor i ze a small sample, and the resultingplume can be investigated by optical or massspectroscopic techniques. Since a laser plumeori ginates from a very well defined focalregion (Hanson and Vanderborgh, 1980), a laserprobe may be combined with scanning microscopictechniques to give a point-by-point analysis ofa material. One application is atomic analysison a microscopic scale. This procedure alreadyhas been applied to biological samples (Suminoet al. , 1980) ~ An advantage of the laser probeover other techniques for microscopic elementalanalysis, such as Secondary Ion MassSpectroscopy (STMS) or electron micro pr obe, i sits ability to work under a normal atmosphere.

An alternative spectroscopic techniquecould be emission or resonance fluoreseeneeanalysis of the plume contents. A promi singtechnique is the use of a tunable laser forboth plume product ion and subsequentmultiphoton ionization of the plume contents.By tuning to one particular atomic r esonance,the ioni zat ion o f the plume will be

Rev. trod. Phys. , Vol. 53, No. 4, Part Il, October 1981


proportional to the presence of a part icularelement in the mater ial (Hurst et al. , 1979).It should be possib le t o map themicrodistribution of trace elements in coal inthis manner.

In addition to the atomic compositionanalysis by such techniques, it is possiblethat chemical str ucture information may beobtained by laser plume spectroscopies. Almostwith the first availability o f lasers, therehas been laser pyrolysis work on coal (Sharkeyet al. , 1964). A high intensit, y laser beamgives rise to acetylene as the pr ineipalproduct (Hanson and Vanderborgh, 1980). Thisis consistent with what is known of thebehavior of organic compounds in high laserfields, where C& r adicals are found as majorspecies. At atmospheric pressures, productscondense out of a plasma cloud, but in vacuumthere may be few subsequent collisions afterthe product chemical is formed. A syst, ematicunderstanding of these effects may give directinformation about such features as theenvironments of heteroatoms. For example, invacuum laser pyrolysis an arnine nitrogen mightcome off pr incipal1 y wi th t he carbon as CN, orwith the hydrogen, as NH or NH . If, then,there are character istic laser "crackingpatterns", one may be able to assay the nearestneighbor environments of important heteroatoms.

Photoaeoustic SpectroscopyPhotoacoust i e measur erne n t s, i n c 1u d i n g t hetechnique of photoacoustic microscopy (Quate,1981), may be helpful in examining thestructure of coal. This form of microscopy,sensitive both to the photoabsorption and tothe thermal transpor t properties of the mediumunder examination, has the potential forcontrast which is enhanced above that normallyseen in traditional optical microscopy forpetrographic analysis. The photoacous tietechnique has the additional advantage ofoperating in air, or in other gases. Since itis inherently a scanning technique, automationshould be straight forward, and development ofsuch instrumentation could be valuable in theroutine classification and characterization ofcoal. Another possibility is IR photoaeousticmieroseopy combined with optical spectroscopy,which should allow mappin g o f chemi c alfunctionality with a spatial resolution on theor der of 10 p m. Shorter wavelengths would showless spectral structure, and hence provide lessdetailed spectroscopic information, but wouldallow spatial resolution on the order of 1 pm.

FTIR with photoacoust, ic detection (Royceet al. , 1980; Rockley, 1979, 1980a, 1980b; andRockley and Devlin, 1980) combines themultiplex advantage of FTIR with the ability ofphotoacoustic spectr oscopy (PAS) to measureabsorption by nearly opaque samples underrealistic conditions. It, is also quitesensitive to changes in surface area, couldpossibly give useful information about, changesin pore structure and has also been used inpreliminary measurements of adsorbed moleculeson zeolite catalysts.

I—ray Scattering anci Fluoreseenee—I-ray diffraction studies have already founda place in the examination of coals (Hirsch,1954; Pavlovie et al. , 1981; Beal1 and Mad ja,1981; and Schmidt et al. , 1981). Large anglescattering indicates the degree to which coalmay be consider ed to be "graphitic" (Hirsch,1954; Menster et al. , 1962). Small angle x-rayscattering (SAXS) indicates the sizedistribution of the pores in coal (Schmidt etal. , 1981; Carnier and Siemon, 1978; Be al 1,1980; Lin et al. , 1978) (See discussion inSection III B). Although this x-ray scatteringwork has been extensive, significant furtherprogress can be expected. For example, byexploiting the high brightness and goodcoll irnation o f synchrot ron rad i at ion, x-rayscatter ing could be carried out on a singlemaceral. More conventional bulk techniquesshould be applied to specimens consisting ofseparated macerals, now that the latter areavailable. This would permit examination ofstructural features (such as pore sizedistribution) in particular macerals, perhapsproviding increased understanding of thebimodal and trirnodal distribution of pore sizesobserved in some coals (Schmidt et al. , 1981).X-rays may be used to examine coal duringr eactions, as has been done in the studies ofintercalation of coals by Lewis acids. (Beall,1980)

Fur ther insight into the meaning ofsur face area and porosity as defined byadsorpt ion measur ements might be obtainedthrough small angle diffraction studies of coalwith its pores filled with liquid N, CO, orother materials. One would chose mater i alwhich would alter the scattering, and theeffect this produces would provide informationon the filling of material pores. This couldbe used to check assumptions made about sizedistributions and aecessibilities of the coalpores, and could provide a cross-check on


Research on Coal

sur f ace area measurements made by gasadsorption or other techn i que s. Contrastenhancement could also be use ful ininvestigating the penetration of reagents intoa coal sample, and the time evolution might bemonitored by X-ray tomography to yield a threedimensional perspect ive of the diffusionprocesses. Such techniques have recently beenexploit ed in X-ray computer i zed tomography o fcoal during pyrolysis (Maylotte, 1981; Kosky et,al. , 1981). Fur ther aspects of SAXS for poresize determinations are discussed in SectionIII B.

Heavy metal labeling of functional sites,either on surfaces or in pores, could providesignificant, information as to the number andlocation of these sites in the material. Moreimpor tantly, labeling would provide a measureof the radial corre1 at ion function for similarand different funet, ionalities. For example,barium loading of acid sites, such ascarbonate, would f ind almost immediateapplicability; other labels should be developedand investigated.

These "contrast, enhancement" techniques offilling coal pores with a fluid, or taggingfunctional sites, should also find use inneutron scattering. One could vary thescattering contrast even for one molecularsubst, ance by changing the isotopic composition.

In addition to diffraction studies usingpenetrating x-rays and neutrons, one shouldinvestigate the utility of using x-rayfluorescence (Botto et al. , 1980; Pr ather etal. , 1979) and neutron act, ivation (Volborth,1979a) for chemical element determination. An

advantage of both techniques is the capabilityof determining element, s in coal or coalproducts on-line in a process stream. Direct,quantitative determination of oxygen (which isotherwise found by di f ferences ) should beimproved and made routine (Vorborth, 1979a).Neutron activation shows promise as a solutionto this problem.

An int r iguing e x t en sion of X-raytechniques is the recent observation o felectronic Raman ef fects in the X-rayscattering from graphite' (Koumelis and Londos,1980) Although the utility of the techniqueremains to be demonstrated for coal, itspotential should be carefully examined.

Mossbauer Speetrescapy -- Oneapplication of the Mossbauer ef feet, thedetermination of iron and the characterizationof its chemical environment, is now a useful

method in coal research, and shows promise as astandard method of assaying pyrite in coal(Huggins and Huffman, 1978, 1979; Levinson,1 981). In

qrecent study, a radioact, ive

pgpcursor ( Co) of a Mossbauer-active nucleus( Fe) has been impregnated into a catalyticmaterial to give infor mation about its localchemical environment. (Clausen et al. , 1979)There is little to preven't an exten sion o fthese techniques to the study of coal itself,using appr opr iate Mossbauer nuclei. Onceincorporated in the coal, these nuclei mayserve the additional purpose of providinginformation during chemical and physicalchanges taking place in the coal. Fromchemical shifts, one ean now reliably determinei f a particular Mossbauer nucleus is in anoxidi zed or r educed chemi cal environment,although determination of the identities of itschemical neighbors is not usually possible.Such in for mat ion is quite useful, especiallywhen one is interested in identifying changeswhich may occur during chemical or physicalpr oeessing. Efforts should be made to exploitMossbauer labels in coal, both naturallyoccurring and artificially introduced. Some

.,further discussion of applications of MossbauerSpectr oscopy to coal research is given inSanction IV F.

d. E lectron probe and electron spectroscopy

Electron probe and various electronspectroscopic techniques ar e well suited forthe characterization of physical and chemicalstructure and composition of surfaces andsur face layers. The thickness of the layeramenable to investigation (~3 mean fr ee paths)depends on the nature and energy of the probingand detected particles and on the angle betweenthe particle's tra jectory and the specimensurface. (e.g. X-rays penetrate to about; 1 p m,while low energy electrons penetrate to only afew Angstroms. ) The limitatio'n of thetechniques to surface layers may seem to be aserious handicap. One should bear in mind,however, that the reactions o f solid coalparticles with various agents occur at thesurface, and depend critically on the chemicaland physical environment ther e. Similarconsiderations apply to coal pr oducts and tocontainer materials (corrosion and degradationprocesses —— See Section IV D). Bulkinformation can also be obt ained by thesetechniques in some cases (beyond a few atomic

Rev. IVlod. Phys. , Vol. 53, No. 4, Part l I, October 1981


layers in the sample, pr oper ties observed ar ethose of the bulk material) either directly, orby successive removal of surface layers.

Because of the microscopical lyheterogeneous structure of coal, the electr onprobe techniques, with their high spatialresolution, may pr ove to be ideal tools ofinvestigation. These techniques r equire avacuum environment and are best utilized incombination to complement each other. They ar enot easily adaptable for on-line qualitycontrol applications. A very importantdeveloping use of these methods is in the areaof model cat, alytic studies (Somorjai, 1979;Goodman et al. , 1980) and the determination ofsurface layer compo sit ion and structure(Somorjai and van Hove, 1979). (See Section IVF. ) The critical questions here concern the wayone should proceed f r om ideal i zed labor ator ysamples .to coal, coal products and componentsof coal conversion systems. The modificationof conventional high vacuum apparatus to allowtemperature contr ol of the sample, theintroduction of high pressure reagents, andsubsequent return to high-vacuum condit ions,already has made possible studies of theprogess of sur face reactions and changes insur face layer str ucture and composition(Somorjai, 1979; Goodman et al. , 1980).

Kl e et ron Nx. eroscopy -- Electr onmicroscopy ut ili zes a highly focused electronbeam and in gener al o f fers excellent lateralresolution ('l00 A or better) and large depth offield (order of micr ons).

Transmission Electr on Microscopy. (TEM) iscapable of a few Angstrom resolution butrequires specially prepared samples. It hasbeen utilized to study surface structure ofcoal and has yielded pore size distribution ingood agreement with low-angle X-ray scatteringresults (Har ris and Yust, 1976; Har r is et al. ,197T) . Strehlow et, al. (1978) investigatedthe relationship between mineral content andcertain maceral types. Monitoring sur facetransformat ion with this technique can alsoyield valuable in formation, e.g. Thompsonutilized a hot stage TEN and obser ved thedevelopment of some mesophase structure oncoal at elevated. temperatur es (Thompson,1981). A similar technique has been used tostudy the catalyt ic gasification of graphite(Baker, 1981).

In Scanning Electr on Microscopy (SEM),the secondary back-scattered electrons aredetected. This method can accept a broad

range of samples and yields a pictur e of thesurface with typical magnifications of 50 to100, 000. SEM has been extensively used (morethan 10, 000 instruments are operating aroundthe world today) for studies of biologicalsystems, miner als, microcircuits etc. Shinnand Vermealen (1979) applied SEM to coalbefore and after leaching with or ganic solventand found it very useful in monitor ing theopening up of the pore structure.

Coal ash is comprised of micrometer sizedaluminosilicate hollow glass spher es. Thesehollows are f i 1 led wi th sma1 1er spher esconsisting of heavy metals and toxic elements.SEM (and TEM) has proved to be useful fordet ermining the speci fic form of thesedeposits and for the better understanding ofthe accessability of the metals to leaching(Gibbon, 19T9). SEM can also be valuable forthe characterization of coal particles (Mozaet al. , 19T9) and for petrographic analysis ofcoal macerals and inor ganic const itutents(Stanton and Finkelman, 1979). Some cautionhas to be exercised in these studies to avoiddamage to the specimen by the highly focusedelectron beam.

In combination with SEN one can alsomonitor the back-scattered primary electrons,Auger electr ons or the X ray signals. Moredetailed discussion of these methods is givenbelow under the appropriate headings.

Photoelectron Spec troseopyPhotoelectron Spectroscopy utilizing UV (UPS)or X-ray (XPS) photons has found wideapplication in the study of solid sur-faces(Siegbahn et al. , 1967; Carlson, 1975; Brundleand Baker, 1979a; Karr, 19T8) and it can a3. sobe applied to liquids and gases. (XPSoriginally wa s c al 1 ed ESCA: Electr onSpectroscopy for Chemical Analysis. ) With .

thi s method, th'e energy (and somet, imesangular) distribution of photoionizationelectrons is determined and correlated withthe char acteristic energies of atomic speciesand their envir onmental influence. One canobtain information on the atomic compositionof the sample, and from the chemical shi ft, onthe atomic envir onment. The method isapplicable to all elements heavier than He.The r esolution is limited by the energy widthof the light source, the natur al line widthsof the levels involved and the electron energyanalyzer. Typically line widths range from afew tenths of an eV to a f ew e V. Theutilization of synchrotron radiation in recent

Re&. Mod. Phys. , VoI. 53, No. 4, Part I I, Qctober 1981

Research on Coal

year s has made possible signi f icantimprovements in line width, in signal to noisecondit ions, and in scanning over a widewavelength range.

UPS provides sensit, ivity to the valence,the conduction and the first energeticallydeeper lying bound atomic elect, rons.Photelectron spectroscopy of the outer shells(PESOS) has been mainly used in studying themolecular orbitals of free or adsorbedmolecules. For solids and clean solidsurfaces, molecular orbitals usually do notexist and the interest lies in the study ofband structures. The method is useful instudying molecules adsorbed on sur faces, buthas not been developed for analyticalpurposes. The importance of angle resolvedultraviolet photoelectron spectroscopy (ARUPS)is further discussed in Section IV F.

XPS allows deep atomic core levelionization. Photoelectron Spectr oscopy of theInner Shells (PESIS) is basically the study ofatomic orbitals. The shift in t, he bindingenergies of the core electrons as a functionof chemical environment can be considered as aperturbation on the atomic system. Theionization energies are characteristic of theatomic composition of the sample and theshifts in these energies indicate the natureof the chemical environment. Typical chemicalshifts for core levels ar e of the or der of afew eV. PESIS can be utilized for qualitativeand quantitative analysis and for diagnosticpurposes. It can identify very small amountsof material on surfaces (on the order of 10monol ayer ) . In quantitative analysis, anaccuracy of about 101. should be possible. Theintensity r at, io of widely separated peaks inthe electron ener gy of the same element givesan indication of the homogeneity of the atomicdistribution within the sampled region(Brundle and Baker, 1979b) . (Electrons ofdi f ferent ener gy sample di ffer ent depths. )The technique has found wide uti1 i zation inconnection with catalysts, polymers, fibers,f ine particles and microelectronics. An

important feature of XPS is that catalysts eanbe examined just, before and just, afterreactions similar to those which take place inplant reactors. This allows correlation to bemade between sur face information dat, a andcatalyst per formance. However, it must berecogni zed that species weakly adsorbed on thesur face would desorb during the XPSmeasurement, .

Frost et al. (1974) investigated theapplicability of XPS to coal analysis.

Carbon, and- total sulphur determinations usingXPS wer e found to be in good agreement withchemical analysis, but no correlation wasfound for oxygen. Attempts to utilize XPS toseparately identi f y organic and inor ganiesulfur in coal wer e not, success ful. XPSshowed two well separated sulfur 2p peaks at163 and 169 eV, corresponding to iron sulfidesand iron sulfate, r espectively, but, noseparate 2p line was found for organic sulfur.This peak would be expected in the 163-165 eVregion (eorr esponding to thiol-, thioether,disulfide- or thiophene-type structures). Itis likely, therefor e, that the organic sul furpeak is overlapped by the pyrite peak.

Extension of XPS investigations to C, N,0, and S ls and to t,race metal core shells andthe study of satellite structures (associatedwit, h simultaneous two and three electr on

' excitat, ions) may prove valuable incharacteri zing the chemical structure of coal.The XPS spectrum changes drastically withX-ray wavelength. This ef feet could beex plo i ted to empha s i ze (deemphasi ze) certainstructures for the purpose of simplifying thespectrum.

huger Electron Spectroscopy ( k ES )In Auger spectr oscopy, the impinging

electron beam (or X-ray beam) causes an atomiccore level ionization. The resulting holestate may decay to a lower ener gy statethr ough an electronic rearrangementaccompanied by the e jection of a low energyelectron and creation of a doubly ioni zedstate. The ejected Auger electron will have akinetic energy which is charaeteristie of theparent atom (Davis et al. , 1976) . Since lowenergy electrons can travel only a fewAngstroms without losing further energy andbecoming tr apped, t he ener gy and to someextent, the shape of Auger features can be usedto identify unambiguously the composit ion ofthe so 1i d sur face. Quantitative analysis maybe accomplished by compar ing peak heightsobtained from unknown specimens with those ofpure elemental standards. This technique hasa sensitivity in the range of 0.02 to 0. 2 atompercent and is applicable t, o all elementsheavier than He. Utilization of thistechnique in combination with SEN as theprimary electron source (SAM: Scanning AugerMicroscopy) is widespread. The utilization ofthe. combined techniques for spectroseopiepurposes (Scanning Auger, ElectronSpectroscopy: SAES) is characterized by highlateral resolution and surface sensitivity,



whi ch i s sur passed only by Field IonMicroscopy. In Auger microanaiysis (100 nm

electr on beam size) a sample 6volume of 10nm, corresponding to about 10 atoms, can bestudied. Further developments are needed,however, to make this technique applicable toquantitative analysis (Van Oostrom, 1979;Goldstein and Yakowitz, 1975), and the dangerexists of locally pyrolyzing the coal' by theelectron beam.

Energy ger Ifave1enghh ) Diaper aiveIran Spectroscopy -- The decay of innercore hole states may also occur via X-rayemission. The X-ray spectrum is characteristicof the atomic composition and is influenced bythe atomic environment similarly to the Augerelectron spectrum. X-r ays, however, havelarger mean free paths and, ther efore, samplea deeper layer than low-ener gy electrons.

Very frequently the electron beam issharply focused and can be scanned over thesur face of the sample (SEM) . With thismethod, chemical analysis of a few tenths of acubic micron volume can be carried ouf with asensitivity of about 'l in 10 (10 grammaterial ) . These electron probemicr oanalytical techniques have been developedfor quantitative anal ysi s o f elements heaviert, han Na (Goldstein and Yakowitz, 1 97 5;Ander son, 1973; Scanning Electron Microscopy,published annually). Their application incoal research is discussed by Raymond andGooley (1980). In another variation, thesample is scanned wi th the electron beam, andthe X-ray signal at a f'ixed wavel ength i smonitored. If this wavelength ischaracteristic of some element, thedistribution of this element over the samplesur face can be displayed with submicr onlateral resolution.

The application of electron microprobeX-ray detection techniques for the

quantitative determination of or ganic sulfurin coal on a maceral level proved successful(Raymond and Gooley, 1978; Greer, 1979).These procedur es were also used to deter mineinor ganic and elemental sulfur. The activevolume excited by the 17 keV electron beamextended to a depth of about 3 to 4 m belowthe sur face and had a width of about 1 m. Thedetermination of trace elements in coal by SEM

micropr obe techniques was also found feasible(Finkelman, 1978; Gluskoter et al. , 197T).

SEM with automatic image analysis wasrepor t;ed very recently (Huggins et, al. , 1980)

for quantitative mineralogical description ofcoal. Ni th thi s method, t he SEM locatedmineral or maceral media and then centered theelectron beam on the particle automatically.An energy-disper sive X-ray spectrum cover ingNa, Mg, Al, Si, P, S, Cl, K, Ca, Ti and Fe wasaccumulated in about 2 seconds.

Quantitative analysi s by e 1 ectr onmicropr obe techniques compares the X-rayintensity produced by the sample to thatproduced by a standard of known composition.The X ray intensity is influenced by X rayabsorption, fluorescence and atomic numbereffects, and thus by the composition of thesample. The critical que stion here,therefore, is preparation of the standard.Further development in this area is neededbefore accurate analysis can be carried out ona routine basis.

Extended X-ray 4.bsorption FineStructur e -- EXAFS studies go back to the1920's. Mith the availability of synchrotronradiation, advances in quantum mechanicaltheories, and developments in signalmonitor ing and data analysis, this techniquehas developed in recent years into a powerfultool for chemical and str ucture analysis(Hodgson and Doniach, 1978). The extendedX-ray fine structure appe a r s a s a mod u 1a t i onof X-ray absorption above the absorptionedges. The modulation ar ises because thefinal state of the photoelectron is pertur bedby the surrounding atoms. The X-rayabsorption can be measured dir ectly, or bymeans of monitor ing X-ray fluor escence,p a r t i al photoelectron yi eld, totalphotoelectron yie 1d

and�/or

Auger electronyield as a function of X-ray ener gy. Thesemeasur ement s yi eld in for mation about theenvironment o f a par ticul ar element in asample, the distance between this andneighboring elements, and the identity ofthese element s. This information can beobtained for trace elements, both in the bulkand on the sur face, and the atoms are notrequired to form a periodic array. Thetechnique is equally applicable t o gases,liquids and solids. The average distancebetween the absorbing atom and its near atomicneighbor can be determined to an accuracy of afew hundredths of an angstrom. The method hasbeen successfully applied to the study ofsupported catalysts (Sinfelt and Via, 1978)(also see Section IV F) and to biologicalsystems (Cramer et, al. , 1 9T8). Application


Research on Coal 843

was extended reoently to the soft X-ray regionwher e K-edges for the important second andthird row elements fall. A method developedat the Unversity of Delaware (Chem. and Eng.News, 1980) utilizes a rotating anodegenerator as its X-r ay source and can produceEXAFS spectra similar to those produced bysynchrotron sources, but with considerablyless investment.

For maeromolecular systems like coal,which have no crystal structure, EXAFS canidentify the eoor dination envir onment ofselected atomic species (heteroatoms, metalcatalysts and trace elements) (Miong et al. ,1981; Naylotte et al, 1981) and follow theirchanges as the decomposition of the sampleproceeds.

X-ray absorption by a monolayer on asurface is so small that it will not show inan X-ray transmi ssion experiment. However,r eoently a new technique, SEXAFS, wasdeveloped to study surfaces by measuring themodulated intensity of absor bate core Augeremission lines as a function of photon energy(Lee, 1976; Citrin et al. , 1977). (Partial ortotal photoelectron yi eld can also bemonitored). The underlying principle is thatthe Auger or secondary electron emi ssion isdirectly proportional to the photoabsorptioncross section, which is modulated because ofinter ference bet we en outgoing and backscattered core photoelectron waves. Anelectron energy resolution of about 2 eV is ingeneral sufficient to discr iminate againstbackgr ound electrons. Furthermore, themonitor ing of low-energy electrons rather thanX-ray fluorescence, greatly reduces thebackground signal, since these electronsoriginate only a f ew angstroms inside thesurface. The sensitivity of this method isabout 1 0 times higher than t ypi eal EXAFS andstatistically mean irL~ ful data can be obtainedfrom as few as 10 absorbers in the X-r aybeam.

The SEXAFS method was applied recently tothe study o f oxygen adsorbed on aluminumsurfaces by using monochromatized synchrotronradiation and part ial electron yieldspectroscopy as the detection technique (Stohret al. , 1978).

The SEXAFS technique should be useful inmonitoring, both simplified model catalysts andcomplex disper sed catalysts. Refinements andapplication of this method to coal catalystsseem feasible. Questions such as whether themetallic elements are combined as homogeneousalloys or in a more complicated binding

arrangement wi th the substr ate oan beanswered. The chemical functionality ofsur face 0, N, S could perhaps be obtained fromSEXAFS studies.

High —Resolution Electro nEnergy-Loss Spectroscopy (HRKKLS)Electrons impinging on sur faces can lose partof their kinetic energy due to variousprocesses which could include excitation ofsurface phonons, plasma waves and electronicsur face transitions (Ibach, 1977). Theinterest here lies, however, in the excitationof the vibrational modes of atoms andmolecules -adsorbed on t he sur face. Theener gies associated with these modes fal. l inthe 50 to 300 me V range and ar e readilydetectable with high resolution (~10 meV)electron impact spectrometer s. The electronssample only a few angstrom layer as opposed toIR photons, which penetrate a depth of t;,heorder of a few micr ons and thus give bulkinformation. Adsorbed species with chemicalbonds perperidieular to the sur face are moreeasily detectable than those with bonds whichare parallel to the surface. The sensitivityof this technique is of the order of 0. 1

monolayer. Chemical bonds characteristic ofthe adsorbed molecule and of the bindingbetween adsorbed atomic and molecular speciesare readily ident i f i ed and can- ser ve as amonitor on the nature of the adsorbed .species.An important aspect of this technique is itsabil i ty to detect hydr ogen, which is notr eadily detected by other techniques. Thisunique capability makes HREELS a veryimportant, tool in studying the sur faoechemistry of hydrogen, hydr ocarbons and otherhydrogen containing molecules. For example,it is easily recognized whether a molecularadsorption proceeds through dissociation ornot. One ean also gain knowledge of theb in ding potent j. al and the a f feet of

.coadsorption. As in the case with othersurface tools, this method is also bestuti1 i zed in combination with LEED, Auger ete.methods.

In recent year s, a number o f HREELSstudies have been carried out on H, CO, 0and hydrocarbons absorbed on metal surfaces(Froitzheim, 1977; Thiel et al. , 1979;Hamilton et al. , 1981). The method yieldsvaluable insight into the mechanisms ofcatalytic r eactions. Combination with SEXwould enable one to investigate small regionsof a sample.

Rev. Mod. Phys. , VoI. 53, No. 4, Part II, October 1981


The application of HBEELS is justdeveloping. The question of selection r ules(the relaxation of which is a major advantagein electron impact spectroscopy as compared toopt ical spectroscopy in the case of gaseousspecies) has not yet been settled and fullyinvestigated for solids. Neither have theangular dependence, resonances and theutilization of spin polar i zed electrons beenextensively exploited in connect ion wi thsurface adsorbed species. One would certainlyexpect important developments in this areawhich should contr ibute to our understandingof the nature of chemisorption.

e. Other techniques

Mass spectroxetry will continue to bea significant tool in the analysis of coal andcoal der ived liquids (Carter et al. , 1979;Guidoboni, 1 979; On do v et, al. , 19'75) .Important techniques such as plasma desor ptionand field desorption (St. John et al. , 1978),f ield ioni zation and chemical ionizationshould complement ordinary spark and electronbeam ionization. A recent development in massspectrometry is the use of laser ionization.A comprehensive review of this technique, asapplied to solids, i s given by Conzemius andCappellen, including a direct application tocoal (Conzemius and Capellen, 1980). In muchof the work with this technique there has beenthe problem of a basic lack of understandingof the processes oceuring in the ion-forminglaser plasma. Becent rapid progress makes itappear likely that laser ionization will be astandard option on mass spectrometers in thenext five year s. Laser intensities highenough to ionize the sample are also highenough to lead to significant cr acking ofhydrocarbons, as evidenced by the large numbero f C& species formed with straight laserionization. For milder ionization conditions,when study of the higher mass region of thespectrum with less cracking is desired, acombination of laser desorption with othermethods of ioni zation shows promise. Recentwor k on laser pyr olysis sho~s that it haspotential for nitrogen and sul furdetermination (Hanson and Uanderborgh, 1980).

Additional advantages accr ue with theability to probe individual macer als in asample. SINS is well suited for this role(Wittmaack, 1979).

Crossed atoxic and xolecular beaxstudies are capable of revealing details o fchemical reactions, and are very important

methods for under standing the mechanism of agiven reaction. - There are several problems,however, in trying to apply these methods tocoal gasi f ication and synthet ic fuelproduction schemes. The beam-beam techniqueis applicable to gas phase processes, while incoal gasi fication and liquefaction, it isheterogeneous catalysis processes, not readilyamenable to study by this technique, thaQ aremost important. The practical usefulness ofbeam experiments in these studies is thereforeuncer tain. Gas phase reactions and flamesassociated with coal utilization involve alarge number o f mutually i nteract ingcomponents and very complex kinetic schemes.Beam/beam studies ar e best suited for thedetailed understanding of specific reactions,from which the complex kinetic scheme, can bebuilt up.

Ion beaxs -- An a lyt i eal techniquesbased upon the use of accelerated chargedparticles to induce prompt y r adiation,elastic scattering, and x-ray radiation haveproved useful for elemental analyses,including spatial sur face distributions anddepth profiling (Pierce, 1 974; Volbroth,1979b; Prather et al. , 1979; Buckle and Grant,1981). Prompt radiation is most useful forthe lighter elements, and the technique hasbeen used for analytical determinations ofear bon, nitr ogen, oxygen, and sul f ur .Proton-Induced X-ray Emission (PIXE) has beenused with somewhat better sensitivity than thecorresponding electron-induced x-radiation fortrace element analysis (Buckle and Grant,1981). In addition, the incident protons canbe focused to beams a few microns in diameterand brought through a window to performmicroprobe analyses with the sp. cimen atatmospheric pressure, something not possiblewith electron micropr obes. This feature couldbe useful in studying elemental variationsover the sur face of whole coal, as well aschanges during mild process conditions.

Elastic scattering of alpha particles hasbeen most use ful in t he thicknessdetermination o f f i lms o f' heavy elementspresent on lighter substrates. The techniqueis complementary to X-ray work, since the verylow atomic mass species produce X-r ays whichare too soft for convenient measurement, whilealpha backseat ter ing shows wor seningresolution for higher atomic masses. Thesurface composition of species such as C, 0,S, Fe are readily resolved (Patterson et al. ,1966).

Rev. Mod. Phys. , Voi. 53, No. 4, Part II, October 1981

Research on Goal S45

Neutrons are useful for activatingtrace elements and heteroatoms of coal andcoal ash. Thermal neutrons from reactors(Steines, 1979) and fast neutrons generated bydeuterons accelerated onto met allic targets(Larsen and Kovac, 1978) each have specificadvantages for di f ferent elements. Fastneutrons, especially. , can provide chemicalcomposi tion information for the importantheteroatoms oxygen and n itrogen. Neutronacti vat ion techniques also have i rnpo r tan tapplications in process monitoring and control(see &ection IU 8) .

Sma 1 1 angle neutron scatter ing,especially i f combined wi th addi tion ofcontrast enhancement flu id s, may pr ov i deuse f ul in formation regarding the sizedistributions of coal pores. As in the caseof x-ray diffraction, there may be mer it intr eat ing the average structure of coal asbeing amorphous, and using di f fractiontechniques to deduce positional correlationfunctions.

Pr el imi nary small angle neutronscattering experiments have been carried out(Maxwell et al. , in press) on coal powders andsolutions in an attempt to use this techniquefor the determination of average molecularweight.

Elastic property aeasureeents incoal are par ticularly difficult because of theheterogeneous character o f the material.Nevertheless, those data which have beencollected have interesting implications. Forexample, Lar sen and Kovac (1978) havesuggested that the t ime-dependent re s po n s e o fbituminous coals to a constant stress showsthat they are cr os s —linked, non-planarmolecules. This is a sufficiently importantpoint in discriminating among proposed averagemolecular structur es of coal that furtherstudies of this aspect of the elasticproperties should' be encouraged.

An ultrasonic experiment (van Krevelan etal. , 1959) obtained velocities of sound forvarious coals, from which carbon aromaticitieswere deduced. The ultrasonic values for thisparameter turn out t o be in rather goodagreement with more direct determinations bythe highly sophisticated resonance methodsdescr ibed in Part 3b, of this Section. Thiscorrelation may also be worthy of furtherinvestigation.

more recent application of ultrasonicsis that of ultr asonic microscopy of coalsurfaces, (Quate, 1981) with a spatial

r esolut ic n of 0.3 m. This technique could beuseful in petrography, the optical microscopyver sion of which depends on var iations amongmacerals in the small ( 1 $) opt icalr eflectivity (Ting, 1979). By compar ison, theacoustic reflectance is 50%, and the wavesalso sample the subsur face of the coal whichis inaccessible to the visible light waves.Additional applications of ultr asonietechniques will be discussed in Sections IU 8and IU F.

4. Recommendations

Samplese A sample bank of carefully selected,

characterized and, stored coal specimensshould be estab li shed where individualsor groups could obtain representati vesamples for analysis or researchpurposes.~ t~ore ext ensive use shoul d be made of

labelling techniques in connection wi thresonance experiments, x-ray and neutr on.scattering.

~ Further study should be carried out toexplore the possibilities of preparingsmall samples of synthetic coal.+Improved petrogr aphic and chemical

characterization methods should bedeveloped. Acoustic and photoacousticmicroscopy techniques seem promising, asdo x-ray tomography and EXAFS studies.

TechniquesNMR —— Bet ter NNR line-narrowingcapability should be developed, e ffortsshould be made to per form resonance (NMRor NQR) measurements on oxygen, nitrogenand sulfur nuclei and attempts should bemade to obtain molecular structur alinformation pertinent to linkages whichplay a role in the pr imary liquefactionstep. NMR techniques applied to zeolitecatalysts are promising and should beextended.ESR and ENDOR studies have importantpotential for determining the influenceof free radical behavior on the rate ofcoal liquef action. Fur ther studiesshould be carr ied out under processconditions, and serious attempts shouldalso be made to car ry out ENDOR studiesat elevated temperatures and pressures.Optical Properties Measurements should

be made of the complex index of

Rev. Mod. Phys. , Vol. 53, No. 4, Part l l, October 1981


0~ M

~ W

E~ W

CCt

~ T~H

0 o

M

0

E

00

Ice

QO

Q5

0V)

(0

CZ'

O

Mj

8N O

g~ crt 0

go

O0O

c ~o0 '

C4 ~0 C

0O

rA

CF'

Mj

~ ~

0

0(0

C

0

~ W

0

Q3

Q

EPEEC3

co ~ O~ Ck

QO O

'gj

a gg~ ~

pH ~ ~H ~

cA

~p.S~ E00U m

.5

0~ o E

c@ ~O

oHQ0E~~

g

at

Gh

w O

E

05

~ W

M)0~ w +

oO

~ ~~ ~

M ~

M

M

MbG

E0

0 ~Q ~

C &.O~ ~

~ H 0

(L)

E s~~&un0 ~

O(D

0C cn

E OQ

0og+owc 0

gg WO

~Per}X

I

0

0

0

C

0A

O~ W

M}

~P,~ ~

Q30 MM4- o ~

0 MO

uaO OQ +W ~ 'H

E &~

~~

0~ ~

E

E0M

zC

0

~ ~

O0

0

0 M

(Q 8O Oat

Gh g0F III

g~ W

CAO

0~ TH

O

~ ~

E

0~ ~

O

CCt

~ W

O

~ W

O0

t5Q

0~ W

E

~ W

0G4

~ H

O~ ~crt

E

.cj~ c0Q

O

gO

~ o(0

M)

0

~ W

0~ W

~ ~

cga

OC40

E

0

~ M

~ om

Q

0Q

~ ~

PR0m

~ EG6

0

0

67O

0cga0

E0

0

0E

0

~ ~(D

0

~ ~

CCt

E(9

O

0~ ~

(D

~ W

2

0~ W

.a

E~ 1H

O

00

G400O

C4

CC

+ IH

G40O

C4

0

G4

0O

O

~ ~

0O

00

M't

O0

S40



C0

aQO

c5

0 as'~ 20M

60

(Q

0

~ ~

O O

GC

CCt

O

O

6~ W

(DbQ

~ ~

z

~ 'IH

M

805)

0

S4E

M

QlH~ tH

0

E

CO

G4

9|=t E

Q

Ptl

m0OQ

0

0

gM

0E

"a~ M

0cn,

74.~

M M

(Q

0

g0crt

h4

Qg

O 0~ ~g4

~ ~

g KM

~ ~ Q

& p

0~ ~

~ M

tD

OQ

0

40~ W

E

0~ ~c5

a"a

0

C3

+

6

IQ

Q~ M

~ '+0O

o R, H

C0

0Q cn

0 0

6 t-0

~ 0+

0~ W

0 +.

0 gO.~

0

0

0

0

0oa ~0 Z~

M

O 0+ 0

Pg0 O

CA

rZO 0

0 G4'C cn

mp 9

~ ~iQ

uC4 W

~ M

0C0

~ ~

E

vi

CA

~ ~ Q

M) WN

O 'C+OV.

g W

g0~ tH

0G4

E0Ch

o

Q

E0

0

E

E

~ IH

O~ W

0~ 'H~ W

0C4E0

~O EE ~0Q

~ ~

~ W

O~ W

E

0~ ~ +mh

~ ~

O~ M

E0

~ W

O

m

O~

~ ~O0

m

.g o

CCt

M

0~ W

O

O~ W

E

O

0~ ~

0CCt

0 4

~ ~ 4

0p

GhQ

&9 P0O

mt0

rA

0~ M

(DN

~ IH

0

0~ ~~ W

0G4

C0O

~ ~

0E

O

0

0~ W

O~ W

S4

0O

O

00

0O

O'~ 'H

E

G40O

0

O D

G40O

63

0O(D

oQ

~ 0

O

G4G6 Q

Q -~

0

O%2

0

C3

+oO

M

M ~™

0

0~ ~

Gh00 +cn

(DP g}„

O

0

O

g cn

0 .m

0 nt

0~ ~

~ ~

0

E

C40

~ peef



refraction of coal o ver a s wi de af requency range as possible, and ther esults used to test features of thevarious aver age str uctural modelspr oposed for coal. Microscopicspectroscopy could be especially usefulin identifying structural differencesamong macerals in the same coal specimen.Vi b r a t i on a 1 Spectroscopy Part i cul aremphasis should be placed on vibrationalspectroscopy, both because o f i tssensitivity to molecular bonding andbecause of its potential for monitoringsubtle changes which take place during1 i que faction. Photoacoustic FTIR andresonance Raman studies could be fruitfulnew techniques applied to coal andcatalysts.Laser Plume Spectroscopy should beexplored for its potential for providinglocal molecular information in additionto its established capabi1 i ty formicroscopic elemental analysis.Synchrotron Radiation sources should beexploited for extended x-ray scat teringexperiments, and contrast enchancementadditive substances should be used inconjunction with both x-ray and neutrondiffr action to study the details of thepore structures of coals.Electron Microscopy shoul d be ut ili zedfor the determination of particle andpore size d istr ibutions and shapes.Development of automated procedures forroutine measurements would be veryuseful.Electr on Spectroscopy Var ious electronprobe and electron spectroscopictechniques should be pursued in acomplementary fashion for charac t er i zi ngsur face structure and composition ofcoal, catalysts and container materials.The quantitative techniques for N, 0, Sand trace element determinations shouldbe refined. SEM combined withenergy-dispersive X-ray analysis seems tobe very promising for fast, automated,quantitative elemental analysis ofmineral and maceral segments in coal.Quantitative analysi s by electronmicroprobe techniques relies oncomparison with standards. Research isrequired on questions concerning thepreparation of standards. SEM, combinedwith Auger Electron Spectr oscopy, SEXAFS,and HREELS are excel lent methods andshould be further developed for the study

of surface adsorbed species.Catalysi s The elect ron pr ob e andelectron spectroscopic techniques shouldbe utili zed in catalytic studies. A

useful approach to such studies is theintr oduction of a high-pressure andhigh-t, emperature r eac t ion chamber into aconventional high vacuum electron probediagnostic system.Neutron Activation analysis should bemore widely applied for concentrationdeterminations of fir st row atoms. Thetechnique should be made generallyavailable to workers in the field, andcoal specimens should be routinelycharacterized for oxygen content.

References

Adams, D. C. , 1970, Ex othermal t'1e tamor phosi sof Coal Precur sors, Ph. D. disser tation,University of Utah.

Anderson, C. A. , 1973, ed. Ni croprobeAnalysis John %bailey and Sons, New York.

Baker, R. T. K. , 1981, Bull. Am. Phys. Soc.26, 200.

Bartuska, V. J. , G. E. Haciel, J. Shaefer,and E. 0. Ste jska1, 1977, Fuel 56, 35M.

Basic Coal Sciences Pro ject Advi soryCommitt ee Report, 1980, Prepared byScience Applicat ions, Inc. for the GasResearch Institute, July, 1980.

Beall, H, 1980, Fuel 59, 140.Beall, H. , and R. J. Mad ja, 1981, in

Chemistry and Physics of Coal Utilzation-1980 (APS, Norgantown) (edited by B. R.Cooper and L. Petrakis, New York), AIPConference Proceedings No. 70.

Bloembergen, N. , 1967, Am. J. Phys. 35, 989.Botto, R. I. , B. H. Mhite and J. H.

Kar chmer, 1980, Fuel 59, '1 57.Br ink, G. 0. , 1980, Opt . Comm. 32, 123.Brown, J.K. , 1955, J. Chem. Soc. (London)

744, 752.Brundle, C. R. and A. D. Baker, 1979a, eds.

Electron Spectroscopy: Theory, Techniquesand Application, Vol. 3, Acad. Press, New

York.Brundle, C. R. and A. D. Baker, 1979b, eds.

Electron Spectroscopy: Theory, Techniquesand Application, Vol. 3, Acad. Press, New

York, p. 309.Buckle, D. C. , arid G. C. Grant, 1981, in

Chemistry Snd Physics of Coal Utilization1980 (APS, Horgantown) (edited by B. R.

Cooper and L. Petrakis, New York), AIP


Research on Goal

Con fer ence Proceedings No. 70.Buss, R. J. R. J. Baseman, P. Casavecehia,

and Y. T. Lee, 1981, to be publ ished.Camier, R. J. , and S. R. Siemon, 1978, Fuel

57, 508-Carlson, T. A. , 19 (5, Photoelectron and

Auger Spectroscopy, Plenum Press, New

York.Carter, J. A. , and D. L. Donohue, J. C.

Franklin, and R. L. Malker, 1978, 1979,in Analytical Methods for Coal and CoalProducts, ed. Clarence Karr, Jr. ', Vols.1, 2, (19T8), vol. 3 (1979).

Cattolica, R. , 1981, Appl. Opt ics 20, 1156.Chem. and Eng. News, July 10, 1980, p. 27.Citrin, P. H. , and P. Eisenberger and R. C.

Jeweitt, 197(, J. Vac. Tech. 15, 449.Clausen, B. S. , and S. Morup and T. Topsoe,

1979, Surface Sci. 82, L 589.Conzemius, R. J. and Capellen, 1980, Int. J.

Mass. Spect. Ion Phys. 34 19(.Cramer, S. P. , and K. O. Hodgson, M. 0.

Gillam and L. E. Mortenson, 19 (8, J. A.Chem. Soc . 100, 3398.

Davidson, Rober t M. , 1980, "MolecularStructur e o f Coal, " Re por t No.ICTIS/TR08, IEA Coal Research, London .

Davis, A. , 1976, in Analyt ical Methods forCoal and Coal Products, ed. ClarenceKarr, Jr . , Vol . 1, pp. 27-83.

Davi s, L. E. , N. C. Mc Donald, P. M.Palmberg, G. E. Riach and R. E. Heber,1 97 6, Handbook for Auger ElectronSpectroscopy, Phys. El. Ind. -Inc.

Depp, E. A. , and M. B. Neuworth, 1958,Br ennst. -Chem. 39, S90.

de Ruiter, E. , and H. Tschamler, 1958,Br enn st -Chem. 39, 362.

Dunning, T. H. , 1981, private communication.Dyer, T. M. , and P. O. Miotze, 1979, Laser

Velocimetry and Particle Sizing, eds. H.D. Thompson and M. H. Stevenson,Hemisphere, p. 267.

Dyrkac z, G. R. , C. A. A. Bloomquist, L. H.Fuchs and E. P Horwi tz, 1981, inChemistr y and Physics of Coal Uitlization—1980 (APS, iilorgantown) (edited by B. R.Cooper and L. Petrakis, New York, 1981}AIP Conference Pr oceedings, No. 70.

Eckbreth, A. C. , P. A. Bonczyk, and J. F.Verdieck, 1978, Appl. Spec. Revs. 13,

Elofson, R. M. and K. F. Schulz, 1967, Am.Chem. Soc. Div. Fuel Chem. Prepr. 11, 513.

Finkelman, R. B. , 19(8, in Scanning ElectronMicroscopy, Vol. '1, SEM Inc. , AMF 0 ' Har e(Chicago), IL 60666, USA, p. 14/.

Fr iedel, R. A. and J. A. Queiser, 1959, Fuel

38, 369.Friedel, R. A. , J. A. Queiser, and G. L.

Garison, 1971, Am. Chem. Soc. Div. FuelChem. Prepr . 15, 123.

Froitzheim, H. , 1977, "Electron Energy LossSpectroscopy" in Electron Spectroscopyfor Sur face Analysis, ed. H. Ibach,

. Springer-Verlag, Berlin.Fr ost, . D. C. , bf . R. Leeder and R. L.

Topping, 1 970, Fuel 53, 206-11.Fu ji i, S. , 1963, Fuel 42, 1 "(; 341.Gibbon, D. L. , 1979, in Scanning Electron

Microscopy, Vol. 1, SEM Inc. , AMF O'. Har e(Chicago) IL 60666, USA, p. 501.

Gluskoter, H. J. , R. R. Rueh, M. G. Ili lier,R. A. Cahill, G. B. Dreher and J. K.Kuhn, 1977, "Trace Elements in Coal:Occur ance and Di st r i bution, " Il 1 inoi sState Geological Survey, Circular 499, p.109.

Goldstein, J. I. , and N. Yakowitz, 1975,Practical Scanning Electron Microscopy,Electron and Ion Microprobe Analysi s,Plenum Pr ess, New York.

Goodman, D. H. , R. D. Kelly, and T. E.Madey, 1980, J. Catal. 63, 226.

Gou1 ard, R. and P. J, Emmer man, 1980, Topi csin Current Physics 20.

Graham, S. C. , A. J. Grant, and J. M. Jones,197LI, AATA Journal, 12 1140.

Grandy, D. ||It. and L. Petr akis, J. Mag. Res.(in press).

Greer, R. T. , 1979, in Scanning ElectronMicroscopy, Vol. 1, SEM Inc. , AMF 0 ' Har e(Chicago), IL 60666, USA, p. 477.

Guidoboni, R. J. , 1978, in Analyt icalMethods for Coal and Coal Products, ed.Clarence Karr, Jr; Vol. 1 (1978), pp ~

421-434.Hamilton, J. C. , N. Swanson, B. J. Maclawski

and R. J. Celotta, 1981, J. Chem. Phys.(to be published, private communication).

Hanson, R. L. , and N. E. Vanderborgh, 1980,in Analytical Methods for Coal and CoalProduct, s, ed. Clarence Karr, Jr. , Vol. 3,p ~ 73-

Harde sty, D. H. , 1980, (privatecommunication).

Harding, L. B. , 1981, ,J. Phys, Chem. 85, 10.

Harr i s, L. A. , an d C. S. Yust, 1979, Carbon17, 307-

Harris, L. A. , and C. S. Yust, 1976, Fuel55, 233-

Harr i s, L. A. , D. Br aski and C. S. Yust, ,1977, Microstruet. Sci. 5, 35.

Harvey, A. B. , and J. M. Nibler, 19(8, Appl.Spec. Rev. 14, 101.

Rev. Mod. Phys. , Vol. 53, No. 4, Part I t, October 1981


Hir sch, P. B. , 1954, Proc. Roy. Soc. ,(London) A 226, 143.

Hodgson, K. O. , and S. Doniaeh, 1979, Chem,and Eng. News, August 21, 1979, p. 26.

Hombaeh, H. B. , 1990, "Fuel, " Vol. 59, No .7, 465 and refs. therein.

Hombach, H. B. , Erdol, Kohle, and Erdgas,1979, Pet r ochem. Br ennst . Chem. , Vol.32, 134.

Howard, C. J. , 1979, J. Phys. Chem. 93, 3.Huggins, F. E. , and G. P. Hu f fman, 1979,

1979, in Analytical .1ethods for Co a 1 andCoal Products, ed. Clarence Karr, Jr. ,Vols. l, 2. , (1978), vol. 3 (1979),Academic Press, Chapter 50.

Huggins, F. E. . . D. A. Kosmaek, G. P.Huf fman, and R. J. Lee, 1990, ScanningElectr on Microscopy, SEM Ine. , AMF O'Hare(Chicago), IL 60666, . USA, Vol. 1 p. 531.

Humphreys-Owens, S. P. F. , and L. A.Gilber t, 1959, Proc. Conf. Ind. Carbonand Graphi te, London, p. 37.

Hurst, G. S. , M. G. Payne, S. D. Kramer, andJ. P. Young, 1979, Rev. Mod. Phys. 51,767.

Ibach, H. , 1977, Electron Energy LossSpectroscopy in Electron Spectroscopy forSur face Analysis, ed. H. Ibach,Spr inger -Ver lag, Ber 1in .

Karr, Clarence, Jr. , 1979, 1979, AnalyticalMethods for Coal and Coal Pr oducts, Vols.1, 2, vol. 3), Academic Pr ess.

Klick, D. , K. A. Marko, and L. Rima, 1991,Appl. Optics 20, 1179.

Kosky, P. G. , E. J. Lamby, D. H. Maylot te,R. L. St. Peters, and C. Spiro, 1991, Am.Chem. Soc. , Atlanta, March 1991.

Koumelis, C. N. , and C. A. Londos, 1990,Can . J. Phys . 59, 1507.

Lane, N. F. , 1990, Rev. Mod. Phys. 52, 29.Lapp, M. , L. M. Goldman, and C. M. Penney,

1972, Science 17 1112.Larsen, J. M. , and J. Kovac, 1979, in

Organic Chemistry of Coal, ed. J. M.Larsen, ACS Symposium Series 71, 36.

Lee, P. A. , 1976, Phys. Rev. B 13, 5261.Levinson, L. M. , 1991, in Chemistry and

Physics of Coal Utilization -1980 (APS,Norgantown) (edited by B. R. Cooper andL. Petrakis, New York, 1981), AIPConference Proceedings No. 70.

Lifshitz, A. , ed. Shock Maves in Chemistryand Chemical Technology (Decker, NY, inpress).

Lin, J. S. , R. M. Hendricks, L. A. Harris,and C. S. Yust, 1979, J. Appl. Crystal.11, 621.

Low, I. J. , 1959, Phys. Rev. Lett. 2, 295.Lucht, L. M. and N. A. Peppas, 1981,

Chemistry and Physics of Coal Utilization-1980, edited by B. R. Cooper and L.Petr akis, New York, AIP ConferenceProceedings No. 70.

Haciel, G. E. , M. J. Sullivan, N.Szeverenyi, and F. P. &iknis, 1981,Chemistry and Physics of Coal Util i zation-1990, edited by B. R. Cooper and L.Petr akis, AIP Conference Pr oeeedings No.70.

Maekawa, Y. , T. Yoshida, Y. Yoshida, M.Imanari, 1 977, J. o f

Jap�.

Pet rol. Inst .

Vol�.

56 351.Makovsky, L. E. , P. Maldstein, and M. H.

Edwards, 1971, Nature (Phys. Sci. ) 231,

Marino, R. A. , H. R. Menk, J. A. Apps and S.M. Kaliner, 1980, J. Molec. Structure 59,

Maxwell, B. E. , H. R. Child and A. H.Nartern, ORNL Solid State DivisionProgess Report, (in press).

Maylotte, D. H. , 1981, Bull. Am. Phys . Soc .26, 201.

Maylotte, D. H. , J. Mong, F. M. Lytle, R. B.Greegor and R. L. St. Peters, 1991,submitted to Science.

Mentser, M. , H. J. O'Donnell and S. Er gun,1962, Fuel 41, 153.

Michael, J. V. , and J. H. Lee, 1979, J.Phys. Chem. 93, 10.

Miyagawa, I. , and C. Alexander, 1981, inChemistry and Physics of Coal Utili zation

1980 (APS, Morgantown) edited by B. R.Cooper and L. Pe tr aki s, New York, 1991},AIP Con ference Proceedings No. 70.

Miyagawa, I. , and C. Alexander, 1979, Nature( London ) 278, 40.

Moza, A. K. , L. G. Austin and G. G. Johnson,Jr . , 1979, in Scanning El ec tronMicr oscopy, Vol. 1, SEM Inc. AMF O'Hare(Chicago), IL 60666, USA, p. 473.

Neavel, R. C. , 1979, Amer iean ChemicalSociety, Division of Fuel Chemi stry,Prepr ints 24, 73.

Neavel, R. C. , 1990, Am. Chem. , Soc. , Div.of Fuel Chem. Preprints 25, 246.

Ondov, J. R. , M. H. Z o 11er, I. Olme z, N. K.Aras, G. E. Gordon, L. A. Ranei tel li, K.H. Abel, R. H. Filbey, K. R. Shah andR. C. Ragaini, 1 979, An a 1 . Chem. , 47,1 102.

Orchin, M. , C. Golumbic, J. E. Ander son, andH. H. St orch, 1951, U. S. Bur . Mines,Bull No . 505.


Research on Coal

Pan, L. S. , 1967, Pressure Ef feet s onLaboratory Coalifieation Processes, Ph. D.dissertation, Univer sity of Utah.

Patt, er son, J. H. , A. L. Turkevieh and E.Franzgrote, 1966, Radiosotope Inst;rumentsin Industry and Geophysics, Vol. 1, p.33'7 ( IAEA ) .

Pavlovic, A. S. , J. M. Cook, J. J. Rent on,1991, in Chemistry and Physics of CoalUtilization -1980 (APS, Nor gantown)(edited by B. R. Cooper and L. Petrakis,New York, 1981), AIP Confer encePr oceedings No. 70.

Petrakis, L. , and D. M. Gr andy, 1979, Anal.Chem. 50 303.

Petrakis, L. , and D. N. Gr andy, 1991, inChemistry and Physics of Coal Utilization-1980 (APS Morgantown) (edited by B.Cooper and L. Petrakis, New Yor k, 'I 98la),AIP Conference Proceedings No. 70.

Pe tr ak i s, L. , and D. M, Gr'andy, 19911,Na t ur e 299, 476.

Pierce, T. B. , 1974, in Characterisat, ion ofSolid Sur faces, eds. P. F. Kane and G. B.Larrabee, Plenum, N. Y. , 1974.

Pines, A. , l1. G. Gibby, and J. S. Haugh,1973, J. Chem. Phys. 59, 569.

Prather, John H. , J. A. Gu in and A. R.Tarrer, 1979, in Analytical Methods forCoal and Coal Pr oducts, ed. ClarenceKar r, Jr . , Vol. 1, p. 357, Academic Press(l 978) ~

Prath'er, John M. , James A. Guin and Ar thurR. Tarrer, 1979, op. cit. Vol. 3, Ch. 49.

Quate, C. F. , 1991, in Chemistr y and Physicso f Coal Ut, ilization -1980 (APS,Mor gantown) (edited by B. R. Cooper andL. Petrakis, New York, 1981), AIPConference Pr oceedings No. 70.

Raymond, R. , and R. Gooley, 1978, inScanning Electron Microscopy, Vol. 1, SEM

Inc. , AMF O'Hare ( Chicago), IL 50666,USA, 1978, p. 93.

Raymond, R. , Jr. , and R. Gooley, 1979, inAnalyt ical Methods for Coal and CoalProduat s, ed; Clarence Karr, Jr. , Vol. 3,Ch. 49, Aaademi c Pr ess.

Retcofsky, H. L. , 1977, Appl. Spectroscopy31, 116.

Retcofsky, H. , 1990, pr ivate communiaation.Retcof sky, H. , M. R. Hough, and R. B.

Cl arkson, 1 979, American ChemicalSoaiet. y, Div i s ion o f Fu e 1 Chemi stry,Preprints 24, No. l, 83 (1979).

Roakley, M. G. , 1990a, Appl . Spect. 34, 405.Rockley, M. G. , 1979, Chem. Phys. Lett, . 69,

455.

Rockley, M. G. , 1980b, Proc. IEEEUltrasonics Symposium, Boston, Nov.1980.

Rockley, M. G. , and J. P. Devlin, 1980,Appl. Spectroscopy 34, 407.

Roy, M. M. , 1957, Fuel 36, 249.Royce, B. S. H. , Y. C. Teng, and J. Enns,

1980, Proc. IEEE Ultrasonics Symposium,Boston, Nov. 5 7, 1990.

St . John, G. A. , S. E. But, t r i 11, Jr . , M.Anbar, 1979, ACS Symposium Series 71,223.

Scann in g Electron Microscopy, SEN Inc. , AMF

O'Har e (Chicago). Published year ly.Sehliak, S, P. A. Narayana and L. Kevan,

1979, J. Am. Chem. Soe. 100, 3322.Schmidt, P. M. , M. Kalliat and C. Y. Krak,

1991, in Chemist, ry and Physics of CoalUtilization —1980 (APS, Morgantown)(edited by B. R. Cooper and L. Pet, rakis,New York 1981), AIP ConferenceProceedings No. 70.

Schuyer, J. , and D. ttit. van Kr evelen, 1959,Fuel 33, 176.

Sharkey, A. G. , Jr. , J. L. Shultz, and R. A.Friedel. 1964, Nature 202, 988.

Shinn, J. H. , and T. Ver muelen, 1979, inScanning Electron Microscopy, Vol 1, SEHInc. , AMF O'Hare (Chicago), IL 60666,USA, p. 487.

Sibener, S. J. , R. J. Buss, P. Casavecchia,T. Hirooka, and Y. T. Lee, 1980, J. Chem.Phys. 72, 4341.

Siegbahn, S. et a1. , 1967, ESCA, Almquist andMiksells Bokt rycker i, AB, Uppsala,Sweden.

Si lbernagle, B. G. , L. B. Ebert;, R. H.Saholosber g, and R. B. Long, 1990, ACSAdvances in Chemistry series, Vol. 92,Chapter 3-.

Sinfelt, J. H. , and G. H. Via, 1979, J.Chem. Phys. 68, 2009.

Smith, J. A. S. , 1980, J. Molec. Str ucture59', 1.

Solomon, P. R. , 1991, in Chemi st ry andPhysics of Coal Utilization — 1980 (APS,Morgantown) (edited by B. R. Cooper andL. Petrakis, New Yor k, 1981), AIPConference Proceedings No. 70.

Somor jai, G. A. , 1979, Surface Science 99,496.

Somor jai, G. A. , and M. A. van Hove, 1979,Adsorbed Monolayers on Solid Sur faces inStr ucture and Bonding Vol. 38,Spr inger -Ver lag, Ber 1in .

Speight, J. G. , 1971, Appl. Spect. Rev. 5,211.



Stanton, S. M. , and R. B. Finkelman, 1979,in Scanning Electron Microscopy, Vol. 1,SEM Inc. ANF O'Har e (Chicago), IL 60666,USA, p. 456.

Steinnes, E. , 1979, in Analytical Methodsfor Coal and Coal Products, ed. Clar enceKarr, Jr . , Academic Press, Vol. 3, Ch.W6.

Stohr, J. , D. Denleyk, and P. Pe r f e t t i,1978, Phys. Rev. 8 18, 4132.

St, rehlow, R. A. , L. A. Harris and C. S.Yust, 1978, Fuel 57, 185.

Sumino, K. , R. Yamamoto, F. Hatayama, S.Kitamur a, H. Itoh, 1980, Anal. Chem. 52,1O64.

Teichmuller, M. and R. Teichmuller, 1966,Coal Science, 55, Advances in Chemistr ySeries, R. F. Gould, ed. , Washington,D. C. , Ameri. can Chem. Soc. , 133-155.

Thiel, P.. A. , M. H. Meinberg and J. T.Yates, Jr. , 1979, J. Chem. Phys. 71, 143.

Thompson, R. R. , 1981, in Chemistr y andPhysics of Coal Utilization -1980 (APS,Mor gantown) (edited by B. R. Cooper andL. Petr akis, New York, 1981), AIPConference Proceedings No. 70.

Ting, F. T. C. , 1978, in Analytical Methodsfor Coal and Coal Products, ed. ClarenceKar r, Jr. , Academic Press, Vol. 1, Ch. 1.

Tsukashima, H. , 1967, Fuel, 46, 177.van Krevelen, D. M. , H. A. G. Chermin and J.

Schuyer, 1959, Fuel 38 483.van Krevelen, D. M. , and J. Schuyer, 1957,

in Coal Science, Elsevier.Van Oostrom, A. , Sur face Sci. , 89 615

( &979) ~

van Vueht, H. A. , B. J. Rietvald, and D. W.van Krevelen, 1955, Fuel 34, 50.

Vidal, C. R. , 1980, Opt. Lett, . 5, 158.Volborth, A. , 1979a, in Analyt ical Methods

for Coal- and Coal Pr oducts, ed. Clar enceKar r, Jr . , Ac ademi c Press, Vol. 3. , Ch.

Volborth, A. , 1 979b, op. c i t . , Vol. 3, Ch.55-

West, P. R. , and S. E. Cannon, 1981, inChemistry and Physics of Coal Utili zation

1980 (APS, Nor gantown) edited by B. R.Cooper and L. Petrakis, New York), AIPCon fer ence Pr oceedings No. 70.

Wittmaack, K. , 1979, Sur face Science 89,668.

Mong. J. , D. H. Maylotte, F. W. Lytle, R. G.Gr eegor and R. L. St. Pet er s, 1981, Bul l.Am. Phys. Soc. 26, 320 (1981)~

Wyatt, P. J. , 1980, Appl. Opt. 19, 975-Ziegler, L. D. , and B. Hudson, 1981, J.

Chem. Phys . 70, 982.


IV. RESEARCH FOR THE TECHNOLOGIES

I V. A. PRIMER ON COAL UTILIZATIONTEC HNOLOG I ES

Empirically based coal technology has along history (Lowry, 1963; Elliott, 1981); itsindustr ial development was well advanced bymid-19th century. In r ecent times the drivefor ever higher ef ficiency in coal firedboilers and the need for petroleum substituteson a large scale have inspired efforts todevelop science-based technologies covering awider range of applications and concerns thanthe traditional ones. The task i s verydifficult —partly because of the inherentlyvariable and ill def ined nature of coalitself, partly because multiphase flows andchemical reactions at solid sur faces areintrinsic features of the technology, andpartly because the hetero-atom and mineralcontent of coal gives rise -- whatever themode of use -- to toxic or harmful residueswhich must be controlled.

Efforts in the United States to developand refine modern coal technology ar e largelyfinanced by industry and by the U. S.Department of Energy. The effort is broad inscope and of substantial size. (Fumich, 1980)Plan of presentation

This section contains brief introductorycharacter i zations o f the most importanttechnologies currently in use or underdevelopment; topics discussed include.

pyrolysisflow conditions in combustion

and conversion processesgasi ficationdirect liquefactionindir ect liquefactiondirect combustion'magneto-hydrodynamic conversion

(MHD)combined cycles and

co-generationhigh temperature fuel cellsmining and coal preparation

Pyrolysis, the general phenomenology ofgasi f ication, and flow conditions arediscussed first since they are of general andrecurrent interest. ' Discussion of these isfollowed by descr iptions of several specificprocesses.

1. PyrolysisHeating of the coal to the point of

structural transformation and par tial

decomposition is an initial step in severalimportant technologies, specifically includingdir ect combustion, MHD, and gasification. Thetransformation and partial decomposition ofcoal upon heating is termed "pyrolysis" or, invarious contexts and with differing emphases,"devolatili zat ion" or "distillation" .

When coal, below the rank of anthracite,is heated (usually in the absence of air)large amounts of gases and volat ile tars andoils are driven of f. The tars and oils arehighly aromatic in character and have astoichiometr y of approximately CH. The gasesinclude methane, hydrogen, carbon monoxide,carbon dioxide, hydrogen sulfide and ammonia.Overall, most of . the hydrogen and, t ypical ly,about 35'$ of the carbon in the original coalis incorporated in these volatiles. Theresidue is termed "char" and is a mixture ofcarbon, mineral matter, and some residualhydrocarbons.

The character of the char and the amountof volatiles driven off depend strongly bothon the original coal and, on the circumstantialdetails of the heating. (With very rapidheating rates achievable in the laboratory,for example, some coals may give off upwar d of,60% of their carbon in the volatiles. )

Changes taki. ng place in the condens edphase during pyrolysis depend markedly uponthe coal used; they may include swelling andsoftening and, in consequence, some degree ofagglomeration or "caking" and the developmentof macroscopic porosity such as is familiar incoke. In fact coke is one specialized kind ofchar, and the pr incipal dir ect industrialapplication of pyrolysis is coke making. Cokemaking for the steel industry is so arrangedthat the volatiles recovered are economicallyimportant, representing about 35/ of theprocess output value.

The coking behavior of a specific coalcan be fairly well predicted from rank andpetrography; for example coals of very highand very low rank exhibit little softening andswelling and are not suitable for coke making.Nevertheless, scientific understanding of thedetailed mechanisms involved in pyrolysis isvery meagre. (Anthony and Howard, 1976;Solomon and Colket, 1977)

2. Synthetic fuels

Direct combustion of coal, though hi ghlydeveloped and in its modern forms quiteefficient, is not very suitable for transport

Rev. Mod. Phys. , Vol. 53, N,o. 4, Part I I, October 1981


power or for other small scale operation.Considerable importance, therefor e, attachesto conversion of coal into clean gaseous andliquid fuels which can serve as substitutesfor petroleum products. That this is possiblewas shown long ago. Germany, in MMII,manufactured commer c i al quantities (smallrelative to present U. S. consumption) ofsynthetic liquid fuel from coal, as does SouthAfrica today.

Historically, new factors now give riseto a need for major improvements of theearlier technology. These factors include thescale on which it is desired to producesynthetic fuels and —partly because of scale

the intense concern for environmentaleffects. (U. S. consumption of energy is nowabout, 3.5 times what it was four decades ago. )Economically, moreover, t here is a greatdif ference between the "cost-no-object"context of a war or embargo situation and theroutine suppl'y of motor fuel for generalpublic use. Therefore, considerable effort isnow going into the cleaning up and scaling upof gasi f icat ion and lique faction technologyand making it more efficient.

In contemplating the development of thesesynthetic fuel technologies one must recognizethat cleaning the coal is a ma jor ob jective,no less important than changing its physicalstate. SRC-I, for example, i s commonlydiscussed as a "liquefaction" process, but ityields a pr oduct (having the appearance ofpitch) which is not, a liquid at ordinarytemperatures. The SRC-I output, however, i smuch lower in mineral mat ter and otherimpur ities and somewhat higher in hydr ogencontent than the starting coal . By the sametoken, many gasifiers yield a product that issuitable neither for storage, nor forshipment, , and is intended rather for immediateon-site use in some further process.

3. Gasification, importance and general features

Gasi f ication plays an essential role inmost processes for transformation or indir ectuse of coal, so it is logical to begin thediscussion with gasification which, in anycase, is at the heart of fuel gas andsubstitute natural gas product ion pr ocesses.For indirect liquefaction pr ocesses,gasification is the key step which producessynthesis gas (CO and H ), while for the

direct liquefaction processes, some kind ofgasi fier is usually relied upon to produce therequisite large quantit, ies of hydrogen. Thus,coal gasification may be addressed to any oneof three dif ferent objectives —— theproduction of CO and H& mixtures ("synthesisgas"), the production of CH& (substitutenatural gas or "SÃG"), or the pr oduction ofhydrogen.

Ec onomi cal ly, t he gasi f ier and i t sassociated auxiliary units are often the mostexpensive part of a coal .conversion plant.Gasi f ier operat, ion is poorly understood, andit seems likely that the potential exists forlarge reductions in cost and increases ine f f iciency. ' The various flow,instrumentation, and materials problems offeran excellent opportunity for physicists tocontribute importantly to a most cr itical par to f coal conversion technology.

A recent, repor t (National Academy ofSciences/National Research Council, 1977)list s and very briefly describe s 37representive processes for the production oflow- and intermediate-BTU fuel gases, groupedinto 8 major categories. All are designed toimplement much the same basic strategy:

1. Pyr olyze the coal.2. React, . the char with steam and oxygen

(or air) -- at high temper ature (800 to 900C) and, in some cases, at elevated pressur e:

C + H 0 + 3l. 38 kcal~CO + H

In addition to being highly endothermic (asnoted), this reaction exhibit s a substantialactivation barrier.

The heat, necessary for the pyrolysis andfor the carbon-steam r eaction is suppliedwithin the gasifier, in most cases, by burninga fr action of the input coal in situ; and itis this which necessitates the feeding eitherof air or of oxygen.

The kinetics of the (solid) carbonreactions with the mixture of steam and othermolecular species pr esent in the gasi f ier arepoorly understood even in the case of purecar bons. In the much more complicated case ofcoal, naturally occuring mineral constituentsplay a significant r ole in catalysis of theheterogeneous reactions. These minerals arediverse and highly dispersed within the coal(Gluskoter, 1975; Jensen, 1975) . The chemicalattack is complex involving a multi-stepprocess such as adsorption, reaction, anddesorption at pr eferred car bon sites (NcKee,197~. 1979) ~

Rev. Mod. Phys. , Yol. 53, No. 4, Part I I, October 1981

Research for the Technologies 855

A general concommitant of the carbon steamreaction is the water gas shift reaction.

CO + H 0 ~ CO + H + 9.8 3 kcal

Kinetically, the shi ft reaction usuallyis in equilibrium in the gasifier. But, ,depending upon ob ject ives, the product mix maylater be deliber ately shifted in a separatecatalytic step which employs the same shiftreaction but, under conditions appropriate to adifferent equilibrium.

Further reactions which can and -- tovarying degrees depending upon oper at in gconditions —do take place in gasifiers are.

C +CO ~ 2CO

C+ 2H ~ CH~2 ~

The essential gasi fier output in any caseis a mixture (primarily) of hydrogen andcarbon monoxide. Operating condi tionsdet, ermine the amount, of residual met, hane,carbon dioxide, and water . Methane, in mostinst, anees, is a minority constituent resultingalmost entirely from the pyrolysis step of theprocess.

Different, gasifier designs oper ating withdifferent feed coals produce a wide range ofproduct gas qualities. In most cases theseare classified as "low BTU" (100 to 200 BTUper SCF, i.e. "standard cubic foot") and"intermediate BTU" (275 t, o 500 BTU per SCF).These heating values may be compared wi thnatural gas which is primarily methane andwhich is considered "high BTU" at 900 to 1100BTU per SCF. Usually the gasifier s fed withair pr oduce gas which is dilut, ed. withatmospheric nitrogen and therefore is low BTU,while those fed wit, h oxygen produceintermediate BTU mixtures (lit, tie or no N ).

It is generally considered economica tot r an s po r t. intermediate BTU gas over distancesup to 100 miles for use as fuel in industrialfurnaces, boilers, power stations, etc. Mostapplications contemplated for low BTU gas, onthe other hand, involve "over -the-fence"utilization in some kind of integr ated or

Note, however, that if the main purpose i st, o produce met, hane, it is possible to ad justthe H /CO ratio by means of the water gasshift react, ion and to follow this by thecatalytically mediated reaction.

CO + 3H -+CH + H 0 + 49. 27 kcalaccomplishing the desired result, .

hybrid process, e.g. in fuel cells, gasturbines, etc. In such cases- obtaining achemically reactive gas or a clean gas i s theprimar y objective of gasificat, ion.

The regime in a gasifier is very complex.Elements other than carbon are present in thechar; numerous reactions take place (some evenproceeding in opposite direct, ions in di fferentparts of the reactor); and many substances arepresent in t, he product, in addition tohydrogen, carbon monoxide, methane, (and, ofcourse, unreacted wat, er and carbon dioxide).

Ash

Every kind of coal util i zation technologymust incorporate some practical means ofcoping with the ash which originates frominor ganie constitutents of the feed coal. Ithas not proven possible within the limitationsof this report to adequately discuss -- oreven to enumerate -- t, he many chemical,physical, and morphological variations of coalash and the associated range of teehn icalproblems. These problems are, nevertheless,of very great importance. For example, one ofthe defining characteristics of a gasifierconfiguration is the method of controlling ashand removing it from the reaction chamber.Reactor temperature may be kept below ashfusion point and the ash removed dry; ash maybe allowed to melt and be removed as "slag",or it may be allowed to agglomerate to somecontrolled degree, and even used as a vehiclefor heat transport.

4. Flow regimes in heterogenous reactors

In characterizing gasifiers and othercoal utilization processes one must considernot only the chemistry but also, and equallyimportant, the physical arrangements by whichthe coal and other reactant, s are heated andkept, in contact and the reaction productsr emoved . Even though the chemist, r y, in broadout 1 ine, may remai n the same, physicalarrangement s can strongly a f feet theefficiency and reliability of the process, thecomposition of the product, and -- mostimportant, perhaps -- the kinds of coal thatcan be used as feed.

Because coal is solid, its use either ind ir ect combustion or for conversion tosynthet, ic fuels requires the contacting of


S56 APS Study Group on Coal Utilization and Syn'thetic Fuel Production

solid coal with flowing gases and liquids.Thus some di scussion of gas-solid,liquid-solid, and gas-liquid-solid flow is animportant par t of any description o f coaltechnology. Multiphase flows, moreover, of fermany int er esting and challenging unsolvedproblems in classical physics. At this point,therefore, we pr esent a quick overview of thevarious gross phenomena that are exhibited inmultiphase flow and the terminology prevalentin their descript ion. Examples of thesephenomena in coal ut ili zation and conversionrecur throughout the remainder of the report.(See, especially, Section IV C. )

The flow systems in chemical andcombustion reactor s used in industry can bedivided into three principal classes . systemsin which the solid phase is fixed; systems inwhich the solid phase moves slowly; andsystems in which the solid phase is in amobile or suspended state due to the motion ofthe gas or liquid phases (fluidized beds).

Fixed beds

The s imp 1est are f ixed-bed gas-solidsystems (Smith, 1970; Carberry, 19T6) in whichgas passes continuously over stationary solidparticles or granules. Such fixed-bed systemsare most f requently encoun tered in coalconver sion technology as beds of solidgranular catalyst through which coal-derivedgases flow for conver sion to more usefulproducts. The heat and mass transportphenomena for these are better understood thanthose for any other gas-solid flow systems.

More complex are three phase systems thathave simultaneous flows of both gases andliquids thr ough a fixed bed of granularsolids. Such systems may be arranged to havethe gas and liquid flow cocurrently downwards,cocurrently upwards, or counter currently toeach other. The hydrodynamics and the heatand mass transfer considerations are differentfor each case ~

A widely used conf iguration forthree-phase systems i s tr ickl e flow in whichthe liquid flows downward through a bed ofgranular solids in the form of a thin liquidfilm and t;he gas flows in a continuous phasebetween the solid particles either cocurrentlyor countercurrently. In the most common modeof operation, gas and liquids flow cocurrentlydownward.

Mhen liquid and gas flow cocurrentlyupwards there are, of course, no thin films,

and fixed bed operation under thesecircumstances normally displays one of threeregimes: bubble flow (at, low liquid and gasrates), pulsating flow (at high gas rates),and spr ay flow. In bubbl e flow, the gas isthe dispersed, phase and the liquid iscontinuous. In pulsating flow, pulses of gasand liquids pass thr ough the reactor . Inspray flow, the gas is a continuous phase, andthe liquid is dispersed.

Moving beds

In moving-bed f1ows the gr anular solidsare allowed to move (slowly) downward undergravity. The flo«of solids in the vesselproceeds in a rod-like fashion wi th therelative positions of the granules in the bedremaining essentially unaltered. Gas flow canbe ei ther upward or downwar d, but in the caseof upward flow the velocity of the gas mustnot be so great as to expand the solid bedinto a fluidized state. The bulk density of amoving bed has been found to correspondclosely to the loosest packing arrangement forfixed beds.

Moving-bed r eactors can be arranged forthree-phase flows similar to those employedwith fixed beds. Mhen the liquid and gasesflow cocurrently upwards, pulsating and spr ayconditions can be reali zed. In some movingbed coal reactors pulsating conditions areused to keep the solid coal from forming plugsin the reactor. The moving bed is much usedin coal gasi fication r eactors and in coalhandling systems.

Suspended particles and fluidized flows

Fluidized flow (Davidson and Harr ison,19T1; Kunii and Levenspiel, 1969) is a ver yinteresting and useful flow condition known tooffer many advantages for chemical reactordesign. It has not yet been used as much asit probably would be with more detailedunder standing, lack of scienti f icunderstanding of the phenomena makes scale-upof reactors uncertain and expensive. Despitethis difficulty, fluidized flow has been usedboth in conversion reactors and in combustion,as well as for the transport ation of solids.This area ~ould appear to of fer physicists aparticularly rich field for making substantialcontributions to coal technology.

Rev. tVlod. Phys. , Vol. 53, No. 4, Part l l, October 1981

Research for the Technologies S57

(I) Cocurrent (2) CocurrentUpf low Deenf Ion

II

e

Zero NetSolids Flow(f luidized)

foal

I

I ~KI

lI

II

l

L

li 'Q

LL. Horizontol

Convoying(3) Countercurrent

Flo~ Gounter currentFlow t g) (o~

Directions of motion in Quid-partic1e systems.

FIG. IVA-1. Flo~ conditions in systems of suspendedparticles (Davidson and Harrison, 1971).

Figur e IV-A-1 (Davidson and Harr ison,1971) indicates the basic flow configurationspossible for systems of suspended particles.Figur e IV-A-2 (Davidson and Har rison, 1971)suggests some of the regimes that can exi st ina fluidized bed and def ines some of the termsassociated with the phenomena.

As shown in Figure IV-A —1, there arethr ee basic ver tical-flow configurations: (1)fluid and solids flowing cocurr ently upwar d,(2) fluid and solids flowing cocurrentlydownward, and (g) fluid flow upward withcountercurrent solid flow downward. Each ofthese three ver ti;cal-flow conditions can occurunder varying degrees of solids concentration,r epresented schematically in Figure IV-A-1.

Homogeneous gas-solids mixtures generallybelong to concentr ation ranges r eferred to asdense or disperse (dilute) . Intermediated en s i t i e s gene r al ly e xhi1 i t non-homogeneousflow regimes involving alternate slugs of gasand solids, or bubbles of gas within afluidized mass. Liquid-solid systems, on theother hand, do not usually exhibit bubbleformation or slugging; and, thus approach morenearly an ideal system in which homogeneousconcentrations from zero to the bulk densityean be maintained.

In gener al, particulate fluidizationwithout bubbling occur s with liquid-solid

OVlI

tll

bc

ParticulateDense

AggregativeL

Bubbling

'.~l'

Il~IIPil~(!

Slugging.'r+"..

~ ~

lJ~c

Le ~ 4 4~ ~0

ParticulateDilute

~ '~ ~

+ ~e~ ~ ~

~ ~

OP

IA

Vl

.uO—lAI

pCO

EA

ParticulateOQf~»" ~ ~

~ ~

~ ~

Concentrations in Quid —solids systems.

FIG. IVA-2. Regimes in a fluidized bed (Davidson andHarrison, 1971).

systems, or with gas-solid systems when thepar ticles are very fine, and in the lattercase over only a limited range of velocities.Practical application of this regime is rare.Aggregative fluidization, however, oeeur s withall other gas-solid systems and sometimeswhen the solids are of -high density -- withliquid-solid systems. Aggr egativefluidization i s char aeter i zed by gas bubblesrising through a bed of minimally fluidizedparticles. How and why bubbles form in suchfluidized-solid systems is little under stood.If bubble sizes approach the diameter of thebed, slugging will occur and one refer s to a"slugging bed. " As the gas velocity isfurther increased, the bubbling or sluggingregime breaks down into a state called"turbulent fluidi zation. " "Fast fluidi zation"is an extension of the turbulent fluidizationregime. The pr incipal distinction between thetwo is that in fast fluidization, fluidi zeddensity is str ongly dependent only uponsolids feeding rate.

Entrained flow or dilute phase pneumaticconveying is character ized by solid par ticlescompletely suspended in the gas-phase. Theflow patterns in either horizontal or verticalentrained flow are r ather complex and areaffected by the solid-to-gas ratios.

Ma jor advantages of fluidi zed bedr eactor s are that they give flexibility ofmixing, of heat r ecovery, and of temper atur econtrol. They allow the use of fine catalystparticles, which minimize the intraparticlediffusion effects. On the other hand, they



may, if not designed properly, give poorconversion due to axial mixing. Andseparation of catalyst from pr oduct mixturemay be difficult.

Table I V —A — t summar i zes t hecharacteristics of various gas-solid systemsand identif ies practical examp1 s ofapplication to coal technology. Table IV-A-2summari zes the same kind of information forgas-liquid-solid systems.

5. Some specific gasifiers

Typical gasi f ier reactor configurationsare depicted in Figures IV-A-3, IV-A-A, andIV-A-5. These ar e, respectively'. entr ainedgasifier, moving bed gasi fier, and fluidizedbed gasifer.

The entrained gasifier sprays a slurry offine coal particles suspended in oil or eater

into a chamber and ignites it in an oxygen (orair) atmosphere. This system behaves ratherlike a combustor, the coal is immediatelypyr olyzed, and the residual char reacts vi ththe steam or burns until the oxygen isconsumed. One advantage o f an entrainedgasi fer is that it operates at such hightemperature ( 1350-1550 C) that all the tar sarising from the initial devolatilization ar econverted to simple gases such as CO, CO

H20, H&, H2S and trace impurities. A

concommx. tant d isad vantage i s an exit gastemperature so high that the ash i s in amolten state ("slagging" operation); and,

' inspite of an inertial separat, ion device (a poolat, the base of the reactor), the exit gas mustbe cooled (by mixing with recycled cooled gas)to solidify molten droplets of slag thatremain. The overall efficiency of the system,moreover, is critically dependent on he4trecovery from the exit gas. (In pr actice a

TABLE IV-A-1. Flow regimes and phenomena in gas-solid systems.

Regime or Phenomenon Characteristic s Practical Examples

Fixed bed

Particulate (homogeneousfluidization

Aggregative or bubbl. ing(heterogeneous) fluidization

Turbulent fluidization

Fast fluldlzatlon

Stationary particles w/gas flow

Particles uniformly expanded, nobubbles (small particles, magneti-cally stabilized bed)

"Open" gas bubbles rising throughminimally fluidized particles

Large eddy motions, absence ofpermanent bubbles or slugs

Low density solids phase; backmixingof particle aggregates

Coal. gasification: Lurgi

Coal gasificationCoal pyrolysis: C OED

Entrained flow Dilute phase transport of particles Coal gasificationCoal combustionCoal pyrolysis

Special Cases

Moving bed Particles move downward in fixed-bedcondition

C oal gasification

Spouted bed

Slugging bed

Vertical gas jet in shallow packedbed forms stable spout, entrainsand recycles solids

Bubbles agglomerate to slugs fill-ing cross-section of bed

Coal combustion



TABLE IV-A-2. Flow regimes and phenomena in gas-liquid-solid systems.

Regime or Phenomenon Charac teristic s Practical examples

Gas-liquid-fixed-bed-solid1. Trickle bed

2. C ocurrent-upf low

Solids in fixed-bed, liquid flowsdownward in thin liquid film,gas flows in continuous phase

Bubble-flow (gas in dispersedphase, liquid in continuous phase)or pulsating-flow (both gas andliquid in pulses) or spray-flow(gas in continuous phase, liquidin dispersed phase)

Various hydroprocessingprocesses including up-grading of coal-derivedliquids.

Coal liquefaction:SYNTHQ'f L process

Fischer- Tropsch process

(Other cases: Cocurrent-downflow, countercurrent-flow, and segmented fixed-bedreactors are less used in coal technology)

Gas- liq uid- su spended- solid1. Continuous slurry or

fluidized-bedSolids in suspended conditions,

continuous flowing gas, liquidand solids

Coal liquefaction:H-coal, SHC, EDS processes

Fischer- Tropsch process

(Other cases: Agitated or non-agitated slurry reactors are less used)

COALSLURRY OXYGEN

COAL

WATER COOLED NOZZLE

RAW GAS

REFRACTORYBRICK

RECYCLE WATER

Cc.

( (

l r

ASH SLURRY oO p 0

RAW GAS TO TREATMENT

BFW

DEVOLATILIZATION

GASIFICATION

BLOW COMBUSTIONDOWN

ASH

SAT. JACKET STEAM

PROCESS STEAM

TO ASH DISPOSAL

FIG. IVA-3. Entrained gasifier. FIG. IVA-4. Moving bed gasifier.



LOCK-HOPPER

GASFEEDER

DRYCOAL

STEAM

RECYCLEGAS

0 M

0 0

RAW GAS

CYCLONE(S)

FLUIDIZEDBED

K2CO3+

COAL'

WATER

CHARQUENCH

CATALYST /SOLUTION &

recover the sensible heat of the productgases. Disadvantages include the carcinogenicnature of the tars produced (they must bedestroyed by recycling back through thegasifier), rheological problems in the bedwith high swelling American coals, and poorability to handle "fines". A related U. S.design over comes some of these meehan icaldifficulties by using an extrusion feeder andmechanical stirrers within the bed (Woodmanseeand Palmer, 'l977).

Several fluidized bed gasifiers have beenbuilt at prototype stage. One (schematicallyshown in IV-A-5) uses a 15$ K CO catalyst to2lower the reaction temperature ana to achievesever al other desir able improvements (Nahasand Gallagher, 1978) including a one-stepproduction of SNG without the need to supply apure oxygen feed.

LIFTGAS

FIG. IVA-5. Fluidized bed gasifier.

severe penalty has been accepted and st, earnpr oduced at a temperatur e of only 550 C. ) Oneengineering advantage of the system is the useof slur ry feed; coal f ines and perhaps"refinery bottoms"" can be fed in this process-- a feature per ceived by the FERblG reviewgroup (Penner, et al. , 1980) to be veryimportant. On t he other hand t he fact thatt, he aqueous slurry brings cold water directlyto the hot combustion zone lower s overallthermodynamic efficiency.

The moving bed configuration (see Fig .IV-A-0) i s well exemplified by the Lurgigasifier. Dry feed coal in sizes ranging frorq3/0" to 2" is lock hopper ed into this reactorat the top, and the gas flow is upward orcounter current. Pyrolysis takes place in thetop increment of the bed and the r esultingchar is gasified in a mixture of steam andgaseous combustion pr oduct, s from the bottom ofthe r eactor . Two advantages of thiscountercurrent moving bed system are. that, itproduces a relatively high quality gascontaining signi f icant methane and otherdirect pyrolysis products, and that becausethis product comes off at r elatively lowtemperature (q, 500 C ), i t i s not vital to

"Heavy tar-l ike residual mater i al r emainingafter purification of crudes.

In situ (underground) gasification

Another process currently under study anddevelopment is underground gasification whichinvolves the same basic chemistry as theprocesses already described but which takesplace "in situ" within the coal seam. Thisidea goes back to the mid-19th centur y and hasat tracted the interest of such notables asMendeleev, Sir william Ramsey, and Lenin(Elder, 1963).

Using wells drilled at su i t ab lelocations, . steam and air are delivered to theburning face o f the coal seam and theresultant mixture of synthesis gas, CO, andnitr ogen returned to the surface. Neither thegeometr y nor the reaction parameters, such aspressure and temperatur e, can be as r eadilycontrolled as in convent ion al pr oc esses;obser va t i on and measurement are extremelydifficult; and in place of materials problemsith reactor vessels one must, deal wi th the

variability and arbitr ariness of geologicformations.

jn situ gasi fication, however, showspotential for using coal deposits that are notrecoverable by known mining techn i que sspecially coal in steeply inc-lined seams.

Envir onmental problems may prove to be lesssevere and cost s lower than f or mor ec on ventional techniques. The process,mor cover, may prove to be mor e adaptable thanothers for really large scale ut il ization ofcoal as an energy resource. Attention ofreaders interested in underground gasi fication


Research for the Technologies

is invited to a recent extensive reviewarticle (Gregg and Edgar, 1978) ~

Molten bath gasification

Gasification may also be accompli shed byfeeding coal, steam, and oxygen (or air ) intoa bath of molten salt (or, in one instance, ofslag ) (Lowry, 1963; Elliott, 1981).Stoichiometrically, the end result is the sameas for other gasi fication processes, and manyimpur itches are r emoved by chemical reactionswithin the bath itself . There are, however,severe materials problems due to corrosion.

6. Liquefaction, general considerations

The over all national economy is heavilydependent upon liquid transportation fuels.Accordingly the conversion of coal to liquidfuel must be assigned a high priority.

As shown in Figur e IV-A-6 (Mhitehurst,1978), coal has a much lower hydrogen contentthan that characteristic of high quali tytransportation fuels like gasoline and dieseloil. For coal the mole (atomic) ratio ofhydrogen to carbon is in the neighbor hood ofO. 8, but for high quality transportation fuelsit is in the neighborhood of 2. Conversion,ther efore, requires that a large amount ofhydrogen be reacted with the coal. In thelong run, there seems little alternative butto obtain the needed hydrogen by gasificationof coal.

Liquefaction processes, themselves, fallinto two general classes -- direct andindirect. The approaches differsub'stantially. In the direct process much ofthe local molecular structure of the organic

part of the coal (i.e. those entities that,were called "sub-units" in Section III-A) ispr eserved in the product liquids. The coal isheated in the presence of a solvent (generallya coal derived liqu'id) and usually under ahydrogen atmospher e. The thermal ruptur e ofthe weaker chemical bonds in the coal givesrise to very reactive chemical intermediates(generally assumed to be free radicals).These will rapidly recombine to formrefractor y chars unless the broken bonds arepromptly saturated or "capped" by reactionwith hydrogen. The hydrogen may come directlyfrom molecular hydrogen in the reactoratmospher e oi fr om a "donor" organic moleculepresent in the solverit. Beyond the hydr ogenrequir ed immediately for stabili zation andprevention of char formation, some additionalamount must be added to bring the hydrogencontent of the liquid up to that desired for aparticular transportation fuel and to removeundesirable heteroatoms, such a s sul fur,nitrogen, and oxygen from the coal liquids.

In the indirect processes, the strategyis entirely di f ferent. The structure of thecoal is completely destroyed at the outset byg a s i f i ca t i on. The resulting syn thesi s gas i sa highly r eactive mixture and can be rathereasily converted into a var i ety o fhydrocar bons and organic oxygenates byappropriate catalytically moderated reactions.

Direct liquefaction is often argued tohave better potential thermodynamic efficiencythan indirect 1iquefac tion because i tpreserves as much as possible of the originalchemical str uctur e of the coal. Such aconclusion is by no means inescapable,however. The net sto'ichiometry for theproduction of hydrocarbon fuels from coal andwater is the same regardless of whether theindirect or direct approach is used. It can

Shale oil,

Premiumtransportationproducts

0.5 0.6 0.7 0.8 0.9 1.0 1.2 1.7 2.0

HYDROGEN/CARBON MOLE RATlO

Hydrogen/carbon ratios for various hydrocarbon sources and end products {Vhi,tehurst,1978)

F&G. Iv&-6. Hydrogen/carbon ratios for various hydrocarbon sources and end products (Whi&ehurs&, 1978).

Rev. Mod. Phys. , Vol. 53, No. 4, Pert l I, October 1981


be written as shown in the following equation

(4 + P)GHu + 2(g-u)H20 ~ (4 + u)CHg+(g-u)CO2

where the average stoichiometry of coal isgiven by CH and of the desired product byCHg. The overall enthalpy of conversion isclose to zero (Mei, 1981), and this suggeststhe possibility, at least, of a highthermodynamic conversion efficiency. But ofcourse, the equation and the enthalpy changedo not tell the whole story. Mith every knowntechnology, the conversion must be carried outin several steps -- some involving hightemperature, elevated pressure, and liquid andgas pumping. These steps involve some degreeof irreversibility. Differ ences in netthermodynamic efficiency between the variousrout es to produce f ue 1 s o f the same f inalcomposition, then, d epend upon theirreversibilities and outright energy lossesin heating, cooling, pressurizing, andpumping, characteristic of the pa rticu1 arprocess schemes being used, (including theefficiency of recovery and utilization ofexcess heat energy generated in driving theprocess kinetically) -- i.e. upon cleverengineering and approach to ideal catalysis.

a. Some important direct processes

The pr incipal dire c t processes dividelargely into three classes -- pyrolysis,solvent processes without addition ofcatalyst, and solvent processes with theaddition of catalyst.

Pyrolysis, which has air eady beendiscussed, is the simplest of the directprocesses; it results in an internaldisproportionation of the natural hydrogencontent of the coal, with the liquid fractioncarrying away the l ion ' s share of thehydrogen. Pryolysis liquids have low mineralcontent and a reduced sulfur level, ' but theytend to be unstable. Although they can beused as a boiler fuel, they have severalundesirable properties in addition to theirinstability. Pyr olysis may be carr ied outunder an atmosphere rich in hydrogen; thisprocess is called hydropryolysis and produces

~Note that this conclusion is not availablefor the free energy change which would beenormously more difficult to estimate, andwould in fact require use of information notpresently available.

a mor e stable liquid with a somewhat higherhydrogen content than that obtained in simplepyrolysis.

In the solvent process without additionof catalyst, an appropriate fraction of coalderived liquid is returned to the reactionzone . Thi s fac i l i-tates the conversion sincethe retur ned liquid acts to some degree as asolvent for the coal. It also facilitates thetransport and exchange of hydrogen so that thereactive broken bonds are hydrogenated. Theprocess is usually carried out in a highpressure hydrogen atmosphere. Some hydrogencan react with the solvent and coal toincrease the hydrogen content of the productsubstance. The two most impor tant pr ocessesof this type are Solvent Be f ined Coal-I(SRC-I) and Solvent, Refined Coal-II (SRC-II)(Sullivan, et al. , 1 979). They differ in theseverity of the processing with the resultthat tne SRC-II pr oduct has hydrogen contentincreased over that of SRC-I as shown inFigure IV-A-7. Although no catalyst i s addedto the solvent in these processes, r ecent work(Granoff and Montano, 1981) has shown thatmineral constituents of the coal play animpor tant, though uncontrolled, catalyt ic rolein the hydrogenation and hydrogen transfer.

The catalytic activity o f the miner alcontent varies from coal to coal and is weakerthan desired. Additional catalytic material,therefore, is added in many of the solventprocesses. Solvent molecules must be chosenfor their ability to transfer hydrogen to thedissolving coal and, subsequently, to bethemselves rehydrogenated under the influenceof the catalyst being used.

Two variations which combine the use ofan added catalyst and this kind of transfermechanism are possible —the solvent can berehydrogenated externally to the vessel inwhich the coal is dissolving, ' or the catalystcan be present in the reaction chambert ogether with the coal and the solvent. Oneprocess of each type i s now under developmentin the U. S.

In the Exxon Donor Solvent (EDS) pr ocess(Fant, 1978; Epperly, 1978, 1979), a hydrogentransfer solvent derived from the productliquids is added to the coal as in the SRCprocesses. The product liquids arecatalytically r ehydrogentated in a separatereactor before being fractionated and (inpart;) recycled to the main liquefier. A

simplified flow diagram of the EDS process isgiven in Figure IV A-8.

In the H-Coal process (Stein, et al. ,



EDS coke liquid products

SRC-I

H-Coal distillate(syncrude mode)

SRC-l I

EDS solvent

Hoal naphtha {eyncrude model

Arab Light crude

EDS vacuum bottoms

H~al fuel oil (fuel otl mode)

COEDH-Coaldistillate-(fuel oil mode)

EDS naphtha

0.5 0.6 0.7I

0.8 0.9 1.0I

1.2 1.3 1.4 1.5 1.8 2.0

HYDROGEN/CARBON MOLE RATIO

Hydrogen/carbon ratios for products of coal lictuefaction pilot plants, with Arab Lightcrude for comparison (adapted from Whitehurst, 1978)

FIG. IVA-7. Hydrogen/carbon ratios for products of coal liquefaction pilot plants, and Arab Light crude for com-parison (adapted from Whitehurst, 1978).

1 977, 1 978; Hydrocarbon Rese ar ch, 1978)catalyst, coal, and solvent are mixed andsimultaneously present in the reactor so thatthe solvent is continuously rehydrogenated inthe presence of the dissolving coal. A flowdiagram of the process is given in FigureIV-A-9. The coal, solvent (recycle slurry

oil), and hydrogen are r eacted in anebullating-bed reactor. The material in thereactor is maintained in a fluidizedebullating state by means of a pump whichrecirculates the liquids. Temperature of theorder of 470' and a presssure of about 3000psi are used. Shen the products ar e

HydrogenGas

ANaphtha

Feedcoal

Recyclesolvent

Coalslurry

C0~ ~VlO

47

cr

$olidsseparation

C0EDC I iquid

and gasseparation

C0~~

CO

C0\ ~VLe

U

Pump

Recycle solvent

llCResidualbottoms

1I

Water1I 8

$ol vent

Recycle coal liquef action process (Epperly, 1979)FIG. IVA-8. Flow diagram of the EDS pilot plant unit (Epperly, 1979).



Recycle hydrogen, GasJL

cleanup Productgases

Separation

Solidsreduction,

Dlstl I I able

liquid products

Fractionation

Reactor

Coal

Feedheater

Recycle slurry oil

Oxygen Steam

Make-up hydrogenPartialoxidation

I(

Ash

H-Coal liquefaction process (Hydrocarbon Research, 1918)

FIG. IVA-9. Flow diagram of 8-Coal process (Hydrocarbon Research, 1978).

subsequently fractionated, material withboiling point above 510C, the ash, and anyremaining unconverted coal are sent to agasifier to produce hydrogen needed for theprocess. The H-Coal reactor is rather complexand presents several interesting problems inmultiphase flow; a diagram ' is given in Figur eIV-A-I Q.

Early rationales for the two processeswere. in the EDS process, that the separatehydrogenation of the solvent permits a moreactive and/or expensi ve catalyst to be usedsince it need not be mixed with ash and coalparticles; and, in the case of the H-Coalprocess, on the other hand, that an intimatemixture of catalyst, coal, and coal liquidspresents a much wider spectrum of molecules onwhich the catalyst carl act.

b. Indirect processes

The synthesis gas produced by a gasifierhas higher f ree energy content than i ts

stoichiometric equivalent of hydrocarbons plusCO and H 0. Thus, it is easy to pr oducehydrocarbons, oxygenated organic compounds,Co, and H 0 from synthesis gas by a varietyof cataly ic processes with the release ofconsider able heat. The ultimate efficiency ofany system incorporating these processesdepends on the effective utili zation of thisheat. There are two general routes that havebeen commercialized for the conversion ofsyngas into high quality transportation fuels

the Fi scher -Tr opsch synthesis and theMethanol route.

In the Fi scher-Tr opsch process(Hoogendoorn, 1973), as presently used bySASOL in South Africa, the synthesis gas isconverted to a broad spectr um of hydrocarbonsand oxygenates over an iron catalyst in eithera fixed bed or an entrained fluidized bed.The hydrocarbons and oxygenates are primarilystraight chained organic molecules with fewring structures. The molecular weight rangecan be quite large, and in the fixed bedsubstantial quantitites of wax may be



pr oduced. Because of the straight chainnature of the Fischer-Tropsch product, verygood diesel fuels are readily made, butconsiderable isomerization is required to makegood gasolines. This situation is improving,however, with the introduction of new shapeselective zeolite catalysts (See Section IV-F)that are capable of modifying the initialFischer-Tropsch products to produce aromaticsand highly branched paraffins.

Considerable efforts are in progress tocarry out Fischer-Tropsch synthesis in slurryr eactors that are able to use synthesis gaseswith l'ow (unshif ted) H&/CO ratios (0.5-0. 7)(Kuo, et al. , 1978). ThPs low r atio synthesisgas can be produced by gasifiers that offerconsider able advantage in cost and efficiency.Indeed, for this reason such slurry (i.e.liquid state) reactors seem to offer the mostimmediate promise for arriving at aneconomically viable synthet ic 1iq Ui d fue 1technology. Processes are being developed forconversion of these products, with high yield(of the order 80$), into high octane gasolinesover zeolite catalysts. The flow conditionsin slurry reactor s o f fer interest ing

possibilities for multiphase flow studies byphysicists, studies which might contributesignificantly to the fur t her development andimprovement of these processes.

The other important indirect route beginswith conversion of synthesis gas into methanol(Schreiner, 1978). A fixed bed reactor isusually used wi th a Cr -Zn -Cu catalyst. Thisis a well established commercial process.Methanol itself can be used as a high qualitytr ansportation fuel. Beyond this, the r ecentdevelopment of processes for convert ingmethanol into high octane gasoline over azeolite catalyst will in'crease the importanceof the methanol synthesis in the production oftransportation fuels. (Meisel, et, al . , 1976).Reactors for conversion of methanol togasoline are being developed in both fixed andfluid bed configurations. Also, in currentdevelopments in slur ry r eactors forFischer Tr o-psch synthesis (discussed in t;hepreceding par agraph), the catalyst, s can besuppor ted on zeolites, so as to incorpor ateshape selective features into a singlereactor.

Catalyst Level IndicatorRadiation Source

Cataiyst Rddit

Outlet

Press. Diff. Indicator

Th er mowel I

P ress. Diff. Indicator

ress. Diff. Indicator

Ebullating Pump

Hydrogen

Slurry Feed

FlG. l&A-10. Ebullating bed reactor (Hydrocarbon Re-search, l 978).

7. Direct combustion

a. Pulverized coal combustion (PGC)

By far the most highly developed andwidely practiced of the energy technologiesrelat ing to coal —at present far outweighingall others in economic importance —is dir ectcombust ion in the boilers of- electric powerstations. This pul veri zed coal combustion(PCC) technology has benefited fr om sever algenerations of incr cmental impr ovement basedupon operating experience at the full scaleindustrial level and currently is the ob jectof increasingly systematic scientific study.

The amount of coal burned in a largepower station (10, 000 tons per day for a 1000NM plant) is cost;ly enough to provide a verystr ong incentive toward higher efficiency.And indeed, just pr ior to the ad vent ofenvir onmentally motivated emission contr olregul at ions, the overall ef ficiency of abase-loaded state-of -the-art plant had reachedapproximately 42/, expressed as the fractionof the enthalpy of the input coal that appearsas electrical energy at the output.

A new technical challenge was posed bythe imposition of emission controls, and muchresearch ef fort has been expended in recentyear s to develop economical systems for

Rev. IVlod. Phys. , Vol. 53, No. 4, Part. l I, October 1981


removal of particulate matter and SO2 from theflue gas of coal fired power stations. Use ofthe best methods known at pr esent entails areduction of several percent in overall plantefficiency. No commercial method of removingnitrogen oxides from stack gas has yet beendevised; measures for reduction of NOXemissions therefore, focus upon controllingthe combustion regime to r educe NO formationfrom the star t. X

One major object of investigation isthe possibility of retrof itting ex i stingoil-fired plants to burn powdered coal or,mor e pr obably, a coal-oil slur ry (See Sec.IV-E); this effort encounters major problemswith handling the ash which may be 100 timesas much as the plant was ori ginally designedto tolerate.

In a modern coal-f ir ed boi 1er the coal i sf inely pulverized ('50 mi c ron med i an ),entrained in part of the combustion air, andfed with additional air into a very largecombustion chamber through burners arrayed

around the walls. See Figure IV-A-11. Thecombustion chamber may be o f the order of 150feet high, and the burners can be as many as50 in number. The upper furnace walls arelined with water tubes.

The stable flame produced in the jet ofgas and suspended coal particles reaches at emperature of about 2000 K. Particleresidence times in the fur nace are of theorder of 0. 1 to 2 seconds. As they r i sethrough t he upper furnace hot combustionproducts cool sufficiently by radiation to thewater tube walls so that the mineral matter'(ash) re-solidifies and can be carried throughthe boiler passages without plugging or"fouling". The time-temper ature-distanceprof il e for coal par t icles in a $00 lM uti 1i tyboiler is shown in Fig. IV-A-12.

The flue gases, before being permitted toexit from the stack, - must be treated to removeash particulates (at, pr esent, this is usuallydone with electrostatic precipitators and/orfabric filters in "bag houses" ) and sulfur

ConvectiveSection

FurnaceOutletRegion

Water TubeWa I ls

(RadiantSection)

Windbox

Holler

1J

I 0

Upper' Furnace '

i EconomizerPr

J. . e. ~ I' 0 ~

E I ectrostaticPrecipitator

8urners

LowerFurnace

I

1~'

lWindbox

— Air heater"

Pul ver I zer i:

) Hopper Forced,Draf tFan

InducedDraffFan

FIG. IVA-11. Schematic of pulverized coal-fired electric utility boiler (Hardesty and Pohl, 1979).


Hesearch for the Technologies S67

3500—I

0 .05 .1 0.25 0. 5 1.0

TIME, SEC.

1.5 2. 0l

2. 5t

3.0

—2000 K

3000— OULING RANGE

2500— FOULING RANGE

2000—

r1500—1000'K

/

1000

500 ~5004K

UJ

0

CLLLICLKUJ

QUJCLCLC3

20 40 F') 0

FURNACE

80

I—

CLUJCL

100 120D I STANCES FT ~

LLI

X:LLII- UJCZ LlUJ

V)

jo

-14 0 160

LX'.

UJ

~ KODUJ

180 200 220

AIR

PREHEAT ER

CLC)

I-

cnUJ oCL

240

FIG. 1VA-12. Time-temperature-distance profile in a 500 MW coal-fired utility boiler (regimes of slagging and foul-ing are crosshatched) (Hardesty and Pohl, 1979).

dioxide. The SO removal techniques currentlymost advanced in development are "scrubbers"of one kind or another in which the flue gasesare treated with limestone sprays.

All of these treatments, of cour se,consume si gni f i cant, amounts of energy(requiring blowers and sometimes evenre-heating of the flue gas, to ma inta infurnace draft) and with a consequent loweringof overall plant efficiency.

The physics of f iltration is veryimperfectly understood and may represent, anunderexploited area of research opportunity(Billings and Milder, 1 970). In particular,electrostatic forces can play an import, antrole because of their strength, r elative togravitat, ional, thermal, or adhesive forces forparticles in the O. l to l micron range. A

potentially important, unexplained observationis that the filtration of electrically chargedparticles can pr oceed wi th only a slight,increase in pressure differences due to filter

loading. No adequate model for the processexists (Lamb and Kost;anza, 1980; Penny, 1977,1978) ~

F'ormation of nitrogen oxides i s verysensitive to the combustion regime (indif fer ent ways for NO arising fr omatmospheric N& and from fuel-bound n itrogen );a single incorr ectly ad justed bur ner canaccount for a large share of the NO emissionsfrom a boiler. (See Sect;ion IU 0 For thisand other reasons it is desirable to controlthe bur ners individually, and there is needfor inst rumentation to monit or flameconditions directly. This is discussed inSections IU 8 and IV C.

b. F luidized bed combustion

A potential alternat, ive to the highlydeveloped and widely used PCC boiler may beoffered by direct combust ion o f coal in



fluidized beds -- a new technology which iscurrently the object of much r esear ch anddevelopment ef fort. In this system crushedcoal is added to a fluidized bed of crushedlimestone or dolomite. Air for combustionenters from below through plenum chambers.The rate of air feed is ad justed so thataerodynamic f orces just neutrali zegravitational ones for the crushed solids; theindividual granules become very mobile ,'andthe bed exhibits some properties of a liquid.Since steam tubes can be situated within thefluidized bed where thermal transfer isfacilitated, the max'imum temperatur es can bekept below those in a conventional boiler(wher e initial heat tr ansfer is radiative) andther mal NO emissions should be more easilycontrolle . SO is removed by chemicalreaction with he bed as soon as it isliberated. (Generally, the amount of airneeded to fluidize the bed exceeds that whichi s use ful for combustion and this lowersoverall thermal efficiency. )

8. Some advanced concepts embodying combinedcycles

a. Pressurized fluidized bed combined cycle

A more advanced direct combustion conceptis that of the "pressurized fluid i zed bedcombined cycle" (Corman and Fox, 1976).Fluidized bed combustion, in this case, iscarried out at elevated pressure, this bringsadvantages but; al so requires substant ialenergy for the pr essur i zation. Energy isextracted from the high pr es sur e/h i ghtemper ature combustion products in two steps.First they are "let down" through a gasturbine which r ecovers enough ener gy to driveboth the compressor and an auxilliary electricgenerator. Exhaust from the gas turbine isstill at a high temperature and is used t, oraise steam for a conventional turbine. Thissystem furnishes an expmple of "combinedcycle" oper at', ion -- in which high and lowtemperature heat engines are cascaded so as tobetter utilize heat from a high temperaturesour ce and to realize a greater overallef ficiency than practical engineeringconsiderations will allo~ in a single cycle.

A ma jor technical problem her e -- andwith several other combined cycle concepts forcoal utili zation -- is that of "hot gasclean-up" . Coal combust ion and gasi fication

products are laden with ash and c or r os i vesubstances (e.g. alkali metal and vanadiumcompounds) which must be removed or reduced toinsignificant levels before being fed into"topping cycles" such as gas turbines or fuelcells. The need to carry out the clean-upwithout excessive loss o f pressure ortemperature raises a number of di f ficultengineering and materials problems. Some ofthese are discussed in Section XV D.

b. Magneto-hydrodynamic conversion "MHD"

MHD conver sion (Petrick and Yashumyatsky,1978; Jackson, 1977) is str aightforward inbasic concept and in it s basic deviceconfiguration, it does not r equire highlystressed moving parts, nor any solid parts atall which are not accessible for externalcool ing, and i t d oe s not demand themaintenance of close dimensional tolerances.For these reasons it is thought possible tobuild an NHD generator to endur e gasconditions far more severe than thosepermissible in a conventional gas tur bine, andsuch a generator should have the potentialboth for increased power handling capabilityand increased reliability. Used as the hightemperature end of a combined cycle, HHD

should af ford the improved efficiencyassociated with a higher Carnot factor. Toreali ze these hoped for advant, ages inpractice, however, will require the solutionof unusual and formidable engineering andmaterials problems. Some of the latter arediscussed in Section IV D.

In an NHD generator the extremely hotgases resulting from direct combustion of coal(with preheated air) are seeded with potassiumor other easily ionized substances to makethem electrically conductive. They are thenexpanded in a very high speed nozzle orchannel. See figure IV A 13. The channel issituated in an intense transverse magneticfield (6 Tesla from a superconduct, ing coil hasbeen proposed). Heing a high speed movingconductor in a transverse magnetic f ield, thestream of (partly) ionized combustion productsbecomes the site of an induced electromotiveforce, and electrical power can be taken offthrough a suitable set of electrodes on thechannel walls. Combustion products exitingfrom the channel are still at extremely hightemperature and are used f ir st to heat theincoming air for combustion and subsequently


h, T,chno og~esResearch or

ent ional ssteamfor a conven i smto

ssible oturbine.

for re

t the HHD

recover s

tures to4

subst, ance.seed su

2400 C an

Besides onare sever e. coel designthe channe

1 reque e1 ctricats b resort

re — -1 3

of metals,ultaneous reqthese simu

ories.1nnsulatorn . s, an

erature fuel ce lls

proveHigh tempera uC.

A prom s

op

yd t 1

—eneratxon

contenthat thehe

r1

conservesynthetic

adily be conve energy zsthis free eobtainehydrogen

1 cells. Theghe 1

assmise 1 tro

thermal, so

fficiency, mesensitxv

11 sys edecentra lized

i s

a table for modu

1 the

highly adapta

are currentlyion.

Two kin sects of ssub s

orle aobjec

1 rbona ee

and mol eac&

rt aal con

s

the exding 1S un

carbon e lectrodeurban settxng

disperseth finelyThey consume

adversely acan be a vmonox 1.de, syn

erati gacid cell's.190 C whicabout

convenienloH gra de e

ric acid cePhosphosulfu

s' ', corrosi or educ

sintering,ism 0mechani

WATER PASSAGE

WATEROUT

MAGNETIFI ELD

INSULA

.WATER

REINFORCEDPLASTIC

0'0, DIFFUSSE'R

IN

FLAMEHOLDER

SALivricaL wsuLaroexeuearsoecour COMBUST ION CHAMBER— —SEED

'c MHD channe .el.- 3. SchematicFIG. IVA-1

I I, October 1981Vol. 6;3 No. 4, Part II, cRev.. Mod. Phys. ,


(relatively inefficient) cathode.Molten carbonate cells operate at much

higher temperatures (approximately 650 C).They are, therefore, suitable for integrationwith conventional steam turbine bot tomingcycles in combined cycle conf igurationspromising quite high overall efficiencies (50%is thought possible). These cells are capableof utili zing the carbon monoxide content ofsynthesi s gas but at present are qu i tesensiti ve to sulfur poisoning.Thermodynamically, it ~ould be attractive tofeed the high temperature product of a coalg a s i f i e r through a hot gas elean-up stage andthence directly to molt en carbonate fuelcells. At pr esent, ho~ever, hot gas clean-upt echnology has not advanced to a level which~ould permit this -- nor can molten carbonatecells tolerate dirty gas. Even more seriousare problems relating to the service life ofmolten car bonate fuel cel 1 components andtrade-of fs between li fetime, e ff i c i ency, andcost.

For further in formation the interestedreader may find it useful to consult: Fickett,1978; Nat ional Fuel Cell Seminar, 1980; andBrowall and Mazandarany, 1980.

9. Mining

10. Coal preparation

Coal -- after it is mined and before itis bur ned, coked or f ed into gasi fiersgenerally receives some preparatory treatment.Grinding, sizing, and washing are common.Flotation methods are used to process about40$ of the coal production. These depend uponspecific gravity di fferences or wet tabilitydifferences between mineral matter and thecarbonaceous components of the coal as mined.

Some treatments ar e designed to reduceunwanted mineral ma t ter, e special. ly inorganicsulfur associated with the ash. Of particularinterest is high-gradient magnetic separationfor coal desulfur ization (Maxwell, 1978).(Also see Section IV-E. ) Others are intendedto control caking behavior, etc. Sometimescoals are blended to reduce variability andbring them to some reduced sulfur content orto achieve a better coke by combining andaveraging the peculiar characteristics ofseveral coals.

For some processes it would beadvantageous to reduce coal to extremely fineparticles (a few microns) if that could bedone economically. The Study Group gaveconsiderable attention to this question of"comminution" or grinding, and it is discussedin Section IV-E of this report.

i

Coal technology begins with mining whicheur r en t1y i s .c la s s i f ied into undergroundmining and sur face (open cast or strip)mining. Primary concerns here are safety,efficiency, and the environmental effects.The technical challenges lie mostly in theprovinces of mining engineering, civil andmechanical engineer ing, operations analysis,and industrial planning. The Study Groupdevoted relatively little attention to miningbut did note several problems that might besolved by improved physical instrumentation.Mell known, of course, is the need to detectand mon i tor toxic and explosive gases. Otherexamples may be found in the need of a"continuous mining" machine to detect largeinclusions of shale or other mineral matterwithin the coal seam and to sense the exactlocation of the coal-shale inter face at theupper or lower boundary of the seam. Thereader interested in mining and preparation ofcoal ean find a general survey in the KeystoneCoal Industry Manual compiled by MiningInformation Services and publi shed byMc Graw-Hi 11 .

References

Anthony, D. .B. and J. B. Howard, 1976, AIChEJ. , 22, 1976, 625.

Billings, C. E. and J. Milder, 1970,Handbook of Fabric Filter Technology, GCAInc. , Bedford, MA.

Browal 1, K. M. and Farhad N. Ma zandarany,1980, Development of Molten Carbona t eFuel Cells for Power Generation, 15thIntersoeiety Energy ConversionEngineer in g Con fer ence, Se a t tl e,Ma shi n gton, August 18-20, 1980.(Available fr om General Elect, ric Corp.Research and Development as Repr int118905. )

Carberry, J. J. , 1976, Chemical andCatalytic React ion Engineering,Me Graw-Hi 11, New York.

Davidson, J. F. and D. Harrison, Editors,1971, Fluidi zation, Academia Press,London.

Elder, J. , 1 9 63, "Th e Under groundGa si ficat ion of Coa 1, '1 Chemi st r y o f Co a 1Uti 1 i zat ion, H. H. Lowry, Ed. ,



Supplementary Vol. 3, John Wiley Sons,Inc . , New York, pg. 1023.

Elliott, M. A. , 1981, Chemistry of CoalUt i 1i za t ion, 2nd Suppl ement, John WileySon s, New York .

Epperly, tht. R. , 1978, EDS Coal LiquefactionProcess Deve lopmen t, Phase 3-B/4. Annua 1Technical Pr ogress Report. washington,D. C. U. S. Department of Energy( FE-2873- & 7) ~

Epperly, M. R. , 1979, EDS Coal LiquefactionProcess Development, Pha se 4. Annu a 1Technical Progress Repor t. Washington,D. C. : U. S. Department of Energy(FE-2983-35) ~

Fant, B. T. , 1978, EDS —Coal LiquefactionPr ocess Development, Phase 3A. FinalTechnical Repor t, Vo 1 3. Mashing ton,D. C. : U. S. Department of Ener gy(FE-2353-20) ~

Ficket t, Arnold, 1978, "Fuel Cell PowerPlants. " Scientific American, December,1978, p. 7o-

Fumich, George, Jr. , 1980, Statement of theAssistant Secretary for Fossil Energy,U. S. Department of Energy, before theSubcommittee on Interior and RelatedAgenc i e s o f the Commi ttee onAppropriations, U. S. Senate, May 8, 1980.

Gluskoter, J. S. , 1975, Adv. Chem. Ser. ,

Granof f, B. and P. A. Montano, 1981, pp.291:398 in Chemi stry and Physics of CoalUtilization -1980, AIP Con ferencePr oceedings No. 70, edited by B. R.Cooper and L. Petrakis.

Gr egg, D. M. and T. F. - Ed gar, 1978,"Underground Coal Gasi f icat ion, " AIChEJ. , 24, 753.

Hardesty, D. R. and J. H. Pohl, 1979, SandiaLaborat or ies Repor t 78-8804.

Hoogendoorn, J.D. , 1973, Experience withFi scher -Tr opsch Syn the si s at Sasol.Clean Fuels f rom Coal Symposium Papers,pp ~ 3&3-365. Chicago, Ill. : Institute ofGas Technology (CFFC-1) .

Hydrocarbon Research, Inc. , 1978, H-CoalIntegrated Pilot Plant. Final Report,Vol. - 1. Electric Power ResearchInsitutue. Palo Alto, Cali f. : ResearchReport Center (EPRI-AF —681, ResearchPro ject 238-1).

Jackson, william D. , 1978, MHD ElectricalPower Generation+ Program Status Report.Scientific Problems of Coal Uti li zation,e d . by B. R. Cooper . TechnicalIn formtion Cente r, U. S. Dept. of Ener gy,

p. 113.Jensen, G. A. , 1975, Ind. Eng. Chem. Proc.

Dev. , 14, 308.Kunii, D. and 0. Levenspiel, 1969,

Fluidi zation Engin eer ing, John WileySon s, Inc . , New York .

Kuo, J. C. ltd. , R. Sh innar, J. Me i, and C. T.D. Mu, 1978, Gasifier Study for MobilCo al to Gasoline Pr ocesses. FinalReport. washington, D. C. : U. S.Department of Energy (FE-2766-13).

Lamb, G. E. R. , and P. A. Kostanza, 1980,Text i le Res. J; 50, 661.

Lowry, H. H. , 1963, Chemi stry o f CoalUti 1 i zat ion, Supplement . John RileySon s, New York .

Maxwell. , E. , 1978, pp. 90 112 in Scienti f icProblems of Coal Utilization, Edited byB. R. Cooper, Technical Infor mationCenter, U. S. Dept. of Energy (availableas CONF — 770509 from National TechnicalInformation Service, Spr ingfield, VA

22161) .McKee, D. M. , 1974, Carbon, 12, 453.McKee, D. tIIt. , 1979, Chemistry and Physics of

Carbon, pg. 16, Marcel Dekker, New York,1979-

Meisel, S. L. , J. P. McCullough, C. H.Lechthaler, and P. B. Meisz, 1976,Chem Tech. , 6, 86-89.

Mining Information Services. Keys tone CoalIndustry Manual. McGraw-Hi 11, New York.

Nahas, N. C. and S. E. Gallagher, Jr . , 1978,1 3th IEC ED, 3, 21 43.

National Fuel Seminar Abstracts, July 14-16,1980, San Diego, Ca. (Copies availableDQE and EPRI)

Nat ional Academy o f Sc iences/Na t ionalResearch Council, 1977, Assessment o fLow- and Intermediate-BTU Gasificat;ion ofCoal, Report FE/1216-4 available fromNational Technical Information Service.

Penner, S. B. (Chairman, Fossil EnergyResearch Working Group, FERWG), 1980,"Assessment of Long-Term Research Needsfor Coal -Li que fa ct i on Technologies, " DOE

Contract DE-AC01-79K R 10007.Penney, G. M. , 1977, Powder Separation 18,

111.Penney, G. 4f. , 1978, "Electr ostatic Effects

in Fabric Filtrat ion", EPA Publ.yE PA-600/7-78-1 42a.

Petrick, M. and B. Yashumyatsky, 1978,Open-Cyc le Ma g net ohydrodynamic Electr icPower Generation, edited by M. Petrickand B. Yashumtatsky. Ar gonne NationalLaboratory Techn ical Repor t DOE-PR-119.



Schr e incr, M. , 1978, Research GuidanceStudies to Assess Gasoline from Coal byMethanol-to-Gasoline and Sasol-TypeFischer-Tropsch Technologies.washington, D. C. : U. S. Department ofEnergy (EF-2447-13) .

Smith, J. M. , 1970, Chemical EngineeringKinetics, Nc Gr aw-Hi 11, New York .

Solomon, P. R. and M. B. Colket, 1977, "CoalDevolati 1i zat ion, " 16th Inter nation a 1Symposium on Combustion, pg. 131.

Stein, T. R. , J. G. Bendoraitis, A. V.Cabal, R. B. Ca 1 len, M J-. Dabkowski, R.H. Heck, H. R. Ireland, and C. A.Simpson, 1977, Upgrading of Coal Liquidsfor Use as Power Generation Fuels.Electric Power Research Institute. PaloAlto, Calif. : Research Report Center(EPRI-AF-444, Research Pro ject 361-2) .

Stein, T. R. , A. V. Cabal, R. B. Callen, M.J. Dabkowski, R. H. Heck, C. A. Simpson,and S. S. Shih, 1978, Upgrading of Coal

Liquids for Use as Power Generat ionFuels. Electric Power Re sear chInsti tute. Palo Alto, Calif. : ResearchReport Center {EPRI-AF-873) .

Sullivan, R. F. , B. E. Stangeland, and H. A.Frumkin, 1979, Re f ining of SRC-I andSRC-II. Paper presented at Re f ining ofSyncrudes Contractors Conference,Richmond, Cali f . , May 7 —8, 1979,sponsored by U. S. Department of Energy.

Hei, J. , 1981, Ind. Eng. Chem. Proceedings(to be published).

ttIth i tehur s t, D. D. , 1978, American ChemicalSociety Symposium 71: Or gani'e Chemistryof Coal, pp. 1-3g. washington, D. C. :American Chemical Society.

Moodmansee, D. E. and P. M. Palmer, 1977,ACS Symposium on Coal Gasi f icationKinetics, 22 (1), No. 1, 158.

Yoon, H. , J. Mei, and N. H. Denn, 1978,AICh E J. , 24, 88&.

Rev. Mod. Phys. , Vol. 53, No. 4, Part ll, October 198'l


IV. B. INSTRUMENTATION AND GONTROL

1. Introduction

Coal utili zation technology stands inur gent need of improved instrumention -- aneed rooted in considerations both of processcontrol and of scientific understanding.

Process contr ol r equi res continuousreal-time or near real-time measurement or"monitoring" whereas scientific understandingis as well served by measurements yieldingtime-dependent, but not i n s tan taneous, data.Cont;rois of coal processing plants must beadequate to insur e. safety, reliability,"compliance" (with environmental regulations),economy and ef ficiency, and adaptability tovariations of output demand and feed-coalproper ties. Requirements in each of thesecategor ies are very exact ing. Safety, forexample, must be assured for plants whichpr ocess enormous tonnages of highly reactivematerials at pressures up to several hundredatmospheres and temperatures up to 2000K; thepotential for fires or explosions endangeringpersonnel and causing damage to major capitalfacilities is ever present. Likewise,"compliance" may require better than 99%effective removal from plant effluents ofnaturally arising pollutants.

Scientific understanding, which isprerequisite to sophisticated plant control,has greater need for preci sion andcompleteness and less need for instr umentaldurabi 1 i ty and immediate availability ofresults than is the case for practical processcontrol. The minimum scientific goal must; bea level of understanding suf f icient to makegood dynamic -models -- models reliable bothfor scale-up of pilot plant processes and forthe design and computer assisted testing ofcontrol systems. Present knowledge does notsuf fice for successful modeling of theresponses of existing pilot plant facilitieseven within the physically attainable andobserved range of operating conditions —anda practical control system must be trustworthyin serious off-normal excursions as well as instart up, shut down, and steady operatingsituations.

Dynamic models, which are based on datafrom diagnostic instrumentation, make possiblea more intelligent choice of instruments forprocess control by quantifying requirements onthese instruments. In addi tion, t;heinformation gained from the models may ease

measurement problems, by permitting the use ofinstruments (which may be already available orat least easier to develop) to measure otherprocess parameters related (through -the model)to the parameter to be controlled.

A measurement intensive and modelingintensive approach to process improvement andscale-up appears necessary i f the desiredt;imetable for coal technology development i sto be realized. Precedents do exist for thesuccess of a trial and error approach toindustrial process chemistry, even forprocesses (e.g. , oil refinery catalyticcrackers) with severe operating conditions.These successes, however, r e suited f romdevelopment extending over several decades ata time when energy was cheap and computersunavailable. Now energy is expensi ve,construction cost;s are astronomical, and theprojected need is for coal utilizationprocesses operating on a large scale and inlarge numbers within ten or fifteen years.Mistakes should not be cast in concrete andsteel where there is any chance that testingof processes and their control can be donemore quickly and cheaply wi th the aid ofcomputers.

a. General nature of the difficulty

Instrumentation of coal pr oce s se spr e sent s un usual d i f f iculties. Advancedprocesses -- 1 i que faction, ga'si fication,fluid i zed-bed combu st;i on, andmagnetohydrodynamics —— involve thetransformation of large amounts of solid coal(crushed to median sizes as small as 40 um) ininteraction wi th other substances attemperatures up to 2000 K and pressures up toa few hundred atmospheres. The particles canbe highly erosive, the process streams ofcrushed solids are sub ject to plugging; andthe liquid products may solidify if they cooldown to normal ambient; temperat ures. Thechemical environment is ext;remely harshwith both oxidizing and reducing conditionssometimes present in different parts pf thesame reaction vessel, molten slag be ingp r od u c e d, and' hydrogen, sul fur, chlorine,sodium, potassium and other harmful speciesunavoidably pr esent. Access o finstrumentat ion to the process material,protection of sensors, and prevention ofplugging are serious problems. In no otherprocessing industry, perhaps, does oneencounter such a range of impediments to



measurement all at the same time.The foregoing are difficulties associated

primarily with design of sensors to penetrateinto or survive the host ile regimes of thefossil energy conversion processes. Otherdifficulties arise with respect to the controlelements (valves, pumps, etc. ) which are to beregulated in accordance with the measurements.Control elements which can operate reliablyand wi th acceptable service life are lacking,some of the associated materials problems arediscussed elsewhere in the Report. Thecontrol schemes which will permit, reliablephysical integration of unit operations withdifferent time constants to produce aspecified output, have yet to be conceived~.

b. Plan of presentation

This section of the report is devoted to asomewhat deta i 1 ed assessment of threeespecially important instrumentation problemsin coal technology. 'multiphase flowmeasurement, par ticul at e mon i to ring, andt emp e r a t ur e measurement. For each there i s areview o f presen t me thods and t heirdifficulties and shortcomings, and a briefdiscussion of some possible new techniqueswhich might be use f ul for making themeasurement s. Following this dis cu s s i on,ther e is provided an il lu str at i vetechnique-ori ented section in which possibleapplications of acoustic methods are suggestedand explored. Before addressing the maintopics, to indicate the scope of the problems,we present brief comments on a few otherinstrumentat, ion n eed s o f possible physicsinterest. Me do not provide a complete orexhaustive catalogue of important, needs. Forthis the reader should consult O'Fallon et al,19S1.

2. Instrumentation and control n"Ms and problems

The process par

arne ters

r equi ringmeasurement and control in coal uti 1i zationprocesses parallel those required in ot herchemical processes, but the conventionalinstruments used in the chemical processing

~For example, in combined cycle electricgenerating plants it may be necessary tofollow changing loads while using a feedstockof varying quality to produce an intermediatefuel which is uneconomical to store.

industry generally fail in the most hostileareas of the coal processes.

a. Level measurement

Reliable measurement of fluidi zed-bedlevels at, temperatures up to 1400K andpressures near 1000 psi is needed, as is alsomeasurement of the level of crushed solids inhoppers. Reliable cont, rol of fluidization andbed level could reduce costs by permittingreduction of freeboard above fluidized beds.In many processes, reliable feed controlrequires measurement of hopper levels.

Fluidized bed depths are now inferredfrom di f ferential pressure measurements.Plugging of the pressure taps presents aproblem which can be alleviated by properdesign of a purging system. Measurement by amore direct method would be preferable. Somehopper level instruments based on vibratingelements, eapaeitance, and acoustic sensingare available.

b. In-situ analysis of solid and/or liquid

process streams

Sulfur, ash, heating value, moisture, andvolatiles content of dense phase streams mustbe monitored for process control in coal-usingplants. The streams may be coal on a conveyorbelt, coal within a pipe, or other solid orliquid streams such as mineral residue or charslurried with water. In many cases the neededinformation ean be inferred from a knowledgeof the elemental analysis of the material.Sulfur, moisture, and HTU content areparticularly import ant parameters of feedcoals. Carbon/ash ratio in residues is animportant indication of process ef ficiency.

Sarnpl i n g f o 1 1 owed by conventionalanalysis ean, in principle, give the necessaryinformation, but not quickly enough for use inprocess control. In addition, sampling isdifficult to achieve with a high temperature,high pressure system, 'and when achievable, ismade representati ve only with di fficulty.Moreover, the composition of whatever samplei s obtained will vary with changes intemperature and pressure in the process andassociated sampling apparatus. For processcontrol, an in-situ technique yielding result, sin minutes or less is needed.

A continuous on-stream automatic sulfur

Rev. Mod. Phys. , Vol. 53, No. 4, Part Il, October 1981


meter has been developed (Stewart et, al, 1974;Bozorgmanesh et al, 1980) . It, is based onanalysis of prompt gamma rays followingcapture of (thermalized) neutrons and hassuccessfully monitored the sulfur content ofcoal feed streams. Multi-element mon itor s forcoal feed streams also have been developed(Gozani et al, 1979; Cekorich, 1979). Theseare currently undergoing tests in coalpreparation and blending plants for a powerstat, ion. Some work has been done on analysisof gamma rays from inelastic neutronscatter ing using neut rons from a 14 MeU.generator (Her zenberg, 1979). Oxygen, whichis not detected with thermal neutron capture,and also carbon respond well to thistechnique. Large angle elastic scattering offew MeV neutrons also shows promise foridenti fieation of important elements in coalstreams (Smith, 1981).

In addi tion, production of short -livedr adio isotopes can yield informat ion onelement, al concentrations and can permit flowvelocity measurement by detection of themoving radioactive tag down stream f rom apulsed neutron source. Other work isconcentrating on extending all of the aboveneutron-based techniques to flows in pi pe s a swell as to transport, on open conveyor belts orin hoppers (Herzenberg et al, 1979). Moisturedetermination by microwave, capacitance, andneutron thermali zation techniques i s beinginvestigated (Brown et al, 1979).

c. Gas analysis

Measurement of molecular composition ofgas and detect, ion of harmful species areneeded at temperatur es up to 2000K or higherand pressures of 1000 psi -- with responset, imes near one second. "Harmful species"include Na, K, and Cl which are injurious toequipment, as well as other environmentallyhazardous substances.

Gasification plants need rapid molecularcomposition measurements of the product gasfor process control. Pressur i zedfluidi zed-bed combustion (PFBC) and ot, hercombined cycle systems in which gas turbinesare used require monitoring of alkali content,of the gas for protection of the turbine,which can be severely damaged in a matter ofseconds (Herzenberg et al, 1981).

High temperatures and pr essur es makedirect access to the gas stream difficult.Sampling is too slow i n many cases -- and

probably unrel i able a s wel 1 —— sinceconst, ituents may condense or undergo chemicalreactions while the gas is being withdrawn andcooled.

Gas ehromatographs have been used formolecular analysis, but suf fer the same kindof sampling problems. In addi tion, themeasurement generally consumes several minutesor more. A mass spectrogr aph system hasrecently become available and has been testedwith encouraging results (EPRI, 1979). It,too, is subject to sampling bias, but it has ashort, response time.

A prototype in-situ infrared gas analysissystem -has been tested with some success onthe product stream of an ent, rained-bedgasifier (Zweibaum, 1978). Other current workis directed to in-situ laser based detectionof gas species, particularly of trace elements(Sullivan and Jensen, 1 981). An isothermalgas sampler has been developed and is awaitingtests on a fixed-bed gasifier (Bump, 1979).

d. Pressure

Conventional pressure measurementsinvolve the introduct, ion of pressure taps intothe system. If the taps become clogged withprocess material, the measurement cannot bemade reliably. One way to allevi ate t hi sproblem i s to pur ge the taps; this, however,may undesirably alter the composition of theprocess stream. Another alternative is toisolat, e the process stream from the pressuresensing line by a movable diaphragm, thispresents problems if the process material isabrasive or corrosive, in which case it mayattack the diaphragm material.

ln fact, the great weakness of head-typeflowmeters (those which measure flow velocityby sensing the pressure differential across arestriction) is that the required pressuremeasurements cannot be made satisfactorily.

3. Multiphase mass flow measurement

Because coal is a solid, t, he task ofaccurately measuring and controlling coal feedinto reactors and combustors is greatlycomplicated. In fact, improved devices formeasuring the flow of solids (usually inliquid slurries or gas entrainment) must beregarded as possibly the most critical andwidely shared instrumentation need of t, headvanced coal technologies now under


APS Study Group on Goal Utilization and Synthetic Fuel Production

development. As has been indicated in SectionIV A of t hi s report, gasi f ication,liquefaction, pr essur i zed flu id i zed bedcombustion, and magnetohydrodynamic powergeneration all require the feeding of crushedor pulverized coal —usually into a reactionvessel at high pressure. There may also beinternal st, reams cont a in ing solids whichrequire monitoring and control.

The tempera tur es and pr essur esencountered may range to 2000K and to a fewhundred atmospher e s . The ma ter ial flowsinvolved may be typically 200 to 2, 000 tonsper hour for demonstration or commercialplants, with pilot plants smaller by a factorof about, 100. Any device inserted into theseprocess streams is exposed not only to thetemperature and pressure of the stream, butalso to the serious ha zard of eros ioncombined, in many cases, with corrosion. Anydevice that obst, ructs the flow, moreover, maycarry with it the threat of plugging. For allof these reasons non-invasive methods ofmeasurement are highly desirable. Finally, insome applications extremely short responsetimes are needed -- as short as millisecondsfor NHD.

For control purposes, the con tr olquantity of interest, is the mass flow rate,pvA; where p is the density, v the flow speed,and A the pipe cross sect ional area. Sincefew sensing techniques respond to mass flowdirectly, it is generally necessary to combineseparate measurements of velocity and density-for a calculated mass flow determinat, ion . Bythe same token, with a multi-phase stream, theproportions of individual constituents mustsomehow be determined. For dry solids atatmospheric pressure, mass flow rate can besatisfactorily measured by a weighing system—either a weighed conveyor ("gravimetricfeeder") in which the rate of loading and beltspeed are controlled to deliver solids at thedesired rate, or a weighed hopper in which therate of weight change is monitored. But if agravimetric feeder is to be used with apressurized system, there is likely to besubstantial and variable feedstock inventorybetween the point of measurement and the pointof injection into the reactor proper. Underthese conditions the measurement may be usefulonly for giving an average over an hour orlonger. The weighed hopper is limited inaccuracy by the tare weight relative to theweight of solids in the tank. In apressurized or hot, system, moreover, it isdi fficult to adequately control and compensate

for forces exer ted on the tank by connectingequipment. In practice these considerations,as well as difficulties due to vibrations,limit the usefulness of the weighed hopper toaverages over an hour or more'.

Positive displacement meters or pumpsprovide a means for absolute monitoring ofvolume flow. They are very accurate for eleanliquids, but the accuracy deteriorates rapidlyin the presence of abrasi ve solids, whichchange the calibration and cause leakage atthe seals. A positive displacement pump oftencan be used to provide a good approximateindication of the flow rate.

Existing devices for fluid flow velocitymeasurement in the chemical process industryare mostly "head-type" meters -- i.e. theymeasure the fluid pressure di fferentialsassociated with presence in the flow path ofsome type of obstacle or venturi. Elementarysingle phase flow theory (Bernoulli 'sPrinciple) predicts that the volumetric flowwill be proportional to the square root of thepr e s sur e d i f fer en t i a l. In . practical systems(multiphase, possibly turbulent flow), it isfound that the ratio of volumetric flow to thesquare root of the pressure differential as afunction of flow speed increases to a maximum,then decreases to a relatively constant valuefor further increase in flow speed. Thehead-type flowmeters are consider'ed to beusable only for flows in the range where theratio is relatively constant, i.e. thevolumetric flow is proportional to the squareroot of the pressure differential. Althoughvarious geometries have been explored, it hasbeen difficult to extend the usable flow rangelow enough for typical slurry applications.Mixed phase media containing solids give riseto plugging problems with the pressuremeasurement taps; and the flow modi fyingobstacle or venturi is itself sub ject to veryrapid erosion. Clearly some alternative meansof measuring slurry flow must be developed.

In the minority of cases where waterbased slurries are involved, the magnet icflowmeter is a good candidate for measuringthe water flow velocity (which will be relatedto the velocity of the solids). This devicesets up a magnetic field transverse to theflow and measures the induced ENF; it has gooddynami c range and accuracy, and isnon-intrusive. It is relatively expensive andhas not yet been developed for high pressur eor high temperature application.

An interesting new device is the Coriolisflowmeter. As indicated in Fig. IVB-1 the

Rev. IVlod. Phys. , Vol. 53, No. 4, Part ll, October 1981


instrument consists of a U-shaped tube or pipesection through which passes the flow to bemeasured. An (oscillatory) angular velocityuis imparted to the U-tube by some suitablemean s, the ax i s of rotation lying in the"plane" of the U-tube, orthogonal to its legs,and at some distance from the bend of the U.Expressed in the (non-inertial) frame rotatingwith the U-tube, the fluid flowing i n the twolegs experiences oppositely directed Coriolisforces which in turn are transmitted to thewalls of the tube. Since the tube i s notper fectly rigid it exhibits a measurabletorsional deflection fr om which mass flow canbe calculated. The device on a small scalehas undergone test s on flow of gas entrainedsolids, and the results were promising(Baucum, 1979). Further tests are indicatedfor solid-gas and solid-liquid systems. Thequestion of scale-up and erosion duringlong-term operation has yet to be addressed.

Acoustic flowmeters are beginning to beused for measurement of slurry flow, althoughcommercially available instruments are limitedto temperatures below 150C or so. Theseinstruments, being non-intrusive, ar e notlikely to cause plugging or to have the sensordamaged by the process material.

Passive acoustic instruments simplydetect the flow noise and interpret changes inthe sound as an indication that something haschanged in the system. The simplest systemsmonitor the total sound emitte~ while moresophisticated devices may perform spectral

FIG. IVB-l. Coriolis flowmeter.

analysis in which di fferent portions of thespectrum are related to various operatingconditions. There are active instruments inwhich a transmitter/receiver pair detects thedi f ference in upstream and downstreamtransmi ssion t ice . Th i s can t hen be relatedto the flow velocity.

The most common active acoustic (orultrasonic) flowmeter is the doppler shiftinstrument in which the frequency, shift ofwaves scattered from the moving particles isr elated to the particle velocity. Ultrasonicflowmeters are fairly expensive, but theirdynamic range is good. Their accuracy is afunction of the signal collection time and maybe brought, within 1'g if a collection time ofthe order of 10 seconds is acceptable. Slurryapplications gen er a 1 ly demand lowerfrequencies than clean liquids -- usual lybelow 500 k Hz —— because of increasedattenuation ascribable to acoustic scatteringby the slurry particles. A high temperaturedoppler slurry flowmeter is under developmentand is operating well in pilot plant tests(Karplus and Raptis, 1981) . Furtherdevelopment may include the use of hightemper ature transducers.

Signal processing by cross correlationcan be used with a variety of signal types toy i e 1 d fl ow ve l oc i ty informat ion. Thecovariance of signals from axially separatedsensors, regarded as a function of time delayon the upstream signal, has its maximum at avalue corresponding to the transit time of thestream from the first sensor to the second,provided that fluctuations in the variablebeing sensed move with the process stream.This technique has been successfully tested ona toluene and char slurry line at the HYGASCoal Gasification Pilot Plant using bothacousti c sensing and capac i tive sensing(Raptis et al, 1978; Nanagan et, al, 1978;Sheen and Raptis, 1979) .

Capacitive instruments measure theeffective dielectric constant of the processstream through capacitive plates imbedded in anon-conducting liner within the pipe. In thecase of a two-phase process stream for whichthe dielectric constants of the individualphases are known, it is possible to deduce therelative amounts of the two phases present.This provides knowledge of the densi ty o f thephase of interest. Segmenting the sensingelectrode in the axial direction can provide asimultaneous flow velocity measurement by thecross correlation technique. Thus a singleset of sensing elements can obtain data for a



computed value of mass flow.Mith tagging methods, time of transit

information can be gained withoutthe complexity of cross-correlations. Thusone technique under development for monitoringflow veloci ty uses Pulsed Neutron Activation(PNA)to induce a radioactive tag in theprocess material. The process stream isirradiated through the pipe with a pulsed 14NeV neutron generator6 to activate susceptibleelements, such as 0, and passag e o f theradioactive tag is detected at a downstreamstation. Initial tests have been carried outusing a slurry flow loop with a 14 MeU neutrongenerator, and the result s are promi sing(Herzenberg, 1979). A syst, em based on thistechnique may prove to be too costly forroutine multiple installations within a plant.However, since it does not, require physicalmodification of the piping, it, may be veryuseful as a movable calibration standard forinstalled flow instrumentation.

4. Monitoring of particulates

a. General discussion

Particulate matter i s involved f rom theoutset in those utilization processes forwhich the first step is comminution to assureintimate mixing of coal with reactants,solvents, or entraining fluids. Dependingupon the process, sub sequent reactors orcombusters change the chemical and physicalproper t i e s o f par ti eulate matter, removeparticles, or produce new. particles. Theobservation of these changes in part, iculatemat ter on a single-particle, real-time basisis fundamental to a deta i led understanding ofthe processes. At the end of the process,ef fluent, s are delivered to the environment.Monitoring of ef fluents f'or . particulates isprimarily for the protection of thee n v i r onment although some process controlinformation may be obtained.

A number of met hods for the measurementof particle size distribution, some of whichare tedious and require care in calibration,are reviewed by Allen (1975). Descriptions ofmonitors used for ambient air or in-stackmeasurements are contained in a comprehensiveworkshop report by Lundgren et al (1979).Techniques for the study o f heterogenouse'ombustion are reviewed in a paper by Chitier(1977) ~

Ad vances in laser and microprocessortechnologies have led to the development ofseveral laser devices, one of which cansimult aneously measure particle size andvelocity on a single-particle, real timebasi s. The appl icabili ty of mod empar ticul ate monitors to a wide variety ofproblems besides those of coal utilization hasstimulated the commercial manufact, ure of anumber of systems (Robinson, 1981).

The major problem which confronts all ofthe monitor systems to a greater or lesserdegree is that of monitoring a process streamof corrosive fluids or gases at hightemperature and pressure. Furthermore thedensity of particulate rnatter and the physicalproperties of the non-particulate matter canmitigate against the full util i zation ofmodern si ngle-particle real-time monitors. A

series of symposia have been organized on thespecial instrumentation and control problemswhich arise in fossil energy processes and theproceed i n g s o f these symposia provideexcellent summaries of the progress in meetingthese problems (Symposia on Instrumentationand Control, l977, 1978, l979, 198O, 1981)-

Table I present s a selected li st ofparticulate monitors in four groups: promptsingle particle count, ing and si zing devices,high-speed photography, prompt sized i str ibut ion analyzers, and part icle si zesorting and collection devices. Acousticalmethods are not listed in the Table, but, alsoare beginning to be used (Robinson, 1980).

Individual particles and sizes aremeasured in real t ime in the f irst group ofmonitors in Table I. In the laser visibilitymethod, both the si ze and velocity aremeasured. Mhere conditions permit in situmeasurements, information i s obtained aboutt he time dependence of the local processstream veloc it y. Hi gh speed photographymeasures both size and velocity, but theoutput i s delayed. Obviously video recordingtechniques can reduce the delay.

The prompt size distribution analyzersmeasure the ef fects of many di f fraction orscattering events and the results are promptin that, the data are analyzed on line withminicomputers. The size range of the photoncorrelation spectrometer is extraordinary, butthe device is vulner able to error due tospurious light.

The particle size sorting and collectiondevices use various properties of particles toseparate different size particles from aerosolsamples. The most commonly used devices in

Rev. Mod. Phys. , VoI. 53, No. 4, Part I I, October 1981


TechniquesTABLE I Particle Monitor Systems

Comments References

iPROMPT SINGLE PARTICLE COUNTING AND SIZING DEVICES

Rat io of laser 1ightscattering at forwardangles

Laser visibility method

Image characterization

~0.1 pm to 5 pm

Laser dopplervelocimeter. Sizeand velocity measure-ment ~pm to 200 pm

Forward scattering ofla ser 1ight with ma skarray

W. D. Bachalo, J.Geffken and G. Meth,1978

M. M. Farmer, 1976W. D. Bachalo, 1980N. A. Ch ig ier, 1976

R. G. Knollenberg1978, 1979

Non- laser Light Scat ter ing &0 .3 pm to &10Range varies withsystem design

D. D. Cooke andM. Kerker, 1975K. T. Whitby andK. Willeke, 1979

~HIGH SPEED PHOTOGRAPHY Size and velocitymeasurement. ~ 5 pm

N. A.W. J.D. R.-J. H.

Chigier, 1976McLeanHardesty andPohl, 1980

PROMPT SIZE DISTRIBUTION ANALYZERS

Laser Dif fraction

Photon CorrelationSpectrometry

1 pm to 1800 pm

Statistical Analysisof Brownian motion.~ 3nm to 3000 nmVulnerable to spuriouslight.

J. Swi thenbank,J. M. Beer, D. S.Taylor, D. Abbot tand G. C . Mc Cr ea th,1976

H. Z. Cummins andE. R. Pike, 1974,1977

yPARTICLE SIZE SORTING AND COLLECTION DEVICES

Incr t ial C la s s if ica t ionI

Electric Mobility

Include s impac tor s,cyclone s, centr ifuge s .Size range depends onsystem, & 0.1 pm

0.005 pm to ~ 1 pm

D. A. Lundgren e t al,1979

B. Y. N. Liu, D. Y. H.Pui an.d Abde Kapadia,1979; K. T. Whitbyand W. E. Clark,1966

Diffusion Battery

Filters

0.002 pm to &1 pmut il ize s dependenceof diffusion coeffi-cient on size

Used for bothmonitoring andcontrol

D. Sinclair, R. J.Countess, B. Y. H.Liu and D. Y. H.Pui, 1979

T. Allen, 1975;C. E. Billingsand J. Milder, 1 970



(Linear Msgnitude Scale)

4

~ ~ I ~ ~

4

~ ~ ~ ~ ~e ~

0 ~ e ~~ ~

~ ~ ~ ~ ~

~ ~ ~ ~

e~ ~

a ~ ~

~ ~

~ ~

, ~0

~ ~~ ~

~ ~ ~~ ~

e ~

~ ~

0~ ~

' ~

~ ~~ ~ e i ~~ ~0

~ ~

~ ~ ~ ~

~ ~ ~~ 0

~ ~

~ ~ ~ ~

~ ~4

~ ~

~ ~

~ 0

e

~ e

~ ~

40

0

~ ~

~ ~~ ~

~ ~

I

~ ~ ~

10

0 ~ ~

~ ~

0 ~ ~, ~

0

0' i 82el

0 82el

4

mat = 1000Di~~ter = 1.0 pn

WN = 13000Diameter = 3.0 p~

EXPANDED FORWARD SCATTER LOBES SHORING COLLECTION ANGLES FOR INTENS I TY RATIOING

FIG. IVB-2. Dependence of forward scattered light intensity (angular distribution) on particle size.

ambient, air or stack gas measurements are theinert, ial classification monitors.

In the remainder o f thi s subsection,three laser dev ices and one incr t ialclassifier are discussed and summary remarksare presented in an overview.

c. Visibility method using split laser beams

Part icles in the size range 2 to 200 pmcan be measured by crossed beam inter ferometry(Farmer, 1973; Bachalo, 1980) . The laser beam

b. Intensity ratio of forward scattered laser light

Particle-LadenGas Flow

Sample Volume

Annular Mask Pair

~5' Scatter

This device uses the dependence of t, hescattered light angular d i stribution tomeasure the particle size (See Fig. IV B-2).Far forward angles are selected t;o minimizethe uncertainties due to the variations inparticle shape and refraction index. Theoptical system described by Bachalo (1978) isshown in Fig. IU 8-3. This method is capableof measuring sizes in the range O. l pm to 5 pmwith partjcle @umber densities limited to lessthorn

10 /cm and count rates less than10 /sec.

Beam Expander/CollimatorI l

Ar gon-ion Laser

SCHEMATIC DIAGRAM OF THE RAT1OLNG OPTICS SYSTEM

FIG. IVB-3. Schematic diagram of the ratioing opticssystem.



Particle-LadenGas Flow

h„.i1

'r

Sample Volume

Bcarn SplitterI /

I

P.M.T.

Argon-Ion Laser

SCHEMATIC DIAGRAM OF THE PARTICLE SIZING

INTERFEROMETER OPTICS

FIG. IVB-4. Schematic diagram of the particle sizinginterferometer optics.

(See Fig. j:V B-4) is split into two componentswhich subsequently intersect to form themeasurement volume shown in Fig. TV 8-5. Si zeinformation is contained in the "visibility"of the signal given by

ax in+ I. .~x min

where I i s the detector photocathode currentmeasured at the maximum and minimum of themodulation shown in Fig. IV 8-6.

Nodulation of the signal results from thepassage of the particle through the bright anddark fringes. Variation of the depth ofmodulation is a function of particle sizerelative to the fringe spacing. Particlessmaller than the fringe spacing will scatterlight to trace out the sinusoidal fringeintensi, y distribution. Particles larger thanthe fringe spacing will af feet the lightscattered into two or more fringes dependingupon the size of the particle. The visibilityas such is the ratio of the scatteredintensity for the particle centered on thebrighest fringe to that for centering on theadjacent fringe. The relationship betweenvisibility and particle size is mathematicallycomplicated, but with the aid of computers itis promptly determined.

The system determines the par ticl evelocity by operating as a laser velocimeter.Consequently this device measures the numberdensity, particle size, velocity and, from thevelocity, information about var i at ion o fveloc i ty or tur bul ence. By acousticallydriving the particle, the aerodynamic size canbe measured (Mazumder et al, 1981). Theaerodynamic size is the size of the particleo f un i t density which would have the samesettling velocity as the particle in question.The environmental dispersion and deposition ofparticles f rom an aerosol are usuallydiscussed in terms of aerodynamic diameter.

ENLARGED CROSS-SECTIONAL VIEWS OF THEBEAIVI CROSSOVER REGION

GAUSSIAN RADIALINTENSITY DISTRIBUTION BRIGHT FRINGES

2 sin (8/2)

1/e2 RELATIVE BEAIVIINTENSITY mh8 sin (8/2)

FIG. IVB-5. Enlarged cross-sectional views of the beam crossover region.



d. Particle sizing using laser diffraction

When i 1 1umi n a ted by paral lelmonochromatic light, spherical particles (ordroplets) produce a diffraction pattern theintensity of which is given by well knownBessel functions. The pattern consists of aseries of bright and dark rings superimposedon the smaller geometric image. By placing a

BIN

T i MF

(A) DOPPLER BURST SIGNAL

lens between the particles and a detector asshown in Fig. IV 8-7, the undi ffracted lighti s focused to a small spot that is the centerof the detector with a much larger diffractionpattern surrounding. Movement of theparticles does not cause a shi f t o f thepattern. Particle diameter is determinedsince the pattern diameter is inverselyproportional to the particle diameter . In asense, the lens functions as a Fouriertransform lens since the di ffraction patternis a Fourier transform of the f ielddistribution over the particle.

Fraunhof er di f fraction, or far-fieldapproximation, works wel 1 for particle si zesgreater than the wavelength of the light.Using a HeNe laser (0.6328um), particles canbe measured with sizes down to 1 pm. Shorterwavelength light and complex Hie theory mustbe used for smaller size particles.

Signal processing from various parts ofthe receiver requi res microcomputercapability. The use of this technique fordroplet and particle size distributionmeasurements is discussed by Swithenbank et al(1976) and recent applications are reviewed byNei enr (1979). This method does not countsingle oarticles but rat her yields ameasurement of size distribution for a numberof particles in the sample.

e. Photon correlation spectrometry

TLUTE

(B) PEDESTAL CONPONENT

Although how diffraction and Brownianmotion are to be taken into account inparticle si zing was understood long ago,devices for prompt si zing awaited thedevelopment of the laser and modernmicrocomputer electronics (Cummins and Pike,1977) ~

, JLA,v 'v

I

TiNK

(C) DOPPL. ER COMPONENT

I

V,

I I

VHearn

Expander

Transmit ter

I

I

I

I

Laserl

I

I

I

I I

I

l

L

Receiver

~ 0

~ ~4 ~

~ e

Lens

Signal Processing andAlignment Electronics

I

I

I

Detectorl

I

I

I

l

I

IJ

D~q)pier burst signal sh&&aving the Dig&pier and pedest;ilec~mt)~~nen ts.

FIG. IVB-6. Doppler burst signal showing the Dopplerand pedestal components.

Computer

Ter rnina I

FIG. IVB-7. Particle sizing using laser diffraction.



The photon correlation technique uses themicr ocomputer to e st, abli sh theauto-correlation functions for light scatteredfrom particles which move in and out of thedetection volume due to thermal motion. Therate and way the particles move i s a directfunction of their size, the larger particlesmoving more slowly than the smaller ones. A

measure of thi s par t, iele movement determinesthe size and other information related to theparticle. In particular, the self correlationin time determines the diffusion coefficient.The diffusion coefficient in turn determinesthe size of the particle through theStokes-Einstein equation in which the size isinversely r el a ted to the d i f fusioncoefficient . The photon correlationspectrometer is not an individual particlecounter, .but rat her obtains stat, isticalinformation fr om a large number of particles(Cummins and Pike, 1974) .

utility in the observation of particles inprocess streams.

The general problem of measur ing thesize, velocity, and temperature of a particlein the frequently hostile environment, of coalcombustion or coal utilization processes hasnot been solved. Sampling trains (USEPA,1977; USEPA, 1976; Chitier, 1979) are oftenrequired either to protect the equipment or toprovide controlled dilut, ion or both.Nevertheless substantial progress has beenmade, most recently due to the combinedappl ication of laser and microcomputertechnologies. Certain of the long establishedtechnologies such as filtration are thesubjects of recently renewed interest. Inparticular, electrification effects in fabricfiltration are being actively investigated(Penny, 1977; Lamb and Constanza, 1980) .

f. Impactors

Se ver al devi ces ut ili ze the inertialproperties of particles (Lundgren et al,1979). The most commonly used device forin-stack monitoring of combustion products isthe impactor. The schematic of a cascadeimpaetor is shown in Fig. IV B-8. By passingthe aerosol stream through sue cess i ve lysmal 1er jet aper tur es with correspondinghigher veloci ties, par tie les o f successivelysmaller size and mass are impacted on plateswhich are subsequently removed and theimpacted material weighed. This device doesnot offer a prompt measurement, much less anindividual particle measurement, but it doesprovide a simple and reliable means fortime-integrated size distribution measur ementwhich is an accepted standard for compliancewith regulations. Units can be mounted in aduct, however, sampling trains are used inmost applications.

g. Overview

STAGE I &

STAGE

STAGE N

r NOZZLE

JET EXITPLANE

IMPACTIONPLATF

Historically the measurement of particlesi ze distr ibutions has been a time consuming,tedious process r equi ring r epeatedcalibrat, ion. A large number of techniques isavailable (Allen, l 975) and only a few havebeen discussed in the foregoing material.Opt, ical microscopy of fers the most directmeans of observing particles but it requirestrapping of the particles and has little or no

AFTERFILTER

FILTER

TO VACUUM PUMPSchematic of cascade impactor with trajectories shown for particles of three different sizes

FIG. IVB-8. Schematic of cascade impactor with trajec-tories shown for particles of three different sizes.

Rev. Mod. Phys. , Vol. 53, No. 4, Part li, October 1981


5. Temperature measurements

Temperatures of gases and mixed pha semedia must be measured under both reducing andoxidi zing condi tions up to about 2500C andunder pressures up to 1000 psi, sometimeswithin a few seconds. Temperature monitoringis critical for safety and process control inall advanced fossil energy processes, sincetemperature is probably the single mostimportant indicator of the operating state ofthe reactor or combustor and can change by 100C per second in 'some processes.

Temperature measurements are needed notonly for process control in full scale plantsbut also for scienti f ic and engineeringdiagnostic purposes in less than full scaleplants, e.g. , pilot plants or even smallerlaboratory models. For diagnostic purposes,on-line r eal -t ime temper atur e mea sur ements,though always desirable, may not be necessary;data can be stored and analyzed later. Forprocess control and safety, on the other hand,on-line real-t;ime measurement ("monitoring" )generally is essential; it is often the casethat failure to hold a process temperaturewithin specified bounds can result in majorstructural damage or even a li fe-endangeringexplosion. Process yields, moreover, of tendepend very sensitively on temperature.

Pilot or demonstration plants, thoughless than full scale, also are not withoutsignificant hazards and must be provided withon-line monitoring adequate for ef fectivecontrol. Temperatures measured for diagnosticpurposes sometimes must be reconstructed asnearly instantaneous functions of time, eventhough the result need not be known to theexper imen ter in r eal t ime. In anentrained-bed gasi f ie r, for ex ampl e, theresidence time of the gas in the reactor isonly a few seconds. To study the complicatedreaction kinetics in such a rec tor,temperature time resolution of a fraction of asecond is highly desirable.

In gener al one expects that many moreparameters must be measured for diagnostic anddevelopmental purposes than are necessary forprocess control. For process control (e.g. ,in a coal liquefier) it may be sufficient tomonitor the temperature at less than a dozenselected points, but for diagnosing theoperation of an experimental liquefier,temperature measurements at hundreds of pointsmay be desirable.

The desire for temperature measurements

for diagnostic purposes may addi t ional lyincrease because conditions at a single pointin a reactor may be characterized by severaldifferent "temperatures. " Unless the systemis in local thermodynamic equilibrium, thekinetic and radiation temperatures at anypoint need not be .the same, and each of thesemay differ, from the molecular vibrationalexcitation temperature, or that describing theatomic level populations. Though it may bereasonable to assume local thermodynami cequilibrium in the interior of a high presureliquefier, this assumption is at leastquestionable in the interior of a gasifier.The gasifier diagnostician therefore may wishto use simultaneously several instruments,each of which measures a di f ferent"temperature. "

a. Thermometry for coal technology

The methods which have been used tomeasure temperature in the coal utilizationindustry have been surveyed recently (Argonne,1980). These methods include thermocouples;optical pyrometry; electrical resistancethermometers, including thermistors anddevices which infer the resistance from themeasured Johnson noise ,'acoustic devices;crystal thermometers, which measure changes inthe crystal resonant fr'equency; "filledthermal elements, " (e.g. , conventionalmercury-in-glass t hermomet er s ); bi-metallicthermomet er s; and even color i ndi cat ors(paints which change color as a function oftemperature). This list by no means exhauststhe methods which have been used or suggested.Probably the most complete surveys oftemperature measuring methods are to be foundin the Reinhold Volumes edited by Herzfeld(1962) and Plumb (1972) . A very recent study(Herzenberg, 1981) di scu sacs stil 1 othermodern temperature measur ing methods, andoffers a very readable and worthwhile analysis—relevant to the discussion below —of theproblem of measuring temperature influidized-bed combustion reactors

It is impractical to discuss all methodshere, ' the following discussion concentrates,for illustrative purposes, on two methodsthermocouples and optical techniques. Thesetwo, and perhaps e lee trical resistancethermometers and acoustic devices, are themethods we believe most likely to prove usefulat the high temperatures and in the hostileenvironments encountered in coal technology.Potentialities of some other methods listed in



the references which have been ci ted may nothave been suf ficiently examined, however.

Thermocouples: Publications onthermocouples and thermoelectric materials arevery numerous; we make no attempt here tosummarize or reference this literature. (SeePlumb, 1972. )

In practice, the Type K thermocouple(chromel-alumel) has been used more widelythan any other type, but various authoritiesreport advant ages for other newly-developedthermocouple types (e.g. , nicrosil-nisil );these are useable to about 1250C. Above12$0C, to about 1700C, use of plat inurn-rhodium(Type B) thermocouples has been recommended,and tungsten-rhenium alloy thermocouples aresaid to be useful as high as 2760C. Theselatter type thermocouples are quite expensive, '

however, particularly in view of the largedimensions of full scale coal utilizationreactors, whose central temperatures often arerequired. It is not unusual to employ -- anddestroy -- thousands of feet of thermocouplewire in a single diagnostic test series.Alternative less expensive thermocouple alloysare much needed, especially for measurementsa t the higher temperatures. Desirablecharacteris tie s obv i ou sly include thecapability of being drawn into long wires andbent without breaking, together with stabilityof calibration over long periods of handlingand exposure to high temperatures.

Probably the most stringent demand placedupon thermocouples by coal technology is theability to withstand the highly corrosiveenvironments prevalent in coal reactors. Todate, no thermocouple type is known that canwithstand this corrosion long enough even fordiagnostic t ests. Even wi th thermocouplesenclosed in corrosion resistant sheaths or"thermowells", this difficulty persists. Forex ampl e, in one experimental gasifier(Pitcher, 1978) thermocouples in siliconcarbide thermowells have rarely survived morethan a f ew days in the ambient mixture ofsteam, methane, oxygen and coal particles at1300C-1600C. Experience with sheathedthermocouples in magnetohydrodynamic (MHD)plasmas has been similar.

Thermowells, even if completely corrosionresistant, introduce characteristic problemsof their own. Response times of thermocouplesin thermowells are rarely shorter than severalminutes. This is far too slow in manysituations. Even in steady-state conditions,moreover, use of a thermowell can introduceartifacts and measurement errors associated

with heat conduction and shunt electricalconduc t ion in t;he the r mowel 1-thermocoupl ecombination. In principle, individua 1calibration of the device can help compensatefor such errors, but reliable calibrationsapplicable to the actual working conditionsencountered in a coal reactor or furnace aredif ficult to achieve. It follows thatresearch seeking to ident i f y improvedthermocouple sheathing materials, capable ofminimizing the aforementioned errors inherentin thermowel 1 use, while simul t aneouslyresisting the corrosive ef fects of the coalutilization industry s hazardous environments,should be strongly -encouraged. Even more tobe encouraged are searches for thermocouplematerials themselves capable of resistingcorrosion, and therefore capable of being usedwithout thermowells. Although at this timesuch searches cannot be said to have more thana small probability of success, it should benoted that high t emperature thermocoupleelemental materials need not be confined tometals or metallic alloys, which do not have amonopoly on coriduc t i'on by e 1ec t r on car r i e r s(such conduction probably is a prerequisite tofinding a measurable thermoelectric emf).Indeed, the dweeb eck e f feet has beendemonstrated in semiconduct', ors (Frederikse,1963), and carbon-graphite high temperaturethermocouples were being investigated at theturn of the century (Kinzie, 1973). However, '

irrespective of their obvious unsuitability incoal-burni. ng circumstances, thesecarbon-gr aphite thermocouples, with theirobvious ductility and fragility problems,i 1 lustrate the advantages of metallicthermocouple elements and the difficulties ofreplacing them.

Optical Techniques' . According to Argonne(1980), temperature measurement by opticaltechniques in the coal util. ization industryhas not extended beyond employment ofradiation pyrometers. The instruments usedhave mainly been infrared pyrometers operatingin the wavelength band from about 1 to 10micrometers. Assuming the source is a blackbody, temperature can be inferred from theintegrated (over wavelength) radiat, ionintensity in the observed spectral band, orfrom the slope (as a function of wavelength)of the radiation intensity in the observedband. The task of measuring the spectralslope typically has been accomplished by meansof a so-called two-color radiometer, whichmeasures the ratio of the received intensitiesin two separate narrow bands.

Rev. Mod. Phys. , Vol. 53, No. 4, Part li, October 1981

S86 APS Study Group on Coal Utilization and Synthetic Euel Production

Depending upon emi ssivity andequilibration, two-color sampling techniquescan occasionally give misleading results. Ifit is reasonable to assume local thermodynamicequilibrium in coal utilization reactors whichoperate at high pressures and/or whose workingsubstances have long residence times, then itis correspondingly reasonable to suppose thatthe radiation is black body in characterespecially since these reactor interiorsfrequently contain many coal particles whichstrongly scatter and absorb radiation. On theother hand, the possibility that the radiationobserved from any given reactor may not beblack body cannot be arbitrarily ruled out.Wherever possible, therefore, spectra shouldbe scanned over a sufficient range to permitslope estimates from more than a single pairof frequenci es; and/or the r ad i ationtemperatur e should be compared wi th theindications from a thermocouple placed in thespatial region whose temperature (supposedly)is being sampled by the observed radiation.

It seems unwise to limit considerationjust to opt ical pyrometers. Line widths andline prof iles also carry temperatureinformation which for many years has been usedin astrophysics, plasma physics, fl arnechemistry, and numerous other research areas.The coal utilization industry has made littlepractical use of these techniques. Argonne(1980) mentions spect, roscopic thermometers,but says of these that they involve"laboratory techniques seldom employedindustrially". This may amount to overlookinga good bet.

F' or d i agnost i c purposes immediatetemperature readout often is not essential, sothat the fairly complicated data processingwhich may be necessary to infer (e. g. ,vibrational and rotational) temperatures frommolecular spectra should not impede use for-pilot plant analysis. Mith modern computingcapabilities, moreover, such measurements maybe practical even for process control in fullscale plants.

In principle, the radiation sensor foroptical temperature measuring techniques canbe located external to the reactor beingmeasured, thereby avoiding the corrosiveenvironments. In practice this advantage ismuted by the fact that many coal utilizationindustry reactors —especially those whichare large, operate at high pressure, orcontain many carbon part icles -- will beoptically thick at visible and infrared wavelengths . Under such ci r cumst ances, optical

measurements of inter ior temperatures requirethe insertion of optical probes in protectiveshe at hs wi th al 1 the at tendant corrosionproblems discussed above for thermowells.

Even when the in ter io r is optically thin,external sensing of the reactor radiationsrequires the construction of optical windowsin the reactor wal 1 s . These may provevulnerable to slagging or coating. Success inkeeping the window clean by "purging" (i.e. byintroducing a gas stream to blow away anypart icles before they can adhere to thesurface) has been reported (Ballard, 1981).Of course, purging can also introduceexperimental artifacts.

Subject to the same problems with windows,reactor opacity, etc. , one may probeconditions within a reactor by means ofradiation introduced from the outside. (Thestudy of flames by absorption spectroscopyprovides a simple and well known example. )Zweibaum et, al (1978) have reported what is inef feet a microwave absorption spectrum for theBI-GAS pilot, plant reactor, in the range 1.6to 5. 5. micrometers. This spectrum showsconsiderable structure, presumably associatedwith various molecular species wi thin ther e ac tor,' the aut hors suggest that, thi sstructure could be cont inuously monitored forprocess control.

Availability o f 1 aser s continuouslytunable over a broad range of wavelengths inthe visible and infrared should make the studyof reactor interiors by means of externallyintroduced radiation much more attractive.Indeed, the measurement of laser Raman andRayleigh scattering as a means of probingreactor interiors was strongly advocated sixyears ago in a study of the role of physics incombustion (Hartley et al, 1975) . It waspointed out that Raman scattering can be usedto measure rotational and vi br at iona 1temperatures of the various species in thereactor, as well as the relative densities ofthose species, and that it even may bepossible to determine their spatialdistributions. Hartley et, al (1975) alsodiscuss other possible laser di agnost ictechniques, (e.g. , laser-induced fluorescence)and suggest a number of specific combustionresearch experiments (mainly applicable toflame research however).

Although the potential use of lasers in,coal conversion diagnostics and processmonitoring has received some attention, wefeel that a more intense and systematic effortis needed and worthwhile. Lasers have proved



to be so versatile a diagnostic tool that webelieve they will open up new avenues for coalconversion reactor diagnostics and processcontrol.

b. Temperature measurements and modeling

As has been explained, under somecircumstances reactor interiors are expectedto be an optically thick distance from thereactor walls, whereas in other circumstancesthe reactor interiors may be at optically thindistances from all walls. In the latter caseit is possible for measured radiationtemperatures at a point to be determined bythe temperatures of "hot spots" remote fromthat point. Simi 1 ar ly, under somecircumstances the interior of the reactor canbe expected to be everywhere in localthermodynamic equi 1ibr ium, whereas in othercircumstances there are likely to bedifferences between the kinetic, radiation,rotational, vibrational, etc. , temperatures at;a given point in the reactor s interior.

Through reactor modeling, who sepotentials and problems are discussed insection IV-C of the report, it may be possibleto estimate in advance the interior conditionsof the reactor (optically thick or thin, etc. )in order that the appropr iate diagnosticsensors and process monitoring controls beincorporated into the design. Reactor modelswill play an im'portant role in interpretingthe temperature and other measurements (e .g. ,species composition) obtained in cases wherelocal thermodynamic equilibrium cannot beassumed. For reasons such as these, Hartleyet al (1975) strongly recommended increasedcombustion modeling ef forts, especially foradvanced coal industry reactors. Me stronglyendorse this recommendation, because of theneed for modeling to properly measure andinterpret reactor temperatures as well as forother reasons. Further, we agree with Hartleyet al that the connection between modeling andd i agnost ics i s a t wo way street. Good5 iagnost ics, which include temperaturemeasurements, are needed to test and improvereactor models.

Reactor models also have an importantrole to play in what may be termed indirectmeasurements of reactor temperatures. Becausereactor temperatures often are so difficult toobtain, there is considerable merit to the

idea of seeking ways of inferring thetemperature from more r'eadily accessible datathan those which are needed for "direct"temperature measurements of the sort thisreport has examined. Of cour se, alltemperature measurements are indirect in asense, but by "indirect" here is meantinferring the temperature from a relationshipwhich has been deduced via reactor modelingand is valid only in the particular reactorunder consideration -- as distinct from somegeneral theoretical relat, ionship (e.g. ,b lack -body radiation intensity or Dopplerbroadening) which is valid for all reactors.

The idea of indirectly measuring thetemperature may be particul arly impor tant inprocess cont rol mon i tor in g o f hazardoustemperature excursions; this possibility ispredicated on the exi stence of a reactorproperty which changes rapidly with reactort;emperatur e (perhaps even changingsignificantly before the temperature excursionis well underway) and which is easilymeasurable with rapidly responding equipment.For instance, to give a purely hypotheticalexample, modeling might indicate that theconcentrations of ethane and other higher(than methane) members of the alkane seriesmeasured at the output stream of a gasifierare very sensitive functions of the gasifierreactor temperature, with these concentrationsdecreasing markedly as the reactor temperatureincreases. In this event, the gasifierreactor temperature could be conv en i en t 1 ymonitored by measur ing the concentrations ofhigher. alkanes in the gasifier output stream.This stream is readily accessible, and thereexist numerous well-est abli shed techniques(mass spectrographic analysis, infraredabsorption spectr oscopy, laser Ramanspectroscopy, ete. ) which offer good hopes formeasuring these concentrations on line. It isnoteworthy that in this sort of indirecttemperature monitoring for process control,the modeling need not be so accur ate orthorough as to enable reliable determinationof the interior reactor temperature from theexternal ly measur ed higher alkaneconcentrations; all that is necessary forprocess control of the reactor temperature isthe much easier modeling task of r el i ab lyestablishing a high correlation between risesin the interior temperature and changes in theexternally measur ed quantit ies, in thishypothetical case the concentrations of theprocess stream alkanes. It i s stronglyrecommended that reactor modeling objectives



include searches for suitable indir eet;measures of reactor temperatures as well asother important process variables.

6. Potentialities of acoustics

The past. two decades have seen remarkableimprovements in acoustic technology,associated with an equally remarkable growthin the applications of acoustics to studies ofall forms of matter, including e.g. , plasmasand biological ma ter i al s . Application ofacoustics to instrumentation and control needsof the coal utilization industry has shared inthis growth. lt may be, nonetheless, that thefull potential of acoustic methods in coaltechnology has yet to be realized.

The remainder of this section speculateson some pot ent ial ly interesting lines ofinvestigation. Ne offer some reasons forbelieving acoustic techniques may, in thefuture, be still more useful in coaltechnology, and then put for th someillustrative possibilities (which we thinkwarrant further investigation) for newapplications in the field. Because knownacoustic applications to the study of matter,like the methods for measuring temperatureexamined in subsection '5 above, already are sonumerous and diverse, the discussion whichfolj.ows can only touch upon a very, few of thepotent ialit ies of acoust ics in the coalutilization industry. No attempt is made hereto thoroughly survey the literature. For thisthe reader is directed to Sessler (1968),Carlin (1964), and other articles in the same"Physical Acoustics" series.

a. Reasons for investigating acoustic techniques

Acoustic and optical techniques share thebasic property of being "non-invasive" or"non —perturbative" under ordinarycircumstances. Acoustic measurements may beeither passive -- listening to the soundsgenerated in the system being monitored —oractive, sending acoustic waves into a systemand monitor ing the backscattere d ortransmitted acoustic signal. Acoustic waves,of course, can be produced or tuned over avery broad band o f f r eque nc i es andwavelengths.

On t he other hand, by increasing theincident power, acoustic as well aselectromagnetic ~aves ean be used to heatsamples and to cause other chemical andphysical changes. That is, they candeliberately be made invasive or evendestructive. Changes intentionally induced bya coustic waves conceivably could beindust r ial ly useful in e . g. , coalliquefaction, or could become the basis forother (not available at lower power inputs)acoustic di agnost i c tech n i ques. One suchdestructi ve acoustic process, having noobvious laser analogue, is "cavitation", theproduction of vapor-filled cavities in liquidsby acoustic waves whose peak t, ensions (peaknegative pressures, taking into account theexternal positive hydrostatic pressure) exceedthe liquid tensile strengths. In practiceliquids can rupture at tensions far less thanthe theoretical tensile strength, presumablybecause of the presence of small bubblescapable of growing to large cavities when thetension exceeds 2 u/r, where n is the sur facetension and r is the bubble radius before thetension is experienced.

In sum, acoustic probing of coalutilization reactors has many of the desirablefeatures associat ed with optical probing,while in other respects being complementary.Because acoustic propagation in a mediumdepends primarily on its visco-clast, icproperties rather than on its electromagneticparameters, acoustic waves can penetrate mediawhich are optically opaque (e.g. metals). Forthis reason alone, acoustic probing techniquesinvite attention. One may expect to encountersituations where sound generated within thereactor could be monitored from the outsidewithout any need to insert a detector into thereactor or to construct, a '-'window" through thereactor wall. The problems stemming fromfouling of the window, described earlier inthe discussion of optical techniques, shouldbe much less troublesome for acoustic wavesthan for infrared or visible light waves.Another possible advantage stems from the factthat, for distances of the order of reactordimensions, the propagation time over anychosen path, from which pulse velocity ean beinferred, is much easier t;o measure foracoustic pul ses than for electromagneticpulses. Acoustic pulse velocities should betemperature dependent; their use to monitortemperatures in reactor interiors presently isbeing seriously investigated, as is explainedbelow.

Rev. Mod. Phys. , Vol. 53, No. 4, Pert I I, October 1981


b. Present applications of acoustics in

coal utilization

Acoustic techniques already are bei ngused, or at least diligently investigated, inthe coal utilization industry. As alreadydescribed, several different applications ofacoustics to the measurement and monitoring ofmultiphase flow (e.g. of coal slurries),involving both passive listening to acousticnoise and active detection of back-scatteredexternally generated sound ~aves, have beenvigorously pursued.

Acoustic Time Doma i n Thermometry .Acoustic time domain thermometry, also calledacoustic time domain ref lectometry or acousticTDR, has been accomplished in nuclear reactorsby inserting into the reactor a wire with oneor several designed acoust ic impedancemismatches. The times taken for an externallygenerated acoustic pulse to run down the wireand be reflected back from the mismatch pointsare then measured. Because the velocity ofsound in the wire depends on the temperature,these time measurements permi t in f err ing thewire temperature. Obviously one can inferonly the average temperature along a segmentof the wire between two impedance mismatehesgenerating reflected pulses; but by makingthese segments sufficiently short, the averageis taken over correspondingly smaller regions.As many as ten wir e segment s reportedly havebeen employed in the case of nuclear reactors(O'Fallon et al, 1976).

The main environmental hazard (aside fromhigh temperature) in nuclear reactors is posedby neutrons, which can and actually dotransmute the wire carrying the acousticsignal. The environmental hazards (againaside from high temperature) in coalutilization reactors surely include corrosion,as been discussed earlier (underThermocouples, subsection 5), and alterationby solid state diffusion. To avoid thesehazards it may be necessary to employprotective sheaths around the wire carryingthe acoustic signal. In this event problemsakin to those arising with thermocouplesheaths may be encountered. The main suchproblem with acoustic TDR ~ould be theincreased response time associated with thefinite thermal conductivity of the sheath.This might nulli fy one of the principalpossible advantages of acoustic TDR, the hopedfor rapid response times. On the other hand,

problems such as shunt electrical conductionin the sheath material would seem to be lesstroublesome for acoustic TDR. Becauseacoustic TDR does not require an electricallyconducting element, it is possible that acorrosion-resistant refractory material notrequiring a sheath could be found for use inacoustic TDR. The question of alteration bysolid state di f fusion would remain to beexplored separately.

It is conceivable (Papadakis, 1976) thatmeasured attenuations of the reflected signalsand obser ved propagation times could becombined to improve the temperatureinformation attainable by acoustic TDR. Theatt enuation of sound waves, like the soundvelocity, , i s usually temper atur e dependent.This idea seems to have been explored neitherfor coal reactors nor (apparently) for nuclearr eactors. Highly temperature-dependentattenuat ion in t, he temperature rangeanticipated may be achievable by takingadvantage of the fact that the acousticattenuation in refractory metal wires has beenobserved to increase sharply at a temperatureequal to about half the melting pointtemperature ( Papadakis, 1976) .

Ultra8onic Fluidized Bed LevelMeasurement: Use of acoustics to measurefluidized bed level height has been proposed(Argonne, 1980; O'Fallon, 1976) . Thi spresumes an - unambiguous reflection at theupper "surface" of the bed, and one can easilyanticipate the sources of error in thistechnique. Echoes fr om points above theactual bed level, for example, could resultfrom high concentrations of particulates inthe gas above the bed, from strong turbulencein the gaseous atmosphere, from violentfoaming at the inter face sur face, etc.

Despite such difficulties simij arapproaches might be considered whenever the1ocat ion of a sur face of discontinuity isdesired. Proper monitoring of the BI-GASpilot plant gasi fier, for example, r equiresmonitoring of the char level in a char hopper.The possible util i ty of ultrasonic levelmeasurements in this and other coal technologyprocesses has been examined (Argonne, 1980).

Acoustic Detection of Valve Leaks. An

already widely employed method for detectingthe development of cracks and other defects inmat erials is the "acoustic emission"technique, also known as AE (Spanner, 1979).In this method, one seeks to detect thecharacteristic transient sounds emitted whenmaterials undergo microscopic deformations or



fractur es; an example might be sliding of agrain boundary. Increases in AE precede thedevelopment of macroscopic flaws. Techniquesfor locating the evolving flaw, usingappropriate processing of the signals receivedfrom several di fferently situated acousticdetectors, have been developed. Increased useof AE in the coal utilization industry hasbeen advocated in a con ference whoseobjectives resembled those of the presentreport (Shewmon, 1978) .

Control valves must operate against highpressure differentials, at high temperatures,and must handle highly corrosive and abrasivematerials. Consequently, on any one day in anoperating fossil fuel plant, many valveshaving important functions are likely to havedeteriorated and to need replacement; failureto promptly replace such deteriorated valvessometimes carries a risk of accident capableof causing serious damage to the plant.

Id enti fying faulty valves by acousticemission presently is receiving serious studyat several laboratories. Although few testsha ve gone beyond the laboratory stage, thereis reason to believe some version of the AE

method will be able to detect valve leaks, as-well as serious deterioration in valves beforeleakage develops (Ellingson, 1980).

As a far more general pr inciple,deviations from normal operating conditions inmany kinds of equipment are likely to beaccompanied by characteristic changes in soundemission, e speci al ly when the operationinvolves fluid flow. While the relativelycrude procedure of simply monitoring the totalsound emission to detect deviat ions fromnormal operating conditions presently is beingused in some plants, spectral analysis andother signal processing techniques can yieldmore detailed information about the operatingconditions. Valve leak studies suggest thatthe valve leak acoustic signal "is basically abroad band random noise signal" (Smith, 1979).Although the specific properties of thisrandom signal (e. g. , spectral slope,correlation functions, etc. ) apparently havenot been invest igated, some elementarytwo-band spec tr al compar i son o f the receivednoise has been found to improve detection ofthe desired valve leak signal (Smith, 1979).

c. Some possible applications of acoustics to coaltechnology

The following topics have not been»researched" in any legitimate sense of the

term, they are offered here in the hope thatsome r ead er s may b e s tirnu 1 a ted toinvestigations both broader and deeper thanours.

Use of High Intensity Sound Waves. Onepossible use of high intensity perturbingsound waves has been recognized in theindustry, namely the control of particleemissions in smoke and other fumes (Carlin,1960) . The smaller particles in theseemissions, those with sizes less than a fewmicrometers, are the most difficult particlesto remove from the gases carrying them.Agglomeration of the smaller particulates bymeans of intense acoustic waves, a procedurewhich has some theoretical basis, is receivingserious examination (Volk, 1976) .

Coal Liquefaction. A mor e d i r ec texploitation of high intensity sound wavesone apparently not yet explored in the coalutili zation industry and of fered here as aspeculation —is the promotion of direct coal1 ique faction (in the presence of hydrogendonors and/or catalysts) . It is known thatultrasonic ~aves can affect the rates ofchemical reactions (Basedow, 1979) .Therefore, it seems worth invest igatingwhether intense ultrasound beams could improveor help to control the process of coalliquefaction. Of course, it is essential thatthe energy expended in producing andtransmitting the ultrasonic beam not be asignificant fraction of the usable energy inthe coal liquids involved. (However, becausecoal liquefaction processes typically requirethat the coal be heated, the ultrasonic energydissipated in the coal will aid the heatingand will not be entirely wasted. )

If the ultrasonic waves traversing theliquefying coal are intense enough to causecavitation (recall the earlier discussion insubsection 6a) very complex chemical andphysical effects may occur e.g. , luminescence(Flynn, 1964). If liquefaction start, s -at theouter sur faces of coal particles, thenexciting cavitation —which typically erodesthe solids at solid-fluid interfaces —mightexpose new sur face and thereby increase therate of liquefaction. On this basis, at temptsto drive the liquefying coal to cavitation viaultrasonic radiation might merit preliminaryconsideration. Unfortunately, production ofcavitation in the highly viscid media of coalliquefaction is likely to be very difficult,i.e. , to require very (perhaps impractically)high acoustic intensities. Any suchdi fficulty will be compounded by the fact that

Rev. Mod. Phys. , Vol. 63, No. 4, Part l I, October 1981


cavitation, once it occurs, greatly impedesfurther transmission of intense acoustic wavesinto the eavitating medium, because ofscattering by the maeroseopie bubbles and theaeoust ie impedance mismatch caused by thebubbles.

Coal Comminut ion . In many coalutilization processes, efficient operationrequires that the coal be fed to the combustoror reactor in the form of comparatively smallparticles (approximately 100 micrometers insize). Customarily the large coal fragmentsobtained from the mining process are grounddown to the se sma 1 1 s i ze s. Gr inding hasproved to be expensive however, in money,t ime, and energy. It has been estimated that'tg or more of the usable energy of the coalcan be expended in the grinding operation (SeeIV E. ) . There have been many attempts toachieve si ze reduction without grinding,primarily chemical fragmentation of coal.Chemical comminution of coal can be achieved,but the chemi cal processes needed areinconvenient and expensive. (Chemicalcomminution is discussed in Sec. IV E. )

The possibility of fragmenting coal intosmall particles by means of high intensitysound waves might well be examined. Soundwaves produce large stress and straingradients in the propagating medium,especially at high frequencies. In effect,therefore, a high intensi ty sound wave mightfunction as a novel sort of grinder. Xf thewater in the coal pores could be excited tocavitation, then the erosive forcescharacteristically accompanying cavitationmight quite effectively pulverize the coal.Even if fragmentation of coal by highintensi ty sound waves proves impractical,cavitation and acoustic streaming might beused to further f ragment already sma&1 coalpart, icles suspended in a slurry (which couldbe a coal-oil slurry, see Section IV E 5);here the liquid surrounding the coal (not t, heliquid in the coal) would be brought tocavitation. In thi s scheme, acousticfragmentation would be employed only to reducepar ticle si zes below the si ze at wh i chcontinued mechanical grinding becomesextremely inef ficient.

Obviously, any scheme involving use ofintense sound waves in a coal uti li zationprocess vill be impractical if the energyrequired to produce and transmit theultrasonic beam is a significant fraction ofthe expected energy yield from the process. A

simple calculation suggests that intense sound

waves are not impractical from thisstandpo int. Suppose, for exampl e, coalparticles suspended in an oil slurry could bef ragmented to desired very small sizes bymeans of acoustic waves. Typical usablecoal-oil slurries contain about 50$ coal byweight . Therefore, taking the energy contentof cog equal to 10, 000 Btu/ib, and using 1.3gm/cm for the density of coal, the coalenergy contained in a volume Ad of thjpcoal-oil slurry will be approximately 2x10Ad ergs in egs units, where A is interpretedas the area of the acoustic beam and d is thethickness of the layer (presumably about equalto the attenuation distance of the aeoustiebeam in the slurry) within which coal chunksvill be fragmented. The rate of radiation ofacyustic energy into the slurry will beAp /2~c', where p is the- maximum pressure, p isthe fluid density, . and c is the velocity o(sound. Using p = 0. 7gm/cm and c = 10cm/sec, typical values for organic liquids ofmoderate atomic weight (e.g. n-pentane), therate of rad jation of acoustic energy will Peabout 7x10 A ergs/sec for p = 10 dynes/cmatmospheric pressure, a pressure sufficientlyhigh to cause cavitation in undegassed waterat moderate frequencies. Consequently theacoustic energy expended in inducingcavitation should be negligible compared tothe energy content of the fragmented coal inthe coal-oil slurry, even if d is no more thanabout one centimeter, the acoustic signal mustbe maintained for t imes of the order of a fewseconds, and the ultrasonic pressures requiredare as high as ten times atmospheric.Nevertheless, though this energy comparisonseems very favorable, the feasibility of thisacoustic fragmentation suggestion ultimatelyhinges on whether a signi f icant fraction ofthe acoustic energy dissipated in the slurrywill go into cavitation-inducing mechanisms.Further investigation is required.

Uses of Low and Moderate Intensity SoundHaves. Present uses of low and moderateintensity sound waves in applied and basicscience, for diagnostic, analytic, and controlpurposes, are numerous and varied. There aremany possible applications of acoustics tovarious coal utilization industry problems.For instance, measurements of . the scatteringof acoustic waves in coal can reveal coali nhomogeneit ies and mierostrueture.Measurements o f acoust ic veloci ties andattenuations might correlate usefully with thearomaticity of coal, and in that case wouldbecome a use ful tool for rapidly determining



coal aromaticity and related pr operties. Itis possible that similar measurements mightprovide a rapid tool for diagnosing theprogress of liquefaction in coal liquefyingprocesses. Acoustic tomography i s beinginvestigated at the present, time (Carson,1977); successful development of thistechnique could make it possible to probe fordefects within reactors, to visualize thegrowth of bubbles in fluidized beds, etc.Even if tomography proves to be impractical,the measurement of acoustic scattering atwell-chosen frequencies could help to diagnosebubble growth and other phenomena withinfluidized beds. Measurements of ultrasonicabsorption and dispersion over a widefrequency range, known as ultrasonicrelaxat ion spectroscopy, have been used tostudy the physical properties of materials aswe11 as reaction kinet ics. Moreover, variousmethods have been devised for extractinginformation about material properties bymaking use of acoust i c c ou pl i ng toelectromagnetic waves. Such methods includephotoacoustic spectroscopy, and acousticnuclear magnetic resonance.

Many of the just mentioned acousticapplications depend for their effectiveness onthe ease with which -- when high intensityacoustic output is not required -- thefrequencies, of acoustic waves can be tunedover very wide bands without much modificationof the equipment being used. A possibleapplication of this acoustic frequency bandvariation is particle size measurement, whoseachievement by optical means has beendiscussed above in subsection IVBA. Mhenparticle sizes get much larger than about 100micrometers, particle sizing by optical meanscan become di fficult, in large part becausesuch particles are too large compared tooptical wavelengths for sensitive dependenceof the scattering on particle size. However,acoustic frequencies easily can be adjusted toyield wavelengths in the range wherescattering by particles of 100 micrometers andgreater is sensitive to particle size.Therefore, acoust ic ~aves may provide aconvenient method of particle sizing over arange of larger particle dimenensions than canbe readily si zed wi th analogous opticaltechniques. The theory of acoustic scatteringby small particles is not identical with theHie scattering theory used to analyze opticalscattering, but is closely related, and can becomputed for whatever parameters (particledensities, sound speeds, acoust ic absorption

coefficients) are desired. One difficultywith this acoustic method is that theattenuation of acoustic waves in air and othergases at wave lengths of approximately 50micrometers is large. An ultrasonic particledetector, , purportedly usable for on-lineprocess control monitoring of particles assmall as one micron in diameter already hasbeen marketed (Robinson, 1981).

7. Recommendations —instrumentation and control

The 1 ack o f dat a n ecessary forfundamental understanding of advanced fossilenergy processes largely i s caused by adeficiency in diagnostic instrumentation tomake detailed measurement s of processbehavior, coupled with a lack of access toexperimental facilities. Such measurementsare essential for model creation andverification as di scussed in IVC on the onehand, and are necessary for interpretation ofinstrument readings and reliable scaleup ofthe instruments on the other hand. To the endof providing these data and ultimately makingpossible the control of the advanced fossilenergy processes (as discussed in theintroduction to this section), we recommendthe following .'1. F' or the purpose of improved processcontrol, present re sear ch directed towardincreasing our ability to measur e -- andthereby monitor -- process parameters underthe harsh corrosive environments of coaluti 1i zation plant s, including temperatur esover 1 QGOG and pressures of 1000 psi, must beexpanded. These parameters include, but by nomeans are limited to. reactor temperatures;mass flow rates and chemical speciescone en tr a t ions in 'mult iphase process streams;particulate numbers and size distributions ingas streams to turbines and to the atmosphere,and fluidized bed levels.2. For improved efficiency and safety,instrumentation researches must be aimed attechniques permitting real-time monitoring ofas many process parameters as possible,especially of reactor temperatures. Often thehar sh reactor environment s and componentinaccessibility wi11 require that thismonitoring be accomplished non-invasively aswell o

In view of the present paucity of reliableinstr-umentation for coal plant processcontrols, a systematic evaluation of available

Rev. IVlocl. Phys. , Vol. 53, No. 4, Part II, October 1981


commercial and developmental instrumentationshould be undertaken, with the clearly statedob jeetives of identifying promising presenttechn ique s and enc ou r a g i n g their impr overnent .Commercial instrument s include, but are not1 irni ted to, state-of -the-art instrumentsdiscussed in this section. acoustic andcoriolis flowmeters; optical par t ieul atemon i tor s; thermocouple, sodium line reversal,and optical pyrometer temperature measuringinstruments', and spectrographic and infraredga s ana lysi s. De ve1oprnental instrumentsinclude. acoustic, capacitive, and nucleardevices to measure high temperature multiphaseflow rates; electromagnetic and acoustic levelmeasur ing instrument s; ne utr on capt ur e g ammar ay t echn i que s for on-line analysi s of densephase streams; imp r o v e d the rmocoup1 e sheaths;and laser -based combus t i on -mon i to r in gi n s trumentation, i nclud ing on-lin ga s'analysis.

To achieve the reliable process controlsneeded for ef ficiency and safety, not onlymust present instrumentation techniques beimproved, but investigations of possible newtechniques —involving physical pr incipl esand phenomena hitherto largely ignored in coalprocess monitoring —will have to be activelyencouraged, even if at first sight highlyspeculative. The se rni ght include reactortemperature measurements vi a conv ent i on a 1vibrational and rotational band spectroscopyand via laser Raman scattering, reactortemperature measurements in ferr ed f r omacoustic time domain ref leetometry or fromJohnson noise; use of crystal thermometers andnovel thermocoupl e element s, e. g. ,semiconductors; acoustic part icul ate si zing,and even ultrasonic aids to particulateagglomeration and to coal liquefaction andcomminution.5. To assist the development of newi n s trumentation and the improvement ofpresently avai lab le in s truments, the fullrange of electromagnetic and acoust icradiations from reactor process componentsshould be carefully studied for possibleextraction of information on ternperatur e,composition, etc. , bearing in mind thatexcursions of these radiations can indicatedeviations from normal operation. In manycases, research on novel signal processingtechniques will be a necessary part of suchstudies.6. Reactor modeling efforts, which up to nowhave been largely i gnored, and have beencomparatively rudimentary when undertaken,

must be greatly increased, with the objectives(inter alia): of improving our abilities toi n ter pr e t and usefully employ instrument, ationreadings; and of finding indirect measures ofreactor temperatures and other impor tantreactor process parameters which are difficultto measure directly.

Ded i cat ed test facilities, forinstrumentation research and for process modeldevelopment, and ver i f ication (using theseinstruments) must be made available. Theseinclude facilities for the study of multiphaseflow as well as combustion and conversioncharacteristics. It is recognized that somefacilities exist, but additional or expandedfacilities will be necessary. Mhile operatingplants can be used for some activiti'es of thiskind, these facilities presently are availableonly on a non-interference basis and are notadequately instrumented for collect ion o fmeaningful process data.

8. References

Allen, T. , 1975, Particle Size Measurements2nd ed i t ion, Chapman and Hall, Ltd .

Amer ican Society fo r Te s t i ng and Mater ials,1977, Standard Method for Sampling Stacksfor Particulate Mat ter, DesignationD2928 —71, in Annual Book of ASTMStandards, Philadelphia, PA, pp. 592-618-

Argonne, 1980, "Overview of Coal ConversionProcess Instr umen tation ", May 1 980,Ar gonne National Laboratory Report,ANL-F E-49628-TM01 .

Bachalo, M. D. , 1980, Applied Optics 19,363-

Bachalo, M. D. , J. Ge f fken and G. Meth,1978, "Qn-Line Particle MonitoringIn strumen t at i on", Pr oc . o f Symp. onInstr umen t a t i on and Con tr o 1 for Fo ssi1Demonstration Plants, ANL-78-62-CONF780656, Ar gonne Na t iona 1 LaboratoryReport.

Ballard, Evan, 1981, Los Alamos NationalLaboratory, Private Communication.

Basedow, Ebert, 1979, "Effects of MechanicalStress on the Reactivity of Polymers'.Activation of Acid Hydrolysis of Dextranby Ultrasound", Polym. Bull. (Germany)Volume 1, pp. 299-306.

Baucum, M. E. , 1 979, "Status of theEvaluation of a Coriolis Effect Mass FlowMeter for Dense Phase Coal Flows", in theProceedings of the 1979 Symposium onInstrumentation and Control for Fossil



Energy Processes, ANL-79-62, CONF 790855,held August 20-22, Denver, CO.

Billings, C. E. , J. Milder, 1970, Handbookof Fabric Filt er Technology, GCA Ine. ,Bed ford, NA.

Bozorgmanesh, H. , et al, 1980, "OperationalCharacteristics of the On-Line Sulfur inCoal Analyzer", in the Proceedings of the1980 Symposium on Instrumentation andControl for Fossil Energy. Process,ANL-80-62, CONF 810607, held June 8-10,1981, in San Francisco, CA, to bepublished late in 1981.

Brown, D. B. , et al, 1978, "MoistureDetermination in Coal: Survey ofTechniques" in the Proceedings of the1978 Symposium on Instrumentation andControl for Fossil Demonstration Plants,ANL-78-62, CONF 780656, held June 19-21,1978, Newport Beach, CA.

Bump, T. R. , 1979, "Design and InitialTesting of Hi gh Temperat, ure/PressureSamplers" in the Proceedings of the 1979Symposium on Instrumentation and Controlfor Fossil Energy Pr oeesses, ANL-79-62,CONF 790855, held August 20-21, 1979,Denver, CO.

Car li n, Benson, 1964, "Applications ofUltrasonic Maves", in Physical Acoustics,edited by M. P. Mason, Vol. I, Par t B,Academic Press.

Carson, E. , 1977, Med. Phys. 4, 302.Cekorich, A. , 1979, "Development of an

Elemental Analyzer for Coal, Oil andSimilar Bulk Str earns" in the Proceedingsof t,he 1979 Symposium on Instrumentat ionand Control for Fossil Energy Processes,ANL-T9-62, CONF 790855, held August,20-22, 1979, Denver, CO.

Chigier, N. A. , 1976, Prog. EnergyCombustion Sci . , 2, 97.

Chigier, N. A. , 1977, Frog. Energy CombustionSci . , 3, 175.

Cooke, D. D. , M. Kerker, 1975, Appl. Optics73$,

Cummins, H. A. E. R. Pike, editors, 1974,Photon Correlation and Li ght BeatingSpect, roseopy, Plenum Press.

Cummins, H. A. , E. R. Pike, editors, 1977,Photon Cor r elation Spec t r ose opy andVelocimetry", Plenum Press.

Ellingson, M. A. , 1980, Materials Technologyfor Coal -Conver sion Pr oc e s s e s, Pr o g r e s sRepor t for Apr il-June, 1980, Ar gonneNational Laboratory Report ANL-8'0-93.

EPRI, Elect r ie Power Research Inst, i tut e,1979, Private Communication.

Farmer, M. M. , 1973, "The InterferometricObservat, ion of Dynamic Particle Si ze,Ve loci ty and Number Dens i ty", Ph. D.thesis, U. Tennessee.

Farmer, M. N. , 1976, Appl . Opt ics 15, 1984.Flynn, H. G. , 1964, "Physics of Acoustic

Cavitation in Liquids", in PhysicalAcoustics, edit, ed by M. P. Mason, Vol. I,Par t B, Ac ademi c Pr ess, pp. 57-152.

Frederikse, H. P. R. , 1963, "Properties ofSemiconductors", in American Institute ofPhysics Handbook, McGraw Hill, pp. 9-53to 9-54.

Gozani, T. , 0. J. Tassicker, R. Bevan, 1979,"Advanced Techniques and Instrumentationfor Real Time On-Line and LaboratoryAnalysi s of Coal" in the Proceedings ofthe 1979 Symposium on Instrumentation andControl for Fossil Ener gy Processes,ANL-79-62, CONF 790855, h e 1 d Au gu st20-22, 1979, Denver, CO.

Hartley, D. L. , 1975, "The Role of Physicsin Combustion", Amer ican Institute of

' Physics.Herzenberg, C. L. , 1979, "Use of Small

Accelerators in Coal Analysis and CoalSlurry Flow Measurements", in IEEETransactions on Nuclear Science, NS-26,pp. 1568-1573'

Her zenberg, C. L. , et, al, 1978, "ProcessStream Composit, ional Analysis in AdvancedCoa 1 Ut, ilization Plants UsingNeutron-Induced Gamma Spectrometry". inthe Proceedings of the 1978 Symposium onInstrumentation and Control for FossilDemon st r ati on Pl an t, s, ANL-78-62, CONF

780656, held June 19-21, 1978, NewportBeach, CA.

Her zenberg, C. L. , K. E. Griggs, R. F.Henry, M. F. Podolski, 1981, "A Study onInstrumentation Needs for Process Controland Saf ety in Coal Fluidi zed-BedCombustion Systems", Ar gonne Nat, ionalLabor at ory Repor t ANL/C EN/F E-80-12.

Herzfeld, C. M. , editor, 1962, "Temperature'.Its Nesurement and Control in Science andIndustry", Vol. 3, Reinhold.

Karplus, H. B. , A. C. Ra pt is, 1981,"Development, and Testing o f A

Hi gh-Temper ature Acoustic DopplerFlowmeter ", ANL-F E -4 69609-TN Ol, t o bepublished in May, 1981.

Kinzi, P. A. , 1973, "Thermocoupl eTemperature Measurement", (Miley) andre ferenc es under the heading"Thermocouples with Graphite or Carbon"p. 208.



Knollenberg, R. G. , 1979a, "An In-St ack FineParticle Spectrometer . A. Discussion ofits Design and Development", Advances inParticle Sampling and Measurement, U. S.EPA Report EPA-600/7-79-065, p. 169.

Knollenberg, R. G. , 1979b, "Single ParticleLight Scat tering Spectrometers", inAerosol Measurement, edit, ed by D. A.Lundgren, F. S. Harris, Jr. , M. H.Marlow, M. Lippmann, M. E. Clar k, M. D.Durham, pub. by Univ. of Fla. Press,Gainesville.

Lamb, G. E. R. , P. A. Costanza, 1980,Tex tile Res. J. 50, 661.

Liu, B. Y. H. , D. Y. Pui, A. Kapadia, 1979,"Electrical Aerosol Analyzer . Hi story,Principle and Data Reduction", in AerosolMeasurement, edited by D. A. Lundgren, F.S. Harr is, Jr. , W. H. Mar low, M.Lippmann, W. E. Clark, M. D. Durham, pub.by Univ. of Fla. Press, Gainesville.

Lundgren, D. A. , F. S. Harris, Jr. , W. H.Marlow, M. Lippman, M. E. Clark, M. D.Durham, ed i tor s, , l 979, AerosolMeasurement, pub. by Univ. of Fla. Press,Gainesville.

Managan, W. W. , J. J. Eichholz, R. W

Doer ing, 1 978, "Field Tes t o f aCapacitative Transduce forDensity/Velocity (Mass Flow} Measurementon the HYGAS Pilot Plant 'Solvent/Co alFeedl inc" in the Proceedings of the 1978Symposium on Instrumentation and Controlfor Fossil Demonst ration Plants,ANL-78-62, CONF 780656, held June 19-21,1978, Newport Beach, CA.

Marple, V. A. , K. Wi 1leke, 1979, "InertialImpactors~', in Aerosol Measurement editedby D. A. Lundgren, F. S. Harris, Jr. , W.H. Mar low, M. Li ppmann, W. E. Clark, M.D. Dur harn, Un i v . o f Fla. Press,Gainesville.

Mazumder, M. K. , R. G. Renninger, T. H.Chang, R. W. Ra ible, W. G. Hood, R. E.Ware, M. A. Smith, 1981, "SimultaneousMeasurement o f Aerodynamic Si ze andElectrostatic Charge of Aerosol Particlesin Real Time and on Single ParticleBasis", Proc. o f Third Symposium on theTransfer and Utili zation o f Par t i cul ateControl Technology, ed. by F. P.Vendi t tie, pub . by Denver ResearchInstitute, Orlando, Fla.

McLean, M. J. , D. R. Hardesty, J. H. Pohl,1980, "Direct Observation of PulverizedCoal Par tie les in a CombustionEnvironment", Central St ates Sect, ion

Meeting, The Combustion Institute,Loui si ana State Un i v . , paper 80-11.Sandia Laboratories Report.

O'Fal ion, N. M. , R. W. Doer ing, C. P.Leiter, 1981, "Review of InstrumentationNeeds for Process Control and Safety inAd vanced Fossil Energy Processes",Ar gonne National Laboratory Report,ANL-F E-09628-TM02, to be publ i shed inMay, 1981.

O'Fallon, N. M. , et al, 1976, "A St, udy onthe State-of -the-Ar t of Instrumentationfor Process Control and Safety inLarge-Scale Coal Ga si f ication,Li que f action, and Fluidi zed-BedCombust ion Systems", Argonne NationalLaboratory Repor t, ANL- f 6-A.

Papadakis, Emmanuel M. , 19 (6, "Ult rasonicVelocity and Attenuation: MeasurementMethods with Industrial Applications", inPhysical Acoustics, Volume XII, AcademicPress.

Penny, G. W. , 1977, "Using ElectrostaticFor ces to Reduce Pr essur e Drop i n Fa b r i eFi 1 ter s", Powder Separation 18, 111,"Electrostat ic Ef fects in FabricFiltration", E PA -600/7-78-1 0 2a E PA,Research Triangle Park, N. C.

Pitcher, N. D. , 1977, "Bi-Gas HighTemper atur e Thermocoupl es De signConsiderations and Operating Experience"in the Proceedings of the 19 f7 Symposiumon Instrumentat ion and Process Controlfor Fossil Demonstrat ion Plants,ANL-78-7, CONF 7 "f0729, held July 13-15,1977, Chicago, Il 1 .

Plumb, H. H. , editor, 1972, "Temperature.Its Measurement and Control in Scienceand Industry", Volume 0, Reinhold.

Raptis, A. C. , et al, 1978, "HYGASExper iment on the Feasibility ofAcoustic/Ultrasonic Flowmeter s" in theProceedings of the 1978 Symposium onInstrumentation and Control for FossilDemonstrat ion Plant s, ANL-78-62, CONF780656, held June 19-21, 1978, NewportBeach, CA.

Robinson, A. L. , 1981, Science 212, 1 06.Sessler, G. M. , 1968, "Acoust, ic and Plasma

Haves in Ionized Gases", in PhysicalAcoustics, edited by W. P. Mason and R.N. Thur ston, Volume IV, Part B, AcademicPress.

Sheen, S. H. , A. C. Rapt is, 1979, "HYGASCoal -Slurry Mass-Flow Measurement s Us i n gUl tr ason ic Cross-Correlation Techniques,Argonne National Laboratory Report,


S96 APS Study Group on Goal Utilization and Synthetic Euel Production

A NL-F E-49622-TM 07.Shewmon, P. G. , 1978, "Materials in Coal

Utilization. Mechanical and ElectricalProperties" in Scienti fic Problems ofCoal Ut, ili zatiori, Con fer ence Proceedingsedited by B. R. Cooper (TechnicalInformat, ion Center, Dept . of Energy) .

Sinclair, D. , R. J. Countess, B. Y. H. Liu,and D. Y. H. Pui, in Aerosol Measurement, ,edited by D. A. Lundgren, F. S. Harris,Jr. , M. H. Marlow, M. Lippman, M. E.Clark, M. D. Durham, pub. by Univ. ofFla. Press, Gainesville.

Smith, A. , et al, 1979, "Acoustic Monitoringfor Leak Detection in Pressurized MaterReactors", in Acoustic Emi ssionMonitoring of Pressur i zed Systems,American Soci ety for Testing andMaterials Special Technical Publicat ionASTM 167.

Smi th, A. , 1981, Argonne NationalLaboratory, Private Communication.

Spanner, J. C. , 1979, "Acoustic EmissionSome Examples o f Increasing Industr ialMatur i ty" in Acoust ic Emission Monitoringo f Pr e ssur ized Systems, American Societyfor Testin g and Ma ter i al s SpecialTechnical Publicat, ion ASTM 167.

Sullivan, J. A. , Reed J. Jenson, 1981,"Laser-Based Spectroscopic Techniques forMe a surement of Containment Species inCoal-Gasi f ication Product Streams",presented at. the American Institute ofChemical Engineer ing Meeting, Houston,TX.

Stewart, R. P. , A. M. Hall, J. M. Martin, M.L. Farrior, A. M. Poston, 1974, "NuclearMeter for Monitoring the Sulfur Contento f Coal Str earns", Bureau of MinesTechnical Progress Report 74.

Swithenbank, J. , J. M. Beer, D. S. Taylor,D. Abbot, G. C. McCreath, 1977, "A LaserDiagnostic Technique for the 1'measurementof Droplet -and Par ti cle Si zeDi st r ibut i onsri AIAA paper No. 76-69,A IAA 14th Aerospace Sciences Meeting,Mashington, D. C. , January 26-29, L976.

Also published in Experiment alDi agnostics in Gas Phase CombustionEngines, Progress in Astronaut, ics andAeronautics, editor B. T. Zinn, 53, 421.

Symposia on Instrumentation and Cont, rol forFossil Eneergy Processes, Proc. Availablef rom NT IS, 5285 Por t Royal Road,Spr ing field, VA, 226161.ANL-78-7, CONF 770729, 1977ANL-78-62 CONF 780656, 1978ANL-79-62 CONF 790855~ 1979ANL-80-62 CONF 800602, 1980ANL-81-62 CONF 810607, 1981

U. S. Environmental Prot, ection Agency, 1976,"Standards of Per formance for NewStationary Sources", Feder al Register,42 (187), pp . 42020-42028.

U. S. Environmental Protection Agency, 1977,"Standards of Per forma nc e f o r NewStationary Sources", Federal Register,43(160), 41776.

Volk, M. , M. Moroz, 1976, "SonicAgglomeration of Aerosol Part icles" inMa ter, Air, and Soil Pollution,(Netherlands), Vol. 5, pp. 319-334.

Meiner, B. B. , 1979, "Particle and SpraySizing Using Laser Diffraction", Soc.Photo-Optical Instr . Eng. Vol. 170,Opt ics in Quality Assurance II, p. 53.

Hhitby, K. T. , M. E. Clark, 1966, Tellus,18, 573.

Mhitby, K. T. , K. Milleke, 1979, "SingleParticle Optical Counters: Principles andField Use", in Aerosol Measurement,edited by D. A. Lundgren, F. S. Harris,Jr . , M. H. Mar low, M. Lippman, M. E.Clark, M. D. Durham, pub. by Univ. ofFla. Press, Gainesville.

Zweibaum, F. M. , 1978, "IntegratedRadiometer System for GasifierTemperat, ure Me a sur ement" in theProceedings of the 1978 Sympo si um onInst rumen tation and Contr ol for FossilDemonstration Plant, s, ANL-78-62, CONF780656, held June 19-21, 1978, NewportBeach, CA.

Rev. Mod. Phy's. , VoI. 53, No. 4, Part II, October 1981


IV. C. MODELING AND FLOW THEO RY

1. Introduction

Modeling of coal conversion processes, is aresearch category. with high potential forproducing results which can dramatically affectthe clean, safe uti li zation of coal. It is anarea, however, into which many physicists havebeen reluctant to enter because of both naturali nclination and background. There are,nevertheless, challen g ing and interestingproblems to which physicists could contribute.

Briefly, we find three ma jor areas inwhich physicists could have significant impact:(a) in the constructio'n of particulate burnmodels; (b) in the linkage between modeling andinstrumentation, and (e) last, but by no meansthe least, in fundamental advances in thetheory of multiphase flow phenomena.

Here, we provide an overview of thevarious theoretical aspects of coal utilizationprocesses with emphasis on those areas where,in our opinion, physicists would feel "at home"and could provide especially usefulcontributions. An example is the relationshipbetween radiat ive transport and temperatureequilibrium in combustion and combustorreactors. On the other hand, we recognize thatthere are other aspects, such as directliquefaction reactor system modeling, which aremore suitably addressed by chemical engineers,because these aspects are totally permeatedwith details of organic chemistry which are notpart of most physicists' working tools. Nehope that our overview provides enough guidancethrough this maze of varied topics to permitphysicists to identi fy the challenges andresearch opportunities available to them.

The major motivation for research intheoret ical modeling in coal utilization todayis the need for more detailed understanding ofthe processes and outputs. Realistic modelsacquire increased importance when details aresignificant. The new significance of detailsis illustrated by the following synopsis ofproblems currently faced by operators of plantsutilizing direct combustion of coal for thegeneration of electrical power. First, arapidly changing energy supply situation hasforced operat or s o f combustion equipmentdesigned for liquid or gaseous fuels toconsider, or in some cases face a requirementfor, switching to solid fuels. The operatormust then determine whether to useconventionally pulverized coal afte r

incorporating dramatic equipment changes, touse the existing equipment at reduced loads, orto fire the solid fuel in an unconventionalmode (ultra-fine particles, oil slur ries,etc. ) . The operator will find the basictechnical information on which to base thisdecision essentially unavailable. Furthermore,predictive tools necessary to ex trapo lateexperience gleaned from one coal type toanother are also non-existent. In addition,environmental considerations require combustionmodification for pollutant emission abatement.For example, nitrogen oxide (NOX) abatement canbe accompli shed economically only throughprevention of their formation, and not throughstack removal. As discussed below, one of thema jor recent accomplishments of combustion andflow modeling has been in providing sufficientunderstanding of process chemistry to devise apractical way of reducing NOX production incoal combustion. The essential point broughtout by this advance is that careful and correctparticulate burn modeling and characterizationare necessary in order to achieve a desiredchange without detracting from the originalpurpose, namely the production of electricalpower.

Modeling of coal conversion processes canbe loosely classified into three overlapppingcategories of scale, necessi tat ing ratherdi f ferent types of theoretical considerations.These are (a) the particulate 3.evel, i.e.combustion or reaction of single coalparticles; (b) zone modeling of an ensemble ofcoal particles under uniform conditions, orconditions which represent some averagebehavior; and (e) macroscopic, or reactorscale, modeling of whole coal reaction chambersof various kinds. Each successive scalesubsumes the previous one, but presents its ownunique problems in modeling. -Me discussexamples from each category in separatesubsections below.

As will be seqn, al 1 of these modelingefforts are linked to the flow conditions ofthe materials involved, requiring details ofmulti-phase flow. This is an important topicwhich pervades all modeling efforts, and one inwhich theoretical progress is -sorely needed.Me discuss this general topic separately inSection IU C-5, because it deserves specialemphasis as an area in which physicistsparticularly can contribute.

The above classif ication scheme is notmeant to discount the importance of theoreticalmodeling of free molecule reactions in coalprocesses. Such processes form an important



par t of the chain of consideration for certainoutput products; we briefly address this topicin the following part of this Introduction.

easily accessible experimentally . .The need tounderstand more det, ails would be well served bysuccessful research along these lines.

Free molecule reactions 2. Modeling at the particulate level

As we will discuss in Section IV C-P, oneof the mechanisms for NOX production is theso-called "ther mal" NOX mechanism, the O~-N~reactions in gas phase. Such reactions areamenable to theoretical treatment. In thisparticular case the reaction is reasonably wellunderstood. Reaction rate modeling can help,however, in assessing the role of intermediatemolecular processes which occur during coalcombustion or gasi f ication. Many of theseintermediate steps are currently no tunderstood. Yet, in some cases, production of"minor" const it uent pollutants may becontrolled by some such intermediate step. An

example of this is the role of the OH radicalin combustion, here a more believabletheoretical model would be extremely usefulbecause the experimental invest i gati on i sdifficult.

A full quantum mechanical treatment ofpolyatomic react io ns at the elevatedtemperatures involved in these processes isintract ible and probably not useful. However,semiclassical models in which the interatomicpotentials are speci fied f rom s pec tra 1information and the mot ions of the atoms aretreated classically are now feasible onmoderate sized computers (Lamb, 1980, Poppe,1980). With this experimentally determinedcharacterization of the interaction potentials,t;he resultant motions of the constituents arecalculated and averaged over init ia 1conditions. Relative probabilities for variousr e arrangements of the atoms can then becalculated, leading to quan ti t a tive estimatesof reaction probabilities. This is a new andpromi sing area of research, and the connectionto quantum theory via the available phase spaceis also being investigated (Reinhart, 1980).Classical estimates o f ioni zat ion andexcitation rates by charged part, icle impacthave long been successfully used in plasmaphysics, and the classical description lendsitself well to characterization of the reactionprobability as a function of temperature.These. successes in plasma physics suggest thatthis type of model may provide usefulexpressions for organization of globalexperimental data, and for charactization ofthe contributions of intermediate steps not

In this section we briefly describe coalparticle modeling. This topic is central toal 1 ex i sti ng combustion and gasificationschemes. Related topics such as surfacereactions and soot formation are included.This is an area which deserves emphasis as oneof special interest to physicists.

N'hen a small coal par tie le is immersed ina hot, oxidant-containing gas environmen t, anywater present is quickly desorbed. This isfollowed by volatili zation of the aliphaticcarbon compounds leaving a porous, solid masscalled char (Solomon, 1980). The volatilizedcarbon compounds oxidize rapidly, and, if thetemperature and oxidant concentration are highenough, completely. The behavior of the charconsists of a more or less r apid reduct ion insize, also depending on the gas temperature,particle size, and oxidant concentration.

The modeling of particulate behavior underthese conditions consists of: (a) identifyingthe physical processes i mport ant tovolatilization, temperature determination, andsize reduction; (b) correctly incorporatingthese sources and sinks in the energy, momentumand mass continuity equations (see Section IVC-5) with appropriate boundary conditions, and(c) solving the resulting coupled equations forparticle temper ature and radius as a functionof time. Analytical solut ions or approximaterepresentations are highly desir able, becausethis description can then be used as a part ofthe macroscopic modeling of reactor systems. A

high level of sophistication is called for insuccessful parameter ization of particulate burnmodels if they are to be simple enough to beuseful . An especially t horny problemencountered is the hand 1 ing of the mineralmat ter in these models, because meltingtemperatures for the mineral matter are oftenreached before burn out, and the distributionof the energy among non-reactive mineralsa f fect s oxidant transport, causing slag tocover the particle surface. It is alsoimportant to have good characterizat ion ofreaction rates. Most particulate models todate deal only with gross aspects of energyrelease and size reduction.

Combustion is the term used to describethe rapid self-sustaining exoergic oxidation of

Rpv. Mod. Phys. , Vol. 53, No. 4, Part II, October 1981


carbon compounds, and arises as a result ofheat loss and generation through reactions(Mulcahy and Smith 1969, Mulcahy 1977). At lowgas temperatures, cooling loss rates (primarilyconvective) are faster than energy gener ationrates, and the equilibrium reached is one inwhich the particle temperature is on3. y slightlyhigher than that of the environment, i.e. , nocombustion occurs. At high temperatures, theenergy generation reaction rates are large, anda new equilibrium is reached at particletemperatures which are much higher than theambient, and which are controlled by the rateof reaction and radiation transport. The charparticles, therefore, may or may not combust,depending on size and temperature (Ubhayakarand Milliams, 1976), again provided that, theoxidant concentrations are adequate. If theydo undergo combustion, size reduction is rapid(0. 1 seconds burning time for 10 um particlesat 2000K). Thus the total dwell time necessaryfor complete combustion of small particles in areactor is not large, a typical number beingone second.

The volati 1 i za t ion of the al i pha t iccompounds mentioned above does not occuri sotropical ly; instead recent experiments(Hardesty, 1980) indicate that these volatilesare emitted in jet-like spurts. This indicatesthat the aliphatic compounds may have beengasified in internal pores, and finally emergeas a jet when the pressure becomes large enough(Hardesty 1980) . Coal particles with lowcontent of aliphatic carbon compound do notexhibit this jetting. Modeling of thisbehavior has not been attempted. This effectneeds to be examined theoretically, anddeterminat ion made as to how it can beincorporated in particulate burn models.

At high temperatures, the volatilescombust completely, producing CO and H 0 only.If the temperature is reduced suffice. ently,however, the combustion of the volatiles is notcomplete, and instead these compounds form sootpar ticles. The nucleation o f t he sootparticles is not understood; it is possiblethat the models discussed in the Introductioncould be used to provide some understanding ofthis situation.

As the above description indicates,surface reactions are very important in themodeling of particulate burning. The reactionsbeing discussed here are qualitativelydifferent from those discussed in Section IV F,associated with catalysis. The lat ter ares i t e-speci fic', only certain lattice sites canproduce the reactions. In the combustion of a

coal particle, all of the atoms are eventuallyinvolved in a reaction, and a global reactiondescription is appr opr iate. Theoreticalguidance is needed for structuring the terms inthe particulate models. Chemi sorption theorymay provide one vehicle for doing this.

Chemisorption of atoms or molecules onsurfaces is a sub ject which is currently ofgreat theoretical interest (Smith 1980) .Sometimes the substrate on which the adsorptiontakes place is treated as a generali zed

Jellium" (Gunn arson, 1 977, Lang 1 978) . Thisis primarily applicable to simple (i.e.nontransition) metals. Given the global natureof the reaction rates of interest in particlecombustion-, it may be that some suchdescription could be useful in these studies.Strictly speaking a jellium description appliesonly when behavior is determined by averageelectron behavior, while, in fact, adsorptionis expected to occur primarily at selectivesi t es in the coal . Nevertheless, the detailsof the reaction sites are likely to varywidely, it may be possible to represent them interms of some generali zed descript ionrepresenting an "average" site in the spirit ofthe jellium calculation.

The standard adsorption theories deal withthe bonding potential of an atom on a surfaceby means of calculation o f the electronicpotential produced by the substrate' and atomicelectrons, either via, a model hamiltonian orusing self-consistent field methods. It may be

, possible to use these theories as a basis formodeling sur f ace reactions, i .e. , thepoten ti als calculated from such treatmentscould be used to estimate activation energiesand reaction rates. This area appear s to holdsubstantial promise of theoretical progress.

As in consideration of free moleculereactions, the important requirement for modelsappropriate to coal burning is to include thedominant physical effects. These models shouldbe based on specific detailed calculat, ionswhere possible, but the speci f ics of theproblems are in general so complicated thatcompletely rigorous calculations do not of fermuch potential, particularly because of thehigh temperature environment in which they mustbe applied. On the other hand, a simplifiedmodel containing the es sent i a 1 physical lyimportant ingredients, may predict trends, suchas temperature dependence, which can be readilyused.

Re-examination of the existing part, iculatemodels is necessary for further progress atthis time. Indeed, modeling at the particular



level is a subtle and sophisticated area ofresearch which of fers substantial challenges,because of the necessity to balance simplicityof description, inclusion of all the importantphysical phenomena, and retention of justenough detail to extract the informat ion onrelevant outputs of a coal combustion system.

3. Zone models"

Two types of reactor models have been usedhistorically. In the s impl est, or "zerodimensional" type, the entire reactor istreated as a single zone and the chemicalreactions occuring are simply treated as energysources or sinks. Mass, energy and speciesbalances are then used to infer meantemperatures and outputs. In the second typeof reactor model, geometrical symmetry i s usedto consider the reactor as a coupled series ofsections, or zones, with energy, momentum andmass balances being used to determine oned imensional prof i 1 es o f the importantvariables.

Essentially all mod cling of industr ialfurnaces and utility boilers has historicallybeen zone type modeling. These reactors aretypically very large volume reaction chambers,in which pulverized (50-200 u size) coal issprayed into (usually also premixed with) a hot,a i r s tr earn under combustion sust ainingconditions. The flow rates are high enough sothat swirling, turbulent flow occurs. Theswirling and turbulence, rather than beingcircumstantial or bothersome, are essentialfeatures of the design. They are required toprovide the air-fuel mixing for completecombustion. Most utility burners are quiteefficient, , with better than 90$ of the energyavailable in the fuel being recovered as steamenergy. Other types of industrial burners arethe fixed bed or traveling bed furnaces inwhich the coal is in a layer and hot air isforced through the bed (Beer 1980).

The modeling is accomplished via theconcept of a "well stirred reactor" (Longwelland Meiss 1955) model. These models ignore thedetails of turbulent fluid dynamics and dealwith spatially averaged properties of thecombustion system, either in layers coupledtogether, or for the reactor as a whole. The"zero dimensional" models are reviewed byMellor (1972). The appropriate averaging ofthe di f ferential equations i nvol ved i s

"Also called lumped parameter models.

discussed in Section IV C-5. In these models,global~ chemical reactions are used to describe

.the combustion.In general, these simple models have been

successful in accounting for energy production,mean temperatures, ma jor constituents of thegaseous outputs, and other averaged parametersof the reactor. This is due primarily to thefact that furnaces and utility boilers areoperated under conditions which lead to fullstoichiometric combustion of the fuel. Untilrecently this has been considered a suf f icientunderstanding of the process.

Nore elaborate one-dimensional zonemodeling for steady flows has also been done.As discussed in Section IV C-5, such modelsrequire aver aging o f t he generalt hr e e -d imensional equations over two spacedimensions (williams 1972; Ishii 1975). Manyearly models used heur i stic forms o f theappropriate equations and took into accountturbulence vi a empi r ical d if fusion andviscosity coefficients. Usually radiat iveenergy transfer is ignored or treated onlygrossly, but some zone models have includedradiative transfer (Lowe, et al. 1977) . Two-and three-dimensional modeling of turbulentflow will be discussed later.

The new constraints involving knowledgeand control of pollutant emissions have turnedup inherent weaknesses of earlier simplermodels. It has become impor tant to understanddetails of the production of minor productconstituents such as NOX and oxides of sulfur(SOX), so that, emi ssion control can beeffected. In fact, the control of nitrogenoxides can be considered one of the practicaltriumphs of modeling. It has been recognizedthat there are two sources of nitrogen oxidesin coal combustion: 1) release of oxides fromthe fuel-bound nitrogen, and 2) the breakup andoxidation of the ni trogen molecules in the hotair used as oxidant. This latter mechanism ist e rme d thermal NOX product ion. The fuel -boundNOX source is potentially four to six timeslarger than the thermal NOX production, underordinary circumstances. Ho~ever, if thevolatilization i s accomplished at hightemperature under oxygen poor conditions, theamount of fuel-bound NOX is dramatically

~Global reactions are those in which initialand final species are specified, assuming thatany intermediate steps are fast enough and

, concentrations adequate to produce the statedr esu 1 t s . F' or minor constituents, thisassumption constitutes a serious drawback.


esearch for the Techno[nO OgleS S]P~

reduced (tjwlendt 1980).experimental data from

acility (Gershman 1977).The figure shows ex eri

the release f g xi ione o nitrogen oxie g xxdes as a functj. ono oxygen ratio, relati

amount requir d fe or full stcombustion of th

s to j.chiometric

1.0) . The data io e fuel (e uiq valence ratio of

a a indicate that asconcentration i

as the oxygen

NOX decreasesxs reduced, the re production of

associated solel 'thases, approachin tg he value

Various flow co d t'e y with thermal NOX production.

but all show the sacon ttions (nswirlsri ) are shown,

e same general trend.odelin g o f various types of feed

configuration fo f has shown thor urnaces

production of NQXo gas dynamic s c an al ter t he

development; s hsubst anti

ave led to m

a 1 ly. These

examination of thmore care ful

particulatee details of

Song 1980)8181rlair 1977, Levy I980

"POX" (small part, ' lar z. culate) roduan volatiles.

iga ed (Neville 1980).Better understanding o f co

systems b o h at the micrxng o coal combustion

macroscopic lemx croscopxc and

evels, is clearl nces x. n reduction o p

maximizationional, more careful mod 1'e zng

800

PULVERIZED COAL SMALL-SCALE HOT TUNNEL

5 x 10 Btu/hr6

LOX SXIRL OMEDIUM SWIRL

HIGH SMIRI

~ 4i30E

C)

100—

1 0 2 1.4PRIMARY ZONE EQUIVALENCE RATIO

1.6 1.8

FIG. IVC-1. Reduction ofNOX emlsslons possiblethrou hg arne aerodynamiccontrol. Resultant NO con-centration (in parts per mil-lion) is plotted as a functionof the fuel to oxygen ratio,referred to stoichiometricvalues. As the oxygen con-centration is reduced thNOe

e0 emission approaches the

thermal NO value (afterGershmann et a/. , 1977).

Rev. Mod. Phys. , Vol. 53, No. 4o. 4, Part II, October 1S81


is a key ingredient in this understanding.The reduction of sulfur emissions is

another constraint, and has also been examinedwithin the existing models. Unfortunately, nosimple alterations of combustion conditionshave been found which have dramatic effects onthis emission, due, of course, entirely tofuel sulfur. As i s we 11 known, current lythese emissions must be removed from theexhaust gases. Advanced systems for utilityboilers deal with this problem by using afluidized bed combustor (FBC) (similar to thefluidi zed bed gasi f iers described in theTechnology Primer -- Section IV A) using coalparticles mixed with limestone particles. Thelime, or other agent, absorbs the sulfur gasesafter emission from the coal particles. TheFBC differs from the gasifiers in operation ini t s temperature and pressure conditions,because ful 1 combust ion, r ather thanparticular gas product production, i s thedesired result. Zone models for such systemshave also been constructed (Sarofim and Beer1978) ~

4. Example of modeling at the coal reactor level

awhile most aspects of modeling at ther eact or level are necessarily intricatelyinvolved in the engineering of the reactor,there are aspects of this problem which couldbenefit from, and provide challenges, to thephysicist. One of these deals withappropriate instrumentation. In particular,there is a close coupling between the need formor e detail, model improvement, theappropriate instrumentation necessitated bythis need, and the potential role o f thephysicist. It is probably reasonable to saythat the physicist's involvement at thereactor level of modeling would be as a teammember. His or her role would lie primarilyin critical examination of assumptions used inmodeling the physicial phenomena occurring anddevising instrumen tat ion appropriate to modelverification.

Since it is well studied, and since it isthe best defined of the various casespreviously cited, the modeling of a fixed-bedgasi fier will be followed in a little moredetail. The methodology which follows is thatof Yoon et al. (1978) . They assume aone-dimensional coal gasi f ier with a steadystate coal bed and a counter-current gas flow

xn thermal equxlzbrxum with ~t.They model the chemi cal react ions

involved using the following global reactions'.

(C) + H 0 m CO + H

(C) + C m 2 CO

(C) + 2H ~~CH~7 (C) + 0 w2(y-1)COCO + H 0 ~ CO + H

2

(a)(b)(c)

(2-Y)C02 (d)(e)

where (C) indicates that carbon in the solidis the reactant; and i s temperaturedependent (Arthur, 1951), but assumed to be aconstant in the Yoon et al. (1978) work.Chemically limited react ion rates arespecified by the Arrhenius expression in termsof an "activation energy" h, E and a r ateconstant K for each of the above reactions(labeled by n ) .

-AE /kTG S

Back reactions (i.e. , right to left aswritten) are accounted for by assuming nearequi 1 ibr ium conditions and an appropriateequilibrium constant. The actual or"apparent" reaction rates include terms forintra-particle diffusion and dependence on anempir ical ly det ermined t urbul ent tr an sportcoefficient in the gas phase. ttiti th distancealong the reactor as the independent variable,steady state differential equations are thenobtained for the temperature andconcentrations of the reacting species, usingconservation of mass and energy (includinglinearized radiative transfer).

These di f ferential equations areinvariably "sti ff", a term used to describeequat ions whose dependent variables havewidely varying time const ants. Such equationspresent well known difficulties for numericalmethods, and must be appropriately handled.In addition, the effects of mineral content,or ash, must be accounted for in terms ofsome, usually simple model.

Finally, in a complete model, the ef feetof pyrolysis must be accounted for. In thiss impl e kind of model, addi tivi ty ofgasi f ication and pyr olysis products isassumed.

Some results of this kind of model aregiven as Figure IV C-2 (Kosky 1978). FigureIV C-2 shows a composite view of a typicalealculaton for local reaction temperature(equal between gas and solid) as a function ofposition in the gasifier. Also forcomparision, the physical model of the



gasifier is drawn to scale, indicating therelative sizes of the various zones in whichdominant processes occur. Gasifier modelersshould beware, however, that result s can bevery dependent on detai led assumptions.Several conclusions can be drawn aboutmodeling gasifiers, the most important ofwhich is that the details of these models areobscure, and the accurate prediction of localeffects is very difficult.

where Pk is the instanteous density of the k .th

phase, 4k the general field quantity ofinterest, + the instant, aneous velocity. '

Theright hand side terms represent surface fluxesand volumes sources, respect ively. Theequat ions governing the rate of change ofmass, momentum and ener gy quations thenfollow if we use for g, k and g thefollowing:

(a) for the continuity equat, ion, set

5. Multiphase flow modeling

COAI

ZONES

RAWGAS

DEVOLATILIZATION

REOUCTION(GAS I FICATION)

000

COM BUSTION

GRATE

TEMPERATUREASH BI ASTGASES

UNBURNTCARBON

FIG. IVC-2. Gasifier model (after Kosky and Floess,1980).

Multiphase matter flows are conceptuallyvery dif ficult to model. These difficultiesare apparent even in their qualitativebehavior as described in the Technology Primerof Section IV A. No model currently exist swhich can exhibit all of the modes of behavioror flow regimes that are observed; in fact,there are no successful models at all for someof the regimes. Furthermore, no completelysuccessful scheme exists for determining theconditions for transition from one regime toanother .

Basic inst antaneous equations of motionfor a two-phase situation can be written interms of continuum field equations for eachphase, constitutive equations within regionscontaining only one phase, and "jumpconditions" or boundary conditions between thetwo phases at the interfaces. The fieldequations can be compactly wr itten as ageneralized conservation equation ( Ishi i,1975) for each phase, labeled by subscript k.

(b) for the equations of motion, set,

~here pk is the pressure, T is theviscous stress tensor ~ and 8 is the bodyforce,

(c) for energy conservation, set

k k k' k k k k k

where u is the internal energy, qk theheat flux, Qkthe internal (reaction) heatsource,

It is then necessary to provideconstitutive equations of state relating u, pk,'rk»qk»Qk to the mechanical variables pthe temperature Tk , and entropy Sk.

k'

Finally, boundary conditions need to beprovided, speci fying the inter -relationsbetween the phases. These must be carefullyformulated (Ishii 1975), and their formdepends on the situation considered. This isanother area -where signif icant contr ibutionscan be made. No completely satisfactoryformulation exists which includes reactivesystems. The type of considerations necessaryfor this difficult, but potentially rewarding,undertaking are very well suited tophysicists specialized skills. It isimportant to note that this requires judiciousand careful modeling to retain the essentialphysical phenomena and still dealrealist&. cally with the varied situations whicharise.

Descriptions dealing with dispersed f3.gw-beyond this instantaneous local level requireaveraging of the equations, either spatially,temporally, or using statistical averages(Ishii 1975). The type of averaging useddepends on the nature of the problem and leads



to at least two versions of averaged equationsof motion. a two-fluid type description wi thaveraged properties, or a diffusion typemodel. Physical interpretation of theresultant averaged terms and association ofcorrect kinematic coef ficients such asviscosity, mobility, di f fusion lengths, andheat conductivity, with these terms is asub stantial undertaking. Several suchformulations for non-reactive flows (Drew andSegel, 1971; Slattery, 1972, Delhaye, 1973),and reactive flows (Squir es, 1979; Davidsonand Keairns, 'l 979; Gr ac e, 'l 971; Men and Fan,1975) exist.

Modeling specific situations using theseaveraged equations remains an unresolvedchallenge, even without including reactions.No clear rules exist for choosing appropriatelengths or times for averaging, and scalinglaws would be helpful in delineating thevarious regimes. Separation of short time,small scale turbulent ef feet s from thelong-range ef fects appears to be necessary,but procedures for doing this are notavailable.

The complexity of the problem has led tot he development of one —, two — andthree-limen s ional computer programs formodeling coal util i zation processes . Theone-dimensional models involve an average overthe cross section of t he device' , thetwo-dimension al ones a s sume a symmetry axe s.The programs utilize —at a minimum -- timeaveraging of the turbulent motion, and arealmost entirely of the Eulerian type. F'orexamples of such programs the reader isreferred to the literature (Launder andSpalding, 1972; Patankar and Spalding, 1973;Denn, Yu and Mei, 1979) . These programsinclude combustion reactions in the form oft ime-dependent glob a 1 r e a c t ions. Some of thefaster reactions are treated as being inequilibrium.

It is in general not feasible to computeall details of fluid motion and follow allreactions in three dimensions, even on today' slarge computers. In addition, the input datanecessary for this much detail do not exist.The equations are invariably "stiff" with timescales varying over many orders of magnitude.Thus, a substantial amount of modeling must bedone in structuring these computer programs.Here again it is apparent that idealizations,of the type that physicists' backgrounds lendthemselves readily to, are important. Thosemodels which are based on the dominantphysical phenomena, and are not cluttered with

details which are present but do not addsignificantly to our understanding, will pointthe way to measurement of a select set ofinput data. Exciting possibilities for theuse of computer "experiments" in producingsignificant advances -in both non-reactive andreactive flow theory are present in this lineof research.

Another type of computer program whichutilizes a combined Langrangian and Eulerianapproach, has been used in modeling combustionengines (Butler et al. , 1980). This techniquecould also prove useful in coal combustion.

It is clear from the above discussionthat there are many research areas in whichsubstantial contributions can be made in coalutilization modeling. The approach and typeof ideal i zat ion wh i ch form physicists 'training seem specially appropriate in thisarea.

6. N"M for model test facility

The above brief survey of modelingaspects of coal utilization serves to showthat it i s not possible at thi s time toreliably model any coal utilization systemstarting from first principles. In fact,modeling of existing ~orking systems is onlypossible when the operating flow regime isknow from observation, if then. Much researchis needed in establishing the validity andlimitations of the various concepts used inthe idealization of both non-reactive andreactive flows.

Under current procedure, instrumentationef forts are devoted almost entirely tooperation and control of working systems (seeSection IV B) . Even for prototype models ofnew systems, the inst rumentat ion i s goaloriented, aimed primarily at producing aparticular result associated with the processbeing considered. Systemat ic collection ofdata on the reliability of any modelingconcept is at best a byproduct of suchresearch efforts.

It is apparent that an experimentalf acility whose primary goal is dataacquisition for the test ing of models andmodeling concepts would contr ibutesubstantially to progress in our understandingo f coal utili zation processes. Such afacility needs to be designed to permittesting of a large variety of flow regimes andconditions. The instrumentat ion needs to beespecially carefully designed to yieldinformation directly pertinent to concept and



model validity, and limitations. This could beaccomplished wi th currently exi stingtechniques normally used in development ofprototype instruments.

Such a model test facility ~ould not onlyyield information use ful in coal utilization,but would also provide data important toprogress in fundamental research in the

.physics of turbulence. It should berecognized that; this cannot be accomplished asa by-product of prototype process development,because of the goal-oriented "boundaryconditions" which are implicit in industrialand government sponsored process research.There is also the question of proprietaryrights which may interfere with disseminationof data pertinent to this research area, andits critical analysis by a community of peers.

7. Theoretical aspects: Summary and

recommendations

1) Me recommend the est, ablishment of anexperimental test facility, appropriately-instrumented, dedi ca ted to research ontheoretical modeling concepts. Validation ofmodels for the var ious flow regimes, andestablishment of the limitations or. conceptsused in the construction of models, are sorelyneeded areas of research. There exists nomechanism currently for funding of suchresearch on a systematic basis. Such afacility would provide information fundamentalto progress in the physics o f turbulentmulti-phase flow, which would also have impacton the understanding of coal utilizationprocesses.

2) Combustion research appears to havespecial institutional barriers to in formationexchange because it is an est abli shed,commercial ongoing effort, with heavy relianceon empirical data for proprietaryconf igurations. At present, in the"combustion community" there is no obviousmechanism for provi d ing feedback betweenexperimental research, theoretical modeling,and system operations. (For gasification andliquefaction, pilot plant design researcheffor ts, laboratory research, etc. providenatural channels for this communication. )Because we view theoretical modeling as beinghighly coupled to both f undamentalexperimental research and to plant operationand design, we recommend setting up amechanism, such as regular workshops or

meetings, e.g. under EPRI sponsorship, for thecommunicat ion of needs and in formation betweenworkers in theoretical, experimental, andprocess design and operation.

3) For both gasification and combustionreactors, current models appear to handleadequately some, perhaps even most, grossaspects of the reactors such as overallefficiency and ma jor chemical outputconstituents. However, new and more stringentrequirements concerning NOX, SOX and POX(small particulate) production require greaterunderstanding of process details and spatialinhomogenities, hence ref inement of currentmodels to include some gr eater detail isnecessary. There are in existence modelswhich have enormous detail compared to thesimpler models referred to above. However, iti s questionable whether enough reliabletheoretical or experimental input data existto make these models viable. There i s a needfor carefully constructed additions to thesimpler model s, t ai lor ed and exper imentallyverified, capable of relating these minorityconstituent outputs to operating conditions.It is envisaged t hat small changes inoperating conditions could be e f fected whichwould not reduce operating efficiency whilea f f ecting the se unwanted outputssubstantially.

4) Further progress in the theory ofsingle-phase turbulent flow would benefit ourunderstanding of both combustor s andgasifiers. Many combustion reactors rely onturbulence as an act ive transport mechanism,but even the steady state gasifier conditionsmust deal with turbulent flow. This isparticularly important in entrained bed and influidized bed gasifiers. The mathematicalmodels for gasifiers include turbulence onlyin terms of transfer parameters, such as t heReynolds number scaling par ameters; forcombustion, semiempirical formulae are usedprimarily. The proper theoretical treatmento f turbulence i s poorly developed. A

theoretically sound appr oach to modi ficationof the Navier-Stokes equations to includeturbulence in a fashion appropriate to thesesituations ~ould be very welcome. Suchef forts should be vigorously pur sued,simultaneously with seeking further advancesin semi-empirical models such as those ofPatanker and Spalding (1973).

5) Another area in which theoreticaldevelopment would be extremely useful ismulti -phase f1ow. The problem o f minimi zingenergy con sumpt ion in water -coal slur ry



transport is an example of two-phase flow;inclusion of particle si ze distributions andstability of suspension are but, two examplesof problems of current interest in the realmof fluid mechanics. Treat ing reactive,t, urbul ent, two-phase flow i s much mor ed i ff ieult . In both the multiphase flow area,and the single-phase flow area of the previousparagraph, both theoretical formulation andfurther computer modeling are needed.

A central problem in multi-phase flow isthe establishment of a "flow regime", i.e. , asteady or temporal spatial conf iguration ofthe phases of the flow, given the flow regimeit is often possible to perform some analyt, icmodeling. In recent years, "averaging theory"has been used to account for interphasetransfer effects. It may be possible incer t ain cases to model some multi-phaseaspects by Monte-Carlo techniques, as theysuggest, a means of decoupling from the initialconditions, based upon the flow regime.

6) Only a little theoretical work hasbeen done associated with direct liquefaction,because of the many unknown aspects of thatproblem. However, it, may be possible to uselimited and qualitative models to aid in theinterpretation of experiments in pilot plantsor separate experimental simulations.

8. References

Anthony, D. B. and J. B. Howard, 1976,A ICh E J. 22, 625.

Arthur, J. R. , 1951, Trans Farad. Soc. 47,

Beer, J. M. , 1980, 18th Symposium(International ) on Combustion, TheCombustion Institute, p. 149.

Blair, D. M. , J. 0. L. brendt and N. Bartok,1977, 16th Symposium (International) onCombust ion, The Combustion Institute, p.

Butler, T. D. , L. D. Cloutman, J. K.Dukowicz and R. B. Krieger, 1980, inCombustion Modeling in ReciprocatingEngines, edited by J. N. Mattavi and C.A. Amann, Plenum Publ. , New York, p. 231.

Davidson, J. R. and D. L. Keairns, editors,1979, "Fluidization, " proceedings of the2nd Engineering Foundation Conference,published by Cambridge University Press.

Delhaye, J. M. , 1973, In. Jour. MultiphaseFlow 1, 395.

Denn, M. M. , tht-C Yu and J. Hei, 1979, Ind.Eng . Chem. Fundm. , 18, 286.

Drew, D. A. and L. A. Segel, 1971, Studiesin Appl. Math 50, 205.

Gershman, B. , M. P. Heap and T. J. Tyson,1977, Proc. Second Stationary SourceCombustion Symp. Vol. V, p. 65 EPAPublication EPA-60017-77-073e .

Grace, J. R. , 1971, A1ChE Symp. Series, No.116, Vol . 67, 159-167.

Gunnarson, 0. , H. Hjelmberg and B.Lundquist, 1976, Phys. Rev. Lett.292; Surface Sci. 63, 348, 1977; SurfaceSci. 68, 158, 1977.

Hardesty, D. R. , 1980, privatecommunication.

Ishii, M. , 1975, Thermo-fluid Dynamic Theoryof Two-Phase Flow, Eyrolles, Paris,France.

Kosky, P. G. , and J. K. Floess, 1980, Ind.Eng. Chem. Proc. Des. Dev. 10, 586.

Lamb, M. J. Jr. , 1980, unpublished.Lang, N. D. , and A. R. williams, 1978, Phys.

Rev. B 18, 616.Launder, B. E. and D. B. Spalding, 1972,

Mathematical Model of Turbulence,Academic Press, NY.

Levy, J. M. , L. K. Chan, A. F. Sarofim, andJ. M. Beer, 1980, '18th Symposium(International) on Combustion, TheCombustion Insti tute, p. ....

Longwell, J. P. and M. A. gneiss, 1955, Ind.Eng. Chem. 47, 1634.

Lowe, A. , T. F. Mal 1 and I. Mc Stewar t, 1977,AiChE 23, 440.

Mc Kee, D. M. , 1974, Carbon, 12, 453.Mellor, A. M. , 1972, in Emissions from

Continuous Combustion Systems, M.Cornelius and N. G. Agnew, eds. , Plenum,NY.

Mulcahy, M. F. R. , 1977, Conferencepresentation (unpublished) ~

Nulcahy, M. F. R. and I. M. Smith, 1969,Rev. Pure and Appl. Chem. 19, 81.

Neville, M. , R. J. Quann, B. S. Haynes, andA. F. Sarofim, 1980, 18th Symposium(International ) on Combustion, TheCombust ion Institute, p. ...

Patankar, S. V. and D. B. Spalding, 'l973,14th Symposium (International) onCombustion, The Combustion Institute, p.605.

Poppe, D. , 1980, Chem. Phys. 45, 371.Reinhart, M. P. and J. Jaffe, 1981, in

Quantum Mechanics and Mat hemat ic s inPhysi. cs and Chemistry, Plenum Press, NY.

Sarof im, A. F. and J. M. Beer, 1978, 17thSymposium (International) on Combustion,The Combustion Inst i t ute, p. 189.



Slattery, J. C. , 1972, Momentum, Energy, andMass Tr an sfer in Cont inua, Mc Gr av-Hi 11 .

Smith, J. R. , 1980, "Theory ofChemisorpt, ion, " edited by J. R. Smith,Topics in Current Physics, Vol. 19,Springer -Ver lag, NY.

Solomon, P. R. , 1980, Proc. of Conf . on theChemistry and Physics of CoalUt, ilization, West Virginia Univ. ,Morgantown, M. Va. (to be published).

Song, Y. H. , J. H. Pohl, J. M. Beer and A.F. Sarofim, 1980, private communication(unpublished) ~

Squires, A. N. , 1979, "Applications ofFluidization in Coal Technology, "Research Notes, published by VirginiaPolytechnic Institute and StateUniversity.

Ubhayakar, S. A. and F. A. Milliams, 1976,J. , Eleetroehem. Soo. 123, 707.

Wen, C. Y. and L. T. Fan, 1975, Models forFlow Systems and Chemical Reactors,Marcel Dekker, NY.

Wendt, J. 0. L. , 1980, Prog. Energy Combust.Sci. 6, 201.

brendt, J. 0. L. , D. W. Pershing, J. W. Leeand J. W. Glass, 1978, 17th Symposium(International ) on Combustion, TheCombustion Institute, p. 77.

%williams, F. A. , 1965, Combustion Theory,Addison-Wesley, Reading.

Woodall-Duckham, Ltd. , 1974, "Trials ofAmerican Coals in a Lurgi Gasifier at,Westfield, Scotland, " NTIS FE-105.'

Yoon, H. , J. Wei and M. N. Denn, 1978,Al Ch E J. 2~, 885.



IV. D. MATERIALS IN COAL CONVERSION ANDU TI L I ZATI ON

1. Introduction

The past three decades have seen greatstrides in materials science. Physics andphysicist s —— e spec i al ly solid st atephysieists —have played an important role inthis development, and a significant fractionof all physicists now work in materialsrelated areas (COSMAT, 1975) . Mater i alascience, it i s well known, is amultidisciplinary enterprise and one whichrequires the closest cooper ation betweenengineers and scientists with a wide range ofspecializations and interests.

Despite the progress realized, it seemsfair to say that the advance of mod emtechnology i s more of ten 1 imi t, ed by theproperties of available mater'ials than by anyother single factor. This is especially tt'ueo f energy conversion technologies in generaland of coal technology in particular. Ofcourse the materials constraints are oftenf ur ther compl i ca ted by economic ones.Stainless steel, because i t is more expensiveand more difficult to fabricate than commonsteel, may not be used in every ease where itper forms better . Solutions to materialsproblems must be economically feasible as wellas scientifically elegant.

Many of the materials problems that arisein coal technology are familiar in othercontexts (e.g. hydrogen at, tack, corrosion,erosion, etc. ), but they appear in coaltechnology in particularly severe form or asinteracting ef feet s. Because of coal 'somnipresent hetero-atom and mineral content,for example, the technology must contend withcorrosion by multiple oxidants or withsimultaneous corrosion and erosion, etc.Other problems arise which are unique to coaltechnology; in NHD generation, for example,t he electr ical transport mechanisms andproperties of slags become important.

Both the general mater ials problems andthose peculiar to coal technology are wellrecognized by the DOE and other federalagencies. The federal government has donemuch to encourage basio and applied researchin almost every important aspect of materialsscience. Still, much remains to be done, andnew scientific developments continually opennew pathways to improved mater ials science.This section is devoted to discussion of a few

urgent materials problems and of their contextin coal technology.

The selected topics are:environmentally promoted

fatigue and fractureerosioncorrosionproperties of structural

eeramiesproperties of refractories

and slagsThese topics may be recognized as involving toa great degree the chemical physics ofsur faces and interfaces, together withdiffusion and microstruct, ure. ttite believe thatr ecent ly discovered or applied scientifictechniques should facilitate much acceleratedprogress in the . study of such phenomena. The .

topios are taken up in turn and in each caseprovided with a brief discussion of thecontext in coal technology.

2. Environmental fatigue and fracture

a. The hydrogen environment

Coal gasifi er s operate a t hightemperature, and one of their most importantproducts is hydrogen, liquefiers operate atmoderately elevated temperatur e and highpressure, and one of the principal reactantsfed to them is hydrogen. (See Section IV-A 5,6) In each case the ef fects of hydrogenattack and hydrogen embrittlement uponreactors and cont ainrnent vessels must bereckoned with. Coal and it, s products,moreover, generally contain alkali salts andother impurities of such a nature that stresscorrosion cr acki ng i s a common pr oblern."Hydrogen attack" and "hydrogen embrittement"denote different phenomena, neither of whichis fully understood.

The deterioration of materials in thepresence of a hydrogen environment can takeplace in two general ways: by chemical attackon one or more phases within the material, andby a physical embrittlement of the lattice.Both lead to embrittlement of a material and,ult, imately perhaps, catastropic crack growth.The first, , called "hydrogen attack", can takethe form of a methane bubble formation fromthe reaction of hydrogen with the carbon inthe steel, or it may be metal hydrideformation. The purely physical at, tack knownas "hydrogen embrittlement" appears to arisebecause interstitial hydrogen may disrupt thelattice structure. The ef feet is strong



enough. so that a material may fail rapidly.The kinetics and f incr details of these

processes are not clearly understood. Forexampl|. , recent work on crack initiation andpropagat ion under hydrogen attack conditionshave indicated that plasticity .is enhanced(flow stress reduced) prior to ultimateembrittlement (Tien et al, 1980) .

In coping with problems of hydrogenattack the chief design tool is the Nelsoncurve (Nelson, 1969), which shows thecondi tions of vessel temperature and hydrogenpartial pressure below which the steel willnot fail by hydrogen attack. Such failuresfrom formation of high pressure methanebubbles by r eact ion wi th the carbidesconcentrated at the grain boundaries are usedempirically to develop the Nelson curves fromoperative experience. Ho~ever, they should bederivable theoretically from the criterion ofthe lowest temperature (at a given hydrogenpres sur e ) for dec ar bur ization or methanebubble formation from hydrogen reacting withgrain boundary carbides. As yet thi s has notbeen done.

"Hydrogen embrittlement" signifies thehydrogen induced r educ t ion of ambienttemperature ductility of mat, erials such asiron and steel' through what appears to be amechani sm more int r i n sic than those ofhydrogen attack. One theoreticalinterpretation of this effect involves the useo f pair potentials both to analyze an ironlattice containing interstitial hydrogen atomsand to model atomic interact ions at a cracktip in such a lattice (Wilson, 1980; Gehlen etal, l 972) . Stress required to cleave t, hehydrogen impregnated lattice is found to bemuch lower than that for plain iron. Morework o f this natur e may improve ourunderstanding of hydrogen and other latticeembrittlement phenomena such as stresscorrosion cracking.

b. Greek nucleation and growth

Physics researchers are beg i nning t oaddress some of the basic questions concerningboth static and dynamic crack characteristics.The basic approach, in contrast to theclassical linear clast icity continuummechanics treatment, is to investigate theinfluence of the discrete atomic nature of asolid on the energet, ics and kinetics of crackpropagation, including the highly non-linearinterations at the crack tip. The ma jor

/

approach so far has been theoretical with themodels varying from analytical treatment, ofsimple models (Fuller, 1978) to ratherelaborate computer simulations (Paskin et al,1980; Wilson, 1980), with emphasi s on theresponse of brittle solids. Several importantpoints have been emphasized: 1) the int;ricateinterplay between crack propagation anddislocation generation (Paskin et al, 1980;Rice and Thomson, 1974), 2) the limitations ofthe Griffith continuum theory (Fuller andThomson, 1978; Paskin et al, 1980), 3) latticetrapping and thermally activated crack growth(Fuller and Thomson, 1978), and 0) thei nteract ion of a crack with an ambientatmosphere (Milson, 1980; Thomson, 1980),e.g. , hydrogen embrit tlement. In connectionwith item 4) two avenues have been opened up:a) application of general rate process theoryand thermodynamics, and b) computersimulations in which the system is built upfrom a quantum mechanically controlled localmetal-hydrogen reaction through the atomicresponse of the metal matrix (via pairpotentials) t o the 1 ar ge scale continuumresponse (Gehlen et al, 1972; Wilson, 1980) .The emphasis has been on the environmentalweakening o f the bonds at the crack tip.However, the possibility of a chemi calbridging reaction increasing the toughness hasalso been suggested. This point is well worthemphasi zing. There is no a priori reason whythe environment should only act to weaken thebonds at the crack tip. Studies are needed tof ind environmental conditions whi ch mi ghtactually strengthen the crack tip, and therebyprevent fur ther cracking. The theories,although very pr el imi nar y, are highlysuggestive and should be pursued actively.

Much less is known theoretically aboutcrack nucleat ion, which is presumablycontrolled by bulk and surface inhomogeneitiesprimarily (e.g. , nucleation and growth ofcavities at grain boundar ies and coalescenceinto cracks, nucleation and growth of methanebubbles, di slocat ion interaction withprecipitates, etc. ). Fracture control ofstructures, based on fracture mechanics (Lawnand Milshaw, 1975; Wells, 1981), utilizesprobabilistic mechanical analysis, fatiguelife analyses, or other forms of stressanalyses based on t he oper ation for which thestructure is intended, combined with knowledgeof the properties of the material of which thestructure is made. One of the key designparameters is K&c, the fracture toughness,which is a measure of the failure stress for a



given crack size. The fracture toughness forunstable crack growth and failure is oftenreduced by environments which promote crackgrowth. Thus the crucial step influencingprobable failure in service is subcriticalcrack growth to the critical crack size(Tomkins, 1981) . Frequently the stressdriving the crack is of a transient natureinvolving thermal stresses during startup andshutdown, suppl erne n ted by residual andoperational stresses. Thus, the subcriticalcrack growth problem has a fatigue nature inwhich the rate of crack growth i s a functionof the stress cycle and the stress.

The base line for fatigue crack growth isa vacuum environment in which minimum crackgrowth is observed. There is considerableef fort on t he phenomenol ogical aspects ofsubcritical growth; and, as indicated above,investigation of the pert inent atomicinteractions are beginning. Little is knownof the mechanisms for the acceleration o f thecrack growth in aggressive environments, suchas hydrogen or acid chlorides. It is herethat physics might have an even mores i gni ficant contribution to make inunderstanding how the lattice responds to suchenvironments.

3. ErosionIn its every facet, - coal technology

requires the direction and control of fluidstreams laden with solid particles. Erosion—often complicated by corrosion -- is anever present problem. Coal itself is aconglomerate of solids .that display a widerange of mechanical properties. The immediateresidue of pyrolysis is char, also a solid.Combustion, gasification, or liquefact ioninvariably leaves a residue of solid ash ormineral matter. Some, at least, of thisresidue is in a finely divided state and isentrained in the product gases or liquids.Slurry pumps are needed to inject coal-oil orcoal-water slurr ies into r eact ion chamber s.Letdown valves are needed to return thereaction products to ambient conditions. Fansor blowers must move particle laden stackgases thr ough bag-houses and scrubber s.Burner nozzles must handle streams of air withentrained pulverized coal. And, if combinedcycle power generation with gas turbines is tobecome a reality, systems must be designed tocope with ash particles entrained incombust;ion products and synthesis gas. Byitself the mineral matter present in coal isextremely erosive, when combined wi th the

corrosive substances in the process stream(e.g. sulfur, alkali metals, chlorides, etc. ),it is destructive to practically allmaterials.

Solid particle erosion is the mechanicalremoval of metal (or ceramic) as a result ofthe impact of soli d part ic les, usuallyentrained in a high velocity gas or liquidstream. Some specific examples of erosionare. erosion of letdown valves in coalliquefaction systems by ash, mineral, and coalparticles suspended in the coal liquids;erosion of gas turbine blades and vanes bysolid particles entrained in the hotcombustion gases; and erosion of fire sideheat exchange tubes in boilers. Erosion isonly one of the forms of wear ', valve faces maybe scored by solid par ticles trapped betweenthe operating surfaces, for example.

There has been a reasonable amount ofbasic investigation of erosion processes [seeFinnie (1979) and Hutchings (1979)j which hasshown that for ductile materials erosion is ata maximum at an impact angle (angle withrespect to tangent) of approximately 3Q, andthat the erosion damage varies with the impactvelocity raised to a power greater than 2.Complementary approaches have analyzed thefluid dynamic determinants of particle impactparameters as carefully as possible, using alargely empirical approach to predict thelocation and extent of erosion damage. It iscustomary to study and measure erosion at highwear rates and thus to characterize theerosion resistance of materials. Me believegreater emphasis should be placed uponstudying threshold effects in erosion (e.g.velocity thresholds, particle size thresholds,etc. ). An understanding of thresholds forerosion could contribute importantly to thedesign of viable systems.

The effect of temperature on erosion isnot understood. At elevated temperature, mostmaterials oxidize, so that in many cases theincident particles are eroding the oxiderather than the met als. The characteristicsof erosion differ for brit tie and ductilemater i a 1s, and the problem of the eros i.on o f acomposite consisting of a hard britt;le thinlayer on a ductile substrate has receivedlittle study. The work of Pettit et al.(1980) has shown that at elevatedtemperatures, there may be a significantinteraction between erosion and corrosion; andqualitative analysis suggests that the effectscould be synergist;ic or antagonistic.

There exists at present only imperfect



understanding of the mechanism of metalremoval during erosion; much more work needsto be done. Erosion mechanisms in morecomplex systems (two phase alloys, thinbr ittle layers on duct ile metal substrates,coatings, etc. ) have not been studied in anyphysical detail. The nature of the erodingparticles varies from coal to coal, and it,would be of value to under stand how theerosiver ess of particulates is related totheir other proper t ies (e.g. hardness,strength, and shape). Existing analyt, icalapproaches are far from being predictive; theyusually do not include the particle propertiesexplicity and are seldom formulated in termsof the properties of the material beingeroded.

In short, the fundamental physics oferosion processes is poorly understood and thefield appears wide open for theoretical andexperimental studies.

4. Corrosion

The use of coal i n ener gy systemsintroduces several problems of corrosion anderosion of the materials of construction. Theharsh chemical environment is provided both bythe nature of the combustion gases of the coalused and by the presence, depending on processtemperature, of liquid ' coal slag or solid ashparticulates. The most severe processes inthis regard are those in magnetohydrodynamicgenerators and the slagging gasifiers. Allprocesses using coal face corrosion anderosion peculiar to, and accelerated by, thespecial nature of coal as a fuel sourcecontaining mineral matter. The subject ofcoal slag is treated in section IV D, 6.

Some problems are peculiar to the variousenergy generation processes. These processesinclude those in the pulverized coal boiler,the fluidized bed combustors, and turbines inthe high temperature combined c yc legasification process.

Pulverized coal boiler

pul ver i zed coal (P. C. ) boiler istoday the mainstay o f the power industry,producing almost half the thermal energy forgeneration of electricity. As the nationconverts from oil and gas to coal, therelative importance of the pulverized coalboiler will become even greater.

Analyses of power plant outages (Edison

Electric Institute, 1975, 1976), show that theboiler is the least r el i ab le component of theentire system, followed by turbines,condensers, and generators in that order. Theshortcomings of boilers are chiefly the resultof corrosion on either the f ire side or thesteam side of the water walls, superheatersand reheaters. Me have only approximateunderstanding of t he corr os ion mechanisminvolved. Corrosion on the steam side perhapsis the easier problem to analy ze f r om ascientific point of view, as variations incoal and firing conditions in various parts ofthe boiler result in a wide variation in fireside corrosion phenomena. Erosion also is amat ter of concern in the combustion chamber,but is less important here than in otheraspects of coal technology, such as in coalgasifiers and liquefiers.

A serious problem in boilers that,adversely af fects the operation of the highpressure and intermediate pressure turbines isthe formation of magnetite (Fe 0 ) scale onthe inner walls of superheaters dna reheaters.The scale spalls during temperature cycles andis swept into the turbine, causing solidparticle erosion. Only rudimentary informationis available on magnetite scale and thestresses generated during thermal cycles thatcause exfoliat, ion to occur (Rehn, 1981).

F luidized bed combustion

The product o f desulfurization influidized bed combustion is -CaSQ, which issolid but corrosive at the reactiontemperature. This in-bed corrosion problemappeared to result from the low oxygen contentin the fluidized bed (Stringer and Erlich,1976). A3 though the overall oxygen content isabout 10 atmospheres (10$ excess oxygen),the oxygen copPent of the bed itself is verylow, about 10 atmospheres, with most of theoxygen contained in the bubbles of air risingthrough the bed. The equilibrium between SOand 0 in the CaO-CaSO system is such tha4the sulphur potential (fugacity) is about 10atmospheres . At th i s sul phur potential,sulfidation of heat exchangers always becomesa matter of concern for metal temperaturesfrom 650C to the bed temperature of 850-90OC.Nickel-based al loys ar e part icul arlysusceptible because of the low melting pointof the Ni-NiS eutectic. Surface melting ofthe eutectic leads to catastrophic corrosion.Ir on-based al loys sur vive longer. By



contrast, nickel was a strong performer in thegasificat ion envi ronment (Perkins, 1 979) .This corrosion problem in the fluidi zed bedboiler is described below under hightemperature corrosion.

The alkali salts in the combustionproducts from a pressurized fluidized bed mayalso be suf ficient to promote hot corrosion ofthe expander turbine, otherwise the in-bedheat exchanger corrosion problem in thepressurized fluidized bed combustion is muchthe same as in the atmospheric fluidized bedcombustor.

Turbines

There are many mater ials problems insteam turbines, but they are not un iquelyassociated with coal as the source of heat.Thus, steam turbines and condensers per sewill not be covered in this report . The useof coal or coal derived fuels in gas turbines,in the simple Brayton cycle, or in thecombined Br ayt on -Rank i n e cyc le, however,brings characteristic problems of its own andwill be discussed.

Use of the high temperature gasificationcombined cycle is necessary in order tojustify the added expense of gasi fying coalbefore combustion, in comparison to burningcoal in the conventional pulverized coal powerplant equipped with a scrubber. The efficiencyof energy conversion in a gasificationcombined cycle may approach 45$ given thedevelopment of suitable advanced gas turbines;~hereas a pulverized coal power plant with ascrubber rarely has a thermal efficiency inexcess of g5$. The thermal efficiency of thecombined cycle plant depends strongly on thecombustion temperature in the gas turbine. Atleast 1200C is necessary (and 1400C isdesirable) for the inlet temperature to theturbine. The exhaust temperature from the gasturbine should be high enough to raise 540Csteam in a waste heat boiler. There are twoknown approaches to the high temperature gasturbine for operation in this range -- anuncooled ceramic turbine or an air-cooled orwater-cooled metal turbine.

Metallic turbines used in combined cycleplants using coal -derived fuels will beexposed to corrosive agents and particulatematter that might adversely a f feet li feexpectancy. The chief agents of corrosion arealkali metals which combine with residualsulfur in the combustion products to produce acondensed salt of sodium or potassium sulfate

on the metallic turbine section. In thetemperature range above its melting point andbelow i ts dew point, , the condensed salt cancause accelerated corrosion. The coal derivedfuel must therefore be purified sufficientlyto permi t the metal lie turbine alloys tooperate without excessive corrosion.

Low temperature corrosion

Coal conversion systems also present anumber of situations, where low temperaturecorrosion may be encountered. The acidcorrosion due to high concentration of sulfuroxides, or the fractionation of impurities toaqueous condensates which contain relativelyhigh concentrations of dissolved salt, caninitiate corrosion on surfaces. These inturn, act as stress raisers and can initiate afatigue failure. Corrosion may also occurfrom the ~ater soluble components of the coal(i.e. slu, rries) or from the presence ofchlorides, organic acids, water, or sulfurcompounds in non-aqueous liquids.

No distinction has been made here betweenthe various modes of corrosion which mayinclude. broad front att ack, crevicecorrosion, pitting corrosion, or the possibleinteractions between corrosion and mechanicaleffects such as. corrosion fatigue, frettingc or r o s i on, str e ss corrosion-cr acking . Infact, stress corrosion-cracking is likely tobe a particular problem because of thepresence of chlorides in coal. The role ofthe physici st here ~ould be to look forsystematic behavior in the growthphenomenology of these corrosion modes as anapplication of the general problem of theinteraction between solid sur faxes and thegaseous or liquid environment in contact.

Mixed oxidant attack and molten salt corrosion

Most of the, hot corrosion problems can begeneralized into 1) mixed oxidant attack and2) molten salt corrosion. In gasifiers, insome regions of fluidized bed combustionsystems, in substoichiometric ("fuel-rich")regions of low NO staged combustors, and insome other special situations, enviqgnmen$qexist, with low oxygen potentials ( 'IO &-10atm) and relatively high sulfur (10 —10atm) and carbon (0. 1-1.0 atm) activities. Inthese environments mixed oxidant attack is aproblem (St,ringer and Erlich, 1976).



The other major form of high temperaturecorrosion involves the presence of moltensalts, usually of the alkali metals, on themetal surface. The gas phase is usually thecombustion product gas, and is typicallyrelatively oxidizing with oxygen potentials inthe range 0. 01 to 5 atm. , This salt-inducedattack is often called hot corrosion, and hasbeen a major problem in oil-fired gas turbinesf or many years. A similar problem isencountered in the radiant superheater ofcoal-fired boilers.

Mixed oxidant attack

Mixed oxidant attack is caused by thetransport of one of the minor oxidants (sulfuror carbon) through the (otherwise"passivat, ing") oxide layer on the alloysurface. The minor oxidant then reacts withthe metallic element responsible for formingthe protective oxide (usually chr omium oraluminum) to form a dispersion of sulfide orcarbide particles within "the depleted oxidematrix. " If these particles are coarseenough, they then oxidize- in situ as the localoxygen activity increases. This prevents themaintenance of the protective oxide. Often,the sulfur or carbon released by the oxidationof the sulfide or carbide does not escape backinto the atmosphere, but makes its way furtherinto the metal, so that the process, onceinitiated, may be self-maintaining. It isthought probable that the transport of theminor oxidants through the external oxidelayer- takes place through low resistancepaths, such as grain boundaries, dislocationcores, pores, or f ilaments of a second phase.There has been a a considerable volume of workon the phenomenology of mixed oxidant attack,principally aimed at the problem of materialsin gasification systems (Hill et al, 1979;Perkins, 1979). Analysis of gasifer corrosionby Perkins reveals evidence of cation motionthrough the oxide as the principal mechanismof sul f idat ion. St r in ger ( 1977) showsevidence that in conventional power plantcorrosion, anion motion is dominant. Clearlysome very fundamental studies are necessary toclarify this problem.

Molten salt corrosion

Molten salt induced hot corrosion in gasturbines has been the sub ject of much study.(See the recent review art, icle of Stringer,1977) . The boiler superheater problem is

discussed at length in the volume by Acid(1971). Essentially, there are two aspect, s tothe reaction. The simplest situation is whenthe deposit is a molten layer of sodiumsulfate. This may act as a diffusion barrierto oxygen so that at the metal/salt interfacethere is a low oxygen activity. Theconsequent dissociation of the sulfate resultsin a high sulfur activity, leading to sulfurattack.

The other aspect is that the molten saltmay be capable of dissolving ("fluxing" ) theprotective oxide layer. Usually, protectiveoxides are stable in the neutral(stoichiometric) alkali sulfate, but, cease tobe stable if the salt becomes suf ficientlybasic (alkali-rich) or acid (SO -rich). Theacidity of the salt, will be a function of thealkali content and the local SO partialpressure in the envir onment, . Fot transportrates to be rapid enough to cause seriousattack, the salt layer must be molten, soconstituents that can generate a low-meltingpoint phase will consequently inereae theextent of observed attack. For this reasonsodium sulfate induced corrosion is observedat temperatures as low as 575C even though theanhydrous sulfate melts at 88WC.

Research issues of interest to physicists

Because little generalization has beenpossible with the- wide diversity o'f hightemperature corrosion problems, it has beendifficult to sift through the details in orderto get a clear picture of research needs. Buti;n most environments of concern in energysystems, virtually all metals of constructionare unstable with respect to their oxides,sulfides, carbides and so forth. Anoxidation-resistant alloy owes its resistanceto the formation of a chemically andmechanically stable protective oxide layer oni ts sur face. Further oxidation then involvesthe transpqrt of one or more of the reactantsthrough the oxide, and if the transport ratesare suffieently slow, the reaction rate is lowenough so that in ef feet the alloy does notdegrade appreciably in times comparable withthe design lifetime of the component. Inpractice, this means that the meehaniealadhesion of the as-grown oxide to thesubstrate must, be good. For hi gh-temper a tur eoxidation resistance, the protective oxides areusually Al 0, Cr 0 or in some cases SiO


S'1 14 APS Study Group on Coal Utilization and Synthetic Fuel Production

although ot, her ox x. des are theoretical1ypossible.

Development of a protect ive oxide on analloy surface involves a number of criteria.Obviously, the protective oxide must, bethermodynamically the most st, able phase inoontact with the alloy. The diffusion of theoxide-forming element (Cr, Al, or Si) out fromthe alloy to the oxide/metal interface must, besuch that a continuous adherent ext ernaloxide, rather than an internal distribution ofoxide particles, forms. Finally, when theprotective oxide is nucleated on the surface,it must spread sufficiently rapidly over themetal sur face to form a cont, inuous layerbefore it is mechanically disrupted by lessstable oxides.

Mechanical adhesion of the protecti veoxide to the substrate is a complex topicinvolving the familiar aspects of inter facialenergies, but also the more qualitative aspect,of the development of an irregular re-entrantinterface ("keying" ) and the establishment ofa growth process which allows the oxide tofollow the retreating alloy sur face withoutthe development of voids between the oxide andthe substrate.

Nuit, icomponent diffusion theory is basicto a full understanding of alloy oxidation.This sub jeot is still at an early stage.Nucleation of oxide phases, the growth of thenuclei across the surface, and the competitionbetween oxides of differing stability and

. growth rates are topics of concern. Work iscurrently being conducted along these lines.A good star ting point to look is Mhittle(1972) and Smeltzer et al (1976).

Transport processes in ox ides, and inparticular, the transport of minor speciessuch as sulfur or carbon, are of interest.Short-circuit transport is of at least as muchinterest as bulk diffusion . .The oxide/metalinterface is not well understood, and indeedis poorly characterized. Much work needs tobe done on the interface in "dynamic- systems"in order to understand the mechanisms ofmechanical adhesion. The detailed structureof the external oxide layer needs a great dealmore work. It is probable that the init ialnucleat ion processes on a metal surfaceexposed to a gas mixture may have a profoundeffect, on the later course of the reaction.The techniques of surface physics shouldcontribute substantially to a solution ofthese problems.

Protective oxides may be either thosenatur ally ocour ing on a material sur face, or

they may be composed of components foreign tothe normal alloy structure. A particularlyintriguing subject for future research isexpanding the area of modification of surfacesto withstand environmental streses such assulf idati on, stress corros ion cracki ng,erosion, sub-critical crack growth, moltensalt at, tack, hydrogen embr it tlement, etc. Inaddition to protective oxide formation on thesur face of the metal or alloy there are alsosurface coatings such as chromiding andsurface modification via diffusion and ionimplantation. There is still much fundamentalwork to be done in understanding the role of,and developing our abi 1 i ty to modify, thesurface chemical and physical structure of amaterial.

Ion implantat ion is in its infancy as atechnological tool and holds much promise forthe future. Progress has already been made instrengthening the wear resistance of bearingsby this technique on a commerical scale andsome progress on an experimental scale made instrengthening a material against, corrosion(Antil1 et al, 1980; Collins et al, 1980) .Bo t h t heoret ical and experimental sur facestudies can contr ibute substantial ly to theextension of the technology of sur f acemodi f ication by fundamental physical studiesof the formation and properties of thesesurface layers.

One type of corrosive at tack at lowtemperatures that deserves the specialattention o f physic i st s, and one that, iscertainly a concern in ad vanced coalconver sion systems i s erosion-corrosion.Erosion-corrosion is a distinct form oflocalized at, tack that occurs on some metals inareas where the turbulence at the metalsurface is high enough to cause mechanical orelect rochemical disruption of a protectiveoxide or other surface film. It is not simplytwo ef fects occuring simultaneously. Theprocess is usually accelerated when abrasivesolid particles are entrained in the water,but solid particles are not a prerequisite forerosion-corrosion. Some aspect s o ferosion-corrosion are understood. Forinstance, because turbulence increases withincreasing velocity, higher water velocitiesfavor the initiation of erosion-corrosion.Flow geomet, ry al so in fluence s turbulenceintensity, and is as important as velocity indetermining whether the protective film willbe disrupted.

There is no agreed upon theory oferosion-corrosion. Many believe in the

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 198'l


"critical shear stress" theory in which theoxide or other sur face film is mechanicallystripped from those areas on the metal sur facewhere the shear stress exceeds a certaincritical value. But it has been pointed outthat the shear stresses actually encounteredin practice are very small ( ~~' '

1 psi) .Others prefer the "electrochemical breakdown"theory, in which the level of turbulenceintensity influences the local i zedelectrochemistry at the metal sur face, andcauses local electrochemical (not, mechanical)breakdown of the protective f ilm. There isalso no clear picture of the role of solidparticles entrained in the water. The size ofparticles, the volume loading, the particleshape, and the hardness all have an ef feet onthe rate of erosion-corrosion but onlyempirical relationships have been derived sofar. .Me need to investigate erosion-corrosionin greater depth in an attempt to eliminatecontroversy, explain the influence of thevarious important var iab les, and come up witha successful theory of erosion-corro sion thati s applicable to particle —free andparticle-loaded waters alike.

5. Structural ceramics

Structural ceramics are a class of highstrength, high density, brittle mater ialsgenerally hot pressed and sintered from thepowder. The most common materials are thecarbides and nitrides of silicon and boron.Many researchers throughout the world areengaged in processing and structure researchto f ind ways of producing desir ab lemicrostructures or of fabricating specializedshapes for use in devices like gas turbines.

The first major effort to apply ceramicsin gas turbines was started by A. R PA/DOD, andis presently continuing with DOE support,mainly for the development o f automotive gasturbines, but also for large industrial gasturbines for electricity generation. Earlyefforts concentrated on developing a designtechnology for brit tie materials andest abli shing a data base for the mostpromising ceramics. Although this proved moredifficult than anticipated, a systematic andgenerally accepted design methodo logy ha semerged, a materials data base for short termapplication has been established, and someceramic parts have been successfully tested intruck turbines and 1ow pr essur e heatexchangers (evangels, 1978; Uy et al, 1978;

Milliams and Uy, 1978) . Long ter m per formanceof ceramic components, however, has not beenproven.

The corrosion/erosion resistance and hightemperature strength of structural ceramics isconsidered to be excellent and far superior tothat of superal loys. F' or temper atur es below1100C this has been fairly well documented(Sims and Palko, 1977; Freimann et al, 1978)but data above 1200C are very scarce and oftenconflicting. The reason for this is that moststructural ceramics contain sintering aids,which may become vi scous at elevatedtemperatures or react with species in theenvironment. This can lead to acceleratedcorrosion and slow crack growth which markedlyreduce the useful service life of ceramiccomponents. Reliable long term serv ice above1100C must be. demonstrated, prior to any largescale component development ef fort. Theinteraction between subcritical crack growthand the environment particularly needs to bedefined.

Performance of structural ceramics iscurrently limited by subcritical crack growthat elevated temperatur es. This can beaccelerated by corrosives in the environment.A thorough study of subcrit ical crack growthas a function of microstructure, grainboundary impur i ties, temperature, andenvironment expected in coal combustion orconversion should be made.

Some studies should be done on singlecrystal materials for reference, especially ifways can be found to grow single crystals ofhigh pur ity silicon carbide and siliconnitride to determine baseline properties suchas subcritieal crack growth, creep, erosionand corrosion resistance. If large singlecrystals having acceptable properties can begrown, single crystal components with superiorproperties might prove feasible.

Non-destructive examination (NDE) refersto a body of techniques used to search forcritically sized flaws in engineer ingmaterials (Coffey and Mhittle, 1981). Thesetechniques are used to study both the bulkmaterial and the welds used to assemble thefinal structure. They most commonly includeradiography, fluorescence, and ultrasonicprobing. Radiography i s conventionalX-radiation photogr aphy; fluorescence uses aspecial coating to expose sur face flaws; andultrasonic scattering can identify the sizeand location of a flaw. Ultrasonic waves canhave wavelengths of a micron ar less.

Cr itical flaw s i ze s in met allurgical


S116I


materials are cracks generally of the order of1-3 cm long, large enough for radiog raphy andfluorescence to be useful. Critical flaws instructural ceramics can be much smaller thanthose in metals. Thus, pr esentnon-destructive evaluation (NDE) techniquescannot be carried over directly from metals toceramic components. Attempts are under way todevelop NDE methods which will detect flaws inthe 50-100 pm range for structural ceramics.This work should be reviewed to determine i fadditional work or a more fundamental approachis needed. Further development of acousticimaging techniques to identify both the sizeand type of flow would be useful. A goodsource of in formation concerning theapplication of NDE techniques is the new EPRINDE Center in Charlotte, North Carolina.

The prevailing design methods for brittlematerials are based on statistical techniqueswhich 'use the Neibull function (Neibull, 1951)to characterize the relation between theinitial flaw size distribution and the failureprobability. Although this function appliesreasonably well in most cases, its validityhas not been checked thoroughly, and it is notknown whether it s application leads toconservative or opt imi stic design. A

fundamental study to validate the physicalbasi s of the present Meibull designmethodology for brittle ceramics is desirable.Mork on statistical models, of failure ofceramics is a noteworthy example of thisapproach ( Evan s and H arne, 1 91'9); andprobabilistic failure mechanics is a field ofstudy in itself (see Hau and Besuner, 1981 andSect ion IV-D-2b ) .

6. Refractories and slags

' For most applications above 'lOOOC, metalsare not practical as structural mater ialsexcept for the refractory metals (tungsten,tantalum, molybdenum, etc. ) which find use instrongly reducing environments. On the basi sof cost and durability, often, the best choiceis to employ nonload-bear ing ceramicr efractor ies to insulate the structuralmaterials from the high temperature corrosiveenvironment, which contains both combustionproducts and mineral matter. Depending upontemperature, the mineral matter can be in theform of a "dry" abrasive ash or a highlycorrosive liquid "slag".

a. Refractories

Refractories are high temperatur ematerials composed of the oxides of silicon,a 1umi num, magnesium, calcium, chromium, iron,sodium and potassium. They are generallyfabricated by mixing the appropriate materials(raw, calcined, or pre-fired and crushed) witha binder and have compositions which oftenvary with the source of the raw materials.Except in speciali zed cases, small impuritiesare not controlled. The exact composition ofa refractory varies with production batchwithin 'certain limits set by the manufactureras a compromise between consistency andeconomy. The structure tends to be a mixtureof crystalline and glassy phases of manydifferent compositions bonded into a structureof varying porosity. Refractories tend to bebrit tie, and fracture with little or nodeformation. At high temperatures they beginto soften and lose strength. Under mechanicalload s they wi 11 c reep a t e1eva tedtemperatures, so care is taken in designingwith refractories.

Commercial development o f re fractoriesfor extreme environments has been hampered bya lack of fundamental research into andunderstanding of the high temperature behaviorand interactions of these materials. The rolewhich the microstructure of a refractory playsin resisting degradation should be studiedusing advanced electron microscopy techniquesdeveloped wi thin the last decade. Theimportance of interfacial phenomena as therate controlling me chan i sms in thedisintegration process need careful study.

Refractory uses can be divided into lowand high temperature applicat ions. Lowtemper at ure applicati on s are t hose belowslagging temperature (about 11000) where coalash is still in the solid form and does notreact with the refractory. Degradationmechanisms are mainly gaseous corrosion anderosion by particulates. High temperatureapplications are generally above 125OC, wherecoal ash i s generally molten and reactschemically with the refractory liner.

Re fractor ies for 1 ow temperatureapplications such as liners of dry a shgasifiers, process piping and cyclones, canoften be modeled a f ter those used in thepetrochemical industry where r efractoryconcretes, both dense and insulating types,are the standard materials. An extensivelaboratory and field test program hasindicated that low cost 50'$ Al 0 concretes



(Bakker, 1978a, 1978b), consisting mainly ofcalcined kaolin grains bonded with calciumaluminate cement, are very suitable materials.They tend to increase in strength during

/service because of the hydrothermal formationof fibrous calcium aluminum silicates.

Refractories for high temper atur eapplications —mainly slagging .pasifiers andMHD air preheate r s —— are in a lesssatisfactory state. Laboratory tests,(Kennedy, 1979; Townes et al, :L981; Pollinaand Smyth, 1981) have indicated thatMg(A1, Cr) 0& spinels should be suitab1e.However, refractory wear in actual pilotplants has often been disappointing. It is atpresent not possible to predict behavior inactual service from systematic la bor a to r ytests. The reasons for this remain to be bediscovered.

Physics related research and developmentin this area would benefit both slagginggasifiers and open cycle NHD power generatingsystems. Combustor temper atur es are highenough to ioni ze the seeded coal combustionproducts so that the plasma is highlyreactive. In most proposed coal-f ired opencycle MHD power generation schemes (Pollinaand Smyth, 1981),the required plasmatemperature is achieved by preheating thecombution air in regenerative heater semploying cored refractory bricks as the heatstorage medium.

Phase diagrams

Interactions between slag s andrefractories are dependent on thermodynamicequilibria, expressed in various phasediagrams. Slags ar e mul ti component oxidemixtures, so phase diagrams are needed whichhave perhaps as many as 9 or 10 components. A.

large number of phase diagrams are available(Matt, 1959; Levin et al, 1964; Muan, 1979;Huan and Osborn, 1965), bu t o f ten t hepertinent information on coal slags is notavailable. For example, the phase diagram forK 0-Al 0 -SiO is largely incomplete. Suchinformation is badly needed. In the system2 2 — . 2

Cr 0 -Al 0 -SiO -FeO for various oxygenpar&i@1 pressures, the effects of CaO, NgO and2 3 2

alkali additives to this systems should bestudied. The experimental work is di fficults inc e thermodynamic measur ements on thesematerials at these temperatures are aformidable task. Theoret ical work is equallyd i ffi cult since it involves computer modeling

the thermodynami cs, and generating phasediagrams with as many as nine dimensions.The experimental data is meager, and such workwould be a significant, long term project forboth theoretician and experimentalist. Someprogress is being made. See, for example,Eliezer et al (1978) and Byker et al (1980).Physicists are particularly equipped to handleboth these aspects; by developing newmeasur ing systems and by so 1 v i n gmultidimensional thermodynamics problems usingsophisticated computer techniques.

Effects of microstructure

Refractories must be able to withstandthermal shock and to resist penetration,dissolution, and erosion by slags. Mi thpr esent re f r act ory technolgy theserequirements are incompatible. Slag-resistantrefractories have a high density and lowporosity, and in consequence they generallyhave poor thermal shock resistance.

The ability of a material to withstand asevere thermomechan ic al environment i sdetermined both by its microstructure and byits chemistry. There is a trade off infabrication and use between the desire to havedense bricks for resistance to corrosiveattack and the desire to have porous bricksfor resistance to thermal stresses arisingfrom rapid temperature gradients. Fusion cast(poured from the melt) bricks have very highdensity with almost no open porosity; and thecorrosion pathways through pores, surfaces andalong grain boundar i es are cons iderablyreduced. However, pores are an important partof the microstruetur e enabling the brick towithstand ther mal s tr esse s . Thermal s tr e s scracks always occur in refractories, but willterminate when the cracks propagate into apore.

Ceramic materials, as a class, are rigidwith high moduli of elasticity allowing verylittle plastic deformation. Cracks forming atpoints of stress concentration are not easilyblunted by yielding of the surround ingmaterial and will tend to propagate in thebrittle fracture failure mode.

Microstresses are developed by phasetransformations and by di f ferent ial thermalexpansion, and occur over extremely localizedvolumes within a mater ial. The microcraekingthat results is localized most commonly alongthe grain boundaries separating the phases.Once formed, these microcracks serve to



relieve microstresses generated in t hesurrounding matrix (Kingery et al, 1976).

The formation of microcracks andmacrocracks dur ing thermal cycling producesirreversible expansion of the material, andthe grains and crystals surrounding the crackstend to push each other apart on heating.During cooling the dimensional changes willnot be totally reversible, and the thermalstresses may lead to further, irreversiblecracking. The result is a gradual increase ofvoid space and bulk volume and a decrease instrength of the bonding matrix. This decreasein strength may ultimately lead to completedisintegration of the material as is sometimesobserved. At present, no theoretical modelsexist that will describe the ef fects ofmultiple thermal cycles on a material. Thelong-t, erm behavior and strength degradat, ion ofceramics must be determined empirically.

Relationship between structure and resistance toenvironmental stress

Physicists might make a contribution bycomputer modeling these materials to determinethe influence of structural parameter s orcomposition changes on the thermal cyclingresponse. Possibly by varying the grain size,or the type of bonding, or the porosity, orthe expansion characteristics of the grains, '

we can substantially improve the situation.Computer modeling can help determine theappropriate microstructure which representsthe. best compromise for the system studied.It should be kept in mind that the results ofany study will be sensit ive to the boundaryconditions and so a set of solutions may begenerated, each applicable to a specificsystem.

However, a complete model must not onlylocate points of maximum internal stress, but,must be able to simulate the introduction,propagation, and arrest of cracks. That therewill be some cracking is inevitable. The goalof the modeling would be to f ind ways tominimize the effect of cracking upon the grossstructural properties of the refractory.

Kinetics of dissolution of refractory oxides in

coal slags

Studies of the dissolution of refractoryoxides in molten slags were carried out in the'1960's for the steel industry; the syst, emsstudied were single-crystal and

polycrystal line alumina in soda-lime glass.Similar dissolution s tud i es us in g s ing lecrystal oxides and spinels present inrefractories and typical coal slags should becarried out now to get enough baseline data tooptimize refractory compositions for coalconversion technology.

The presence of a hostile chemicalenvironment, presents a range of problems.There can be granular, intergranular, or grainboundary attack by one or more elements in theseeded coal slag, leading to complete chemicalpenetrat ion and failure. This chemi calpenetration may be accelerated by the thermalcycling which can ser ve to open and closepores so that the corrosive slag has readyaccess to the interior. New pores (micro-ormacrocracks) may also be created and becomeavenues of penetration. Long dur ation atelevated temperatures and at too low an oxygenpart, ial pressure may lead to the growth of newphases within a multiphase brick —with thesubsequent crack generation and failure asdescribed above. Finally, vaporization of themore volatile constituents will also lead tomechanical failure especially if they are thestablizing additives to the material (as aloss of calcia in calcia stabilized zirconia).

The extensive microcracking which hasbeen observed in a rebonded magnesia chromematerial suggests the presence of microstressproducing mechanisms. Leonard aAd Her r on( 1972) report a permanent volume expansion andst ruct, ural damage in magnesia-chromerefractories due to temperature or atmosphericcycling. They contend that the primarymechanism is a cyclic reduction and oxidationdue to a changing equilibrium oxygen content.The reduction occurs with a loss of oxygen andthe collapse of,lattice vacancies to form newclosed pores.

Physical properties of refractories

The experimental i st can develop newtechniques to measure physical properties ofthese mater ials a t e le v a ted temperatures.Items needed by designers are thermale x pan sion, thermal conductivit, y, speci ficheat, creep, elastic moduli, and fracturetoughness. It would be extremely helpful- ifsome of these properties such as the lastthree could be monitored in situ. Typicallythe thermal diffusivity is measured by thelaser-flash technique (Bates, 1970), the heat



capacity guessed on the basis of simple mixingof known oxides, the density calculated fromexpansion data, and the thermal conductivitycalculated from all of the above.

The problem with this approach istwo-fold: a) The laser flash technique worksbest on specimens such as metals, glasses, anddense polycrystalline mat', erials. It does notwork on materials with large grains or por es.b) The heat capacity of silicate materialscannot be evaluated reliably by assumingsimple mixing since this assumption has beenshown to be invalid in some cases (Howald etal, 1979). While the error may be tolerablefor engineering design, it cannot beoverlooked by those working on phase diagrams.Obviously any discrepancy is of crucialimportance to the phase chemi sts andphysicists.

There are few reliable di f ferential'scanning calorimeters usable with any accuracyabove about 800C; most designs break down inthe radiation regime. The few commerciallyavai lab le drop calorimeters are quiteexpensive' , while the construction of anaccurate drop calorimeter for high temperatureapplication requires major financial outlaysand labor intensive effort using presentconcepts. It should be investigated whetherother thermal techniques using bulk specimensand conventional (ohmic) heat sources in therange 1200-2000C might prove to be useful inmeasuring thermophysical properties o fmaterials.

Creep is easily measured at hightemperature but the elastic moduli and thefracture toughness ar e not, . For cerami cmaterials which may have many weak bonds orinter nal cracks, t here i s a- need forstress-strain curves under both tension andcompr ession at elevated temperatures. Fewmeasurements have been done under tension incoarse grained materials such as refractorieswhich have intrinsically low tensile strength.Thermal cycling conditions can produce hightensile stresses in materials.

The goal, of any materials testing programis to determine.

Mhich environmental characteristicsare degrading the material?

Mhat phase of the materialis being attacked?

Mhat is the detailed failuremechanism?

Mhat property or properties inthe material and/or environmentcan be changed to enhance resistance?

b. Slags

All coals contain mineral matter of from5$ to 40$ by weight of the coal. This mass ofmaterial must be carried through the processand removed for disposal. Some fly ash isused in concrete manufacturing but the ash haslit tie ec onomi c value gener al ly . One hopesthat such a plentiful and inexpensivesubstance would find more uses in the future.Although the. character, quality, anddistribution of minerals may vary considerablywithin the coal seam, this variation tends tobe masked dur ing the mining, transporating,comminuting and other operations which turnthe coal into a usable form. Lar gedifferences exist in the ash characteristicsof coals from different regions. In general,however, it is composed chiefly of compoundsof silicon, aluminum, iron, and calcium withsmaller amounts o f magnesium, sodium,potassium, and titanium. These elements arepresent, as oxides, sulfates, and complexalumino-silicates. Below 1100-1300C the ashoccurs as an abrasive solid particle. Abovethis temperature range the ash is in the formof a liquid "slag" which moves through thesystem as droplets which coat the walls andlead to fouling (constriction of mass or heatflow). Above the fusion temperature, theliquid ash is highly corrosive and can attackany or all of the phases present in arefractory material. Certain components ofthe ash, notably some sodium and potassiumcompounds, are volatile at coal combustortemperatures and will condense on the coolerwalls downstream, leading to the hot corrosiondiscussed earlier.

Slag penetration and wetting of refractories by slags

The wet ting of refractory oxides byvarious slags is poorly understood. Thephysical mechanism leading to slag interactionin a porous refractory in the presence of athermal gradient should also be studied andclarified.

Magnetohydrodynamic (MHD) e'nergy conversion

The magn. t~riydrodynamic cycle is a highlyefficient energy conversion method, where anelectrically conducting high temperature gasis driven through a high magnetic f ield

Rev. Mod. Phys. , Vot. 53, No. 4, Part II, October 1981


(~6-7T), generating electric power in theprocess.

The MHD generator i s a large combustorwhich burns coal at 2200C using preheated air,followed by an electr ical generating chambercalled " "channel" in which the conducting.exhaust gases traverse the magnetic field. Toenhance the conductivity of the combustionproducts, a potassium (seed) compound (such asK CO ) is added before the plasma passesdownier earn into the gener at i n g channel . The2

channel is ericased in a large superconductingmagnet; and two opposing walls are lined withelectrodes which transmi t the c u r r en tgenerated by the induced EMF. The rate ofenthalpy extraction for a full scale generatoris high enough for the channel plasma to be ina non-equilibrium state.

In the most general case, the path forcharge conduction in an NHD channel includesmetallic or ceramic electrode materials onwhich there is a layer of solid (mixedglass/polycrystalline) and/or liquid coalslag. Above the slag layer is the gaseousplasma. Current flow may be diffuse, or theremay be arcing. Since the conduction mechanismin slag is largely ionic, while electrodes areelectronic oonductors, polarization layerswill develop; and electrochemical degradationcan occur. To eliminate problems of arcingand electrochemical degradation we must findelectrode mat rials which are compatible withslag conduction mechanisms or which can resist,arc erosion. Not all potential electrodecandidates have been adequately characterized;and the charge carrying species in slags havenot even been adequately identified. Nor havetheir "transferance numbers" and di ffusioncoef fioients been measured. This is an areawhich is especially appropr iate for work byphysicist, s.

On the cathode wall, the intrinsicthermionic work function of a slag layer maybe too high to allow needed operating currentdensities (1-2 amp. cm ) except at operat, ingtemperatures higher than anticipated for thechannel. There is now strong evidence(Anderson et al, 1981), however, that thepresence of potassium vapor pressure in theplasma layer over the electrode can lower thework function suf ficiently to allow therequired current densitites. Nork needs to bedone to measure the actual thermionic workfunction of these materials under an appliegpartial pressure (fugacity) of potassium ( 10

'IO atm) in order to simulate the MHD

environment.

Also on the cathode wall, simultaneouslywith the development of a polarization layer,a spatially periodic shor ting occur s betweenelectrodes along the wall (Demir jian, 1978) .There is not sufficient information about thislayer to indicate whether the shorting is dueto the concentration of cations in thepolar i zation layer, or to a more complexinteraction between an elect rode and itsneighbors, the magnetic field, and the plasmacurrent . But concentrations of these speciesin a normal slag have been shown to enhancethe slag conductivity substantially (Pollinaet al, 1980a, 1980b).

In addition to identifying the conductingspecies in the slag and finding suitableelectrode materi al s, the slag layercon d u c t i v i t, y m u s t b e i n a n e l e c t rj. c a 1 1ysuitable design range (typically 10 to 10ohm N ) for the generator to functionpr'operly (Rosa, 1971). Thus it is necessaryto measure the conductivity of a slag layer asa function of composition and temperature. Inaddition to the study of the bulk property, iti s also use ful to know how the conductivitywould change at the boundary layers betweenelectrode and slag, where the chemi calcomposition might be changed because of ionicconduction.

Finally, the slag proper ties may bemodified to be more favorable to the MHDsystem by the a/dition of oxides for thepurpose of changing either the conductivity orthe conducting species.

On the slag-plasma inter face (~ 1500C ) onthe anode side of the channel, i t, has beendemonstr ated that an ion cur rent of K is+

easily obtained going from the slag into theplasma layer. At the anode-slag inter face,however, the polar i zation boundary layerproduces a very low conductivity region in theslag because of the removal of cations such asFe and K from that layer. This lowconductivity layer is thought to be the causeof arcing on the anode. If so, i t wouldexplain why anode arcing has been such anintractable problem so far. This is furthersubstantiated by the fact that anode arcing isgreatly reduced by the use of iron anodes(Koester, 1 978) which act as sacrificialelectrodes in contr ibuting themselves ascharge carriers in the slag, thus eliminatingthe low conductivity iron-depletion layer fromnear the anode. This, however, is not asatisfactory solution to the problem, becausethe electrode would be quickly exhausted.

Qne can see from the above that the

Rev. Mod. Phys. , Vol. 53, No. 4, Part I l, Qctober 1981


problem of electrical conduction through alayer of molten oxides is formidable, but onewhich is well suited to work by physicists.The ex per imental needs are for thermionicemission, charge carrier identification, andtransference number measurements. Theseshould be carried out to simplify the searchfor suitable electrode ma ter ials. Once thesemeasurements have been done, then thesolutions to the electrochemical equations canbe found, and the nature and behavior of thepolar i zat ion layer and i'ts e f feet onelectrochemical degradation of the electrodemater ial investigated. The goal is to findelectrode materials having li fetimescompatible with the normal outage schedule ofa power plant, and which can operate almostcontinuously at 1 amp/cm current densitieswithin that lifetime.

The development o f adva n c edsuperconducting magnet technology is necessaryto attain the high magnetic field strength(6-7T) over the volume of an MHD channel. , 14 m

long by 3 m diameter (Narston and Thome,1980).

7. Recommendations

~ Me recommend expanded andstrengthened effort on computersimulations (at the atomic level) ofinteractions between the 1att ice and"foreign" atoms. Theoretical modeling ofFe-H lattice interactions is a case inpoint.

Me recommend surface science studiesof the manner in which hydrogen gainsentry to the lattice.

~Me recommend further theoret icalinvestigation of crack nucleat ion andgrowth, taki n g i n to account chemicalef fects. The possibility should bei n ve st i. g a ted that some types of chemicalenvironment might strengthen the materialat a crack tip.

~ Me recommend intensi f ied study ofall aspects of solid particle erosionincluding the mechanisms of materialremoval, the dependence on kinematicalvar iables and temperature, and thedependence upon particulate properties.Me invite attention especially to theimportance of studying threshold effects,as they may provide the key to viablesystems design.

Me recommend broadened andsystematic studies of the transport of

minor oxidant s and substrate atomsthrough prot ective oxide films, withspecial effort to understand the pathwaysinvolved.

~Me recommend the u se ofmulti component di f f usion theory tounderstand the cooperative relationshipbetween sulfide and carbide formation atthe substrat e-f i 1m inter face and theirsubsequent oxidation and consequentdisruption of the protective film.

+Me recommend use of the methods ofsurface physics to study the mechanismsby which alkali sulfates attack metalsurfaces and their protective films andt he mechanisms of scale formation a'ndspallation.

«Me recommend systematic study,including use of sur face physicstechniques, of phenomena that can beinduced by deliberate sur f acemodi f ications such as coatings, ionimplantation, and other chemicaltreatments.

~ We recommend fundamental studies toe luc i date t he nature of"erosion-cor ros ion" which at present ispoorly understood.

o Me recommend an ef fort to improveand extend the techniques ofnon-destructive evaluation for structuralceramics —improvement, for example, ofacoustic analysis techniques.

+Me recommend fundamental studies ofsubcritical crack growth, including theuse of thermodynamics and rate-processtheory.

+Me recommend statistical modeling offailures based on fundamental principlesin order to provide a physical basis tothe present empirical design methodology.

Me recommend incr eased efforts tocomplete the phase diagrams that arecritical for understanding refractoriesin slagging envi r onments . These aresystems o f 4 — 9 components; anddevelopment o f computational techniquesto deal with such systems should beencouraged and enhanced.

ice recommend that research on thekinetics of refractory dissolution byslags be strengthened and encouragedwith specific attention to di f fusionpathways or mechanisms, and to therespective roles of grain boundaries andmatrix phase components in resist ingd i ssolution. In the same vein, we



recommend a quantit, ative study of thedi f fusion through the s t r uc tur e, andsubsequent evaporaton and loss, ofessential chemical species (i.e. binder,et, c. ), and an analysis of theconsequences for the int, egrity of thecomposite re fr act ory under long termexposure to the process environment.

~Me r ecommend st, u d y o f t heinteraction of refractory microstructure-- specifically, porosity -- with thegrowth or arrest of cracks.

+Me recommend systematic r esear chinto the phenomenology of refractorymodification and failur e at, tendant uponthermal cycling.

~Me recommend research to identifythe charge carriers and to determinetransport parameters for slags, as wellas studies o f electro-chemi cali nteractions in the slag-electrodesystem.

~Me recommend research to establishthe systematics of wetting o frefractories by slags.

~Me recommend that efforts beencouraged to impr ove quantitat ivemeasurement, s of physical properties (heatcapacity, thermal conductivity, vaporpressure, etc. ) of refractories at hightemperatur e under realistic or in situconditions, as well as. under laboratorycondit, ions.

8. References/

Ant, ill, J. E. , M. J. Bennett, R. F. A.Carney, G. Dearnaley, F. H. Fern, P. H.Goode, B. L. Myatt, J. F. Turner, and J.B. Marburton, 1980. Corros. Sci. 16, 729.

Ander son, J. , M. F. Anderson, M. A. wilson,and G. J. Lapeyre, 1981, "ThermionicEmission o f Some Syn thet i c hand Na t ur a 1Slag s", Montana State University,publi shed by Mont ana Ener gy In sti t ute,Butte, Montana 59701.

Bakker, ttlt. , 1978a, "Refractory Applicationsin Gasi fiers", 10t,h Synthet ic PipelineGas Symposium, pp. 307-336, American GasAssoc. , Chicago, Ill.

Bakker, W. , l978b, "Refractory Applicationsin Coal Gasifiers" in 3rd Annual Conf. onMat erials for Coal Conversion andUtilizat ion, Oct . 10-12, 1978.Gaithersburg, Md. , available from NTIS asDOE pub 1., CONF -781018.

Bates, J. L. , 1970, Thermal Conductivity of

"Round Robin" Uranium Dioxide, BN'AIL-103,Battelle, Pacific Northwest Laboratories,Richland, MA 99352.

Byker, H. J. , R. E. R. Craig, I. Eliezer, N.Eliezer, R. A. Howald, P. Uiswanadham,and R. Moodr i f f, l 980, "Acti vi tyCoef ficients in Seeded Coal Slag andTheir Ef fects on Vapor i zation andCrystallization Equilibria" in 7t, hIn t, e r n a tional Con f er ence on MHDElectrical Power Generation.

Cof fey, J. M. and M. J. whittle, 1981, Phil.Trans. Roy. Soc. (London) A 229, 93.

Collins, R. A. , S. Muhl, end G. Dearnaley,. l980, Phil. Trans. Roy. Soc. (London) A

295 33l ~

COSMAT, 1975, Materials and Man's Needs,Summary Report of the Committee on theSurvey of Mat er ials Science andEngineering, Nationa l Academcy o fSciences, washington, D. C.

Demir jian, A. , S. M. Petty, and A. Solbes,l978, 17th Symposium on the EngineeringAspects of MHD, C. H. Kruger, Ed. ,Stan ford Un i ver si ty, March 2729, 1978,p. D. 'l. 1, available from Dept. of Mech.Eng. , U. Miss. , University, Miss. 38677.

Edison Elect r ic Insti t ute, "Report onEquipment Availability for the Ten YearPeriod", 1965-1975 AND 1966-1976, report875-50 and 7685, Ed i son Electr icInst, it,ute, washington, D. C. 20036.

Eliezer, I. , N. Eliezer, Reed A. Howald, M.C. Uerwolf, 1978, J. Phys. Chem. 82,2688.

Evans, A. G. and A. Rame 1979»Hi ghTemper at, ur e Failure Mechanisms inCeramics", Lawr ence Her ke ly Laboratory,Report LBL 8692.

Finnie, I. , 1979, Con f er ence onCor r os ion/Erosion of Coal ConversionSystem Mat, erials, National Association ofCorrosion Engineer s, Houston, TX, 1979,pp. 029-443.

Freimann, S. M. , J. J. Mechol ski, M. J.Mc Donough, and R. M. Rice, 1978, "Effeetof Oxidation on the Room TemperatureStrength of Hot Pressed Si N -MgO andSi N -Z 0&" in Ceramic) or High2.Performance applications II, Brook HillPubl. Co. , Chestnut, Hill, MA, p . 1069ff,J. Burke, E. Lenoe, and R. Nathan Kat z,Eds.

Fuller, E. R. Jr . , and R. M. Thomson-, l 978,in Fracture Mechanics of Ceramics, Vol.

Ed . by R. C. Br adt, , D. P. H.Hasselman, and F. F. Large, Plenum, N. Y. ,

Rev. Mod. Phys. , Vol. 53, No. 4, Part II, October IS81


pp. 507-548.Gehlen, P. C. , J. R. Beeler Jr . , and R. I.

Jaf fee, 1972, Eds. , "Interatomi cPotenti al s and Simulation of LatticeDefects", Plenum Press, N. Y.

Hill, V. L. , et al. , 19 f9, Symposium 89,Amer i can Chemi cal Society, Ma shing ton,D. C.

Howald, Reed A. , Isaac Eliezer, P.Vi swan adham, 1979, J. Am. Cer am. Soc. 62,134.

Hutchings, I. M. , 1979, Conference onCorrosion/Er osion o f Coal ConversionSystem Materials, National Association ofCorrosion Engineers, Hou s ton, T X, 1 9 "f9,pp. 343-428.

Kennedy, C. R. , 19 f9, "Refractories forApplication in Slagging Gasifiers" in 4thAnnual Conference on Materials for CoalConver sion Ut ili zation, Oet. 9-1 1, 19"f9,Gaithersburg, Nd. , available from NTIS asDOE publ. , CONF-791014.

Kingery, M. D. , H. K. Bowen, and D. R.Uhlmann, 1976, Introduction to Ceramics,2nd ed. , J. Riley and Sons, Inc. , N. Y.

Koester, J. K. , and R. A. Perkins, 1 9 "f8, i nFour t h US/USSR Colloquium on MHD,publi shed by DOE, a va il ab le fr om NTIS,washington, D. C. , pp. 603-634.

Lawn, B. R. and T. R. Mi 1 shaw, 1 9 75,"Fracture of Br i t tie Solids", CambridgeUniverity Press, Cambridge.

Leonard, R. J. and R. H. Herron, 1972,Cer amie Bull . 51, 891.

Levin, Ernest M. , Carl R. Robbins and HowardF. MeMurdie, 1964, 1969, 19"f 5, Eds. ,Phase Diagrams for Ceramist s AmericanCerami cs Soc i ety, Columbus, Ohio, 1964and Supplements (1969 and 1975) .

Mangel s, J. A. , 19 f 8, "Development ofInjection &olded Reaction Bonded Siin Ceramics for High Per forma ceApplications II, Brook Hill Publ. Co. ,Chestnut Hill, MA, edited by J. Burke, E.Lenoe, and R. N. Katz, pp. 113-130.

Marston, P. G. , and R. J. Thome, 1980, inProceedings 7th International Conferenceon MHD Electrical Power Generation, J. F.Louis, Chairman, MIT, Cambridge, NA, June16-20, 1980 pp. 450-458.

Muan, A. , and E. F. Osborn, 1965, PhaseEquilibria Among Oxides in Steelmaking,Addison-wesley, Reading, MA.

Muan, A. , 1979, in 4th Annual Conference onMaterials for Coal ConversionUtil i zation, Oet . 9 —11, 1979,Gaithersburg, MD. , available from NTIS as

DOE publ . , CONF -791014. p IV-44f f .Nelson, G. A. , 1969, Action of Hydrogen on

Steel at High Temperatur e and HighPr essures in Interpretative Repor t onEf fect of Hydrogen in Pressure-VesselSteels. Melding Re sear ch CouncilBulletin 8145, N. Y. , Oct. , 1969, pp.33-42.

Paskin, A. , A. Gohar, and G. H. Dienes,1980, Phys. Rev. Letters, 44, 940-943 andreferences therein.

Perkins, R. A. , 19 f9, Materials Problems inFluidized-Bed Combustion SystemsI V-Co r r os i on Chemi stry in Low-OxygenAetivi ty Atmospheres, ReportEPRI-FP-1280, Electr ical Power ResearchInstitute, Palo Alto, CA.

Pettit, F. S. , R. H. Barkalow, and J. A.Goebel, 1980, Mater ials Problems inFluidi zed-Bed Combustion Systems, HighTemperatur e Erosion-Corrosion byHigh-Velocity (200 m/sec) Particles inReport EPR I-CS-1448, Electr ic PowerResearch Institute, Palo Alto, CA.

Pollina, R. J. , Raymond Lar sen, and AlKumn i ek, 1980a, "Coal Slag . ItsElectrical Properties and Some NechanicalEffects on MHD Ceramics" in Energy andCeramics, Ed. , P. Vincenzini, Elsevier,N. Y.

Pollina, R. J. , Raymond Larsen, and DavidMestpfahl, 1980b, J. NoncrystallineSolids 40, 197.

Pollina, R. J. , and R. Smyt h, 1981,"Refractories for High Temperature MHD

Applicat ions" in Proceedings of theInternational Colloquium on the Use ofRe fr ac tory Ox ides for Hi gh TemperatureEnergy Conversion Processes, Toronto,Canada, July 1 6-1 9, 1 979, Hi ghTemperature Science (in press).

Rau, C, A. , and P. M. Besuner, 1981, Phil.Trans. Roy. Soc. (London) A 299, 111.

Rehn, I. N. , 1981, "Review o f' RecentDevelopments Concern in g Seal e Forma t ionand Exfoliation of Steam Generator TubeMaterials", EPRI Repor t on Pro ject 644-1(in press).

Reid, M. T. , 1971, Ext ernal Corrosion andDeposits, Elsevier, N. Y.

Rice, J. R. and R. N. Thomson, 1974, Phil.Nag. 29, 73.

Rosa, R. J. , 1971, "Design Considerationsfor Coal-Fir ed MHD Gener at or Duets", in5th Inter nat ional Con fer ence on MHD

Electrical Power Gener a t, ion, Munich,1971 ~



Sims, C. T. , and J. E. Palko, 1977, "Sur faceStability of Ceramics Applied to EnergyConver ion Syst ems in Proceedings on theWorkshop on Ceramics for Advanced HeatEnergies, Jan. 24-26, 1977, Or lando,Florida, ERDA, CONF 770110, availablefrom NTIS.

Smeltzer, W. W. , D. P. Whittle, and D. J.Young, 1976, in Properties of HighTemper atur e Al loys, El eetrochemiealSociety, pp. 134-153.

Str inger, J, and S. Er lich, 1976, ASNE Paper76WS/CD-4.

Stringer, J. , 1977, Ann . Rev. Natl. Sci.477.

Thomson, R. N. , 1980, J. Natl. Sci. 15,1014.

Tien, J. K. , O. Buck, R. C. Bates, wnd S.Nair, 1980, Seri pta Net 14, 583.

Tomkins, B. , 1981, Phil. Trans. Roy. Soc.(London) A 299, 31.

Townes, H. W. , R. J. Pollina, R. Larsen, K.Kiddo, R. Smyth, G. E. Youngblood, and W.C. Seymour, 1981, "Operability andMaterials Per formance of a RegenerativeHeat Exchanger During 1000+ Hours ofOpera t ion" in 19th NHD Symposium,Tullahoma, TN, June, 1981.

Uy, J. C. , R. N. Wi 1 1 i ams, and N. V.Goodyear, 1978, "Study of the Processingand Bond Strength of Hot-Pressed Si N4",in Ceramics for High PerformanceApplications II, Brook Hill Publ. Co. ,Chestnut Hill, NA, edited by J. Burke, E.Lenoe, and R. N. Katz, pp. 131-150.

Watt, J. D. , 1959, BCURA Bull. 23, 49-69.Weibull, W. , 1951, J. of Appl. Neehanies,

18, 293-297.Wells, A. A. , 1981, Phil. Trans. Roy. Soc.

(London) A 299, 19.Whittle, D. P. , 1972, in High Temperature

Corrosion of Aerospace Alloys, Ed. by J.Stringer, et al. , Proceedings of theAdvisory Group for Aerospace Research andDevelopment, AGARD-C P-1 20, pp. 1 7 3-199,available from NTIS.

Williams, R. N. and J. C. Uy, 1978, "CeramicMaterial Characterization", - in Ceramicsfor High Per formance Application II,Brook Hill Publ. Co. , Chestnut Hl ll, NA,edi ted by J. Burke, E. Lenoe, and R. N.Katz, pp. 151-178.

Wilson, W. D. , 1980, in Arpa Workshop, LaJolla CA, July 16-18, 1980, availablefrom NTIS.



IV. E. SIZE-DEPENDENT PHENOMENA

1. Introduction

Coal util i zation eritail s particle si zereduction (usually by grinding) as well asinteractions with fluids, whether these arenative to the coal or are added for thepurpose of transportation or during sizereduction. Some fraction of the energy inputf or coal processing and utili zation isexpended in inc ff icient gr inding processes.The under standi ng o f the f un damentalmechanisms involved is rudimentary. There isa dearth of measurement on stress-strainrelations and on the swelling behavior withvarious fluids (cog. H 0, methanol), while therole of grinding aids is not understood.Little is understood about fractur e mechanismsor about crack initiation and propagation inthe coal particles under grinding stresses.Chemical comminution, currently an expensiveparticle size-reduct ion technique, is poorlyunderstood. Similarly, the interact ionbetween coal particles and fluids such aswater or methanol in slur ry formation has notbeen sufficiently studied. ' Especially lackingare data relative to high concentrationslur ries. Reduction of coal particles tosizes of the order of a few microns, in a costef fective manner, woul d al low the directutilization of coal in boilers designed forresidual fuel oil, without major retrofitting.On the other hand, the ability to control coalparticle size better could improve efficiencyby the production of fewer coal "fines", i.e.particles of a few microns in typicaldimension. F'inally, the ability to separatecoal particles by techniques not dependent ondensity differences i s highly desirable,especially in the finer particle sizes (seebelow).

2. Coal preparation

Current "coal preparation" practices inthe U. S. involve size reduction, mechanicalcleaning, and thermal drying of coal for thepurposes of maximi zing recovery of coal heatcontent in boiler s; washing coal f ines;optimi zing usage of scarce water resources;and reducing sulfur in the final product. Thecoal preparation sequence involves four steps.(1) size reduction -- (Progressive sizereduction of coal particles is a most

s igni f icant process in t he over al lpreparation or "beneficiat ion" o f coal pr iorto its use. The most important procedures forsize reduction are the following. primarybreakage of mined coal (8" to 3"), secondarycrushing to a maximum size (1 —1/4"),pulverization ( 1/8" ) and grinding (~1 mm) .);(2) "sizing" of coal, i.e. , screening to adesired range of sizes; (3) cleaning to removedust and non-coal mater ials by dr y o r we twashing techniques -- (Dry washing involvesblowing air for removing small particles. Metwashing may utilize floating the coal in a"dense medium" water/magnetite suspension~ inwhich the impurities sink; or froth flotation,whereby chemicals are added to increase thehydrophobicity of coal, which attaches itselfto air bubbles and finally is skimmed of f thetop as a froth. It should be noted that allwet washing methods are undesirable becausethey add moisture to the coal, an ener gyexpensive process since the water will have tobe removed later. ); (4) drying of coal fortransport, ation or other end uses -- (Dryingnormally utilizes hot air streams in variousconfigurations including fluidized beds.Oxygen may also be used, promoting at the sametime oxidation of the inorganic sulfur (Lowry,1945; U. of Oklahoma, 1975) . )

Coal pr epar ation, or bene fici ation, aspracticed currently is generally recognized tobe an "energy-inefficient" process and,therefore, sub ject to possible majorimprovements. Breaking, si zing and washingprocesses for coal consume 0. 5-1.0% qf theenergy produced (4 x 10 Btu's per 10 Btu'sprocessed) (U. of Oklahoma, 1975) . Thequestion of defining the efficiency of sizereduct ion, ho~ever, is in itself a difficultone and warrants further research. Thedefinition as currently used (Gaudin, 1939)utilizes the increase in surface energyinvolved in creating smaller particles tospecify the useful energy input, and definest, he remaining (~99%) of the energy input thatgoes into "heating" the sample as beingwasted. However, there is a fundamentalquestion as to whether at least some of the"wasted". energy is involved in unavoidable

~Typically, finely ground magnetite is mixedwi th water to make such a dense mediumsuspension with a specific gravity that allowscoal to float, while heavier mineralimpurities sink. The magnetic moment of themagnetite is used to recover it for reuse.


S'1 26 APS Study Group on Coal Utilization and Synthetic Fuel Production

deformation andchemical bond breaking withinthe coal. (See discussion on pp. 48-49 of Orr(1966).)

Currently coal is "prepared" by variousmechanical means for gross cleaning andpreparation of feeds of maximum particle si zefor power generation. However, coalpreparation or beneficiation for the purposeof creating particles free of deleter iousimpurities, in an efficient and cost-effectivemanner, is becoming even more important. Sizereduction, i f carried far enough, breaks theparticles to small enough size to liberatemost of the organic matter from pyritic sulfurand other undesirable inorganic components.(Of course, the problem remains that, after"liberation" one must be able to separate theparticles).

In addition to breaking apart the organicand inorganic matter, grinding may also effecta separation of the various carbonaceouscomponents of coal, the macerals. If one thencould separate the constituent macerals fromeach other, one could have the means o fpreparing coal feeds of highly selectivecharacteristics that not only may be free ofenvironmentally del et er ious componen ts, butalso may be better feeds for gasification,liquefaction or direct combustion. It isestablished that different macerals not onlyhave di f f erent physical properties(friability, density, hardness, etc. ) but alsogreatly differing chemical reactivities. Forexample, whole vitrinites and resinites mayunder go 1i que fact ion quite readily-; fusiniteson the ot her hand ar e r es i stant toliquefaction under ordinary conditions(Petrakis, 1981). However, the liberation andsubsequent separation of various carbonaceouscomponents from each other as well as from thelighter inorganic components is a difficultproblem. This is true because the primaryproperty (i.e. , density) used for separationis not very di fferent among the macerals andlighter inorganics. Fur thermore, since thepar tie le si ze wi 1 1 be very small and thesur face chemistry o f macerals is unknown,froth floation is not ye t a pr ac t icalseparat ion technique. Thus, there is achallenge to develop techniques that willseparate the macerals from each other and thelighter inorganics once they have beenliberated from the coal.

Size reduction by grinding is mainlycarried out in inefficient ball mills.Liquids (e.g. , H 0) are often used as grinding

aids. In the long run this may increase theinefficiency of the overall process, since thewater may have to be removed again before theground coal can be used.

Another problem is that all comminutionprocesses -- crushing, grinding, pulverization—produce a wide range of particle sizes thatinclude ultrafine particles, (&1 p m) . This isundesir able because the ul tr a f inc particlesare lost (washed away) in the preparationprocess. It has been estimated that, 10'$ ofprocessed coal may be lost due to the makingof particles & Q. 5 mm in si ze dur ing coalpreparat, ion (Soo, 1976). There exist, on theother hand, some arguments in favor ofproducing uniformly microscopic coal particles(Peters, 1965; Herbst, , 1978). Such particles,it is. claimed, can be burned directly in adiesel engine since coal particles & 5 p m

behave like oil droplets (Foo, 1979)". Such adevelopment has also the distinct advantagethat it would allow the direct combustion ofcoal in oil boilers wi thout ma jorretrofit ting. A phenomenon o f practicalimportance, that is of particular interest tophysicists, is that .at a size of $0 micronsand less, grinding rates slow abruptly. Thehypothesis is that there is some cohesi veforce acting between the particles that causesthem to agglomerate. The physical questionsinvolved seem to be evocative of quest ionsthat arise in polymer physics for forceswithin collections of particles in this sizerange. For discussion of such ef fects incoal see Austin and Bagga (1979). ( Ingrinding of aluminium powder, a rewelding maybe involved (Neloy, 1981).)

Opportunities exi st for significantcontributions to the better understanding andimproved practice of size reductiontechnology, which is generally regarded as notvery sophisticated. The physics community hasthe opportunity to assist several aspects ofcoal utilization through specific studies thatmay include the following.

o Define the physical structureof coal and relate si gni f icant

~In the New York Times of June1981, a report appeared that GeneralMotors had developed a experimentalvehicle using a turbine eng ine toburn coal particles of three micronaverage si ze ( J. Holusha, page D4,New York Times, June 4, 1981).


Research for the Technologies S'i 27

proper ties to size reductionprocesses. This is a particularlyimportant, area for the solid statephysics community. After years ofneglect, only now do there appear tobe some serious efforts to use incoal research the power fulexperimental and theoret, ical toolsthat, were developed in conjunctionwith more homogeneous, better definedsolid state systems. The challengehere will stem pr imar ily from thehighly complex and i nhomogeneou snature o f coals. Yet, anunderstanding, for example, offracture initiat ion and crackpropagation might be of much help indesigning ener gy ef ficientsize-reduction processes. (See Sec.IVD2b for a general discussion ofcrack nuele at ion and growth inmaterials. ) There appears to beeither uncertainty or lack ofinformation as to the brittleness andplastic'ity of coal, the nature ofelasticity, t, he nature and role ofint, ernal and surface inhomogeneities(occlusions, grains, etc. ) and thesigni fieanee o f local de formation.One immediate and very far-reachingimplication of the betterunderstanding of the fracture physicsof coal would be the design ofenergy-efficient mills (Luckie,1980).

o Define the nature of theadhesion between organic and mineral

matter as well as between macerals; relate thenature of these bonds to size reductionprocesses. There is little information on thenature of the bond between the inorganic andcarbonaceous component s of coal or on bondsbetween the various macerals. Di f ferentmacerals are known to have quite di f ferent,physical characteristics (hardness,friability, swelling) which can affect thegrinding process. But, these characteristicsare neither defined precisely nor is it knownexactly how they affect size reduction.Empirical correlations have been attempted,mostly with the gross parameter of coal rank.Impr oved understanding of the physicalvariations among coal constituents and oftheir roles in coal ut ili zation i s verydesirable. For example, (Oder, 1980) it hasbeen observed that very light components are

erueial in determining certain engineeringvariables. Plots of volatile matter vsspeci fie gravity o f a sh-f ree carbonaceouscomponents of coal are critical phenomena typecurves, i.e. the plot of volatile mat terversus specific gravity is similar to that ofordered moment versus temperature for aHeisenberg ferromagnet, dropping of f sharplyat some er itical specific gravity. This typeof behavior suggests t, hat light components ofcoal may be the critical f r act ions indetermining engineer ing behavior that isdependent on volatile matt, er . This behavioris one of several examples which indicate theimportance of finding relat, ionships betweenchemical or physical structure of coal andsize reduction, and which also indicate thepotential value of developing techniques toprepare feeds for industr ial processes thatconcentrate selected components (e.g. readilyliquefiable macerals) as well as accomplishingsize reduct, ion. Another very suggestive fact(Petrakis, 1981) is that the concentrations ofparamagnetic centers (organic free-radicals)in residual fr ee-radical concentrationscorrelate quite well with the propensity ofthe maeerals to liquefy. If free-radicalchemi stry pr oves to be the key t oliquef iability, then the development ofcombined s i ze -r eduction maeer al -se gr egat ionprocesses could be a major contribution toliquefaction t,echnology.

o Develop energy-efficient, grindingprocesses. All current sophisticated grindingtechniques are inefficient. Mhat are thefundamentals of grinding —— in c oa1speci fieally? Can processes be developed thatallow the select ive liberat ion of theinorganic mineral from the organic matter, orsuch that there is a minimum of ultrafineparticles? Similarly, can selective gr indingbe carr ied out that selectively pul ver i zesmacerals of a single kind rather than allmacerals indiscriminately. Can the efficiencyof breaking-up of coal be increased by novelapproaches such as explosive shattering, shockwaves, high intensity acoustic energy (seeIVB6), or electric shock? An ad hoc committeeof the National Academy of Sciences hascollected data on e f f iciency of grindingprocesses (Herbst, 1978), and a criticalevaluation of the data is being prepared.Modeling of such data could lead tosignificant theoret ical and experimentalexamination of this problem.

o In summary, size reduction is likely



to increase in importance as an aspect ofcoal preparation prior to end use. Upon finegrinding, mineral matter is separated from theorganic matrix. For larger particl es,separation is effected by taking advantage ofthe density differences between organic andinorganic matter. This i s the ext ent ofcurrent coal preparation practices. Growingdemand for environmentally acceptable coalfeeds, as well as for feeds tailored tospeci f ic coal conversion processes or directcombustion, make i t impor tant that we improveour fundamental understanding of the phenomenarelating to size reduction.

3. Mechanical Properties of coals

The mechanical properties of coal play asignificant role in all aspects of coalutilization, and yet there is a dearth ofreliable measurements and ve ry 1 i t tl efundamental understanding of the relationshipbetween the mechanical properties of coal andits physical and chemical structure. Thedeterminations of mechanical properties ofcoal have been limited in number and resultsappear to differ according to the methodologyemployed (Larsen, 1978).

Conventional testing for propertiesrelevant to coal preparation is ratherprimitive, and it involves such tests as freeswelling index (ASTN test for a given coal'ssuitability for coking or other uses);friability (ASTM test to measure sizereduction during handling operation); and theHardgrove Gr indab i 1 i t y Test (an empiricalcorrelation with ease of grinding) (&oughman,1978}. The existing theoretical considerationof the dynamics of coal fracture requiresinformation on how di fferent particles arestressed by forces in a mill; the frequencywith which particles are stressed; and tensileregions produced in each particle. Theincomplete unders tan d i ng of the phenomenainvolved has led to the following observation:"@hen one considers the complexity ofdescr ibing the dynamic applied forces in(various mills), the complexity of thedistribution of flow sizes in solids and thedi ff icul ty of solving stress-strain equationsfor irregular shapes and glancing blows, it isnot, surprising that the designer (of grindingequipment} has to rely largely on empiricism

in the prediction of mill performance"(Boughman, 1978).

Ther e i s general agreement thatbituminous coals are polymeric in character(Larsen, 1978) and that they contain a numberof smaller (NM 1000) extract, able molecules.Both invasive and non-invasive techniques havebeen uti1 i zed t o i nve stigate themacromolecular structure of coals. Invasivetechniques, utilizing solvents such aspyridine, include extract ion, separation,diffusion and swelling of coals. Non-invasivetechniques include mostly mechanical testingsuch as stress-strain behavior at lowdeformation and ultrasonic determination ofshear and compressive moduli (Larsen, 1978;Lucht and Peppas, 1981). The at tempt toquantify these data and determine molecularparameters from elasticity measurements hasnot yet succeeded.

The nature of the macromolecular networkhas been discussed recently; and bituminouscoal is thought to consist of covalentlycross-linked macromolecules r ather thanmacromolecules that are held together by vander Waals forces or H-bonding (Larsen, 1978) .This conclusion is based on the observationthat strain vs time curves level out after aninitial increase, the leveling value beingdetermined by the nature of the cross links.If the macromolecular structure were notcross-linked, then one woul d pr ed i ct tha tsince there is no internal limit to the flowdue to a constant stress -- the deformationwould continue to increase wi th time. Thusthe polymeric picture of coal is strengthenedby consideration of its bulk plasticproperties. But there is di sagreement aboutthe si ze. d i s tr i. bu tion of the constituentmacromolecules, the degree of cross-linkingand the topology of the macromolecul arnetwork, especially as it is affected by thepores o f the overall physical structure(Larsen, 1978) . The need for, and theimportance of, extensive and reliablemeasurements of the mechanical properties ofcoals and coal constituents seem obvious.

The actual stress-strain measur ement s oncoal are quite limited, and they have beenobtained with large samples (Uan Krevelan,196$}. As such, they have been critici zed inthat cracks and other macroscopicinhomogeneities may have in fluenced theresults. Nevertheless, data exist which showthat strain-time curves for bituminous coalsdo reach a constant value; for both bituminous



coals and pure clarains", only 1$ or less ofthe total strain is irrecoverable, as would beexpected from a 3-dimensional macromolecular,covalently bonded structur e (Larsen, 1980) .But such data are sparse.

The area of mechanical properties ofcoals pr e sents many opportunities forcontr ibutions f rom the physics r esearchcommunity. Among the many opport unit ies orneeds, we consider the following as the mostimportant.

o To improve fundamentalunder st anding of the mechanicalpr oper ties of coal . Such anunderstanding i s impor tant not onlyfor si ze reduction processes asdiscussed earlier, but also in miningand transportation (support. strengthof pi 1 lar s i n "long-wa 1 1" and"room-and-pillar" mining patterns) .Understanding the mechanicalproperties of coal would also aid inthe development of abrasion resistantvessels and equipment in all aspectso f coal utili zation, from mining togrinding to transportation andconversi on. There i s need tounder stand how the mechanicalbehavior is af fected by, and relatesto, the chemi cal stru c tur e(heteroatomic species, bridgingmoities between macromolecules, polar.chemical funct, i on a 1 i ties ) and theoverall physical structure (porosity,permeability of coals ) . Solvent,(especially water, methanol, ammonia)diffusion and interaction with coal,along with solvent swelling behavior,should be studied; the behavior understress should be s tudied as well,since these data can be used to inferthe cross-link parameters. The mostimport ant par

arne ters

are (a ) thenumber average molecular weight percross link, M, and (b) the crosslinking dense ty p expressed asmoles of cross links per unit volume.Perhaps something can be learned fromthe correspondi ng physics ofsynthetic and natur al polymers.Appropriate stat, istical theories of

WW ~~~~~~~Br it i sh "type" classi fication for br i ght,finely laminated material in bi tuminous coal(Stopes, 1919).

polymer network elasticity must bedeveloped, accounting for bothentropic factor s and intermolecularforces. The relative contributionsof macromolecular entanglements willneed to be examined as they effectelastic restoring forces.

o To de.ter mi n e she ar and othermoduli and their dependence on coalrank, maceral and coal primary(chemical ) st ructur es; to determinethe applicability and utili zation ofvarious techniques (e.g. , soundvelocity measurements) to thedetermination of such modul i . Morkis needed to bridge the discrepancyof several order s o f magnitude in N cas determined by solvent swe1 1 i ngcompared to that determined bystress-strain measurements. Thediscrepancy has been attr ibuted tod i f f er ence s in exper imentalprocedures; various arguments alsohave been advanced to account par tlyfor the di screpancy as being dueeither to intermolecular obstructionsin the network or to swollen networkeffects.

o To-examine the statisticaltheory of coal network elasticity andto inquire what statistics the coalmacromolecules may obey, because coalmacromolecules may be neither longenough nor flexible enough forGaussian stati sties to apply.Entanglements, intermolecular forces,cooperative network effects, and therole of small coal ex t r act ablemolecules or solvents entering thecoal, are important effects in themechanical behavior of coal and needto be accounted for (Larsen, 1978;Lucht, and Peppas, 1981). There is aparticularly signi ficant need forexperimental work to reexamine theextraction and s~elling behavior ofbituminous coals, emphasizi ng thethermodynamic interaction parameter,the cross linking parameters N andp and the molecul ar we i gh t.cdistr ibution of the "extractablecoal" portion.

o To investigate the mechanismof fracture initiation andpropagat ion in coal. Conventionaltheory suggests fraetur ing would be

Rev. Mod. Phys. , Vol. 53, No. 4, Part II, O~cober 1981


determined by critical stress, rateof loading and temperature (amongother factors) . However, there isvery little information ava i 1 ablethat is specific to coal.

4. Chemical comminution

Chemical comminution of coal i s analternative an/ conceivably attractive processfor size reduction and concommitant liberationof mineral matter. This involves thetreatment of coal with a r eagent such asconcentrated aqueous ammonia solution. Thistreatment brings about the selective breakageof coal particles along the boundaries betweenmacerals and mineral components, as well asalong bedding planes.

There seems to be general agreement thatchemical comminution to a given particle sizegenerally produces gr eat er liber at ion o fpyr i tie sul fur than mechanical grinding,although the two processes appear to yieldcomparable liberation of mineral matter thatproduces ash. Ex per iments wi t h an IllinoisNo. 6 bituminous coal have shown chemicalfracture (4. 5/150 micron') liberatesconsiderably more pyr it ic sul fur thanmechanical crushing even to a considerablyfiner size (22/150 microns) (14heelock, 1977) .

Another possible advantage o f chemi calcomminution appears to be a narrower particlesi ze distribution, wi th smaller production ofultrafines. Thus the economics of such anapproach to size reduction might be favorable,although there are other factors (processingtime and cost o f r eagents ) that mitigateagainst favorable economics (Datta, 1976) . A

reagent cost as low as $10 or $15 a ton isquite significant compared to the price of aton of coal (say $25 to $50 a ton). Also,there is some evidence that —at least withammonia as the comminuting agent —there maybe a reaction taking place which increases theorganically bound nitrogen content of thecomminuted coal, an adverse effect.

Treatment of coals with chemicals, suchas ammonia, fragments coals selectively alongbedding planes and at boundar ies betweenmacerals and mineral matter. Theunderstanding both of the phenomenology and ofthe mechanisms involved and, therefore, of thepotential o f chemi cal comminution fortechnological exploitation on a cost effectivebasis, is rudimentary. But the potential may

be quite signi ficant. One would need tounderstand the nature of the various bonds andinteractions that are affected in thisprocess. Particularly important is the needto identi fy those characteristics of thereagents that are effective in comminution.In addition to ammonia, there are others(methanol, amines) that also comminute coal,although not as ef fectively. Al 1 thesecompounds are highly polar, with oxygen orn i tr ogen functionali ties and a lonenon-bonding pair of electrons. Such types ofcompounds can. swell and dissolve in coal evenat low temperatures.

o Di ffusion of the comminutingagents in coals, the s~elling ofcoal, the nature of the sur faceinteraction between the coal and thecomminuting agent, and the .resultingcrack initiation and/or propagation,all need to be investigated boththeoretically and experimentally.Some preliminary theoreticald iscussion o f the comminutionmechanism does not account for thecomminut ion by ga seous ammonia(Mheelock, 1977, 1981) . Thetechniques of solid state physics,particular ly emphasi s on suchtechn i ques a s electr on microscopy,could shed light on this importanttopic; these techniques also mightprovide some explanation of the factthat chemical comminution is stronglyaf fected not only by the nature ofthe comminuting agent, but also bysuch parameters as coal moi stur e,particle si ze of coal, pressure,pretreat, ment (partial drying orevacuation), maceral content,porosity, and permeabili ty of coal.A most si gnificant question iswhether there exists a combination ofchemical comminution and ot her si zereduction procedures that a) may bemore energy efficient than mechanicalcrushing alone and b) may result in abetter controlled particle sizedistribution.

5. Goal-oil mixtures

Coal-oil mixtures for combust ion inboilers are receiving considerable researchand development attention. Efforts are under

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 19S1


way in several countries not only todemonst rat e the engineer ing feasibi 1i ty ofutilizing coal-oil mixtures in utility andindustrial boilers which were designed forother fuels, but also to develop a fundamentalunderstanding of the behavior of coal-oilmixtur es. The research and developmentef forts address such aspects of theutilization of coal -oil mixt ur es, aspr epar at ion, storage, transportation, andcombustion control (Botsaris, 1979; Smith,1979) ~

Coal-oil mixtures have been exami nedper i od i cally in the past, mostly on atrial-and-error basis. Much technicalprogress was made in achieving slur rystability (i.e. particle suspension andpumpability) through viscosity adjustment. A

1942 program at an Atlantic Company refinerydemonstrated that stabilization of suspensionto prevent coal settling could be accomplishedby using very finely ground coal ( 1$ coalparticles ~ 70 microns). Up to 40 wt $slurries were prepared with No. 6 fuel oil,the slurry retaining a viscosity that allowedpumping with prehea'ting (250-350 cp I 100 F) .To achieve suspension stability requires coalpar ticles & 10 gm or, alternatively, viscositycan be increased by "gelling" with soaps orlime-resin greases. Use of such additivesresults in thixotropic slur ries that aresuf ficien tly vi scou s to suppor t coalparticles. Much of that early work related toslurry formation and stability was empirical,with minimal effort devoted to delineating thefundamental rheological properties and surfaceinteractions of the constituents involved.

ln addition to the degree ofpulverization, other factors which a f feetcoal-oil slurries are introduction of water asa third major component in the slurry;interaction between coal particles andsolvent/carrier, reactivity of lignites; thespeci fic fuel oils used as slurry vehicles;and the coal rank, maceral content, andmineral matter contained in the coal. Thee f f ect o f these latter specificcharacteristics of the coal may be especiallyimportant in the burning characteristics ofthe coal-oil slurry.

Grinding in an oil medium (as opposed todry grinding) and then dispersing pulverizedcoal to make oil slurries is receivingincreased attention. The scheme appears tooffer several advantages in that it mayminimize surface oxidation during grinding and

thereby enhance dispersion, reduce settling,eliminate explosion hazards due to dust beinggenerated in dry grinding, and simplifygrinding equipment (e.g . , Sz ego grindingmill) . Finally, consideration has been givento transportation of coal in oil slurries;this may be particularly attractive i f iteliminates the need for using scarce water, orif coal must be further reduced in size aftertransportation but prior to combustion.

Propert ies of coal-oil slur ries areaffected by the nature of the interactionbetween coal and the carrier oil. Someimportant questions r elate to stability,determination of vi scosity, pump i n grequirements, and determination and monitoringof composition and stability. Needs existhere also for contributions from physics. Forexample, difficulties have been reported inattempted viscosity measurements with No. 2fuel oil mixtures, because these slurries havetended to channel around the viscometerspindles in a Brookf ield vi scometer.Measurements of these high viscosity slurriesare additionally complicated when they areattempted at high temperatures (300F) (Smith,1979)-

o Prepar at ion and stabili zationof high-concentration slurries (50-70wt $ coal), heat transfer rates, flyash particle properties andcombustion mechani sms ar e importantproblems. Smi th et. a 1 . haveprovided a brief introduction to thepertinent fluid mechanics relevant tomixing and slur ry stability.Viscosity, evaporation losses, mixingand settling, mixing and powerr equi rement s, and maintenance ofslurry stability through the use ofadditives (dispersants orsurfactants) are important topics fori n ve s t i gation. Physical mearis ofmaintaining the stability of a slurryhave also been considered (ultrafinepulveri zation and ultra soni cagitat, ion).Rotary viscometric studies have shown that

the rheology of slurries below 30 vol g coal isNewtonian. The behavior of the moreconcentrated slurri es i s not particularly wellunderstood, although i t appear s to depend onshear stress and to be thixotropic. Thickeningcan require a large shear stress to restorefluid motion, and pr esents a hazard intransportation ( pi pel inc blockage ) a s we 11 .

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I; October 1981


Emulsification with water appears to amelioratethis problem and also improves combustion.

o Sediment, ation and filtrationtheory need to be investigated athigh loadings of solids in highviscosity liquids. Thecoal-oil -wat er ternary mi xtures areparticularly important and stand inneed of experimental and theoreticalinvestigat, ion. The relativehydrophobicity (or-- oleophilicity) andhydrophilicity of coals are believedto in fluence t he e f feats ofaddi ti ves, a nd therefore theproper t ies of t he coal -oi 1-s1u r r i e s,but t, he detailed mechanisms are notunderstood.

6. Coal/water systems

Fundamentals of the interaction of waterand coal merit considerable investigation, thisinteraction is important in most branches ofcoal technology. For example. '

Dewatering is an important aspect of coa1preparation. Water (either native or added asa grinding aid) is present in coal which hasbeen ground, but must be removed prior to use.Thermal, mechanical, or chemical means may beused to effect water removal. The interactionis complex and recent, ly has been the object ofmuch study. The di f fusion of water, itsinteraction with coal sur faces and the natureand strength of the bond that forms wi th coalclearly are important matters for investigation(see discussion in Section III B). Thesynergistic or antagonist, ic ef fects of waterand surfactants (or other additives) are alsoof interest.

Mater is particularly significant in theutilization of lignite. Lignite and other lowranked coals may contain up to 40% water. Thishigh water content gives r ise to severalproblems, among them transportation problemssuch as "excess baggage of water", freezing, oreven spontaneous combustion. It has beenestimated that lower ing the water content toabout 10$ would make lignite economic to use at,much greater distances from the mine.

I

7. Beneficiation/high gradient magnetic separation{HGMS}

o It is highly desirable topursue investigations i n novel

separ ation technologies for coalbene f ic i ation, t hat d epend onproperties other than density. Thereare many possibilit, ies that, can bepursued, and reviewing them all isbeyond the scope of this report.However, the possibilities may beillustrated by discussing the case ofHigh Gradient Magnetic Separation(HGMS) (Maxwell, 1 978; Petr aki s etal, 1980).High gradient magnetic separat ion isa recently developed technique in whichmicron-si ze part ic les cont ained in aflowing fluid medium can be extracted fromthat medium. The t echnique can also beextended to non-magnetic moities providedthat they can be made to associate withmagnetic seeding materials.

Magnetic separations have been usedin mineral processing for many years andin several forms. The basi s of theseparation process is the force actingupon a magnetic part, icle in a magneti cfield due to the magnetization of theparticle and the magnetic gradient, in theregion about the particle. Typicalmagnetic separators that have been usedcommercially, as well as the values forthe magnetic gradient contained wit, hineach separator, are as follows: wet drumseparator employing permanent, magnets forprocessing magnetic ores (gradient 0. 1kG/cm); suspended elect romagneticseparator s for t ramp i ron removal(gradient of 1-10 kG/cm); wet highintensity separators for processing weaklymagnet, ic iron ores (100-1,000 kG/cm).HGNS r epre sents. an extension of thesegradient values, i.e. , the gradientcontained within an HGMS unit possesses avalue 1000-10,000 kG/cm. It is theincreased value for the gradient within anHGMS separator that permi ts, inter alia,the separation of paramagnetic micron sizeparticles from a diamagnetic medium on alarge scale. A second att, ribute of anHGMS separator that is significant is thatis permits separation rates that are muchgreater than those obtained by standardfiltration methods. HGMS has already beenapplied successfully for beneficiation ofcoal and coal liquid, showi ng somecons i der able pr omi se of success (Maxwellet al, 1981).

Rev. Mod. Phys. , Vol. 53, No. 4, Part iI, October 1981


8. Recommendations

1. Research i s needed to improveunderstanding of the physical structure ofcoal and to relate measurable/describableproper ties to the mechanics of si zereduction processes.

2. It is desirable to investigatethe nature of the "bond" (adhesion?)between organic and mineral matter, aswell as between macerals, and to relatethe strength of these bonds to sizereduction proces'ses.

3. Invent ener gy -e ff i c ient gr indingprocesses. Especially develop models t oexplain the decrease in grinding rate forparticles below 50& diameter.

Attempt a rational descriptionand classificat ion of the mechanicalproperties of coal.

Determine shear and other moduliand their dependence on coal rank, maceral

, and coal primary (chemical) structure; anddevelop new techniques (e.g. , soundvelocity measurements) for t, hedetermination of such moduli.

6. Examine the statistical theory ofcoal network clast icity and investigatethe statistics coal macromolecules mayobey, because they may be neither longenough nor flexible enough for theGaussian statistics of polymer theory tobe appropriate.

7. Investigate the mechanisms offracture initiation and propagation incoal and in coal particles of var iousSl Ze ~

8. Di f fusion of the chemicalcomminuting agents in coals, swelling, thenature of the sur face interaction betweenthe coal and the comminut ing agent, andthe r e su 1t in g cr ack init, iation and/orpropagation, all need to be investi'~atedtheoretically and experimentally.

Mixing and stabilization, heattr ansfer rates, fly ash particleproperties and combustion mechanisms areimportant, problems for the development anduse of high-concentration slurries (50-70wt $ coal).

10. Sedimentation and flit;rationtheory need to be investigated at highloading of solids in high viscosityliquids.

11. Investigate applicabilit, y o fnovel separat ion/benef iciat iontechnologies that depend- on propertiesot, her than density (e.g. high gradient,magnet, ic separation).

9. References

Austin, L. G. and P. Baggy, 1979, November,preprint.

Baughman, G. L. , 1978, Synthetic Fuels Data- Handbook, Cameron Engineers, Inc. ,Denver, CO.

Bot sar i s, G. , et al, 1979, SecondInterna t i onal Sympo si um on Coal -OilMixture Combustion, NTIS Con f-791169,Vols. I and II.

Dat ta, R. S. , P. H. Howard, A. Hanehet t,1 979, NCA/B CR Coal Conference andExposi tion I I I. Co al: Energy forIndependence.

Foo, O. K. , et al, 19"(9, DOE Report Contract,No. ET- (B-C-01-3179, The Mi t re Corp . ,Mc Lean, VA.

Gaudin, A. M. , 1939, see page 184 inPrinciples of Mineral Dressing,Mc Gr aw-Hi 11, New York.

Herbst, J. A. , 19 (8, "Energy Requirementsfor the Fine Grinding of Coal in anAttritor", U. S. Dept. of Energy, FinalReport Contract No. EY-77-S-02-4&60.

Howard, P. H. , and R. S. Dat ta, 197 (, CoalDesul fur i zation' . Chemi cal and PhysicalMethods, T. D. Mheelock, ed. , ACSSymposium Series No. 6W, Mashington, D. C.

Larsen, J. M. , T. K. Green, P. Choudbury,and E. M. Kuemmerle, 1981, Advances inChemistry, Vol. 192 edited by M. L.Gorbatty and O. Ouchi, American ChemicalSociety, washington.

Larsen, J. and J. Kovac, 1978, "PolymerStructure of Bituminous Coals" in OrganicStructur e of Coal, J. Lar sen, ed. ,Symposium Ser i es 71, Amer i can Chemi calSociety, Washington, D. C.

Lowr y, H. H. , ed. , 1945, Chemistry of CoalUtil i zation, J. Mi ley, NY.

Lucht, L. M. , and N. A. Peppas, 1981, pp.2848 in Chemistry and Physics of CoalUt i 1i za t i on-1980, edited by B. R. Cooperand L. Petrakis, Am. Inst. of Phys. ,Proceedings Series No. 70.

.Luckie, P. T. , and L. G. Austin, 1980,F E-2375-25, DOE Coal Gr inding Technology—A Manual for Process Engineers.



Maxwell, E. , 1978, pp. 98-112 in ScientificProblems of Coal Uti li zation, edi ted byB. R. Cooper, Dept. of Energy SymposiumSeries, available a s CONF-770509 fromNational Technical Informat ion Series.

Maxwell, E. , D. R. Kel land, I. S. Jacobs,and L. M. Levinson, 1981, in Chemistryand Physics of Coa 1 Util i zation-1980,edited by B. R. Cooper and L. Petrakis,A I P Con ference Proceedings No. 70,American Institute of Physics, New York.

Heloy, T. P. , 1981, private communication.morgan, M. T. A. , and N. B. Terry, 1958,

Fuel, 37, 201.Mraw, S. C. , and B. G. Silbernagel, 1981,

Physics and Chemi stry o f Co a 1

Utilization-1980, B. R. Cooper and L.Petr akis, eds. , Am. Inst. of Phys.Proceedings Series No. 70.

Oder, R. R. , 1980, private communication.Orr, C. , 4r. , 1966, Particulate Technology,

MacNi 1lan, New York.Par i sh, 8. , 1974, A Bibliography of

Comminution, Inst. Chem. Engineers,London.

Peters, M. , and H. Juntgen, 1964, 1965,Brennstoffe-Chemic, 46, 56, 46, 175.

Petrakis, L. , P. F. Abner, and F. E. Kiviat,1980, p. 1260 in Fine ParticleProcessing, Vo 1. I I, e d i ted by P.Somashundar an, AINE.

Petrakis, L. , and D. N. Grundy, 1981,Nature, 289, 476- .

Petraki s, L. , and D. M. Grundy, 1981, Fuel,60, 120.

Petrakis, L. , and D. M. Grundy, ProceedingsRoyal Society, accepted for publication1981.

Smi th, R. S. , et al, 1979, SecondInternat iona 1 Sympo si um on Coal-OilMixtures Combustion, Dept. of Energy,Conf . 791160.

Soo, S. L. , et al, Annual Review of Ener gy,J. M. Hollander, ed. , Vol. I, 1976.Stopes, N. C. , 1 91 9, Pr oc. Roy. Soc.

(London) 90B, 470.University of Oklahoma, The Science and

Public Policy Program, 1 975, EnergyAlternatives. A Comparative Analysis,U. S. Gov't Publication PREX 14.2: EN2,washington, D. C

Ursu, Ioan, ed. , 1975, Energy and Physics,Proceedings o f the Thi rd GeneralConference o f the Eur opean Physi ca 1Society, Bucharest, Romania.

Van Krevelen, D. N. , 1965, Fuel, 44, 229.Mhee lock, T. , ed. , 1977, Coal

Desul fur i zation, Symposium Series 64,American Chemi cal Soci ety, washington,D. C.

Mhitehurst, D. D. , M. Farcasiu, and T. 0.Marshall, 1976, EPRI Report AF-252.

Yeager, K. E. , 1980, Coal Clean-upTechnology in Annual Revie~ of Energy,ed. by J. H. Hollander, vol. 5.



IV. F. CATALYSIS

1. Introduction

The development of appropriate catalystsprobably provides the greatest opportunity forexerting leverage on synthetic fueltechnology. The advantages o'f per forming t henecessary reactions rapidly under " mild"conditions of temperature and pressure areobvious. What a physicist approaching t, hesubject for the first time should also bear inmind, however, i s that the key desirableattribute of a catalyst is productselectivity. To put it quite simply, onewishes to produce a desired end product suchas high octane gasoline. Catalysts that allowone to obtain that en/ product mostexpeditiously, and with the least amount ofless valuable products, are most desirable.

In thi s section, we will discuss severalareas of catalysis offering interesting andtechnologically impor tant r esearchopportunities. These deal with: (1) shapeand size se lee ti ve zeoli t e catalysts,especially the synthetic zeolite ZSM-5; (2)mineral matter catalytic e ffects in directliquefaction; (3) sur face science related totransition metal heterogeneous catalysts forsynthesis gas conversion; and briefly withboth (4) catalyzed gasification of carbons andcoals and (5) denitrogenation anddesulfuri zation catalysts.

As discussed in the Introduction and inthe Technology Pr imer of Section IV A, thereare two main technological strategies for coalliquefaction that are being actively pursuedat present. One is indirect liquefaction.Here, one first gasifies the coal to a mixtureof CO and H&, and then converts this synthesisgas ("syngas") catalytically to the desiredhydrocarbons. This gives rise to standardclassical problems in heterogenous catalysis.The other technology i s direct liquefactionwhere a mixture of coal, process-derivedsolvent, and hydrogen is "cooked" at hightemperature and pressure in a reactor. Topics(&), (3), and (4) above are relevant toindirect liquefaction, and (2) and (5) arepertinent; to direct liquefaction.

The traditional approac h to ind i r ec tliquefaction is to synthesize the desired endproduct s directly thr ough Fi scher-Tr opschprocessing using transition metal catalysts,

such as iron. The problem here has beenproduct selectivity -- especially the factthat the product distribution extends tolinear hydrocarbons heavier than the C -Cgasoline range. An alt ernative approach,developed recently by the Mobil Co. , i s toproceed from syngas to gasoline in two stages.First is the catalytic conversion of syngas tomethanol (CH OH); this is an establishedcommercial technology. Then follows catalyticconversion of methanol into high octanegasoline using a recently developed si ze andshape selective synthet ic zeolite. Currentefforts attempt to carry this type of approacheven furt;her, by combining transition metalFischer-Tropsch catalysts (e.g. , Fe, Ru) withshape-selective zeolites (e.g. ZSM-5) toeffect direct conversion of synthesis gas togasoline-range hydrocarbons (Chang et al,1979; Huang and Haag, 1980; Rao and Gormley,1980) .

awhile zeolite catalysts are very widelyused in the petroleum industry, being bydollar value the most important of catalystsused commercially, the unusual propertiesoffered through their unique type of crystalstructure are not well explored by physicists.Thus they combine the possiblity of novelty inphysics with practical significance. Mineralmatter catalytic effects [topic (2) above],while not being so evocative of novel physics,constitute a relatively new area and one wherephysics technique has already been usefullyapplied. Discussion of topic (3), the surfacescience related to transition metalheterogeneous catalysis for fuel synthesis, isvery important technologically and interestingscientifically, and general interest in thisarea is already well establi shed among solidst at e physics. Remarks here will highlight afew questions of technological importancewhich deserve special emphasis in research.Brief remarks on (4) and (5) deal with areasof est, ablished chemical interest which arerelatively unknown to physicists.

2. Zeolite size and shape selective catalysts

Zeolites (Smith, 19(6; Eber ly, 1976;Kerr, 1981) are high surface areaalumino-silicates containing exchangeable,change-balancing cations. The crys t alstructure gives rise to cavities and channels



(l3.3l A)

CANC R I N I TE

SUPEACAGE

HEXAGONALPRI SM

I

/IIIII

I

\

I

1III

C(15.18 A)

FIG. IVF-1. Line drawing of erionite. The drawingdepicts the framework structure made up of corner-linked tetrahedra in which small atoms (denoted Tatoms) lie at the centers, and oxygen atoms lie at thecorners. The. T-atoms in erionite are Al and Si. (In syn-thetic- zeolites, neighboring atoms in the periodic table,such as Ga, Ge, and P, might be incorporated. ) The linesin the drawing connect the T-atoms. The large sorptioncage (super cage) in erionite is enclosed by rows consist-ing of alternating units of smaller cages, called cancrinitecages, and hexagonal prisms. The sorption cavities consti-tute a three-dimensional pore network. Each cavity has alength of 15.1 A and a free cross-sectional diameter of6.3 to 6.6 A. Sorbate molecules can enter the cavitythrough six elliptical openings (minimum and maximumdiameter of 3.5 and 5.2 A respectively) formed by rings ofoxygen (after T. E. Whyte, Jr., E. L. Kerr, and P. B. Ven-uto, J. Catalysis 20, 88 (1971)).

on an atomic scale, and of well defined shape.(See Fig. TU F-1 for an illustration of such astructure, that, of the zeolite erionite. ) Theproperties of zeolites that are important forcatalytic applications are: (1) large sur facearea; (2) shape and size selectivity; (3) verystrong acidity —(By acidity we mean that thezeol i tes are proton donors. The availabilityof protons is an important factor in thepractical importance o f zeoli tes. Inc atalytic "cracking" a proton breaks acarbon-carbon bond, thus breaking largehydrocarbons into smaller molecules. In acommercial "cat-cracker" the hydrogen in thezeolite would be replenished by some feedingprocess. ) --; (4) the presence of strong

internal electrostatic fields; and (5) thepossibility that zeolites can be used to makedual-funct ion catalys t s by being used t osuppor t ultrafine particles of transitionmetals. Since the acidity affects theparticular organic chemi c al. r eac t ion sinvolved, physicists will probably not be verydirectly concerned with this aspect of zeolitebehavior . On the ot her hand, physicistsprobably will be interested in gaining furtherunderstanding of the shape and sizeselectivity which arises from the fact thatthe pore structure determines the size andshape of the organic molecul es t hat can passthrough the zeolite.

The zeolite of most current interest inthe context of coal utilization and syntheticfuel production is a syn'thet ic zeolitedeveloped by Mobil Res. and Dev. Corp. , anddesignated ZSM-5. In a process recentlydeveloped by Mobil, ZSM-5 is used to catalyzeconversion of methanol (CH OH) into highoctane gasoline (Derouane atA Uedrine, 1980;Derouane and Gabelica, 1 980). So far as theproduction of gasoline i s concer ned, theimportant factor is that the molecular sizerestrictions for di f fusing through the ZSM-5channel structure discourage the formation ofundesired larger hydrocarbons, as occurs inFi scher -Tr opsch processes . ( ZSM-5 has twoorthogonal channel syst ems wi th anintermediate sized access diameter, relativeto other zeolites, of 5. 5 to 6 A) .

The quest ion here of most immediateinterest to physicists probably i s how tounderstand the dynamics of molecular diffusionthrough t he zeol i te channel structure. Dataobtained by traditional bulk sorption methodsmay be quite misleading, since uptake behaviormay be determined by phenomena ot her t hanintracrystalline diffusion (Dubinin et al,1980). The most promising tool for achievingthe necessary understanding is hi gh powerpulsed NMR (Karger and Caro, 1977).

Motion of diffusing methanol, or otherorganic molecules will strongly affect theirNMR behavior . Nuclear inter actions areaveraged by molecular motion~ occurring ontime scales shorter than 10 —10 seconds(motjonal 8nar rowing) . Motion s on the1 0 —10 second time 'scale producefluctations which cause ener gy exchangebetween the nuclei and their environment(spin-lattice relaxation) . sideline NMR

measurements of the shape and width of the NMR

absorption, and tr ansient NMR measurements of



the spin lattice relaxation time (T&

) probethese motions. The methods of nuclear spinrelaxation analysis permit determination ofcorrelation times for molecular motion andmean interaction energies for absorbatemolecules (Pfeifer, 1972) . Pulsed fieldgradient NNR measurements permit determinationof the diffusion coefficients directly (Kargerand Caro, 1977; Karger et al, 1978; Karger andVo 1kmer, 1980) .

Besides using ordinar y or ganicmolecules, it may be useful to work withdeuterated ones. Re sonance properties ofdeuterons are dominated by the interaction ofthe nuclear electric quadrupole moment withelectr ic field inhomogenei tie s coming fromtheir bonding in the molecule. This is a muchstronger, and more local, interaction than themagnetic interactions determining the NNR

behavior of protons, and thus providescomplementary information on molecular motion.One can extract information on correlationtimes and activation energies for moleculardif fusion, and their temperature variation,from analysis of the WR linewidth behavior.

On the theoretical side, the suggestionhas been made (Rabo, 1980) that in thedi f fusion of organic molecules throughzeolites, the zeolite acts as a strong solidelectrolyte, interacting strongly with thepolarizable molecules. This raises thequestion ' as to whether there are usefulanalogies to the theory of ionic conductivitywhich might be developed.

As in the case of the studies on water incoal (Mr aw and Si lbernagel, 1981),complementary measurements of equilibriumproper ties, especially speci fic heat, maypr ove helpful, although the situation,involving reactions of organics, is obviouslymore complex. Heat capacity measur ementsshould reveal long time, macroscopic averagebehavior for molecules r esident, in zeolites.Such measurements could provide valuableinformation about adsorption, if the organicfluid resident in the pores had a phasetransition in its free state within thetemperature range investigated (i.e. one couldr ecogni ze change in the nature of thetransi t ion through the heat c apacitybehavior).

There is great current interest, in usingzeolites as supports for very small transitionmetal particles (Cusamono et al, 1981; Changet al, 1979; Huang and Haag, 1980; Ninachevand Isakov, 1976; Gallezot, 1979; Naccache et

al, 1 977 (also see some earlier studies citedin Naccache et al ) ) . The particle can besupported on the surface or (with greaterdi f ficulty, but of more scientific -- notnecessarily more technological -- interest)imbedded in the zeolite, i.e. using the cagestructure to stabilize very small dispersedmetal particles or cluster s in the range7-10A. The practical interest in suchsystems, as discussed in the introductoryremarks of I VF' 1 above, i s to gain thecapability to effect the direct conver sion ofsynthesi s gas to gasoline-range hydrocarbons.It would be interesting to explore thepossible di f ference in behavior betweentransition metal particles on the surface of azeolite and particles embedded within thecages of the zeolite structure. Undertakingsuch studies hinges on being able to embedsuch par ticles or clusters. Work by Naccacheet al (1977) for platinum catalysts shows thatsmall part icles of 8 —10A size arehomogeneously dispersed inside the supercagesof Linde Y zeolite. Once formed, thepar ticles are too 1ar ge to escape through thesmall 7. 2 A port windows (i.e. largest,atomically def ined openings from thesupercage).

Study of such ze ol i te -me tal par ticlesystems could of fer several interestingpossibilities for physics work of practicalimport. In the making of such systems onemight consider the use o f solid statechanneling phenomena ( i .e. par ticularlyconsidering the natural channel structure ofthe zeolite itself), such as are used in ionimpl anta ti on technique s for semiconductors.One might be able to control both compositionand location of tnet, al cluster s. EXAF S(extended x-ray absorption fine struct, ure)could be applied t o quest ions of smallparticle structur e (interatomic dist, ances andparticle shapes) (Renouprez et al, 1980) andof metal-zeolite interactions (the particlesmay be significantly electron deficient)(ideber et al, 1980) . EXAFS can also beapplied to investigate the local coordinationof cations in zeolites (Morrison et, al, 1980a,1980b), and can be used to investigatemetal-clustering phenomena. The whole arrayof modern surface physics theoretical andexperimental techniques for studyi n g theelectronic structure of sur faces could be usedto study the charge transfer between metal andzeolite. This is important for determiningthe "acidity", and hence the way in which



particular reactions occur. XPS (X-rayphotoemission spectroscopy) and associatedtheoretical calculations of electron structureshould be valuable. Study of quantum -sizeeffects by susceptibilit, y aed NNR measurementsshould also be useful. Besides examining theelectronic structure per se, one would want tostudy vibrational spectra, perhaps by Raman,x-ray, or neutron scattering, in order toreveal effects of the metal particles bondingto the zeolite internal surface.

3. Mineral matter catalytic effects in direct liquefaction

There is substantial evidence (Granoffand Montano, 1981; Montano, 1980) that thenative mineral residue in coal plays ani mpor tan t catalytic role in directhydroliquefaction processes. This effect isthus far neither scientifically understood norcontrollable. It is important to learn whichare the catalytically effective minerals, andhow they work.

Certain naturally occurring minerals incoal such as pyrite (FeS ) and iron bearingclays have been shown t, o enhance the liquidyield and product quality (i.e. selectivitytoward desirable weight r ange o fhydrocarbons) . Similar effects have beenobserved when additional pulverized pyrite wasdeliberately added to coal, and this suggestst he possibility of using inexpensivedi sposable catalysts for direct coalliquefaction.

In direct 1i que fac t ion pr oc e sses, coal,process-derived solvent, and hydrogen are fedfirst into a preheater, where liquefactionbegins, and then into a reactor where furtherconversion of the initially "solubilized" coaltakes place. It was found that r ecyc lingmineral matter residue enhances the yield of"distillate" products. The nature and extentof these mineral effects should be assessed asone step toward optimum conversion of coal toliquid products, and perhaps toward thedevelopment o f r el a t i v el y inexpensivedisposable catalysts. Understanding of therole of the mineral matter is also desirableto develop direct liquefaction processes thatare not too narrowly restricted in regard tofeed coal type.

Observation of mineral matter effects hasthus far consisted primarily of noting the

correlation of minerals present with productcomposition. The key result has been thatpyr i te enhances conversion, and that theeffectiveness of different pyrites (i.e. theterm "pyrite" is often loosely used to meaniron sulfides of varying stoichiometry andcrystal structure) appear to varyconsiderably. Under the t ypi c a l hi g ht emperature and pressure condit, ions ofliquefaction, pyrite is partially transformedinto a non-stoichiometric iron sulfide knownas pyrrhotite (Fe

&

S) . It has been suggestedthat pyrrhotite is really the active catalystfor coal liquefaction. It is important,therefore, to learn the role of pyrrhotite incoal conversion.

Ho ssbauer spectroscopy has beenintroduced (Granof f and Montano, 1981;Montano, 1 980) to study iron-containingspecies, such as pyrrhotites, in coalliquefaction residues. In par ticular, thevarious Fe

&

S compounds can be distingui shedt hr ough t he magnet ic hyper fine spl i tt ing.(This split, ting depends on the number ofvacancies in the neighborhood of an ironsite. ) Through correlation of the Mossbauerinformation with liquid yields and properties,these exper iment, s have shown that anincreasing number of vacancies (i ~ e. greateriron deficiency in the pyrrhotite) isassociated with better conversion of coal tothe desired liquids.

To further define the role of pyrrhotitesas catalysts, it will be valuable t, o monitorthe pyrite to pyrrhotite transformation insitu, i.e. under actual coal liquefactionconditions. Fortunat;ely, because i t d epend supon gamma rays, Mossbauer spectroscopy iswell sui ted to observing exter'nally thepyrrhotite transformation as it occurs in areactor, i.e. the signal is able to penetratethe reacting medium, which would be opaque tolonger wavelength radiation.

It i s desirable to further study therelationship between vacancy concentration andpyrrhotite catalytic activity. Apparently thevacancies act to facilitate atomic migration.Systematic studies of catalytic activity inpyrite materials with varying vacancy contentare worthwile. Recent work (Graham et al,1981) indicates the potential value ofshock-wave activation (Davison and Graham,1979) to generate vacancies in pyrites forsuch studies. The star ting and shockactivated material were analyzed by X-raydi ffrac t ion and magnet i zat ion mea sur ement s.

Rev. Mod. Phys. , Vol. 53, No. 4, Part I l, October 198'I

Research for the Technologies $139

The measurements revealed the creation ofcrystal defects and the presence of a fewpercent pyrrhotite after shock activation.Ion bombardment is another technique that mayprove useful for inducing defects in pyrites.It is important to develop other analyticaltools, in addition to Mossbauer spectroscopy,to investigate mechanisms such as the pyriteto pyrrhotite transformation -- particularlybecause one may wish to investigate mineralsyst ems based on other metals, such as nickelor molybdenum, which are not suitable forMossbauer experiments. An att r ac t i vepossiblity, that might be useful for "in situ"i nvesti gation s under actual reactionconditions, ~ould be some sort of .acousticspectrocopy, based say on magnetoelasticinteraction.

4. $urface science related to transition metalheterogeneous catalysts for gas synthesis

Fundamental surface science studiescontribute to the research base for technologydevelopment, because they provide guidance indealing with problems of practical catalysis(Somor jai, 1979). Two categories of suchproblems are: (1) attempts to find newcatalysts, and therefore to recogni ze andunder stand electronic and crystal-structureeffects involved in catalyt ic behavior,including metal-support interactions and metalcluster size effects on electronic behavior(.always bearing in mind the importance ofproduct selectivity); and (2) maintenance ofthe effectiveness of catalysts, i.e. problemsof deactivation and "poisoning. "

Pr ac tie al t r ansition metal catalytic' systems typically consi st o f small particles,say 1 g-'l 00 A diameter, d i sper sed on hi ghsur face area supports, e.g. refractory oxidessuch as silica (Si0 ) or alumina (Al 0 ). Itis difficult to apply those surface alieneetechniques involving high vacuum (i. ,e. thosemeasuring emitted electrons spectroscopicallyor using bombardment with a stream ofparticles) directly to such systems. (This isbecause of the state of cleanliness necessaryto operate in high vacuum. ) Moreover, thedifficulty of direct examination of operatingcatalytic systems i s compounded by the factthat the usual operating conditions alsoinvolve high .pressure. Thus sur face science

studies pert inent to transition metalcatalysts tend to f all into two ratherdis joint categories, those on "real i stic"d i st ributed systems of supported smallparticles, and those on "model" systemsinvolving single crystals of significant, ar ea(~1em ) and epitaxially grown single-crystal,or polycrystalline, films. Surface scienceexperiments involving high vacuum play acentral role in studying such model systems.Such studies on model systems usually dealwith chemisorption and binding energies ofreactive species and the associated steric andcrystallographic behavior (locations andpostures of adsorbed species) . Interest insuch model system work is already widespreadand intense among physicists, chemists, andmaterials scientists. The present discussionis meant to identify a few questions t hatdeserve special emphasis in research becauseof their technological implications.

The tools of high vacuum surface scienceare used to character i ze modeltransition-metal catalytic systems throughcharacterizing t he sur face structure andadsorbed species present, and probing thedetailed electronic structure, including themapping of the electron charge densitydistribution (Somorjai, 1979; Somorjai and VanHove, 1979; Van Hove, 1980) . Among thesetools are low energy electron diffraction(LEED); sur face-sensitive extended X-rayabsorption fine structure (SEXAFS); low —,medium-, and high-energy ion scattering (LEIS,MEIS, and HEIS); high resolution electronenergy loss spectroscopy (HREELS); Augerelectron spectroscopy (AES); and angularr e so 1 ved ul tr av i o let photoem'i ssionspectroscopy (ARUPS).

Surface science studies of additives tothe elean metal surface (i.e. additives otherthan the molecular species involved in theparticular reactions of interest) deservespecial emphasis. Such additives play asignificant role -in controlling the chemistryof catalytic reactions, and may be harmful(poisons) or helpful (promoters) .

Comm'on poisons are carbonaceous deposits,sulfur, and nitrogen. These may come fromimpurities pr esent in the reactants (e.g.sulfur present, in coal used to make syngas) ormay develop as products of side reactions(e.g. in hydrogenation of carbon monoxide oniron metal, the catalyst loses effectivenessthrough buildup of a carbon layer). A

question of great importance, naturally, ishow the poisoning occurs. Several mechanisms



are possible, and i t i s o ften not clear in agiven case which mechanism applies. Simplemechanisms for poisoning or deactivation arecoverage of the active catalyst site or, inthe case of small particle systems, sinteringtogether of the particles with consequentreduction in active sur face area. Moreinterest ing from the physics point of view arepossibilities such as sur fac'e structuraltransformations or a change in the electronicstates of the per tinent sur face metal atoms.Questions of interest include how tounder stand any corr clat ions between sur facerestructuring (rearrangement of surface atoms)and poisoning of catalytic activit, y; and doesa poison work by blocking si tes on aone-to-one basis, or does the poison operatevia a long-range electronic interaction viathe metallic substrate? The same spec i e s maybe either a poison or a promoter depending oncircumstance. For example, potassium acts asa poison for the hydrogenation of carbonmonoxide by iron metal; but potassium improvesthe catalytic activity of an oxidized ironmetal surface.

The way in which poisons act can beexamined by a combined study of sur facemorphological changes during adsorption of thepoison (e.g. site coverage and/orreconstruction as ment, ioned above) and ofdetailed electronic structure, i.e. electroniccharge density studies. The most powerful ofsur face science tools to probe electronicstructure is angularly resolved ultra-violetphotoemission spectroscopy (ARUPS). ARUPS isvaluable, of course, not only for studyingdetailed working of poisons but also forgeneral probing of effects associated withchemi sorption, such as sur face states andadsorbate energy bands. ARUPS allows one tolook directly at charge rearrangement at, thesur face. Such studies pr omise to be mostpowerful in conjunction with self-consistentcalculations of surface electronic structureand chemisorption. Such calculations are in avery acti ve and advancing stage ofdevelopment. + These calculations, which can

~ Some references pertinent, to self —.consistentsurface electronic structure calculations are.Appelbaum and Hamann, 1978; Arlinghaus e t a 1,1980a, 1980b; Feibelman and Hamann, 1979;Goddard, 1981; Jepsen et al, 1978; Kasowskiand Caruthers, 1980; Kleinman and Mednick,1981; Kr akauer and Cooper, 1977; Louie et al,1976; Na et al, 1981; Posternak et al, 1980,%fan g and Fr eeman, 1979.

yield complete charge density mappings, aid inunderstanding the site selection of anad so r bed species, including the importantquestion of whether t he adsorbed moleculepenetrates the sur face layers of the hostmetal.

"Model" system studies of adsorption oflarge organic molecules on transition metalsurfaces are quite limited as yet. Fewexper imental studies have been done usingtechniques other than LEED (Somorjai and VanHove, 1979). One would also hope to extendself-consistent surface electronic structurecalculations to be able to deal withchemisorption o f 1ar ge molecule s. Recentlydeveloped technique s f o r ab —i n i ti ocalculations for electronic structure of largemolecules (Kaufman et al, 1979) should provevaluable in this regard. Nevertheless, thisis a formidable problem.

Tradit ional sur face sc i ence stud i es andactual catalytic conditions (even for "modelsystems) are seIarated by an enormous pressuredisparity (10 torr compared to 10 torr).An important recent development is havingr eaction chambers, i.e. small enclosures,incorporated into vacuum apparatus, in such away that a sample can be characterized underhigh vacuum before reaction; then the reactionchamber can be isolated and the surfacereaction carried out under "realistic"conditions; and finally the high vacuum- can berestored, and the state of the sample can beexamined after the reaction (Somorjai, 1979;Goodman et al, 1980). Thus the whole study isdone in situ without removing the sample fromthe cont;rolled atmosphere enclosure. (In thethird stage of t,his procedure, of course, somecaution in interpretation is necessary becauseo f possible changes in chemisorbed structurescaused by pumping down to very low pressure. )This technique was used recently to examinethe effect of surface chemical composition onthe kinetics of the cat alytic methanationreaction over a single crystal Ni (100)catalyst (Goodman et al, 1980).

For st, udying electronic aspects ofchemisorption of tr ansi t ion metals underincreasing pressure of the reac ti vea tmosphere, one should invest igate thecapabilities of electromagnetic (optical)reflectance spectroscopy as a substitute forelectron emission spectroscopy, eliminatingthe need for high vacuum. It is true thatreflectance spect, roscopy i s not usual 1. yregarded as sur face effeet selective. Perhapsa grazing angle technique could be used to



overcome this. Careful feasibility estimatesfor a range of spectral regions would seem tobe desirable.

In addition to the work on model crystalsystems, cr y s tal —s tr uetur e andelectronic-structure work on the disper sedmetal par t icle systems that ar used inpractice is called for. Sinfelt (1969, 1977)has explored the properties of bimetalliccluster catalysts. Such a catalyst system,consi sting of very small metal clustersdispersed on the surface of a carrier such assilica or alumina, is made by impregnating thecarrier with an aqueous solution of salts ofthe two metals, drying, and then reducing thesalts in a stream of hydrogen at elevatedtemperatures. One obtains mix d clusters oftwo metals even for cases, such as Ru-Cu orOs-Cu, which show very low miscibility in thebulk. Moreover, the properties are markedlyaf fected, on going to very small particles, bythe tendency for surface segregation, e.g. thesur face becoming very Cu r ich in the Cu-Rusystem leading to a situation with a layer o fcopper on a core of ruthenium. Such catalystso f ier improved selectivity for certainreactions. A start has been made on usingEXAFS to examine the structure and compositionof such systems (Sinfelt, 1977) . (Note thatthis is very similar .to the use, suggestedabove, of EXAF'S to investigate metal particlesinbedded in the zeolite ZSN-S. ) It should bepossible to parry out sel f -con si stentcalculations of the electronic structure ofsuch systems (Na et al, 1981).

Systems that are of interest in modelingbimetallic cluster catalysts are epitaxiallygrown films, where one deposits a monolayer,or a few layers, of one transition or noblemetal on top of another metal (Biberian andSomor jai, 1979; Richter et al, 1981; Ma et al,1981). Such epitaxial deposition might evenbe relevant to practical catalysts, i f thedeposition is on an appropriate supportsubstrate (e.g. , if substrate for bimetallicfilm structure is a transition metal oxide).

Morphological change s o f small metalpar ticle catalyst systems in' r eac t i veatmospheres have recently been studied usingmodern sur face science techniques (Schmidt,1981). Mith STEM (scanning transmissionelectron microscopy) and XPS, the behavior ofPd-Pt alloy particles of 20 to 200 A dimensionwas observed. Sintering in an- oxygenatmosphere, and the r.eduction back to metalpar ticles in a hydrogen atmosphere were

observed; and cracking processes to getsmaller particles and reannealing to get backta larger particles could be followed.

Metal-support interactions and thephenomena, already ment ioned, of sur facesegregation (Brundle and Mandelt, 1981;Abraham, 1981), are important questions in theuse of di sper sed-par ticle tr ansi tion-metalalloy catalyst systems. Consideration ofinterfacial energy should relate to thecurvature (wetting angle) of small metalparticles sitting on a support. Formation ofcracks in such particles, again should beunderstood on the basis of inter facialeffects. XPS and SINS (secondary ion massspectroscopy) are among techniques that havebeen used to study surface segregation inalloy systems (Brundle and Mandelt, 1981).The variation of this ~ segregation wi thr eac tive atmospheres i s a par ticul arlyinteresting featur e. Theory should beextended to cover this.

cwork has recently been re po r tedindicating the possible importance of metallicglasses, made by splat cooling, as catalyticsystems (Smith et al, 1.981). Pd/Si glasses,made by splat cooling, ar e activehydrogenation catalyst syst ems. During thehydrogenation with deuterium ofcis-cyclododecene these syst ems produce lesstrans insomer ization, more dideuto-saturateand less extensive exchange than crystallinePd. Such study of amorphous catalysts must beregarded as being at an early exploratorystage, but the potential scientific andpractical interest warrants encouragement offurther work of this type.

In closing this subsection on surfacescience related to transition metalheterogeneous catalysis, we Aish to emphasizethe value of self-consistent sur faceelectron i c structure calculations, incon junction with experiment, for understandingsuch important aspects of chemisorption aslocations, binding energies, and postures ofad sorbed spec i e s . Such calculations,employing power ful iterative computationaltechniques, allow one to obtain a realisticpicture of the complex charge distributionpertinent to the chemisorption involved incatalysis. These calculations are alsopertinent to understanding the key question ofwhe ther chemi sor pt ion o f a- molecul e ofinterest (e.g. CQ or H&) on a specific surfaceof a particular material will be dissociative,yielding atoms in a state suitable for

Rev. Mod. Phys. , Vol. 53, No. 4, Part Il, October 1981


reactions yielding the desired hydrocarbons.Conversely, these calculat ions ean giveguidance to under standing detrimental"poisoning" effects, e.g. , the binding energyo f an atom such as sul fur may allow it topreferentially occupy a site that otherwisewould be the preferred site for chemisorptionyield ing desired intermediate species forcatalytic synthesi s. Incorporation into suchcalculations of methods for simulating theef fects of small amounts of additives(promoter s ) that enhance catalyticeffectiveness, would be highly desirable.

S. Catalyzed gasification of carbons and coals

It has been known for many years thatsmall amount s of a wide variety of inorganicimpur ities can e f feet ively catalyze thereactions of carbon with gases such as oxygen,carbon dioxide, steam, and hydrogen (NcKee,1981) . Extension of such studies to coalitself is impor tant. The most practicalcatalysts for catalytic direct methanation ofcoal are salt s, e.g. , carbonates, of alkalimetals and alkaline earths. Understanding oftemperature ef fects i s impor tan t. Thecatalytic gasification goes up strongly onapproaching the melting of the salt, and onewould desire an ef fective catalyst at lowertemperatureS. (Nethanation is exothermic andhence thermodynamically most efficient at lowtemperature. ) Spectroscopic experimentsshould be developed to study the chemicalstate of the catalytic species. Onepossibility i s Na nuclear quadrupo leresonance, but practical problems arising fromdistributions in the field gradients wouldmake detection and analysis dif ficult;moreover, the in t er pret ation, g i ven theabsence of knowledge of the coordinatingatoms, would be hazardous. NMR and EXAFSmight be better probes for this purpose.Measurements pertinent to migration of thecatalyst particles through the pore structureof coal also are important.

A striking feature, apparently central tothe mechanism of catalytic gasification inmodel experiments on graphite is the cuttingof channels by catalytic microparticles (e.g. ,see NcKee, 1981). Baker (1979) has developeda . controlled-atmosphere transmission electronmicroscope (CAEN) technique enabling him tofilm the channel cutting process as it occurs.

This technique promises to be particularlyvaluable in identifying the cr ucial steps inthermal activation of catalytic gasification.

6. Denitrogenation and desulfurization catalysts

7. Recommendations

3 0

It is important to understand thedynamics of the di f fusion processt hr ough the zeoli te channelstructure. —Nideline and transientNMR, including pulsed magnetic fieldgradient stud ies, are pertinent.Isotopic experiments with deuteratedmolecular species could pr ovideadditional information.Me s tr on g ly r ecommend making andstudying systems having very smalltransition metal par tic les orclusters embedded in zeolites. Bothexperimental and theoretical studieson such systems should provevaluable.

Work should be done to systematicallystudy the role of vacancies in thecatalytic activity o f pyr i t icmater ial s. In such studies, shockwave acti vation or ion bombardmentmay be useful for creating vacanciesin a controlled way.

It is important to develop analyticaltools o ther t han Ho s sb auerspectroscopy, to investigate changes

De sul fur ization and deni trogeni zationreactions, which require appropr iatecatalytis, are an important step in theproduction of liquid fuels from coal (Grange,1981) . The catalyst, s that appear mosteffective for t hese r eactions aremolybdenum/cobalt, /nickel oxides in slurries.Nader n physics techniques might play asignificant role in the development of highlyselective catalysts, by providing informationabout the sur face/solid state propertiesresponsible for the catalytic effectivenss ofthese materials. Such properties includeoxidation states of the transition metals,their distribution, crystal fields in thesupport material, and mechanical integrityafter long term use in a hostile environment.

Rev. Mod. Phys. , Vol. 53, No. 4, Part ll, October 198't


in mineral matter, such as t;hepyrite to pyrrhotite transformation,under actual operating conditions.Acoustic pr obe s should beconsidered.

5. An important problem in catalysi s i sto understand the way in which"poisons" act, . For t his purpose,surface science studies, on "model"systems (single crystals,epitaxially grown films) should beuse ful. Ex ami nation o f sur facemor phological chan ge s, and o fdetailed electronic structure usingangular 1 y r e solved ul tr a-violetphotoemission spectroscopy (ARUPS)and associated sel f-consistentcalculations of the electronics tr uc tur e, should be valuable inthis regard.

6. Work should b e done on f ur therdevelopment of technique for "ins i tu" ex aminat;ion of "model"systems, involving both high vacuumcharacter i zation o f the sampl e andthe car rying out of reactions ofinterest, without removing thesampl e f r om the c ont r ol ledatmosphere enclosure. The po ssib leadvantages (and disadvantages) ofelectromagnet i c r e flectancespectroscopy, possible at hi gherpressures, should be considered.There i s interesting and usefulresearch to be done to gain furtherunderstanding of bimetallic catalystsystems, i nc iud ing wo rk onepitaxially gro~n "model" filmsystems. Such "model" bimetallicfilm systems of fer much opportunityfor productive t, heoret ical workinvolving self-consistent sur faceelectronic struc tur e calculations.Effects of particular int;crest fordispersed supported small particletransition metal alloy syst ems aremetal-support interactions and thephenomenon of sur face segregation ofcomponents.

8. References

Abraham, F. F. , 1981, Proceedings of t he27th National Symposium of the AmericanVacuum Society, 1990, Journal of Vacuum

Science and Technology, March-April 1981.Appelbaum, J. A. and D. R. Hamann, 1979,

Solid State Commun . 27, 881.Arlinghaus, F. J. , J. G. Gay, and J. R.

Smith, 1980a, in Theory of Chemisorption,edited by J. R. Smith, Springer-Verlag,New York.

Arlinghaus, F. J. , J. G. Gay, and J. R.Smi t h, 1980b, Phys . Rev. B 21, 2055.

Baker, R. T. K. , 1979, Catal. Rev. — Sci.Eng. 19, 161.

Biberian, J. P. , and G. A. Somor jai, 1979,J. Vac. Sci. Technol. 16, 1073.

Brundle, C. R. , and K. Mandelt, 1981,Proceed ings o f the 27th NationalSymposium of the American Vacuum Society,1990, Journal of Vacuum Science andTechnology, March-April 1981.

Chang, C. D. , M. H. Lang, and A. J.Silvestri, 1979, J. Catal. 56, 269.

Cusamano, J. A. , D. L. King, and R. L.Gar ten, 1981, to appear in CatalysisReviews —Science and Engineering, Vol.21.

Davison, L. , and R. A. Graham, 1979, PhysicsReports 55, 255.

Derouane, E. G. , and J. C. Vedrine, 1990a,J. Molec . Cat al . 8, 479.

Derouane, E. G. , and Z. Gabelica, 1980b, J.Catal . 65, 486-

Dubinin, M. M. , V. A. Gorlov, and A. M.Voloshchuk, 1980, in Proceedings of theFi f th j:nt ernational Conference onZeolites, edited by L. V. Rees, Heydenand Son, London .

Eberly, P. E. , Jr. , 1976, in Zeoli teChemistry and Catalysis, edited by J. A.Rabo, ACS Monogra phy 171, AmericanChemical Society, Ma shington, D. C. , pp.392-436.

Feibelman, P. J. and D. R. Hamann, 1979,Solid State- Commun. 31, 413.

Gallezot, P. , 1979, Catal. Rev. Sci. Eng.20, 121.

Goddar d, M. A. II&, 1981, Proceedings of the27th National Syrnpo si um of the AmericanVacuum Societ;y, 1980, Journal of VacuumSc i ence and Technology, March-April,1991.

Goodman, D. M. , R. D. Ke1 1ey, T. E. Madey,and J. T. Yates, Jr. , 1980, J. o fCatalysi s 63, 226.

Gr aham, R. A. , B. Morosin, P. M. Richards,F. V. Stohl, and B. Granoff, 1981, inChemistry and Physics of Coal Utilization

1980, edited by B. R. Cooper and L.



Petrakis, AIP Conference Proceedings No.70, AI P, New York.

Grange, P. , 1981, Catal. Rev. 21, 135.Granof f, B. , and P. A. Montano, 1981, in

Chemistry and Physics of Coal Uti1 i zation1980, edited by B. R. Cooper and L.

Petrakis, AIP Conference Proceedings No.70, A IP, New Yor k.

Huang, T. J. and M. 0 Haag, 1980, presentedat The Second Chemical Congress of theNor th Amer ican Cont inen t, La s Ve g a s,August, 1 980. (Cited in Chemical andEngineering News, Sept. 8, 1980, p. 44).

Jepsen, 0. , J. Madsen, and 0. K. Andersen,1978, Phys. Rev. B. 18, 605.

Karger, J. , and J. Caro, 1977, J. C. S.Faraday, Trans. I, 73 1363.

Karger, J. , M. Bulow, P. Struve, M. Kocirik,and Z. Zikanova, 1978, J. C. S. FaradayTrans. I 74, 1210.

Karger, J. , and P. Volkmer, 1980, J. C. S.Faraday Trans. I, 76, 1562.

Kasowski, R. V. , and E. Caruthers, 1980,Phys. Rev. B 21, 3200.

Kaufman, , J. J. , H. E. Popkie, and P. C.Hariharan, 1979, pp. 415-435 in ComputerAssisted Drug Design, edited by E. C.Olson and R. E. Chr istof fen, ACSSymposium Series 112, Amer. Chem. Soc. ,washington, D. C.

Kerr, G. T. , 1981, to appear in CatalysisReviews — Science and Engineering, Vol.23.

Kleinman, L. and K. Mednick, 1981, Phys.Rev. B. 23, 4960.

Krakauer, H. and B. R. Cooper, 1977, Phys.Rev. B 16, 605.

Loui e, S. G. , K. M. Ho, J. R. Chel ikowski,and M. L. Cohen, 1976, Phys. Rev. Lett.37, 1289.

Ma, C. Q. , H. Krakauer, B. R. Cooper, and I.Kulikov, 1981, Proceedings of the 27thNational Symposium of the American VacuumSociety, 1980, Journal of Vacuum Scienceand Technology, March-April, 1981.

McKee, D. M. , 1981, in Chemistry and Physicsof Coal Utilization — 1980, edited by B.R. Cooper and L. Petrakis, AIP ConferenceProceedings No. 70, AIP, New York.

Minachev, Kh. M. and Ya. I. Isakov, 1 976, i nZeol i t e Chemi str y and Catalysis, editedby J. A. Rabo, American Chemical Society,Mashington, D. C. (ACS Monography 171).

Montano, P. A. , 1980, to be published inAdvances in Chemistry, "Recent ChemicalApplications of Mossbauer Spectroscopy, "

edited by J. Stevens and G. K. Shenoy.Morrison, T. I. , A. H. Reis, Jr. , E.

Gerbert, L. E. Iton, G. D. Stucky, and S.L. Suib, 1980a, J. Chem. Phys. 72, 6276.

Morrison, T. I. , L. E. Iton, G. K. Shenoy,G. D. Stucky, S. L. Suib, and A. H. Reis,Jr. , 1980b, J. Chem. Phys. 73, 4705.

Mraw, S. C. , and B. G. Silbernagel, 1981, inChemistry and Physics of Coal Utilization

1980, edited by B. R. Cooper and L.Petr aki s, AIP Conference Proceedings No.70, AIP, New York.

Naccache, C. , N. Kaufheer, M. Dafaux, J.Bandiera, and B. Imel i k, 1977, inMolecular Sieves-I I, ed i ted by J. R.Kat zer, American Chemical SocietySymposi um Ser ies, pp . .538-548.

Pfeifer, H. , 1972, in NMR-Basic Principlesand Progress, Vol. 7, edit ed by P. Diehl,E. Fluck, and R. Ko s f e 1 d,Springer-Verlag, Berlin.

Posternak, M. , H. Krakauer, A. J. Freeman,and D. D. Koelling, 1980, Phys. Rev. 821, 5601 .

Rabo, J. A. , 1981, to appear in CatalysisRe v i ews — Sc i en ce and Engineer ing, Vol .23-

Rao, V. U. S. , and R. J. Gormley, 1980,Hydrocarbon Processing, November, 1980.

Renouprez, A. , P. 'Foui 1 loux, and B.Moraweck, 1980, in Growth and Propertiesof Metal Clusters, edited by J. Bourdon,Elsevier, Amsterdam.

Richter, L. , S. Bader, and M. Brodsky, 1981,Proceed ing s o f the 27th Nat ionalSymposium - of the American Vacuum Society,1980, Journal of Vacuum Science andTechnology, March-Apr il, 1981.

Schmidt, L. D. , 1981, Proceedings of the27th National Symposium of the AmericanVacuum Society, 1980, Journal of VacuumScience and Technology, March-Apr i1 1981.

Sinfelt, J. H. , 1969, Rev. Mod. Phys. 51,569.

Sinfelt, J. H. , 1977, Science 195, 641.Smi th, G. V. , N . E. Br ower, M. S.

Matyjaszczyk, and T. L. Pet tit, 1981, toappear in the Proceedings of the 7thInt ernationa 1 Con g r e s s on Catalysi s,July, 1 980.

Smith, J. V. , 1976, in Zeolite Chemistry andCatalysis, edited by J. A. Rabo, ACSMonograph 171, published by the AmericanChemical Society, ttwtashington, O. C. , pp.1-79-

Somor jai, G. A. , 1979, Sur face Science 89,



496.Somor jai, G. A. , and M. A. Van Hove, 1979,

Adsorbed Honolayers on Solid Sur faces,Volume 38 of the Structur e and BondingSeries, Springer-Verlag.

Van Hove, N. A. , l981, Proceedings of the27th National Symposium of the AmericanVacuum Society, 1980, Journal of Vacuum

Sc i enc e and Technology, March-Apr il,1981.

Wang, C. S. and A. J. Fr eeman, 1979, Phys.~ev- B- l9, 793-

Meber, R. S. , N. Boudart, and P. Gallezot,1980, in Growth and Properties of MetalClusters, edited by J. Bourdon, Elsevier,Amsterdam.

Rev. Mod. Phys. , Vol. 53, No. 4, Part II, Octaber 1981

CO2 Emissions S147

APPENDIX I. CO2 EMISSIONS: COIVIPARISONSOF THE USE OF COAL FOR ENERGY PRODUCTIONBY VARIOUS ROUTES

Introduction

In thi s append ix, we compar e var ious coalutil i zat ion methods as sources of carbondioxide. We delineate, where possible, howphysics research could elucidate or af feetthi s aspect of the state-of-the-art of coalutilization.

In the analysis which follows, it hasbeen necessary to use reports and otherpubl i shed data that refer .to speci ficcommercial (or proposed commercial) processes.However, no comparison is intended inter sein the first instance the relative degree ofcommercialization of these proposed processesare not equivalent; therefore the flow sheetswill be subject to change. It is further tobe emphasized that the plant efficiency (andhence its inherent CO production) is subjectto each specific flow sheet, modification, evenwithin a given-named pr oce ss significantdifferences can exist. Because of thissensitivity to engineer ing detail, differencesin carbon utilization derived in this appendixindicate only general trends. We indicate anapproach to this subject, and extract generalconclusions relating to efficiency rather thanconclusions specific to particular utilizationschemes.

A. second f und amen ta 1 pr ob lem wh i choccur s i s the fact that the fuel value of theproducts of coal use ar e not equal by severalmeasures, thermodynamic but also non-technical—e.g. , the societal values of electricityvs. natural gas or synthetic gasoline. Inthis appendix, we have disregarded what manyhave rightly considered important in energyanalysis.

We r estr ict our di'scussion toquantitative assessments of the CO outputs ofthe technologies considered. He do not broachthe sub ject of possible ef fects of CO& fromany sources. This latter problem has receivedmuch recent attention (e.g. , NacDonald, et al,1979, 1980; Charney, et al, 1979; AAAS 1980;Council on Environmental Quality, 1981; andAbarbanel et al, 1980) . Our discussion isrelevant to an assessment of the trade-of fsand penalties associated wi th various fuelstr ategies which could be adopted; and shouldbe viewed as providing a method ' for

quant itative analysis of this one limitedaspect, with the pr imary ob jective ofdelineating the physical basi s, and researchavenues, pertinent to CO& release.

To these ends, we first consider thegeneration of electricity using various fossilfuels, and define an index appropriate to acomparison of these fuels as sour ces of CO

including the ef fects of such constraints asscrubbers, etc. Me proceed to an analysis ofCO emission from synthetic fuel production.Thx. s leads naturally to definition of a carbonefficiency index, and to a comparison of CO&release associated with various ways of usingthe same primary resource fuel (i.e. coal).

The analysis leads to some conclusionsabout relevant research directions.

First, the incorporation of sourcecarbon into a synthetic fuel product(i.e. , the desirable complement, ofCO effluent) increases directly asthe thermal efficiency of a processfor fuel synthesi s increases. Thusresearch directed toward improvingthe thermal ef ficiency of syntheticfuel processes will have theconcomitant e ffeet of reducing theamoun t o f CO r el ea sed dur ingproduction of (he synthetic fuel .Second, i t i s possible tosubstantially reduce CO emi ssionsfrom synthetic f uel plants bypr ovid ing a non-fossil source ofprocess energy. This reduction canamount to great, er than 75% of the C02released during convent ionalprocessing. Thus, insofar as CO canbe reduced either by improvements inthermal efficiency or by use ofnon-fossil process energy, interestedmembers of the physics community havea number of alternative s a ndpossi bili ties for researchopportunities in this area.

- We also corroborate previous estimatesused in the CO&-ef fects studies, thatconversion of coal to a synthetic fuel andutilizing that fue1, gives a net, CO& productionof about twice that for the same energy outputusing a comparable natural fossil fuel.

CO2 evolution from direct conversion of fossil fuel toelectricity

It is convenient to define a criterion tocompare ef ficienci es o f d i fferent electric



CW

q hHC

kgC (released as CO )kg -b,r of electricity producede

where q is the first; law efficiency(electricity produced/combustion heat in coal),and AH is the enthalpy of combustion per unitmass o coal.

The utility of this index is that itcompares the evolut, ion of CO with the usefulproduct, , i.e. units of elec rical energy.There are very reliable estimates of powerplant efficiencies, q , available in the Energye'Conversion Alternat. ives Study (Corman and Fox,1976). A selection of cycles of interestinclude conventional steam plant, conventionalsteam plant wi th sc r ubber s, atmospher icfluidized bed (AFB), pressurized fluidized bed(PFB), integrated gasification combined cycle(IGCC), and open cycle magneto-hydrodynamicgenerator (MHD).

po~er plants on a carbon-evolved basis. A

use ful generalization can be made of acriterion used by MacDonald et al, (1980),i.e. , carbon dioxide evolved in ratio todelivered energy.

If w is the weight, $ of carbon in thecoal then a suitable index of performance is-defined by.

Figure Cl is a composite figure showingthe index I as a function of power plantefficiency p for a variety of fossil fuels.Anthracite 1s not usually considered an6

appropr iate fuel for a power plant, system andis sho~n only for comparative purposes. In anycase, it is the least promi sing fuel on a.carbon-evolved basis. awhile lignite is aslightly better fuel on this basis, its lowcarbon content does not, in fact, - compensatefor its low heating value. Furthermore, thenet heating with lignite is diluted by its highmoisture content.

There are two deleterious ef fects whichhave been accounted for in the calculationssummarized in Figure Cl: a) The need t;o supplythe latent heat of evaporation of the containedwater. b) The reduction in boiler efficiencydue to the increase in exhaust stack partialpressure of water vapor. Effect (a) isaccounted for by reducing the heating value ofthe fuel accordingly; effect (b), by reducingthe boiler efficiency progressively from 88$ to81$ on going from bit;uminous coal to naturalgas (Brown, 1980), since stoichiometricconsiderations demand an increase in theexhaust eater vapor content.

Figure Cl shows that hi gh volat i 1 eb i tuminous coal, sub-bituminous coal orresidual oil behave very similarly on a (carbon

0.5

0.4

CB

+ 0.2

0.1—

020 30

I I

40 50

FIG. Cl. Relative carbon release rates for several power plants using selected fossil fuels.I

Rev. Mod. Phys. , VoI. 53, No. 4, Part II, October 1981

CQ2 Emissions

0.5

0.4

g 0.3

E3CA

0 2

as TurbineC}Open Cycle PHD

I 1085 C}

020

l

30I

40pe%

50 60

FIG. C2. Relative carbon release rates for advanced bituminous coal power plants.

evolved)/(output power) basis. Shown on thiscurve are the state-of-the-art (SO%) dat, apoints for an oil-fired generating plant and ahigh-sulfur coal powered generating plant with302 scrubbers. "The lowest curve is that for a natural gasplant, where its SOA ef ficiency point isplotted. There is a decrease of net CO2evolved by a factor of approximately 2 betweenthe fuels giving the extremes. For example, at35'$ power plant e f f iciency, anthracite yields0. 285 kg C/kM -hr, l i g n i t e 0. 28, b i t,uminous andesub-bituminous coal 0, 25 and nat ur al gas 0.1 0kgC/kM -hr. Reality does, ho~ever, reduceethe actual power pl ant, ef ficiency as, thehydrogen/water content incr eases for thereasons previously noted, so that an assumedf ixed power plant ef ficiency i s asimpli fication.

Figure C2 summar i zes expected values ofthe power plant efficiency for those plantsconsidered by Corman and Fox, (1976), andt, hose for which updated data are available(Pomeroy et al. , 1980; Shah et al. , 1980) .These data all refer to the same fuelbituminous (or sub-bituminous) coal as wasused in the calculations of Figure Cl.

CQ2 Evolution from direct conversion of fossile fuel toelectricity

It is significantly more difficult todevelop simple indices o f per formance forsynthetic fuel plants than for elect, r icalgenerating plants. The reason is that theformer produce a plethora of products ~hosethermodynamic worth is obscure. For example,methane i s a by-pr oduct of many lique factionschemes, is its fuel value to be included inthe products, or is it a waste material?Furthermore, what is its "fuel value" ? Thereare several candidates. The higher heating

SO scrubbers "consume" coal by i) the needto calcine limestone or dolomite with coal heat-- e.g. , CaCO = CaO + CO, where the 3 ime(Ca0) is used 2s the sulfur a%sorbant, and ii)the requirement of internally consumedelectricity for the gas handling equipment inthe scrubber . A net CO2 evolution o f 2-6 molesmay be produced per mole SO scrubbed dependingon the sulfur content of the original coal.

Rev. Mod. Phys. , Vol. 53, No. 4, Part II, October 198'I


values (HHV~) and lower heat, ing values (LHV"")can be used to define First Law efficiencies.

HHV HHV of productsHHV of input coal

(%)

LHV of productsLHV of input coal

Such definitions can be very misleadingi f there is an input of auxiliary fuel or anexport of work from the process. A furtheref ficiency which is needed is a carbonefficiency here defined as

C C in productsC in input coal

Including latent heat of condensation.~~Excluding latent heat of condensation.

in which the products may be furthersubdivided into di f ferent useful classes,e g. , gaseous and liquid, to account for t heco-generacy in many processes.

There are several classes of conversionprocesses to consider . These are listed withsome specific examples of each in Table Cl.4(i t h each acronym the proprietary name isgiven for the developer of each process.These are but a few of those that can be foundin the literature. Combinations are possibleas shown in Table C2. In these data(Harrison, 1980; Shih, 1981), the productgYqlds (on a lower heating value basis, i.e,

expressed in %) are given for a range ofproducts in order of increasing molecularweight, viscosity, density, etc. For example,a gener ic Fi scher -Tr opsch process usingBritish Gas Corporation Slagger gasi fier swould have 14$ LPG and 17$ gasoline (by energycontent). Note SRC-I primarily makes what itsname implies, solvent, refined coal, 49. 5$ to62$ of its input coal energy. The more recentSRC-II has a lighter spectrum of products withits emphasis on fuel oil. Likewise, two modesare quoted for H-Coal with different productspectra. The mechanism for these di fferencesi s in the recycle of hydrogen fromcarbonaceous residue or coal. Furtherdifferences in each process are due to

slightly di f ferent flow sheets, emphasizingdifferent products.

It is important to r eal i ze the greatsignificance of gasification of coal (toproduce H ) even in liquefaction of coal. Thequantity of H recycled (or equivalently ahydrogenated donor solvent, ), the makeup H

and the loss of H2 by heteroatom capture hePpdetermine the pr oduct d i str ibution. Hhi leundigested-coal residue can be burned, it isprobably optimum to gasi fy it as a fur theroption, but the use of this residue obscuresthe specification of conversion efficiency andproduct spectra. Having noted all theseambiguities, the bottom line message of TaffyC2 is that the conversion efficiency on a 0basis ranges from 34$ to 74$ for a range ofprocesses.

Is there any r ational basis for astandard against which these proces se s(ranging from laboratory to commercial) can becompared? In the absence of a completethermodynamic treatment, Mei (1980) has shownhow simple considerations based on C-H-Ostoichiometry can lead to limits on conversionefficiency. He has considered the theoreticalliquefaction of graphite to 1-octene by fourdi fferent pr oce sse s. 'Ai th the r e str ictionsimposed by the stoichiometry conditions, allfour processes lead to essentially the samemaximum utilization of carbon in the liquidproduct, viz 50-60$, the remainder beingvented as CO2. Of course, coal is notgraphite, real processes are not ideal, and1-octene is not the sole product in actualliquefaction. Nevertheless, the utilizationof carbon ig the graphite to 1-octeneconversion provides a simple bound which canbe used as a zeroth order yardstick on carbonefficiency.

Substantial data do exist in the reportliterature (e.g. McNamee et al. 1978;Schreiner et al. 1978; and Shinnar et al.1978); these particular references areselected for their per tinence to the presentstudy. In particular, for example, McNamee etal (1978) pr ovide process description oftheoretical 50, 000 barrel/day coal refineriesproducing a variety of products. Two of theseprocesses ar e labeled as "catal yt i chydroliquefaction" and "solvent refi, ned coal;"these processes are herein interpreted as ageneric H-coal-type and a generic SBC process.The flow sheet s in each o f the quotedreferences, and indeed for all chemicalprocess plants, are rather specific. Theseflow sheets reflect, in general, a plant


CO2 Emissions

cQ

ocQ

CP,,

)I

gPE'CP

C- '

,40 '

CJ,

I

CP

cG

0

UCPL

~Oalg

oOCP

SO

e ooL

CP

i'Cl

ttd ~O~UCP ow

~ C4CP

.o~e5 U

COgc~~cQ

'0g$ w~ e$& o

tp~~x cd ~~gp~ ~M~a

g$ fQ ~ 4A

+O AJ

fO ep a-e

Q i

p ~-~X~CJ

Li) cR W L

CP,oOCCg

CPOt ~~~+ CACO

CO 6$CO

fQ UQ.~ x

LgQ LQ

II II II II II II II II

UU

LP'

CP

~~A

l,'

4h fQ ~ U UI gg l3 Cn I- ~ UX m LQ cn W D. X cQ LQ



e ~ I~ ~~ i

KC0U

I

0

OQI ~ I

I I

I

CV I l

O0

I ~ I I

~ +UQ

00Q

zC)WO0~

@ ~ OO

Q ~~~icG+ x0COU

QC

Q

Q(p 0

~Q~

Q,C

~ ~0cf)09

C)

O

I ~ I ~ I

™

I ~ ~ M I

QP

e~

aj

o

OQ

U0

JDAj

C~ ~Q

~ ~QU

E

QV


CO Emissions S153

optimized in some way sub ject to inherentand/or applied constraints such as economicswould impose. The pertinent conclusion isthat under a different range of constraints arather di f ferent flow sheet could emerge, andas one possible consequence, the thermal andcarbon balances of the process would bechanged. The reader is, therefore, cautionednot to draw canonical conclusions with respectto .any of the named processes -- these aredemonstratively subject, at least, to revisionfor engineering improvements as they nearimplementation.

These 1 i que faction processes producenaphtha, fuel oil and residual fuel ascondensed phase products, plus a fuel gas,mainly hydrogen with hydrocarbon impurities C

or less. Of these products, naphtha isgasoline like, fuel oil can be used forpurposes from running diesels to space heating(depending on further treatment), and residualfuel (or residuum) boils above 400 C.Residual fuel is used exclusi vely in heavyutility operations and, except where a cleancoal based substitute is required, i s theleast desirable of the products. In thegeneric SRC process, residual fuel amounts to86. 5 wt. g of the condensed phase products;1 ikewi se, in the generic catalytichydroliquefaction process, it amounts to 67. 2wt. $ of the condensed phase product . For thi sreason, these generic processes show a highthermal e f f iciency and a high carbone f f iciency. The e f f i ciency values arecertainly too high for a working H-coal orSRC-II plant producing "light" products.

The e f f iciency data for the genericcatalytic hydrolique fact ion and SolventRefined Coal processes are sho~n in Figure C3as two sets of adjacent bars for concomitantthermal (HHU basis) and carbon efficiencies.Similar data exist for the principal indirectroutes —— i.e. , Fi scher-Tropsch andmethanol-to-gasoline. The former, an oldertechnology, is based upon the gasification ofcoal to synthesis gas, i.e. , CO/H mixturesand their r eact ions ori ruthenx. um/ironsupported catalysts. The action of thetraditional Fischer-Tropsch catalysts is toshunt hydrocarbon building blocks on toexisting units and to allow two basic reactionpaths, one to desorb a stable specie s and t hesecond to attach an additional building blockfor subsequent reaction. In consequence,there is a geometr ic decrease in theprobability of finding higher hydrocarbons inthe resultant product; methane is the primary

product which has to be considered either as aco-generated product, recycled to a steamreformer to produce additional syngas, or as a"waste" material t o be used elsewhere in theprocess as a heating source. The latter ispar t.i cul ar ly unde si r able since Fi scher -Tr opschsynthesis is exothermic, producing an excessheat source. In fact, Hoogendoorn (1977)indicates that SASOL II (the Fischer-Tropsehtype process in commercial use in SouthAfrica) would have a thermal efficiency of35-38$ i f based on recycled methane, and of60$ if methane were sold as a gaseous product.

Schreiner et al. (1978) give a comparisonof a Fischer-Tropsch process and amethanol-to-gasoline process. Considering theformer for the moment, one may follow itsvarious recycles and processes to determinethe fate of the original coal carbon step bystep. From these data, it is possible toconstruct bar graphs as shown in Figure C3for the carbon and thermal efficiencies (HHU)basis for that process; again, separate barsare included for the liquid and gaseousportion of the product, spectrum (here thegaseous spectrum is considered to be C& orlower). A similar analysis ean be carried outfor the coal-met hano 1 -ga sol inc route wi thzeolite catalysis, and is also shown in FigureC3.

An evaluation can also be made ofgasi f ication processes. Pflasterer et al.(1980) discuss a high BTU plant using a"standard" Illinois N6 bituminous coal, insuf ficient detail to produce the "High BTUgasification" entry in Fig. C3.

The dat a shown in Figure C3 are takenfrom the several quoted sources. The datashown in Fi gur e C4 were provided by Shih(1981) from additional sources. Inparticular, the bar graphs for H-Coal and SRCwere derived from Fluor (1976) and SRI (1979) ~

The efficiencies for these processes are.considerably lower than those quoted in FigureC3. The reason is that the spectrum ofproducts (see Table C2) contains more "light"pr oduc t s than tha t quo ted by NcNarnee et al,1978, for the corresponding generic SRC andcatalytic hydroliquefaction processes studiedby them.

Me have previously asserted that the CO

effluent from a synthetic fuel plant is flowsheet dependent; we have shown that the carbonefficiency and the thermal efficiency of aprocess are directly related. In Table C3 wequote thermal ef ficiencies from a variety ofsources for the same titular process.


APS Study Group on Coal Utilization and Synthetic Fuel ProductionS154

100 ~

Thermal

Thermal

CarbonCarbon

Thermal

Thermal +Thermal60—

CO ~

Carbon g'

massa

& CarbonCO ~

Carbon

40

CS

CO CO

W 'U

CF CF

CO CO

aCF CF

CO

U;20—

CO CO

CF CF

CO CO

~CF CF

~ ~

QQO

qOpie~. g,G

QO+

~e&'"

100

80—

GiiLiiiii

iii

iiLii

LLL

!L:iiiLLLiLLL

iiL

iiiiiH-CoalFuel Oil

LLLLLL

iiiLLLLLL

LLLL

LLL

iiLLLL

LLLL

LLLLLLLLLL

~OO

:L:.

SRC-Il

LLLLLiLLLL

iiLLLiii~LL

H-CoalSyncrude

0LLLL

iiiLiiLiiLL

iLSASOL-II

:L:

MTGGasoline

iiLLLiL

LLiiL

MTGSNG+

Gasoline

iiii.'L

iii

FT

pL

HighBTU

Gasification

SRC-I

FIG. C4. Thermal efficiencies of some coal conversion processes (after Shih, I981).


jgG ~QS

~gC

FIG. C3. Represeotative thermal and carbon efficiencies of several synthetic fuel routes.

CQ2 Emissions S155

Table C3:

Process Thermal Efficiency Source

Quoted q for several named processes.HHV

SRC- ISRC-IISRG- ISRG-II

Generic SRC

H-Coal

H-Coal( )

H-Coal( )

H-Coal( '

Generic CatalyticLiquefaction

71.565.8

74 (LHV)

69 (LHV)

59.2

60.0

76

Harrison, 1980

Harrison, 1980

SRI, 1979

SRI, 1979

NcNamee et al. , 1978

Harrison, 1980

Harrison, 1980

Fluor, 1976

Fluor, 1976

NcNamee et al. , 1978

Fuel Oil NodeSyn Crude Node

Significant differences emerge among them, thebar chart, Figure C3, must be viewed in thelight of these differences.

The sources of CO2 from these processesprimarily arise from the need for plant power,for steam to run the various auxiliaries andchemical purification steps required in achemical plant, and for production of H . Allcof these components of n are signi meant.The hydrogen production step causes CO

emission for two reasons: l) CO shift, tohydrogen; 2) Excess steam to effect thisshift. That is to say, hydrogen is made fromsyngas by shifting according to CO + H 0 ~ CO

+ H2 in which 1 mole of C02 is released permole- of H2 formed; but, in addition, excesssteam is needed to drive the reaction to theright and this involves additiona1 combustionof fuel.

It should now be evident that the thermaland carbon ef ficiencies of a large number ofprocesses are directly related. There i sroughly a linear relationship between theseef f iciencies. In so far a s improvements in

thermodynamic ef ficiency ar e d i r ec tlyrelat;able to product; costs, studies directedtoward thermodynamic/economic goals will havethe benef icial by-product - of decreasingexpected CO2 emissions from synthetic fuelplant s, a point no t speci fic'ally notedpreviously

Estimate of total CO2 releases

For the pur pose of eempar ing CO emi ssionfrom synthetic fuel production to that fromcoal powered electricity plants, we use thyonetime DOE announced goal of 2. 5 x 10—barrels/day of synthetic fuel. Fr om thepreceding discussion in this Appendix, we seethat a typical value of the carbon efficiencyof rP: 50$ is reasonable. We then find thatthe rate of CO2 emission for this rate ofsynthetic fuel production at this carbonefficiency is the same as that from thirty1 000 me gawa t; t coal-t o-electr icity plant shaving a thermal e ff i c i ency o f 35'$ and being



run at an energy output of 90$ of their annualcapability. The total ef fluent for bothexamples is 60 megatonnes of carbon or 220megatonnes of CO . For purpose of comparison,this amount of C

2 is about 0. 008$ of the 1975atmospheric C02, and equals 43$ of the COf rom flare gas burned of f worldwide a trefineries and oil fields in 1975. Mhile thisnumber of power plants produces the same CO

as the quoted amount of synthetic fuel oiproduction with a carbon efficiency of 50$,end-utilization of the synthetic fuel willcause an additional equal emission of CO

It i s also possib 1 e to compare ourestimates of CO evolution for di f ferentfuels, natural and synthetic, with those ofMacDonald et al. (1980) . These estimates, asshown in Table C4, refer to the total carbonr el eased as C02 in both processing andutilization.

The agreement between the two sets ofestimates is good for the synthetic liquidsprocesses (H —Coal and CatalyticHydrolique faction) and for coal toelectrici ty. There i s a sign i f icantdi f ference between the MacDonald et al.estimates for HYGAS and our s for genericsynthetic natural gas (SNG) production. Thisis due in the main to the (high and assumed)50'$ carbon efficiency in the case of high BTU

gasification (i.e. SNG entry for this reportin Table C4) compared to the actual & for thecprototype HYGAS process. The higher q

e

corresponds to an increase in thethermodynamic ef f iciency that may beachievable by the introduction of one-stepdirect high BTU gasification processes such asa catalytic gasifier (Nahas and Gallagher,1978) rather than using an inherentlymulti-step methanation process as in the HYGAS

technology.The most important conclusion of this

section is the observation of the correlationbetween thermal and carbon efficiencies. Theexistence of this correlation should serve toreinforce the value o f a research programdirected toward the general improvement ofsynthetic f uel plant ef ficiency. Theinterested reader i s again referred topreviously quoted sour ces for g lobalimplicat ions of this C02 loading to theecosystem.

Use of non-fossil supplementary sources of process energ~for synthetic fuel production

awhile

it i s clear that it should bepossible to reduce CO production by use ofnon-fossil sources of process ener gy in

Table C4. Total (production R utilization) carbon release rates comparedto estimates of NacDonald et al, 1980

Converted FuelNacDonald et al.'

198Q

kgC/10 Btu

This Appendi~

H-Coal

Catalytic. Hydroliquefaction

39.2

39.1 (q =5Q%)

HYGAS( Coal - to -SNG )

42. 3

SNG-Generic Process

Coal to Electricity(Conventional Plant)

28.3 (q =50%)

73.2 (q =35%)

Gasoline

Biomass

14.2

19.218.3

Rev. lVlod. Phys. , Vol. 53, No. 4, Part I I, October 1981

CO2 Emissions 8157

CO

O

QCO

C0) tO

Q)

~ IL

0) CQ

~ c5C

C ~S CQ)

CCOC5~O—E+oIL IL

SocOOE0IL

0O

100

80—

60—

40—

20—

0l

20 60 80 100

Coal Feed to Syn Fuel Plant Using Alternative Energy SourcesCoal Feed to Conventional Syn Fuel Plant

FIG. CS. Reduction of CO2 emission in synthetic fuelproduction by use of a supplementary non-fossil energysource (basis: same carbon in product).

synthetic fuel production, the magnitude ofthi s e ffeet i s n o t we l l ~kn own . Conceptual 1ysuch alternative power supplies can be anynon —fossil based source. For example,electrical energy input could be a substitutefor a portion of the total process plantrequirements. It might also pay to have hightemperature process heat available forgasi f ication and hydrogen production. Thenumber of possible sources of process heatwould be r educed to those capable of supplyingheat at the pertinent conditions. Pflastereret al. (1979) give data which indicate therequirements for high temperature ( &700 0)heat and for low temperature (& $00 C) heat(steam and electricity) are approximatelyequal in coal gasification. Lesser amounts ofhigh temperature heat would be needed for coalliquefaction, and for supplying the hydrogenvia gasification.

Irrespective of the source of alternativeenergy, one can easily estimate the reductionin ef fluent CO& achievable by r educing thecoal feed for a specified quantity and qualityof product. Figure C5 shows the results ofsuch a calculation. The ordinate is the CO&ef fluent using a supplemental alternate(non-fossil ) energy source comp'ared to that

for a conventional plant with the same output.The reduction in CO i s achieved by reducingthe coal feed, here given as the abscissa interms of the ratio of coal usirig alternateenergy sources to that for the conventionalplant of the same output.

For instance, Kosky and Flock (1980) haveshown coal gasification efficiency can beincreased by substitution of high temperaturenuclear heat from a gas-cooled reactor in aconventional coal gasification process. Infact, the total coal consumpt ion can bereduced by 44'$, of which 20$ i s in thegasifier proper, and 24'$ in process steampr oduction. Thi s has the ef feet ofappreciably increasing the utilizable fossilcarbon. This point is also plotted on FigureC5. Lesser reduct ions in coal feed areexpectd for liquefaction plants (Pflasterer etal. , 1979).

Economics dictate that the implementationof such coupling o f technologi. s is mostlikely in countries in which fossil f uelprices are high, and conventional energy is inshort supply. For example, the FederalRepublic of Germany has an 'on-going aggressiveprogram to match high temperature gas cooledreactors to coal gasifiers (von der Decken,1975) . However, Dietrich et al. , (1975) havesuggested that CO& emission reduction would bea concomitant benefit to be derived, a pointvery recently reiterated by Haefele et al. ,(1980) in an IIASA study.

It is tempting to suggest that this is anarea, which may be of interest to the physicscommunity. It is problematic as to whetherthe U. S. with its abundant coal suppliesshould consider thi s option. Fur thermore,consideration of Figure C5 indicates that CO&can be reduced per unit of output in two ways—by reducing the coal feed to a fossil plantusing substitute non-fossil energy sourcesand/or by increasing the carbon efficiency ata fixed coal feed —.the latter achievable byincreasing the thermal efficiency of synthet;icfuel plants ~

Conclusions

CO release dur ing synthetic fuelproduction is almost as ubiquitious as it isfrom utilization of fossil fuels. The amountof CQ evolved in the synthesis process isinversely related to the thermal efficiency ofthe, conversion process. For each process one

Rev. Mod. Phys. , Vol. 53, No. 4, Part l l, October 198't


may quote a range of valid t hermalefficiencies based on product, distribution andon the particular flow sheet chosen by thesynthetic fuel plant designer. Conclusions asto the preferred synthetic fuel route based onCO evolution alone are probably premature atthzs time.

It appears that roughly as much CO willbe produced dur ing t he manufacture of asynthetic fuel as will be produced during theuse of that fuel. The amount of CO2 produceddur ing synthesis is 220 megatonnes/year for 21/2 million barrels/day. This is equilvalentto about 30 x 1000 MM coal-fired electricgener ators or A3$ of refinery and oil fieldgas flares est imated in 1975 on a worldwidebasis ~

CO evolution from synthetic fuel plantscan be reduced by either increasing thethermal ef ficiency of the process or by theuse of non-fossil energy supplements to theprocesses. However, even for an infeasible100% efficient synthetic fuel process, the CO

environmental impact que st ion st, i 11 revolvesaround the overall global use of fossil fuels,whether natural or synthetic, 'a determinationof the extent to which such use leads to anincrease in atmospheric CO eoneentration, andfurther determination of the environmentaleffects of such an increase. This vitalquestion therefore falls out, side the scope oft,his study.

References: Appendix I

A. A. A. S. , 1980, "workshop on Environmentaland Societal Consequences of a PossibleCO -Induced Cl imat, e Change, " Con f . ;Annapolis, MD, 4/2 — 4/6, 1 979, USDOE,

2

CONF: 704l 43, UC-ll .Abarbanel, H. , et al, 1980, "JASON Study

The Carbon Dioxide Problem. DOE Programand General Assessment" SRI Project 5793,Rept. JSR-80-06.

Brown, D. , 1980, private communication.Charney, J. G. et al, 1979, "Carbon Dioxide

and Climate. A Scienti fic Assessment, "National Academy Sc iences Rept, . t oClimate Research Board (NRC).

Corman, J. C. and G. R. Fox, 1976, "EnergyConver sion Alter n at, i v e s Study-G. E. PhaseII Final Report, , " 1, (Executive Summary)NASA-CR 134949.

Co un c i 1 on Environmental Quality, 1 981,

"Global Energy Futur es and the CarbonDioxide Problem", January, 1981.

Deeken, C. B. von der, et al, 1975, "SpecialIssues; High Temperatur e Reactor forPr ocess Heat Applicat ion, " Nue. Eng.Des. 34, gl .

Dietrich, G. , H. G. Eiekhof f, F. Niehaus,and H. F. Niessen, 1975, "Motivation forand Possibi 1 i ties for using NuclearProcess Heat, " Mucl. Eng. Des. , 34, 3.

Fluor Engineers and Con structors, 1976,"H-Coal — Commercial Evaluation" DOERept . FE-2002-12.

Haefele, M. , et al, 1980, "EnergyAlternatives for a Fi n i te Norld",International In st it ut e for Appl i edSystems.

Harrison, J. , 1980, National Coal Board,St oke Orchard, Glous. England, Pr ivateCommun ication .

Hoogendoorn, J. C. , 1977, "Gas from Coal forSynthesi s of Hyd roc arbon s, Status ofSASOL I I, " Panel di scussion, 9thSynthetic Pipeline Gas Symp. , (11/1/77).

Ko s'ky, P. G. , and J. It'll. Flock, 1 980,"Nuclear Process Heat -- Application toCoal Gasi fication" General Electric C.Repr t. 80CBD151.

NacDonald, G. , et al, 1980, "JASON StudyThe Long-Term Impac ts o f Incr easingAtmospheric Carbon Dioxide Levels, " U. S.DOE Contr act, , ME-AM-03-76SF -00115, SRIPro ject, 5793, Rept. JSR-79-04.

MacDonald, G. et al, 1979, Ibid, U S 0 0 E.Contract EY-76-C-0115, p. 136, SR IPro ject 5793, Rept. JSR-789-07.

Mc Namee, G. P. , N. K. Pa tel, T. R.Boszkowski and G. A. white, 1978,"Process Engineering Evaluation ofAlternative Coal Liquefaction Concepts, "EPBI AF-741, 1.

Nahas, 'H. C. , a nd S. E. Gallagher, Jr . ,1978, "Catalytic Gasi ficat ionPredevelopment Research, " 13th IECEC, 3,p. 2143 (San Diego, CA).

Pflasterer, G. B. , et al, 1979, "HTR-SynfuelApplication Assessment, " USDOE Rept.COO-4057-1 2.

Pomeroy, B. D. , J. Fleck, M. Marsh, D. Brownand B. Shah, 1978, "Comparative Study andEvaluation of Advanced Cycle Systems, "EPR I-AF 664, 2, pt, . 1 .

Schreiner, M. et al, 1978, "ResearchGuidance Studies to Assess Gasoline fromCoal to Me thanol-to-Ga so 1 inc andSASOL-Type Fi scher-Tropseh Technology, "USDOE Re pt, . FE2447-1 3.

Rev. Mod. Phys. , Vol. 53, No. 4, Part ll, October '1981

CO2 Elnissions 8159

Shah, R. , D. Ahne r, G. Fo x and N. Gluckman,1980, "Per formance and Co s tCharacter i sties of Combined CyclesIntegrated wi th Second GenerationGasification Systems, " ASIDE Gas TurbineConference, Paper 80GT-106 (3/11/80) .

Shih, S. S. , 1981, pr ivate communication.Shinnar, B. , et al, 1978, "Gasifier Study

for Mobil Coal to Gasoline Process, "USDOE Rept. FE-2766-13.

Stanford Research Institute, 1979, Energy,Economics and Technology Program, 14,"Coal Li quef action" .

Mei, J, 1981, to be pub 1i shed in Ind. Eng.Chem. , Proc. Des. Dev.

Rev. Mod. Phys. , Vol. 53, No. 4, Part il, October 1981


Rev. Mod. Phys. , Vo(. 53, No. 4, Part I I, October 1981

Acknowledgments S161

APPENDIX I I. ACKNOWLEDGMENTS

Me wi sh to express our particular thanksto J. A. Burton, Treasurer of the AmericanPhysical society, for his interest, and aid inall aspects of this study. The administrat;iveassistance of a number of other people at theAmerican Physical Society i s gratefullyacknowl ed ged .

The Study Panel benefited greatly fromthe encouragement and advice provided by theAPS Review Committee ehaired by H. Brooks.The membership of the Review Commi ttee islisted at the beginning of this Report.

Support and comments from many of thepersonnel of t, he U. S. Department of Energywere valuable in initiating and carrying outthe st, udy. Many o f these people ar eacknowledged below for supplying technicalinformation to the Study Panel. However, wewould like to particularly thank those peopleat, DOE who worked with us in initiating thestudy and in maintaining contact during itscour se . J. R. Power s, C. D. N.. Thornton, andJ. A. Snow of the Of f iee of Energy Research;R. Roberts and I. Mender of the Office of theAssistant Secretary for Fossil Energy. Mewould also like to thank the Director of theOf fice of Energy Research, E. A. Fr iemen andsubsequent;ly N. D. Pewitt, and the Assistant

Secretary for Fossil Energy, G. Fumieh andsubsequently R. M. A. LeGassie for the supportof their Offices in the conduct of this study.

Me are grateful for the hospitality ofMest, Virginia University which as. , providedfacilities -for the study headquarters. M. E.Vehse —Chairman of the Physics. Department andR. Koppelman — Vice-Pr esident for Ener gyStudies, Graduate Programs, and Research, havebeen helpful in many ways; and the MestVirginia Universi ty Foundation has provideduseful administrative services. Our specialgratitude goes to the Study secretaries —toC. A. Flagg, who bore the brunt of the effort,and to K. J. Bukr im and J. A. Cooper .

The Study Panel and Review Committeebenefited from the hospitality of a number ofinst;itutions at; which our meetings were held.Me are particular ly thank ful to B. C. Frazerand his colleagues at the American PhysicalSociety editorial of fices on Long Island fortheir hospitality and consideration during avery hectic two-week workshop in August, 1980.Me also appreciate the courtesy extended to usduring shorter meetings of the Study Panelhost ed by Mobil Research and Development, theLawrence Berkeley Laboratory of the Universityof California, the U. S. Department of Energy-Germantown, MD, and the headquarters of theAmerican Physical Societ;y in New York City.

Me are indebted to the following formeeting with and 'providing technicalinformation to members of the St,udy Panel:

E. M. Alb aughR. A. AlpherL. G. AustinM. BakkerJ. BeerJ. BirkelandE. BurwellH. R. ChildR. Clas senS. J. DapkunasR. DattaA. DavisA. DietzS. ErgunM. FarthingR. FischerB. C. FrazerJ. GethnerP. GivenD. M. GoodmanF. J. Grunthaner

Gulf Research 6 Development CompanyGeneral Electric Research 6 Development. CenterPennsylvania State UniversityElectric Power Research InstituteMassachusetts Institute of TechnologyU. S. Department of Energy — GermantownU. S. Department of Energy — GermantownOak Ridge National LaboratorySandia National LaboratoriesU. S. Department of Energy — GermantownGulf Research 6 Development ComapnyPennsylvania State Universi. tyU. S. Department of Energy — GermantownLawrence Berkeley LaboratorySouthern Research InstituteU. S. Department of Energy — GermantownBrookhaven National LaboratoryEXXON Research 5 Engineering CompanyPennsylvania St.ate UniversitySandia National LaboratoriesJet Propulsion Laboratory

Re&. Mod. Phys. , Vol. 53, No. 4, Part I I, October 1981


D.R.L.J 4

T 0

B.H.p.R 0

D.T.L.R.P.M.A.H.M-.

K.p.J.p.S.p.B.N.GD.pJ.N.

' J.pP.R.R.

R. HardestyHildebrandItonM. LarsenK. LauLewisL ~ Lov'el 1T. LuckieNarkleyM. NcKeeP. NeloyMillerR. OderPainterPetersC. Rapt isL. RetcofskyK. RhimM. SchatzM. Scil.dtR. Schrief ferScottN. Shap iroG. ShewmonG. SilbernagelSluyterSomor jai,K. StevensD. StevensonStringerStronginA. SullivanVastolaL. Malker, Jr.T. MoodA. Young

Sandia National Laboratori esU. S. Department of Energy — GermantownArgonne National LaboratoryUniversity of TennesseeU. S. Department of Energy — GermantownJet Propulsion LaboratoryPennsylvania State UniversityPennsylvania State Universi. tyU. S. Depar tment of Energy — GermantownGeneral Electric Research 6 Development CenterMest. Virginia UniversityU. S. Department of Energy — GermantownGulf Research 6 Development CenterPennsylvania State Universi. tyPittsburgh Energy Technology Center — DOEArgonne National LaboratoryPittsburgh Energy Technology Center — DOEJet Propulsion LaboratoryMobil Research 6 Development CenterUniversity of Missouri — ColumbiaUniversity of Calif ornia — Santa BarbaraU. S. Department of EDergy — GermantownBrookhaven National LaboratoryOhio State UniversityEXXON Research 5 Engineering CompanyU. S. Depar tment of Energy — GermantownUniversity of Calif ornia — BerkeleyU. S. Depar tment of Energy — GermantownU. S. Department of Energy — GermantownElectric Power Research InstituteBrookhaven National LaboratoryLos Alamos National LaboratoxyPennsylvania State Univer sityPennsylvania State Univers ityGeneral Electric Research 6 Development CenterUn iver s ity of Ar izona

Me are indebted to the following forcritically reviewing drafts of this ' Report inpart or in its entirety.

R.R.M.R.M.H.A.M.J.N.D.H.L.K.

AlpherBacastowA. BaumS. BerryBeezholdR. ChildDamaskA. EllingsonE. EppersonGr eenspanR. HardestyHeinemannItonJordan

General Electric Research 6 Development CenterUniversity of Calif ornia — San DiegoFlorida State UniversityUniversity of ChicagoSandia National LaboratoriesOak Ridge National LaboratoryQueens Gollege of the City University of New YorkArgonne National LaboratoryArgonne National LaboratoryNational Bureau of StandardsSandia National LaboratoriesLawrence Berkeley LaboratoryArgonne National LaboratoryUniversity of Pit tsburgh

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October 'l981

Acknowledgments S163

J. J. Kauf manL. KlemmA. LevyD. R. Lide, Jr .P. LuckieG. MacDonaldD. M. NcKeeT. P. NeloyP. A. NontanoM. A. NierenbergM. PetersR. PerkinsE. H. PiepmeirA. C. RaptisH. L. Retcof skyR. Rot tyJ. SchooleyB. G. SilbernagelP. L. Malker, Jr.M. H. MeinbergI. MenderJ. 0. L. MendtR. T. MoodT. Yule

Johns Hopkins UniversityUniversity of OregonLaurence Berkeley LaboratoryNational Bureau of StandardsPennsylvania State UniversityMITRE CorporationGeneral Electric Research 6 Development CenterMest. Virginia UniversityMest Uirginia UniversityUniversity of California — San DiegoPittsburgh Energy Technology Center — DOELockheed Research LaboratoriesOregon State UniversityArgonne National LaboratoryPittsburgh Energy Technology Center — DOEOak Ridge Associated UniversitiesNational Bureau of StandardsEXXON Research and Engineering CompanyPennsylvania State UniversityCalifornia Institute of TechnologyUniversity of PittsburghUniversity of ArizonaGeneral Electric Research 6 Development CenterArgonne National Laboratory


Glossary S365

APPENDIX III. GLOSSARY

AES: Auger Electron SpectroscopyANTHRACITIC: the highest (most completely

metamorphosed) ranks of coal with highfixed .carbon percentages (86$ or more)and low volatile content

ALI PHATIC: organic hydrocarbonschar aeter i zed by straight- orbranched-chain arrangement o f theconstituent carbon atoms —composed ofthree subgroups: (1) paraffins (alkanes),(2) olefins (alkenes), and (3) acetylenes(alkynes)

AROMATIC: organic hydrocarbons incorporatingthe basic (ring) structure of benzene

ARUPS: angle r e so 1 v ed ultr avioletphotoelectron spectroscopy

BAG-HOUSE: an emission control system of apulverized coal fired boiler in whichcooled stack gases (combustion products)are passed through fabr ic f ilter s toremove particulates

BASE-LOADED: specifies an electricgenerating station operated cont, inuouslyat full eapaeity in order to takeadvantage of its high efficiency andlower cost . Other mor e expensivegenerators are used for peak demand(turbines fired with kerosene forexample) ~

BITUMINOUS: i n termed i ate r anks o f coalwith lower fixed carbon percentages thanan t hr ac i te and considerable volat ilematter (1$$ to more than 30$)

BOTTOMS: heavy tar-like residual material(hydrocarbons, other high molecularweight compounds and carbon ) remainingafter pur i ficat', ion of er udes or otherdistillates

BTU: "British thermal unit"; t;he heat,necessary to raise the temperature of 1pound of water at it, s maximum density by1 F; 1 Btu = 252. 0 Calorie = 1054 J

CAKING: the propensity of coal to form asolid mass when the (pulverized) materialis heated in the absence of air (as int he st, andard volatile matterdetermination)

CARBON EFFICIENCY INDEX (n ): the percentageof carbon in the products of a processrelative to that in t,he input coal

CARBON YLS: i nor gan ic carbon compounds

containing CO groupsCARS: coherent anti-Stokes Raman

spectroscopyCAT-CRACKER: catalytic cracker (chemical

reactor for breaking up large hydrocarbonmolecules into smaller ones in thepresence of a catalyst)

CHAR: the solid r esj. due resulting fr ompyrolysis or other conversion reaction ofcoal -- it has high fixed carbon and lowvolatile content, and contains residualcoal minerals

CO-GENERATION: utilization of re jected heatin a steam generation plant so that theplant, generates both electric power andheat for industrial processing or spaceconditioning

COKE: a char of high porosity and goodmechanical strength -- thereforeimportant in steel making

COMBINED CYCLE: a scheme for the generationof electric po~er in which two generatingsystems are operated in cascade so thatone (the "bottoming" cycle) uses as input,t, he heat, rejected by the other("topping") cycle

COMMINUTION: breaking a solid up i nto apowder or pieces (it, is most commonlydone by grinding or crushing but, in thecase of coal, can be aecompl i shedchemically and perhaps even by shockwaves. )

CYCLONE SEPARATOR: a centrifugal device forseparating suspended solid (or liquid)particles from a gas by inducing a gasflow of high vortieity resulting in walldeposition and fall-out collection

DE VOLATILIZATION: an alt ernate ter m fo r thepyrolysis or destructive distillation ofcoal by heating in the absence of air

DIRECT LIQUEFACTION: a liquefaction processin which coal is dissolved or d i sper sedin a solvent, and reacts with hydrogen toproduep liquids without breakup of thecoal structure to produce low molecularweight intermediates (such as H2, CO, orCH OH)

DONOR SOLVENT: an organic solvent (used incoal liquefa'ction) usually a coal-derivedliquid, which i s capable of beingdirectly hydrogenated (usually in t, hepresence of a catalyst, ) and is capable,in turn, of transferring hydrogen tochemically active sites in coal

EBULLATING BED: a multi-phase system inwhich gas and/or a liquid passes upwardsthrough a bed of so 1id s, thereby keeping

Rev. Mod. Phys. , Vol. 53, No. 4, Part I I, October )S8't

S]66 APS Study Group on Coal Utilization and Synthetic Fuel Production

the particles in motionENDOR: electron nuclear double resonanceENTRAINED FLON: a eo-current flow o f

part, iculates in a continuous phase fluidEPRI: Electric Power Research InstituteESR/EPR: electron spin resonance/electron

paramagnetic resonanceETHERS: organic oxides —compounds in which

a single oxygen atom is linked to twoseparate organic groups

EXAFS: ext ended X-r ay absorpt, ion f inest, ructure

EXINITE: (See Liptinite)EDS or EXXON DONOR SOLVENT: a coal

liquefaction process in which the donor. solvent (a coal-derived organic liquid)i s catalyt ical ly hydrogenated in asepar ate reaction chamber before beingbrought into interaction with the coal

FBC: fluidized bed combustion (or combustor)FINES: very finely pulverized coal produced

incidental to grinding or crushingintended to produce a coarser product

FISCHER-TROPSCH: a eommerc ial process forcatalytic reactions of synthesi s gas(hydrogen and carbon monoxide) into(mostly) str a ight and branched chainhydrocarbons

F LOTATION: var ious met, hod s o f"bene f iciation" (separating, orpurifying) of mixed pulver ized solids(mineral ores or coal) which depend upondifferences of density and/or wettability

FLUIDIZED BED: an arrangement in whichcrushed or pulverized solids in a bed arelevitated because the gravi tationalforces are just balanced by hydrodynamicforces exerted by gas or liquid fed frombelow and flowing upward through t, he bed

FLUIDIZED FLOM: a mode of gas-solid andliquid solid flow systems in which thesol id par tieles are suspended in the flowf i eld of the cont, inuous phase, andexhibit fluid behavior

FREE RADICAL: a molecule with an odd numberof electrons or, more gener ally, achemical configuration containing more orless localized unpaired electrons

FRIABILITY: a tendency to break up orcrumble under handling -- a quantitativeindex is defined by an ASTM testinvolving measurements of average sizesbefore and after one hour of beingtumbled.

FTIR: Four ier transform in f r a —r edspectroscopy

H-COAL: a coal liquefaction process (under

development by Hyd rocarbon Research,Inc. ) in which coal, donor solvent, andcatalyst are all mixed together in theprimary reaction chamber under a hydrogenatmosphere, and the r eactor is anebullating bed

HEIS: high energy ion scatteringHETERO-ATOM: atom of an element other than

carbon or hydrogen incorporated in thearomatic structure (as distinct from themineral content) of coal -- most commonare oxygen, sulfur, and nitrogen

HH V or HIGHER HEATING VALUE: heat ofcombustion of a hydrocarbon fuel assumingthat fuel-bound hydrogen ends up as ~aterin the liquid st, ate

HREELS: high-resolution electron energy-lossspectroscopy

HYDROCRACKING: catalytic cracking (of largehydrocarbon molecules) in a hydrogenatmosphere to yield smaller molecules ofhigher hydrogen content

HYGAS: Instit, ute of Gas Technology processfor conversion of coal to subst;it, utenat ur al gas by hydrogasifieat, ion,involving direct hydrogenation of coal inthe presence of hydrogen and steam underpressure in sequential fluidi zed bedreactor stages

HYDROGEN DONOR SOLVENT: an organic solventwhich is capable of transferring hydrogento chemically active sites in coal

INDIR ECT LIQUEF ACT ION: a 1 i que fact ionprocess which proceeds through theintermediate step of first makingsynthesis gas (H and CO), andsubsequently synthesizes liquidhydrocarbons from t,his gas

INERTINITE: a group of macerals which isusually inert or semi-inert during normalcarbonization or hydrogenation processesin a retor t or reactor; they are opaqueto transmitted light and bright inr e fl ec ted 1 i ght, and der ived fromcharring of plant, tissues or possibly asthe result o f intensive b ioehemi ea1processes.

LEED: low energy electron diffractionLEIS: low energy ion scatteringLHV OR LONER HEATING VALUE: heat of

combustion of hydrogen containing fuelassuming that, fuel-bound hydrogen ends upas water in the vapor phase

LIGNITE: the lowest (least metamorphosed)rank of coal with low fixed carbon (lessthan 69$) and low calorific value (8, 300Btu/lb or less)


G lossary S167

LIPTINITE (EXINITE): a maceral group havingthe highest hydrogen content,comparatively low index of refraction andref lectivit, y, and derived from resinousand waxy material of plants, includingr e sins, euticles, spore and pollenexines, and algal remains

LOCK HOPPER: a hopper for feeding dr ymaterials through a pressure differenceby input admission at one pressure(usually low) with output, doors closed,and subsequently emptying after the inputdoors are closed and the pressure hasbeen equalized with the output pressure(usually high)

LURGI: (trade name) a commercially availableand eommer i c a 1 1 ly u sed batch coalgasifier of fixed bed configuration

MACERALS: the most elemen t a r ypetrographical ly d istingui shable (bymicroscopic examination) constituents ofthe organic part of coal (the organicanalog of minerals)

MEIS: medium energy ion scatteringMETC: Morgantown Energy Technology CenterMETHANOL-TO-GASOLINE: a process developed by

Mobil Research and Development Corp. inwhich a ZSM-5 catalyst is used to convertmethanol (CH OH) (which can besynthesi zed through i nd i r ec t e oalliquefaction) into high octane gasoline

MHD: Magneto-hydrodynamic electrical powergeneration

MM : "megawatt eleetrie" —unit employed inspeci fying the capacity of an electricgenerating station

NM : "megawatt thermal" —unit employed inspecifying the heat input to an electricgenerating station

NATUR(L GAS: high ptu fuel gas (about; 1 x10 BTU/1000 ft. ) of natural originconsisting mostly of methane (CH&)

NMR: nuclear magnetic resonanceNO : one (or a mixture) of the oxides of

xnitrogenNQR: nuclear quadrupole resonancePAS: photoacoustie spectroscopyPCC: pulverized coal combustion (or

combustor)PESIS: photoelectron spectroscopy of the

inner shellsPESOS: photoelectron spectroscopy o f the

outer shellsPETC: Pi t, tsbur gh Ener gy Technology Cent erPETROGRAPHY: t he stud y and d i fferentiation

of components of rocks and minerals byvisual observation with a microscope

PHOTO-ACOUSTIC (MICROSCOPY, ETC. ): detectionand measur ement of acoustic signalsemitted as a consequence of opticalirradiation or excitation of the specimenor system under study

PNEUMATIC CONVEYING: transport of pulverizedsolids which are entrained or suspendedin a flowing gas and carried along withit

PROXIMATE ANALYSIS: characterization of coalin terms of moisture, ash content, fixedcarbon, volatile matter, and ealorifievalue

PYRIDINE: aromatic compound with hexagonalcarbon r ing havin g one subst itut ednitrogen atom (C H N)

PYRITES: loosely used to deno te i r onsulfides with a range of stoiehiometryand crystal structures, an importantmineral constituent and source of sulfur1n coal

PYROLYSIS: decomposition by heating (in theabsence of air)

RANK: a classif icat ion scheme for coalsbased primarily on their heating values(per unit weight, ) in dir ect combustion,and increasing with per cent carbon

REGENERATIVE HEATER: a heat, exchanger inwhich the heated and cooled fluids occupythe same space by t urns -- respectivelyreceiving heat from and imparting heat toa storage medium

SAM: scanning Auger microscopySANS: small angle neutron scatteringSASOL: South Af r ican Syn thetie Oi1 -- t he

world ' s only oper at ing commercial plantfor production of synthetic liquid fuelsfrom coal (by the "indirect"Fiseher-Tropsch process -- located inSouth Africa)

SAXS: small angle X-ray scatter ingSCF: "st,andard cubic foot"; the amount of

gas which occupies one eub ic foot atnormal atmospheric pressure and 60 F

SCRUBBER: a sub-system of a pulverized coalf ired boil er in which stack gases(combustion products) are treated withalkaline aqueous sprays to remove sulfuroxides

SEM: scanning electron microscopy (detectsthe secondary back scattered electrons)

SEXAFS: surf aee extended X-ray absorptionfine structure

SIMS: secondary ion mass spectroscopySINGLE PARTICLE MONITOR: a device which

detects individual particles of anaerosol or other aggregate and, from a



s tatistically large number of suchevents, provides direct information aboutnumber concentration, and distribution insize or other physical properties, of theparticles

SLAG: ash in the molten state -- usually amixture (mostly) of oxides and complexsilicates

SLUGGING: the formation of large particulatevoids or bubbles in multiphase fluidizedflow which grow to emcompass the entirechannel perpendicular to the flow

SNG: "sub sti t ute natur al gas"manufactured or man-made high Btu fuelgas consisting mostly of methane (CH )

SO : one (or a mixture) of the oxides ofsulfur

SBC I and SRC II: "solvent refined coal", Iand II, —processes for cleaning up andhydrogenating coal using a coal-derivedsolvent and hydrogen atmosphere, butwithout added catalyst. SRC-I producespi tch-1 i ke sol id fuels; SRC-II producesliquid fuels.

"STIFF": (of di f ferential equations)equations expressing the coupl ing orinter action of systems or processeshaving very disparate time-constants

SUPERHEATER: the highest temperature stageor tubing bank of a steam boiler

SYNGAS (SYNTHESIS GAS): the mixture of

hydrogen and carbon monoxide produced bythe carbon steam reaction (and itsconcommitants) in a coal gasifier

TDR: time domain ref lectometryTEN: transmission electron microscopyTHIXOTROPIC SLURRIES: gels (pseudofluids

containing suspended solids) for whichthe shear stress i s not linearlydependent upon the velocity gradient

TYPE: a classi f ication scheme for coalsaccording to maceral content

ULTIMATE ANALYSIS: an analytical scheme thatis more complete than proximate analysis,including a detailed chemical contentstated in terms of elemental constituents

$C, gH, $0, etc.ULTRAF INE PARTICLES: coal part icles of

average size &1 pmUPS: ultraviolet photoelectron spectroscopyVITRINITE: the most abundant and important

groups of macerals in coals, consistingof coalified woody tissues derived. fromstems, roots and vascular tissues ofleaves

XPS: X-ray photoelectron spectroscopyZEQLITES: a family of alumino-si 1 ieates

whose (unusually open) crystal structureis characterized by cages and channels ofwell defined shapes, giving high internalsur face area —— somet imes called"molecular sieves"


i=&or::o'"ie!-'. - Physical Review Link Manager

Documents

Transcript of i=&or::o'"ie!-'. - Physical Review Link Manager