Applied Energy xxx (2012) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/locate /apenergy

Technologies for energetic exploitation of biodiesel chain derived glycerol:Oxy-fuels production by catalytic conversion

Carlo Beatrice a,⇑, Gabriele Di Blasio a, Maurizio Lazzaro a, Catia Cannilla b, Giuseppe Bonura b,Francesco Frusteri b, Francesco Asdrubali c, Giorgio Baldinelli c, Andrea Presciutti c, Francesco Fantozzi d,Gianni Bidini d, Pietro Bartocci d

a CNR-IM, Istituto Motori, Via G. Marconi 4, 80125 Napoli, Italyb CNR-ITAE, Istituto di Tecnologie Avanzate per l’Energia ‘‘Nicola Giordano’’, Via S. Lucia sopra Contesse 5, 98126 Messina, Italyc CIRIAF – Centro Interuniversitario di Ricerca sull’Inquinamento da Agenti Fisici, Università degli Studi di Perugia, Via Duranti, 67, 06125 Perugia, Italyd CIMIS – Consorzio Interuniversitario Macchine Impianti e Sistemi per l’energia, l’industria e l’ambiente, c/o Dipartimento di Ingegneria Industriale, Via Duranti 97,06125 Perugia, Italy

h i g h l i g h t s

" A glycerol conversion in an oxy-fuel for diesel engine has been proposed." Hyflon� based catalysts supported on spherical silica have been developed." The impact of the oxy-fuel on the engine emissions and performance has been described." LCA of the oxy-fuel/diesel blend shows a lower impact that the one of neat diesel.

a r t i c l e i n f o

Article history:Received 23 February 2012Received in revised form 27 July 2012Accepted 9 August 2012Available online xxxx

Keywords:Biodiesel derived glycerolFuel catalystOxygenated diesel fuelsDiesel engine emissionsLCA

0306-2619/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.apenergy.2012.08.006

⇑ Corresponding author. Tel.: +39 0817177186; faxE-mail address: c.beatrice@im.cnr.it (C. Beatrice).

Please cite this article in press as: Beatrice C et aalytic conversion. Appl Energy (2012), http://dx

a b s t r a c t

The paper describes the results of a co-operative research project aimed at the energetic exploitation ofthe biodiesel derived glycerol. The research objective is the definition and validation of a synthesis pro-cess for conversion of glycerol in an oxygenated fuel additive (glycerol alkyl–ether) suitable for blendingwith diesel and biodiesel.

The employed facilities, methodologies and first results from engine tests are described, including apreliminary Life Cycle Assessment (LCA).

Etherification of glycerol with tert-butyl alcohol and isobutylene allowed to identify the suitable exper-imental conditions required to obtain a mixture predominantly composed of higher glycerol ethers(GEM). Novel catalysts, based on perfluorosulphonic ionomers as active species and spherical silica assupport, were found to be stable and easy reusable allowing to obtain mixtures containing very lowamount of monoethers and oligomers considered as undesired products.

Engine tests with blends of GEM in diesel fuels revealed good combustion and emission performanceswhen compared both to neat diesel fuel and a blend diesel/biodiesel.

LCA analysis indicates that the GEM/diesel blend generates an environmental impact which is lowerthan the one of neat diesel.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Pending the large-scale production of second generation biofu-els [1], in the near future the biodiesel production should increasein order to approach the EU target of biofuel consumption for theyear 2020 [2]. As a result, also the production of glycerol, whichis the main by-product of the transesterification process, will grow

ll rights reserved.

: +39 0812396097.

l. Technologies for energetic ex.doi.org/10.1016/j.apenergy.201

proportionally. So, the question arises from a cost-effective andefficient use of such a glycerol surplus. The combustion of glycerol‘‘as it is’’ would represent a desirable solution. Unfortunately, be-cause of its detrimental physical and chemical properties, rawglycerol is hardly usable in conventional energy production plants,such as fuel burners or internal combustion engines (ICEs). How-ever, a proper design of the burner and refractory combustionchamber could allow to effectively burn crude glycerol, althougha high particulate matter concentration is detected at the exhaust[3]. Alternative routes for glycerol recovery are represented by its

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-2.08.006

Fig. 1. Experimental set-up for catalytic etherification of glycerol.

2 C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx

energetic upgrading into high value products, such as synthesis ofgases [4] and oxygenate additives for gasoline, biodiesel and dieselfuels [5].

In order to increase the energetic exploitation of the biodieselproduction chain, a co-operative research project was started in2010, aimed at recycling the biodiesel-derived glycerol in energygenerating systems. The project was partially founded by MiPAAF,the Italian Ministry of Agriculture, Food and Forestry. It consists ofthree research lines: (i) conversion of the raw glycerol in a fuelsuitable for ICEs; (ii) pyrolysis of biomass soaked with glycerolfor syngas production; (iii) combustion of glycerol in a flamelessoxy-burner.

The first research line, described in the present work, is focusedon the development of a synthesis process for converting glycerolin a fuel suitable for ICEs. The final goal should be the definitionand validation of an etherification process of raw glycerol for theproduction of an oxy-fuel that can be blended with diesel and bio-diesel. An industrially relevant route for the conversion of glycerolinto oxygenated chemicals involves the etherification to tert-butylethers [6–8]. In particular, since mono-tert-butyl ethers of glycerol(MBG) have a low solubility in diesel fuel, the etherification of glyc-erol, taking place with a consecutive reaction path, should bemainly addressed towards the formation of a mixture of di-(DBG) and/or tri-ethers (TBG) [6–9].

The project aims at a complete evaluation of the productionprocess (in laboratory scale), as well as the validation of the Glyc-erol–Ethers Mixture (GEM) of DBG and/or TBG as fuel in experi-mental engine tests, with two main advantages: (i) the increaseof the energy and CO2 Well-To-Wheel (WTW) factor of the biodie-sel supply chain; (ii) the addition of a high-quality oxy-fuel in themineral fuels which contributes to the reduction of pollutant emis-sions from diesel engines.

The GEM production process for glycerol exploitation was alsosubjected to Life Cycle Analysis (LCA), from which the impact tohuman health, quality of ecosystems, and resources has been eval-uated. The present paper illustrates the employed facilities, meth-odologies and the results of the research activity.

2. Materials and methods

2.1. Catalytic systems for the conversion of glycerol into diesel fuelcomponents

Several Hyflon�-SiO2 catalysts (c.a. 20 wt.%) were prepared bythe conventional incipient wetness method using an alcoholic solu-tion of Hyflon� Ion S4X perfluorosulphonic ionomer (H730, SolvaySolexis) as acid precursor. As carriers, two spherical silica, MS3030and ES70Y, characterized by surface area of 300 and 295 m2 g�1

respectively, were supplied by PQ Corporation (Liverpool, UK). Inaddition, a fumed silica, LM50 with surface area of 150 m2 g�1

was supplied by Cab–O–Sil Division-Rheinfelden. After impregna-tion, the catalysts were dried in air at 120 �C for 12 h. As a reference,Amberlyst� 15 ‘‘dry’’ resin (A-15) from Rohm and Haas was used.The investigated catalysts were characterized by different tech-niques in order to determine textural and morphological structure,thermal stability, loading of active phase and acid capacity.

2.1.1. Plant and catalytic processIsobutylene 3.0 (IB) by GHC Gerling (Hamburg, Germany), tert-

butyl alcohol (TBA) and commercial anhydrous glycerol (purityP99.5%) were used as reactants, deferring the problem of usingraw glycerol after a first validation of the catalytic process. Theetherification reaction was carried out in liquid phase in a lab-scalestainless steel ‘‘jacketed-batch’’ reactor of 300 cm3 (see Fig. 1)under a stirring frequency of 1200 min�1 in order to limit the

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.20

influence of external mass transfer phenomena. Experiments wereperformed by operating at different tert-butyl alcohol/glycerol(RTBA/GLY) or isobutylene/glycerol (RIB/GLY) molar ratio (2.5–5.0)and by employing a catalyst amount of 0.3–7.5 wt.% referred tothe glycerol weight (Rcat/GLY) at variable reaction time (6–17 h). Awell-defined amount of glycerol and dry catalyst were loaded intothe reactor and, before the addition of TBA or IB, the reaction envi-ronment was flushed with nitrogen to remove the air. A knownamount of tert-butyl alcohol or liquid isobutylene, previously fedinto a graduated tank, was injected into the reactor under 5 barof nitrogen pressure. At the end of the run, the reactor was cooleddown by an ice-bath, thus allowing all the gas phase compounds tobe condensed.

2.1.2. The experimental engine for the GEM validationThe preliminary validation of the fuel was carried out with a sin-

gle cylinder diesel research engine, specifically designed for com-bustion studies. All engine components represent the state of theart of the engine technology. The engine, developed at the IstitutoMotori, has a modern combustion system design, 0.47 l of displace-ment and four valves, derived from a reference Euro 5 compliantfour cylinder engine for passenger cars. The engine is equipped witha common-rail injection system and a fully programmable elec-tronic control unit (ECU) to allow the modification of the engineoperating parameters. The test bench is fully instrumented for mea-suring indicated signals and analyzing exhaust gases.

Auxiliary systems for boost, cooling, lubrication, etc., are notcoupled to the engine in order to allow for a maximum control ofthese parameters without influencing the load conditions. This ap-proach offers flexibility pertaining to operating conditions, withoutforfeiting state-of-the-art technology. The engine layout issketched in Fig. 2, while the main design characteristics are listedin Table 1. Additional details are described elsewhere [10].

In each test points, the thermodynamic conditions at enginesystem boundaries (temperature, pressure and gas flow rate) werestabilized running the engine for 10 min, then fuel consumption,emissions, operating parameters (temperature, pressure, etc.) and160 consecutive indicated cycles were recorded.

2.2. Life Cycle Analysis

The environmental impact of the Gycerol Ethers Mixture (GEM)production, as mixture of DBG and TBG compounds, was evaluated

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-12.08.006

Fig. 2. Single cylinder research engine layout.

Table 1Main characteristics of the single cylinder research engine.

Engine characteristics

Bore (mm) 82.0Stroke (mm) 90.0Compression ratio 16.5Displacement (cm3) 475Valve number 4Combustion chamber shape x shapeInjection system Common railInjector Centered 7 holes microsac

C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx 3

on the basis of a LCA approach. As previously mentioned, GEM canbe employed as an additive of diesel in internal combustion en-gines. The study focused on the etherification process of glycerinat laboratory level, with regards to two different processes, oneusing tert-butanol (TBA) as additive, the other one using i-butene(IB). Etherification reaction condensation/filtration, decantation/centrifugation and distillation processes were considered for thetwo procedures.

According to EN ISO 14040 and EN ISO 14044 [11,12], a Life Cy-cle Assessment includes four main steps:

– Goal and scope definition.– Life cycle inventory.– Life cycle impact analysis.– Interpretation of the results.

The Dutch code SimaPro 7.2 [13] was used to perform the LCAand three different methods were implemented: Eco-indicator99(H) v2.03, IPCC 2007 GWP 100a v1.02 and Cumulative EnergyDemand v1.07 [14].

IPCC (International Panel on Climate Change) method classifiesthe different emissions according to their contribution to thegreenhouse effect; the indicator used is Global Warming Potential(GWP), which is measured in kg CO2-eq. The energy used during

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.201

the entire life cycle of the building is calculated by the method ofCumulative Energy Demand (CED), the unit is MJ-eq. The catego-ries are: Non-renewable, fossil; Non-renewable, nuclear; Non-renewable, biomass; Renewable, biomass; Renewable, wind, solar,geothermal; Renewable, water. Finally, Eco-indicator 99 is one ofthe most complex and complete methods. It allows the evaluationof the emissions and the use of resources according to 11 impactcategories (carcinogenic substances, respiratory diseases, climatechanges, ozone depletion, radiation that causes ionization, acidifi-cation/eutrophication, ecotoxicity, land use, mineral resourcedepletion and fossil fuels), grouped into three damage categories:

(1) Damage to human health, expressed as the number of yearsof human life lost or spent suffering from disease.

(2) Damage to the quality of ecosystems, expressed as the lossof living species in a certain area over a certain period oftime.

(3) Damage to resources, expressed as the surplus of energynecessary for the further extraction of minerals and fossilfuels.

Differently from the other parts of the paper, in LCA simulationsraw glycerol was considered, to verify the environmental benefitsof the valorization of this byproduct of biodiesel chain.

3. Results and discussion

3.1. Catalytic etherification of glycerol

The preparation of heterogeneous acid catalysts for the etherifi-cation of glycerol by tert-butyl alcohol or isobutylene was aimed atthe development of a system characterized by high surface area,high active site accessibility, no-leaching of active species and easyseparation from liquid reaction medium. The main physico-chemicalproperties of the catalysts are summarized in Table 2.

Compared with commercial A-15 resin, all the silica supportedcatalysts exhibited in all cases total acidity about 10–25 times

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-2.08.006

Table 2Catalysts and physico-chemical properties.

Catalyst Active phase (wt.%) Surface area Brønsted acidity(m2 g�1) (mmol H+ g�1)

Hyflon�/LM50 R-SO3H (16) 29 0.23Hyflon�/MS3030 R-SO3H (18) 177 0.14Hyflon�/ES70Y R-SO3H (19) 204 0.17A-15 R-SO3H (100) 45 4.75

4 C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx

smaller, due to the high acidity of bulk A-15 resin(4.50 meqH+ gcat

�1).In view of a possible use of glycerol ethers as oxygenated addi-

tives for (bio-)fuels, the performance of the investigated catalystswas compared in the etherification reaction of glycerol either withtert-butyl alcohol (TBA) or isobutylene (IB). In such reactions, therole of the temperature is essential, because, further to influencethe kinetics of glycerol conversion, the reaction temperature canpromote phenomena of thermal decomposition or recombinationof the intermediate products with consequent negative results interms of selectivity and total conversion of glycerol. On this account,as already reported in previous papers [6–8], the suitable experi-mental conditions allowing maximum yield to higher ethers are ob-tained operating at 70 �C, using a catalyst amount of 7.5 wt.% withrespect to the glycerol weight and alcohol (olefin) to glycerol molarratio of 3. Fig. 3 shows the preliminary catalytic results obtainedover the reference A-15 sample under these reaction conditions.

In terms of glycerol conversion, Fig. 3A shows that at 70 �C, byusing TBA, the reaction tends to reach the equilibrium after c.a.10–15 h. As regards the product distribution, a progressive de-crease in selectivity to mono-substituted compounds (MBG) mir-rors a progressive increase in the selectivity to di-substitutedcompounds (DBG), demonstrating that the reaction proceeds witha consecutive reaction mechanism. However, even after 30 h ofreaction, only about 25% of mono-ether is converted into di-ethers.Moreover under the reaction conditions investigated, after 24 h ofreaction the formation of the tri-substituted compound (TBG)becomes observable (�1.5%).

In presence of IB, Fig. 3B clearly shows that the etherificationproceeds at a reaction rate significantly higher than that observedusing TBA: the conversion of glycerol is close to 95% just after 1 h ofreaction, with significant selectivity values to DBG and TBG com-pounds (c.a. 60% and 10% respectively). It is also important tounderline that, after 30 h of reaction, glycerol attains a completeconversion, although approximately 10% of MBG compounds arestill present. As expected, the poor catalytic pattern observed inthe etherification of glycerol with TBA is ascribable to the forma-tion of water during the reaction, which involves the poisoning

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

1 2 6 24 30

DBG TBGMBG

RTBA/GLY , 3 mol/mol

reaction time (h)

Con

vers

ione

of

glyc

erol

(%

)

Selectivity to glycerol ethers (%)

(A)

Fig. 3. Glycerol conversion at 70 �C by A-15 cat

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.20

of the active sites onto the catalyst surface, not allowing the con-secutive reaction pathway to go on with production of discreteamount of TBG compound. For this reason, the catalytic activityof the silica supported Hyflon� systems was evaluated in the ether-ification reaction of glycerol with IB, so promoting the formation ofpoly-substituted products, very soluble in the traditional fuels.

In the attempt to evaluate the influence of the amount of cata-lyst to be used on glycerol conversion, preliminary tests in theetherification reaction with IB were carried out at different Rcat/GLY

(0.3–7.5 wt.%) using A-15 catalyst. The results obtained revealedthat it was not necessary to operate at Rcat/GLY higher than 5 wt.%,since neither glycerol conversion nor product selectivity signifi-cantly change. It is also important to highlight that the increaseof catalyst concentration only slightly affects the formation of di-isobutylene (DIB), considered as an undesired gummy product.On the contrary, both reaction time and isobutylene/glycerol molarratio significantly affect the product distribution. In particular,once the equilibrium condition is reached, reaction no longer goeson towards TBG formation; so, it appears useless to extend thereaction time which, rather, could favor both dealkylation of higherethers to MBG and DIB formation. Moreover, under over-stoichi-ometric conditions (RIB/GLY > 4 mol/mol), oligomerization reactionbecomes also important and DIB can reach very high concentration(>10 wt.%).

Once identified the suitable experimental conditions requiredto obtain a mixture containing low concentration of MBG (Rcat/GLY,5.0 wt.%; reaction time, 6 h; RIB/GLY, 4 mol/mol), the catalytic behav-ior of the prepared catalysts was compared with A-15 commercialcatalyst. As reported in Fig. 4A, over spherical silica supported per-fluorosulphonic catalysts (Hyflon�/ES70Y), mixtures of ethers con-taining very low content (<2 wt.%) of MBG ethers can be obtainedin one step by operating at relatively low reaction temperature(70 �C). Furthermore, as shown in Fig. 4B, such a catalytic systemwas found to be much more stable and easily reusable than others,maintaining high glycerol conversion (XGLY, 99%) and poly-ethersyield (YDBG+TBG, �85%). The superior activity and stability of Hyf-lon�/ES70Y catalyst was attributed to a strong ionomer–silicainteraction (preventing the leaching of active species during reac-tion), higher active sites accessibility (due to the large surface ex-tent of spherical silica) and surface hydrophobic properties.

In an attempt to further optimize the experimental conditionssuitable to obtain a mixture of glycerol ethers (GEM) containingMBG compounds at a concentration lower than 5 wt.%, the possi-bility to treat the reaction mixture containing mono-, di- andtri-ethers with fresh catalyst was also considered. In particular,as reported in Fig. 5, a GEM composition containing a high concen-tration of MBG (28.3 wt.%) was obtained after 22 h of reaction byusing a RIB/GLY = 3 mol/mol and Rcat/GLY = 7.5 wt.% (STEP A).

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

1 6 17

DBGMBG TBG

30

RIB/GLY, 3 mol/mol

reaction time (h)

Con

vers

ion

of g

lyce

rol (

%)

Selectivity to glycerol ethers (%)

(B)

alyst: etherification with TBA (A) or IB (B).

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-12.08.006

0

10

20

30

40

50

60

A-15 Hyflon(R)/ES70Y

MBGDBGTBGDIB

Pro

duct

dis

trib

utio

n (w

t.%)

50

60

70

80

90

100

0

10

20

30

40

50

60

70

1° run 2° run 3° run

regenerationsfreshcatalyst

ΧG

LY (%

)

Product distribution (w

t.%)

(A) (B)

Fig. 4. Etherification of glycerine by IB: (A) catalytic behavior at total conversion of glycerol (TR, 70 �C; Rcat/GLY, 5.0 wt.%; RIB/GLY, 4 mol/mol; time, 6 h); (B) Hyflon�/ES70Ycatalyst activity after regeneration cycles (TR, 70 �C; Rcat/GLY, 7.5 wt.%;c, 3 mol/mol; time, 17 h).

A B C D1

D2

D3

D4

0

10

20

30

40

50

60

70

80

90

100

6 h

6 h

15 h

6 h18 h @ 70 oC

RIB/MBG=3.0 mol/mol; Rcat/MBG=10 wt.%

TR

=34

3 K

x 2

2 h

RIB

/GLY

=3.

0 m

ol/m

ol;

Rca

t/G

LY=

7.5

wt.

%

15 h @ 70 oC

DBG

MBG

TBG

PROCESS STEP

Gly

cero

l eth

ers

dist

ribu

tion

(wt.%

)

70 oC 70 oC 70 oC

Fig. 5. Sequential process steps to achieve high tri-tert-butyl glycerol ethers yield.

C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx 5

Then, this mixture (without any purification) was used for newetherification runs with IB. So, at 70 �C and high catalyst/MBG ratio(Rcat/MBG = 10 wt.%), the concentration of MBG decreased from 28.3to 9.9 wt.% in 18 h (STEP B) and to 6.4 wt.% after 15 h of reaction(STEP C). However, in spite of the progressive increase in TBG con-tent up to 28.9 and 36.0 wt.% in the step B and C respectively, theMBG concentration still remained too high (>6 wt.%). Under thesame conditions, a further treatment of the mixture C with freshcatalyst did not allow to obtain a mixture with a MBG amount low-er than 9.3 wt.% (STEP D1). At this point, the reaction temperaturewas lowered, maintaining a high Rcat/MBG (10 wt.%). From resultsobtained under such conditions (STEPs D2–D3), it can be seen thatthe temperature plays an important role; in fact, as the reactiontemperature decreased, the amount of MBG decreased from 6.3to 3.2 wt.%, while TBG increased from 36.0 to 46.8 wt.%. Then, pro-cessing the mixture C at 50 �C for 15 h of reaction (STEP D4) MBGdistribution dropped to 2.4 wt.%, while TBG reached a value of56.1 wt.%. This result can be explained as a consequence of the sig-nificant dependence of the equilibrium concentrations on the reac-tion temperature. This means that the dealkylation of higher ethers

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.201

(with re-formation of MBG and isobutylene) proceeds as a backreaction only at temperature higher than 50 �C.

With the aim of using a mixture of ethers as oxygenated addi-tive for diesel fuel, 5 L of GEM were prepared and measurementsof solubility have been made too. From the results obtained, itwas seen that ethers mixtures at content of MBG as low as5 wt.% were perfectly soluble in diesel up to a concentration of20 vol.%.

3.2. GEM characteristics and the fuel selections

Engine tests were carried out with GEM obtained by etherifica-tion with IB on Hyflon�/SiO2, whose composition is 1.6%MBG + 39.2% DBG + 58.3% TBG.

To evaluate its combustion characteristics in a diesel engine,two fuel blends were prepared: a blend of 10% v/v of GEM, indi-cated as GEM10, in a commercial EN590 European diesel fuel,and a blend of 40% v/v biodiesel/diesel (B40) having the same oxy-gen mass content. The biodiesel was a rapeseed-methylester (RME)meeting the EN14214 European standards for biodiesels.

In this first screening, the choice of the GEM10 was due to avoidsignificant alterations of the fuel blend characteristics during theexperiments thus ensuring the engine reliability. Moreover, sincebiodiesel derived raw glycerol is about 10 wt.% of the primary oil,GEM10 could represents a preliminary (optimistic) blending target,in the hypothesis that all the raw glycerol is converted in GEM.

The main physical and chemical characteristics of the pure GEMand the blends are listed in Table 3.

3.3. Diesel engine tests

In order to evaluate the fuel impact on different engine combus-tion conditions, the two fuel blends (GEM10 and B40) and the ref-erence diesel fuel were tested in three engine operating conditions:

� 1500 revolution per minute (rpm) of engine speed and 3 bar ofIndicated Mean Effective Pressure (IMEP) as representative of alow speed and low load operating conditions, indicated in thefollowing as 1500 � 3.� 2000 revolution per minute (rpm) and 5 bar of Indicated Mean

Effective Pressure (IMEP), as critical operating point for soot andNOx emission control, indicated in the following as 2000 � 5.

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-2.08.006

Table 3Main characteristics of the tested fuels.

Property GEM Neat diesel GEM10 B40

(A/F)st (–) 6.6 14.54 13.7 13.67Low calorific value (LHV) (ASTM D240) (MJ/kg) 32.240 43.26 41.28 40.73Carbon (ASTM D5991) (%, m/m) 54.95 85.22 83.19 83.21Hydrogen (ASTM D5991) (%, m/m) 5.77 13.03 12.68 12.39Oxygen (%, m/m) 39.28 1.45 4.13 4.32Cetane number (ISO 5165) – 51.8 49.3 51.2Cetane index (ISO 4264) 18.2 – – –Density (ASTM D4052) @ 15 �C (kg/m3) 877.4 829.1 833.9 857.4Viscosity (ASTM D445) @ 40 �C (mm2/s) 4.190 3.141 3.457 3.51Atmospheric distillation (ASTM D86) IBP (�C) 95.5 158.9 172 189.8

10 vol.% (�C) 207.5 194.3 202.7 244.250 vol.% (�C) – 267.6 253.3 311.390 vol.% (�C) – 333.4 329.8 338.595 vol.% (�C) – 350.0 345.2 345.5FBP (�C) 208.0 360.9 355.9 354.2

6 C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx

� 2000 rpm and 10 bar of IMEP as medium load operating point,where the fuel combustion characteristics in terms of thermalefficiency, noise and impact on engine functionality can beput in evidence better. This point is reported in the followingas 2000 � 10.

Neat diesel and GEM10 tests were repeated three times in orderto decouple the test-to-test measurement deviations with respectto the fuel effects. Since the repeatability accuracy for biodieseland its blends with diesel fuels, in line with neat diesel onthe same engine, was validated by a lot of past experimentalstudies, in the case of B40, the test repetition was not taken intoaccount [15].

Under 1500 � 3 and 2000 � 5 condition start of injection (SOI)and the Exhaust Gas Recirculation (EGR) rate were changed in or-der to evaluate the GEM impact on combustion process for a differ-ent operating parameter set.

The Fig. 6A reports the comparison of cylinder pressure evolu-tion and heat release rate from combustion (HRR) versus enginecrank angle (c.a.) for neat diesel, GEM10 and B40 in the test point2000 � 5. For this test case, the engine was operated with singlefuel injection per cycle and no EGR. To get the same combustionpattern, the SOI of GEM10 was slight advanced in the compressionstroke, in line with its lower cetane number (see Table 3). However,the Fig. 6A shows an almost complete overlap of all displayedcurves (both the cylinder pressure and the HRR) indicating thatno significant drifts from the diesel and B40 were observed.

0

10

20

30

40

50

60

70

-40 -20 0 20 40 60Engine crank angle [°]

Cyl

inde

r pre

ssur

e [b

ar]

0

5

10

15

20

25

30

35

Rat

e of

hea

t rel

ease

from

com

bust

ion

[%/c

.a.]

DieselGEM10B40

Ignition delay GEM10

Ignition delay neat Diesel & B40

(A) (B

Fig. 6. (A) Cylinder engine pressure and HRR traces for all tested fuels at 2000 � 5 with1500 � 3 and 2000 � 5. Variation bars on diesel values represent the measured test-to-

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.20

Such result was evidenced also for the other operating parame-ters explored in the test campaign. It is noteworthy that, in real en-gine work conditions, a pilot fuel injection before of the main one isgenerally adopted for controlling the ignition delay and the com-bustion noise. Nevertheless, even though test conditions with sin-gle injection are not representative of actual engine calibration,they become necessary to highlight the effects of the fuel auto igni-tion properties on the cylinder combustion process. The Fig. 6B dis-plays the fuel ignition delay time for the test point 1500 � 3 and2000 � 5 with engine working with single injection per cycle butvarying SOI and EGR values. In particular, at 1500 � 3, SOI waschanged between a conventional value for engine calibration(�10 c.a. before top dead center) and an early value in order torealize an almost total premixed air/fuel charge (�15 c.a. beforetop dead center).

It is observable that the GEM10 gives unappreciable incrementin the ignition delay time in all operating modes, summarizingthat, within the chosen blending ratio, the addiction of GEM to neatdiesel does not reduce the autoignition quality of the blend.

At 2000 � 5, the typical EGR sweep by means of the step by stepincrement of the EGR rate was also performed adopting the normalreleased engine calibration. Such kind of experiments is normallychosen by test engineers in order to evaluate the benefits fromthe fuel quality, or from the engine configuration, to improve thetypical diesel soot-NOx trade-off. This analysis is particularlyimportant for operating conditions, as 2000 � 5, where NOx andsoot emission becomes critical for emission regulations [16].

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1500x3 singleinjection w/o

EGR_conventionalcalibration SOI

1500x3 singleinjection w/o

EGR_early SOI

2000x5 singleinjection w/o EGR

2000x5 singleinjection with EGR

Igni

tion

dela

y [m

s]

DieselGEM10B40

)

single injection per cycle and without EGR; (B) fuel ignition delay for all fuels attest variation.

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-12.08.006

0.0

0.5

1.0

1.5

0 2 4 6 8 10 12 14 16NOx e.i. [g/kg]

Soot

e.i.

from

sm

oke

mea

sure

men

t [g/

kg]

neat Diesel fuelGEM10

B40

Effect of fuel oxygen addiction

Increasing the EGR rate

Fig. 7. Soot-NOx trade-off by EGR sweep on the research diesel engine at 2000 � 5.

C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx 7

Fig. 7 shows the evolution of the soot and NOx trade-off of dieselengines when the EGR method is employed to reach the best com-promise between the two pollutant species. In the diagram, thetraces of the three tested fuels are displayed and the emission val-ues are expressed in emission index (e.i.) as grams of pollutant spe-cies per kilogram of burned fuel. Soot mass emissions arecalculated from exhaust smoke emission trough a consolidate cor-relation [17]. It is notable the significant NOx emission reductionwhen EGR rate is increased.

Generally, in the diesel engine calibration, the interception ofthe trace with the diagram bisector is chosen as the best compro-mise in terms of soot and NOx emissions. It is clearly evidenced inFig. 7 that the GEM10 blend shows the lowest trace, indicating thelowest soot production at same NOx exhaust emission level, partic-ularly al high EGR rate. Such result, that is typical of the oxygen-ated fuels [18], appears favorable for GEM10 with respect to B40,despite of the same oxygen content between the two blends. Sucha result seems in line with other literature results [18,19]. Thesestudies showed that not only the intramolecular oxygen contentinfluences the combustion but also the nature of the oxygenatedcompound plays a role in preventing soot formation. Even thoughthis hypothesis requires further validation in the future with spe-cific investigations, the results in Fig. 7 highlight the benefits ef-fects of GEM as soot suppressor in diesel combustion.

The diagrams in Fig. 8 display the comparison among the testedfuels in terms of cylinder pressure signal and in-cylinder heatrelease rate from combustion (HRR) (diagram A), and the

0

10

20

30

40

50

60

70

80

90

-50 -30 -10 10 30 50Engine crank angle [°]

Cyl

inde

r pre

ssur

e [b

ar]

-2

0

2

4

6

8

10

12

Rat

e of

hea

t rel

ease

from

com

bust

ion

[%/c

.a.]

DieselGEM10B40

(A)

Fig. 8. (A) Cylinder engine pressure and HRR traces; (B) regulated e

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.201

corresponding pollutant emissions and efficiencies (diagram B),at the test point 2000 � 10.

The comparison of both the engine cylinder pressure and HRRtraces in Fig. 8A evidences the good combustion quality of theGEM when blended in diesel fuel. No significant drift of the overallcombustion pattern was observed for GEM10 with respect to boththe neat diesel and B40. The comparison in Fig. 8A confirms thesuitability of the GEM for diesel fuel blending in terms of requiredcombustion characteristics.

In terms of emission performance, i.e. Indicated Specific FuelConsumption (ISFC) and fuel conversion efficiency (gfuel) displayedin the right plot of Fig. 8B, the comparison among the fuels givesquite expected results. Indeed, as evidenced in other literature data[20], the main effects of the oxy-fuels are on the unburned com-pound reduction (mainly HC and CO) and on the soot emissionsuppression. Generally, oxygenated fuels tend to increase NOx

emission, but in the present case no significant variations were ob-served among all the fuels. Such result is due to two factors: therelatively low oxygen content in the blend, and mainly the strongimpact of EGR on NOx formation that is widely used in the moderndiesel engines. Since it is the main driver of the NOx emissions con-trol in current diesel engines, its massive use makes less significantother control parameters, including fuel characteristics [16]. In thepresent case, an EGR rate of about 20% was set in the test point2000 � 10, masking possible slight effect of the GEM on NOx

formation.It is also noteworthy that notwithstanding the oxygen fuel

content reduces the LHV and increases the ISFC, the gfuel, calculatedas (see [21]):

gfuel ¼1

ISFC � LHVfuelð1Þ

remains almost equal for all the fuels. This result means that theISFC increment depends only on the reduced LHV of both GEM10and B40, but within the blending level tests in the present work,the oxygenated fuels have no penalties on the combustion processefficiencies.

These preliminary results have evidenced a positive behavior ofthe GEM mainly of exhaust soot emissions, without significantimmediate detrimental effects on the other pollutant species andengine performance in about twenty hours of running tests. Fur-ther research work is required for a ‘‘wide spectrum’’ analysis ofthese new fuels on the long term engine functionality.

0

10

20

30

40

50

HC NOx CO Soot x 100 ISFC/10 fuel

[Em

issi

ons:

g/k

g of

bur

ned

fuel

; ISF

C:

g

/kW

h; η

fuel

: %]

Diesel fuelGEM10B40

(B)

missions and efficiencies, for all the tested fuels at 2000 � 10.

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-2.08.006

Fig. 9. (A) LCA of GEM production at laboratory level: comparison between the total impact for the two processes using TBA and IB additives (Eco-indicator 99H/H); (B)comparison between LCA of the production and combustion of GEM/diesel blend (5%) and of neat diesel (Eco-indicator 99H/H), with reference to the same mass unit.

8 C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx

3.4. LCA results

In this first part of the project, LCA analysis was limited to theGEM production chain. For the sake of brevity, only the results ob-tained with the Eco-indicator 99 methodology are presented.

Fig. 9A shows the comparison between the impact (expressedin Eco-points) of the production of one mass unit of GEM fromtert-butyl alcohol (TBA) and isobutylene (IB) respectively. The re-sults show that the two processes are almost equivalent, sincethe absolute values of the two impacts are approximately thesame (+1.7% TBA vs. IB). The slight difference is mainly due tohigher values of damage of TBA in Ecosystem Quality category.Moreover, the results highlight for both addictives that the distil-lation process has the greatest impact due to fossil fuelconsumption.

Another LCA analysis was performed to compare the impact of a5% GEM/diesel blend (GEM5) with the one of neat diesel, consider-ing production and combustion processes. As far as the combustionproducts, GEM data are gathered from [22], while neat diesel dataare taken from Ecoinvent data base. The analysis was carried outwith reference to the same mass unit (kg).

Fig. 9B shows that the GEM blend has a lower impact thanneat diesel (�4.5%, expressed in Eco-points); in particular the dif-ferences are due to the Ecosystem Quality category (�27%, lowerimpacts for acidification and climate change), to the humanhealth category (�15%, mainly for the decreasing of effects ofthe inorganic respiration (�7%)), and to the Resources category(�3%, for the lowest employment of fossil fuels). The LCA analysiswas performed assuming that glycerine is a waste of the biodieselproduction chain; therefore the impacts due to land use and fossilfuel consumptions for the biomass cultivation were not includedin the system boundaries. The fact that GEM blend has a LHVlower than the one of neat diesel was taken into account in theLCA too.

The next steps of LCA will be focused on the use of raw glycerolin a oxy-reactor and on adding glycerol to a pyrolysis plant for syn-gas production; in both cases, the results will be compared withconventional combustion solutions.

4. Conclusion

The preliminary results of a co-operative research project aimedto the energetic exploitation of the biodiesel chain derived glycerolhave been discussed. It was outlined that:

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.20

� New Hyflon� based catalysts supported on spherical silica arepromising systems for carrying out the etherification reactionof glycerol with isobutylene, offering the possibility to operateat low temperature (70 �C) with total glycerol conversion to GEMand very low mono tert-butyl ether glycerol and di-isobutyleneformation.� The obtained GEM is effective for in-cylinder soot suppression

without significant impact on the in-cylinder combustion pro-cess and on the other regulated pollutant emissions (NOx, HCand CO).� LCA indicates that the GEM/diesel blends (both with TBA and IB

additives) generate in a combustion process a much lowerimpact than the neat diesel.

The described results represent a first global evaluation of theopportunity in the production and use a glycerol-derived syntheticin a diesel engine for automotive application. It is implied that fur-ther investigations are required to assess:

� Issues related to the use of raw glycerol in the production chain.� GEM compatibility with all required fuel standards.� GEM impact on engine reliability.� Accurate LCA based on a larger database.

Acknowledgments

We thank the Italian Ministry of Agriculture, Food and Forestryfor partially funding of this work through ‘‘TERVEG’’ Project (Na-tional funding call ‘‘bando bioenergetico DM 246/2007, GU No.94 del 27/11/2007).

References

[1] Sims R, Taylor M, Saddler J, Mabee W. From 1st to 2nd generation biofueltechnologies: an overview of current industry and RD&D activities.International Energy Agency. OECD/IEA, November 2008. <http://www.iea.org/papers/2008/2nd_Biofuel_Gen.pdf>.

[2] Directive 2009/28/EC of the European Parliament and of the council of 23 April2009 on the promotion of the use of energy from renewable sources andamending and subsequently repealing Directives 2001/77/EC and 2003/30/EC[Text with EEA relevance].

[3] Bohon MD, Metzger BA, Linak WP, King CJ, Roberts WL. Glycerol combustionand emissions. Proc Combust Inst 2011;33:2717–24.

[4] Valliyappan T, Bakhshi NN, Dalai AK. Pyrolysis of glycerol for the production ofhydrogen or syn gas. Bioresour Technol 2008;99:4476–83.

[5] Rahmat N, Abdullah AZ, Mohamed AR. Recent progress on innovative andpotential technologies for glycerol transformation into fuel additives: a criticalreview. Renew Sustain Energy Rev 2010;14:987–1000.

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-12.08.006

C. Beatrice et al. / Applied Energy xxx (2012) xxx–xxx 9

[6] Frusteri F, Arena F, Bonura G, Cannilla C, Spadaro L, Di Blasi. Catalyticetherification of glycerol by tert-butyl alcohol to produce oxygenated additivesfor diesel fuel. Appl Catalysis 2009;A 367:77–83.

[7] Melero JA, Vicente G, Morales G, Paniagua M, Moeno JM, Roldán R, et al. Acid-catalyzed etherification of bio-glycerol and isobutylene over sulfonicmesostructured silicas. Appl Cataysis 2008;A 346:44–51.

[8] Karinen RS, Krause AOI. New biocomponents from glycerol. Appl Catalysis2006;A 306:128–33.

[9] Deutsch J, Martin A, Lieske H. Investigations on heterogeneously catalysedcondensations of glycerol to cyclic acetals. J Catalysis 2007;245:428–35.

[10] Beatrice C, Avolio G, Guido C. Experimental analysis of the operatingparameter influence on the application of low temperature combustion inthe modern diesel engines, SAE technical paper 2007-01-1839; 2007.

[11] En ISO 14040, Environmental management – life cycle assessment – principlesand, framework; 2006.

[12] EN ISO 14044, Environmental management – life cycle assessment –requirements and guidelines; 2006.

[13] SimaPro 7.2, LCA software, PRé Consultants, <www.pre.nl/content/simapro-lca-software>.

[14] Althaus HJ et al. Implementation of life cycle impact methods, Data v 2.1(2009), Ecoinvent report no. 3, May 2009.

Please cite this article in press as: Beatrice C et al. Technologies for energetic exalytic conversion. Appl Energy (2012), http://dx.doi.org/10.1016/j.apenergy.201

[15] Beatrice C, Capaldi P, Del Giacomo N, Guido C, Mancaruso E, Vaglieco BM.Analysis of impact of diesel fuel/biodiesel blends on a modern dieselcombustion system performance by means of injection test rig, optical andreal SC engine experiments. SAE technical paper 2009-01-0484; 2009.

[16] Beatrice C, Guido C, Di Iorio S. Experimental analysis of alternative fuel impacton a new ‘torque-controlled’ light-duty diesel engine for passenger cars. Fuel2010;89:3278–86.

[17] AVL application note AT1007E, Rev. 02, 2005. <www.avl.com>.[18] Beatrice C, Bertoli C, Del Giacomo N. New findings on combustion behaviour of

oxigenated synthetic diesel fuels. Combust Sci Technol 1998;137:31–50.[19] Pepiot-Desjardins P, Pitsch H, Malhotra R, Kirby SR, Boehman AL. Structural

group analysis for soot reduction tendency of oxygenated fuels. CombustFlame 2008;154:191–205.

[20] Litzinger T, Stoner M, Hess H, Boehman A. Effects of oxygenated blendingcompounds on emissions from a turbo-charged direct injection diesel engine.Int J Engine Res 2000;1:1.

[21] Heywood JB. Internal combustion engine fundamentals. McGraw-Hill; 1989.ISBN-10: 0071004998.

[22] Jaecker-Voirol A, Durand I, Hillion G, Delfort B, Montagne X. Glycerin for newbiodiesel formulation. Oil Gas Sci Technol Rev IFP 2008;63(4):395–404.

ploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-2.08.006