Available online at www.sciencedirect.com

www.fuelfirst.com

Fuel 87 (2008) 1050–1062

Effect of catalysts in the quality of syngas and by-products obtainedby co-gasification of coal and wastes. 2: Heavy metals, sulphur

and halogen compounds abatement

Filomena Pinto *, Helena Lopes, Rui Neto Andre, I. Gulyurtlu, I. Cabrita

INETI-DEECA, Estrada do Paco do Lumiar, 22, 1649-038 Lisboa, Portugal

Received 7 May 2007; received in revised form 15 June 2007Available online 19 July 2007

Abstract

This paper analyses the formation of sulphur and halogens compounds during co-gasification of low grade coals with different typesof wastes that include: pine, petcoke and polyethylene (PE) with the aim of taking profit of waste energy value with the smallest possibleimpact on the environment. The influence of different types of catalysts or sorbents was studied: calcined dolomite, dolomite enrichedwith nickel, olivine, nickel and magnesium oxides, zinc oxide and cobalt and molybdenum oxides. The presence of dolomite led to thelowest HCl, HF and H2S concentrations in the syngas produced. ZnO also gave rise to significant decrease in H2S concentration, thoughhigher concentrations were obtained than those with dolomite. It was found that when catalysts or sorbents were used, a significant frac-tion of sulphur, halogens and heavy metals were trapped in the solid residue left behind in the bed and cyclone. Higher temperaturesincreased the volatility of some metals and sulphur. Leachability assays of solids showed that small quantities of SO2�

4 and Cl� couldbe released and most metals were not leachable, although the use of dolomite slightly increased metals leachability, because of the higheralkaline nature of residual solids produced.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Co-gasification; Coal; Sulphur compounds abatement; Halogen compounds abatement; Heavy metals abatement

1. Introduction

Most EU countries, like Portugal, are very much depen-dent on imported petroleum for energy production, todecrease this dependency, EU directives encourage theuse of renewable energies, alternative fuels and the devel-opment of new technologies to convert wastes to energy.These new technologies should be environmentallyfriendly. Most wastes usually contain nitrogen, sulphur,chlorine and metals contents higher than those present inbiomass and coal, hence a large amount of these elementspresent in the feedstock might pass to the gas phase. There-fore, the environmental implications of the utilisation ofwastes should be carefully evaluated, because, in many

0016-2361/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.fuel.2007.06.014

* Corresponding author. Tel.: +351 21 092 4786; fax: +351 21 716 6569.E-mail address: filomena.pinto@ineti.pt (F. Pinto).

countries the clean air legislation imposes emission limitsto minimise the impact on the environment. In addition,high concentrations of these elements in the fuel may causethe formation of compounds in the syngas that could giverise to corrosion and to deposition, hence increasing main-tenance costs.

Gasification could be a technology suitable to processwastes, as it appears to be more environmentally friendlythan combustion. Gasification technologies to produceelectricity with high thermal efficiency can be combined sat-isfactory with a gas-steam turbine cycle (IGCC), or a fuelcell (MCFC). As co-gasification is performed under reduc-ing conditions, the release of sulphur, nitrogen and chlorineinto the gas phase occurs mostly as H2S, NH3 and HCl, allconsidered pollutants. The gas leaving the gasifier willrequire cleaning to remove particulate matter, metals andundesirable gaseous pollutants, such as sulphur, nitrogen

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1051

and chlorine compounds, before its final use for combus-tion or in a fuel cell. Generally, when high sulphur coalsare gasified, the preliminary sulphur retention in the gas-ifier with the ash may be advantageous, as the gas pro-duced is cleaner. The effect of adding Ca-based sorbents,like limestone, dolomite or olivine has been studied by sev-eral authors, who have shown that they have a positiveeffect on sulphur retention [1–3].

Drift et al. [4] used a circulating fluidized-bed gasificationunit to study different biomass fuels: demolition wood (pureand mixed with sewage sludge and paper sludge), vergegrass, railroad ties, cacao shells and different woody fuels.When the feedstocks contained low sulphur concentrations,these authors measured H2S contents of about 230 ppmv inthe fuel gas obtained at 800 �C. When feedstocks presentedmeasurable sulphur contents, the percentage of H2S/S-fuelvaried between 39% and 50% [4]. On the other hand, Kura-mochi et al. [5] reported that due to the higher sulphur con-tents of sewage sludge, its gasification produced H2Scontents of around 4000 ppmv. In another study, Khanet al. [6] showed that sulphur distribution between char, tarsand the fuel gas depended on the form in which sulphur waspresent, as sulphur existed in both organic and inorganicforms in the fuel. The addition of CaO significantly reducedthe release of H2S into the gaseous phase, whilst the pres-ence of CaCO3 had relatively low influence, because of sig-nificant calcination of CaCO3 to CaO did not occur underthe conditions used by Khan et al. [6]. The presence ofMgO and SiO2 did not affect H2S formation, but the addi-tion of 10% of dolomite allowed great reductions of H2S inthe gaseous phase, usually higher than 90%.

The release of sulphur compounds both in tars and gasphase could be diminished and controlled by the use ofinorganic additives containing Ca, Fe, Mg and Si com-pounds [6]. Park et al. [7] also used additives based onmetal oxides containing Fe with Zn, Ni and Co for hotgas desulphurization and stated that it was difficult toremove sulphur contents completely from the syngas,because of several secondary reactions. The metal sulphides(MS) produced could react either with CO, which is amajor component of syngas, to produce COS or withH2O to produce SO2. Xu et al. [8] also reported that higherretention of sulphur in the solid phase could be achieved inthe presence of CaO, probably due to the formation ofCaS. This was also proven by Sage and Welford [1], as dur-ing co-gasification of sludge with coal, the sulphur reten-tion reached 96% in the presence of dolomite. Similarresults were obtained by Paterson et al. [2], who achievedaround 87% sulphur retention in the solid phase with theaddition of dolomite.

Several authors have reported that HCl contents in thegas obtained during gasification depended on several fac-tors, most important ones being the form of chlorine (inorganic or inorganic compounds) and its content in thefeedstock, gasification operating conditions: temperature,oxygen and steam contents and the presence of other ele-ments or compounds that might react with chlorine, lead-

ing to its retention in the solid phase. Drift et al. [4]gasified different biomass fuels in a circulating fluidized-bed reactor and reported that for most tests, less than15% of the fuel-bound chlorine was converted to HCl. Coalpyrolysis studies by Li et al. [9] showed that chlorine vola-tility in the presence of CaO was lower than that withoutCaO. The efficiency of slowing down chlorine volatilitywith the addition of CaO on chlorine volatility was greatlydependent on the pyrolysis temperature. Furthermore,these authors also stated that HCl formation was alsodependent on the content of other elements such as: Si,Al, K and heavy metals. Wei et al. [10] also predicted thatthe HCl release was also affected by Na content and by itsrelative concentrations in relation to other elements. Theinformation available so far has shown that the formationof chlorine compounds is a complicated issue that dependson so many factors, whose influence is not yet clear, there-fore, further research work is still needed.

The presence of heavy metals in fuels may present prob-lems [11–13]. Metals may remain in the solid phase, how-ever, some of them may get volatilized at relatively lowtemperatures, especially under the typical reducing condi-tions of the gasification processes [14–16]. The presenceof metals in the syngas produced by gasification is highlyundesirable, because of corrosive effects of these metalson the final end-use equipment, such as turbines and fuelcells and also because of emissions of polluting gases. Onthe other hand, the retention of metals and other contam-inants in the residual solid phase, being beneficial todecrease gas phase concentrations, may restrict its re-util-isation or may lead to increased landfill costs.

The aim of this work is to control the release of sulphurand halogen compounds to the syngas during co-gasifica-tion of coal blends with different types of wastes to improvethe process viability and to take profit of energy value ofwastes, whilst minimising emissions to atmosphere. Low-grade coals are usually difficult to gasify, but their negativeproperties, like high ash content and low volatile matter,could be counterbalanced by the presence of wastes, suchas pine, PE and petcoke. This paper mostly focuses onunderstanding the behaviour of different fuel blends duringco-gasification process, with the aim of decreasing sulphurand halogens compounds emissions, for which the effect ofadding different catalysts or sorbents to the gasifier wasalso studied.

2. Experimental part

Fig. 1 shows a schematic representation of the installa-tion used, which was previously described [17]. The gasifieris a bubbling fluidised bed electrically heated, circular incross-section with an inside diameter of 80 mm and witha height of 1500 mm. The gasifying/fluidising agent was amixture of steam and oxygen, introduced through a gasdistributor located at the bottom end of the reactor.

The fuel gas first goes through a cyclone to remove par-ticulates and then through a condensation system to retain

Fig. 1. Schematic diagram of INETI gasification installation.

1052 F. Pinto et al. / Fuel 87 (2008) 1050–1062

tars and condensable liquids. The gas was finally filtered,before being injected into CO and CO2 on-line analysers,used to monitor gasification process. The gas obtainedwas also collected in bags to be analysed by gas chromatog-raphy to determine CO, CO2, H2, N2, O2 and hydrocar-bons composition as previously described [17].

In Fig. 1, it may be also observed that there were sam-pling points for sulphur and halogens compounds analysis.The chlorine release as HCl was sampled using a method-ology based on method 26 of Environmental ProtectionAgency (EPA). Chlorine was retained in deionised waterand analysed by Capillary Ion Electrophoresis. Chlorineand fluorine retained in the condensation system were alsoanalyzed using Capillary Ion Electrophoresis. The sulphurrelease in the form of H2S was sampled using method 11 ofEPA. Sulphide was retained in an absorbing solution ofCdSO4 and then analyzed by iodometry. Sulphur held inthe condensation system was also analysed, as SO2�

4 , usingCapillary Ion Electrophoresis.

After each experiment, the residual solids, retained inthe fluidizing bed were first separated from the sand bysieving and then were analysed, as well as the solids cap-

tured in cyclones. The content of heavy metals in bed res-idues was analyzed after dissolution with HF, HNO3 andH3BO3 in a microwave oven. Then they were analysed onan Atomic Absorption Spectrometer. Hg, C and S in thesolid matrix were determined using automatic analyzers.Leachability of the residues was assessed in accordancewith EN 12457, using ultrapure water at L/S ratio of 10,without pH adjustment.

A high-ash coal from Puertollano mines in Spain waschosen for this work, which was mixed with pine, PE andpetcoke. In Table 1, elemental and proximate analysisand mineral composition of coal and wastes are given.The operating conditions used were the following: reactiontemperature �850 �C, steam flow rate �5.0 g/min, O2 flowrate �2.9 g/min, fuel flow rate – 6.0 g daf/min, particle sizeof coal, pine and petcoke �1250 to 2000 lm, PE particlesize 65 mm. In some experiments only silica sand (meanparticle size 350 lm) was used as bed material, but in oth-ers silica sand was mixed with a catalyst or sorbent in aproportion of 75/25 by weight. In Table 2, the characteris-tics of the different catalysts or sorbents tested are given.Apart from natural minerals like dolomite and olivine,

Table 1Feedstock elemental and proximate analysis

Puertollanocoal

Pine Petcoke PE

HHV (MJ/kg daf) 31.11 20.20 34.83 46.1Proximate analysis

Fixed carbon (% w/w) 28.9 13.6 85.9 –Volatiles (% w/w) 20.6 74.5 13.2 –Ash (% w/w) 42.5 0.3 0.4 –Moisture (% w/w) 8.0 11.6 0.5 –Elemental analysis

Carbon content (% daf) 77.8 50.6 87.2 85.7Hydrogen cont. (% daf) 4.7 6.4 4.0 14.3Sulphur content (% daf) 2.4 0.2 6.2 –Nitrogen content (% daf) 1.4 0.2 1.4 –Chlorine content (% daf) 0.06 0.07 <0.02Oxygen content (% daf) (estimated

by difference)13.7 42.5 1.2 –

Ash forming elements (as received):

Calcium (% w/w) 0.53 0.113 0.004 –Aluminium (% w/w) 4.5 0.002 0.01 –Iron (% w/w) 2.5 0.033 0.01 –Sodium (% w/w) 0.11 0.024 0.11 –Potassium (% w/w) 0.68 0.089 0.0007 –Magnesium (% w/w) 0.26 0.040 0.002 –Manganese (mg/kg) 246 32 1.0 –Lead (mg/kg) 196 0.2 0.7 –Cadmium (mg/kg) <5.0 <0.2 <0.5 –Chromium (mg/kg) 46 1.1 1.3 –Copper (mg/kg) 40 0.9 0.5 –Zinc (mg/kg) 329 2.6 6.5 –Fluorine (mg/kg) <0.02 <0.02 <0.02 –Nickel (mg/kg) 42 1.0 262 –Mercury (mg/kg) 0.334 0.021 0.013 –

Table 2Characteristics of the catalysts or sorbents used

Sbet

(m2 g�1)aVp

(cm3 g�1)bXRD analysis

Calcineddolomite

12.3 4.30 · 10�5 Magnesium oxide carbonate –Mg3O(CO3)2

Lime – CaOPortlandite – Ca(OH)2

Olivine 9.8 8.64 · 10�5 Forsterite – Mg2(SiO4)Iron oxide hidroxide –FeO(OH)Forsterite – (Mg,Fe)2SiO4

Talc-2M – Mg3SiO4O10(OH)2

Nacrite-1Md – Al2Si2O5(OH)4

Calcinedolivine

4.7 9.51 · 10�5 Forsterite – Mg2(SiO4)Forsterite Ferroan –(Mg1.44Fe0.56)(SiO4)Enstatite – MgSiO3

Ni–Mg 45 – Nickel aluminum oxide –NiAl2O4

Magnesia – MgOBunsenite – NiO

Ni–dolomite

33.2 2.87 · 10�4 Bunsenite – NiOLime – CaOPortlandite – Ca(OH)2

G-72D 27.5 5.32 · 10�4 Zincite – ZnONordstrandite – Al(OH)3

C 49-TRX Cobalt oxide – Co3O4

Molybdenum oxide – Mo3OAluminum – Al2O3

a Sbet – BET surface area determined at P/Po from 0.5 to 0.3.b Vp – total pore volume determined at P/Po a 1 · 10�5.

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1053

two commercial catalysts, G-72D and C49 TRX, were alsoused. While all the other catalysts were used as small par-ticles, both G 72D and C49 TRX were used as pellets,which also fluidise. The pellets of the first catalyst had adiameter of 5 mm and a maximum length of 12 mm. C49TRX pellets were 2–3 mm in triangular section and had amaximum length of 9 mm. A catalyst prepared in accor-dance with the description of Garcia et al. [18] and referredto as Ni–Mg was also tested, as well as a dolomite enrichedwith nickel, named as Ni–dolomite, which was prepared bythe method described by Wang et al. [19].

3. Results and discussion

No significant changes to the existing coal gasificationfacility were required to co-gasify blends of coal with thewastes presented in Table 1, as shown by the experimentalwork done so far. However, the feeding system of theexperimental gasifier was water cooled to avoid the clog-ging of the feeder caused by pyrolysis of some wastes, likePE, prior to their entry to the gasifier. The nitrogen flowused in the feeding system also helped to ensure smoothflow of solid material. The presence of wastes was shownto be advantageous to co-gasify low-grade coals. Some ofthe wastes used counterbalanced some of the negative

effects of low grade coals, such as low volatile contentsand high ash and sulphur contents. However, petcoke alsopresented high sulphur content and low volatile content.By adjusting the composition of the blends, it was possibleto take advantage of them, regarding char reactivity andgaseous pollutant emissions.

The reproducibility of experimental results was checkedby repeating the same test run under equal experimentalconditions several times. When variations higher than 5%in experimental determinations were observed, more testswere carried out to check the results and to make sure thatthe degree of uncertainty was lower than 5%.

3.1. Sulphur compounds abatement

Sulphur release into the gas phase was analysed as H2S.Sulphur retained in the condensation system, analysed asSO2�

4 , presented values always lower than 200 ppmv. SO2

was also analysed in the gas leaving the condensation sys-tem, but the presence of this compound was found to betoo low to be detected. In Fig. 2, H2S and SO2�

4 concentra-tions obtained for different coal and waste blends are pre-sented. The percentage of the total sulphur released intothe gas phase in relation to the sulphur amount in the inputfuel, referred as S(g + c) is also shown. Though sulphur

0

1000

2000

3000

100%Coal

80%Coal 20%PE

80%Coal 20%Pine

80%Coal 10%Pine 10%PE

75%Coal 10%Pine 10%PE 5%Petc

60%Coal 20%Pine 20%PE

55%Coal 20%Pine 20%PE 5%Petc

0

20

40

60

80

100H2S gas(ppmv) SO4 cond. (ppmv) S (g+c) %

H2S

, SO

42-

(ppm

v)

S (g

as +

con

d.) (

%)

Fig. 2. Effect of waste type and content on H2S formation during co-gasification of coal mixed with wastes at 850 �C. Other experimentalconditions were: steam flow rate – 5 g/min, feedstock flow rate – 6 g daf/min and O2/fuel ratio – 0.58 g/g daf.

0

1000

2000

3000

No Catalyst

DolomiteOlivine

Olivine Calcin.

G-72 D

C49 TRX

0

20

40

60

80

100

H2S gas(ppmv) SO4 cond. (ppmv) S (g+c) %

H2S

, SO

42- (

ppm

v)

S (g

as +

con

d.)

(%)

Fig. 3. Effect of catalyst or sorbent type on H2S formation duringgasification of coal at 850 �C. Other experimental conditions as shown inFig. 2.

0

1000

2000

3000

No Catalyst

DolomiteOlivine

Olivine Calcin.

NiMg synt.

Ni-Dolomite

G-72 D

C49 TRX

0

20

40

60

80

100

H2 S gas(ppmv) SO4 cond. (ppmv) S (g+c) %H

2S, S

O42-

(pp

mv)

S (g

as +

con

d.)(

%)

Fig. 4. Effect of catalyst or sorbent type on H2S formation during co-gasification at 850 �C of 80% of coal mixed with 20% of PE. Otherexperimental conditions as shown in Fig. 2.

0

1000

2000

3000

No Catalyst

DolomiteOlivine

Olivine Calcin.

NiMg synt.

Ni-Dolomite

G-72 D

C49 TRX

0

20

40

60

80

100

H2 S gas(ppmv) SO4 co nd. (ppmv) S (g+c) %

H2S

, SO

42 -

(ppm

v)

S (g

as +

con

d.) (

%)

Fig. 5. Effect of catalyst or sorbent type on H2S formation during co-gasification at 850 �C of 75% of coal mixed with 10% of pine, 10% of PEand 5% of (w/w) of petcoke.

1054 F. Pinto et al. / Fuel 87 (2008) 1050–1062

retention in the solid phase depended on sulphur content offuel, the sulphur release to the gas phase varied between20% and 30%, therefore retentions higher than 70% wereusually obtained, even without the addition of any catalystsor sorbents.

As may be observed in Fig. 2, H2S release was depen-dent on sulphur content in the feedstock and its rise inthe blend resulted in greater H2S formation, being thechanges proportional to the amount of sulphur input.Probably, due to the high sulphur content in coal, H2S con-tents of 2700 ppmv were obtained during coal gasification.Pine sulphur content was much lower than that of coal, asmay be observed in Table 1, therefore, gaseous sulphurrelease during co-gasification of coal mixed with 20% (w/w) of pine was 36% lower than that obtained when onlycoal was gasified. As the presence of sulphur was notdetected in PE wastes, co-gasification of coal mixed with20% (w/w) of PE led to an even lower H2S release, around1560 ppmv. The sulphur release into the gas phase, whencoal was co-gasified with 10% of PE and 10% (w/w) of pinelaid between those obtained when only one of these wasteswas blended with coal. Due to the high sulphur content ofpetcoke, as shown in Table 1, its presence in the blendsresulted in higher sulphur release into the gas phase, asboth H2S and condensed SO2�

4 contents increased. For thisfeedstock composition H2S concentration was 1920 ppmv,against 1611 ppmv, when the same wastes were co-gasified,with the exception of petcoke.

Experimental results were coherent and agreed withthose reported by several authors, even if different fuelswere studied [4–6]. However, Sciazko et al. [20] reportedlower sulphur retentions in the solid phase, as 60% of thetotal sulphur was released into the gas phase, with theremainder being retained in the char, this could beexplained by different mineral matter compositions.

In Figs. 3–5 it may be analysed the effect of different sor-bents or catalysts in sulphur compounds content in fuel

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1055

gas. Ni–dolomite and Ni–Mg were only used with coal andwastes blends, because they are more effective in reducingtars and hydrocarbons contents, which were formed inhigher amounts in presence of wastes. Figs. 3–5 show thatthe presence of dolomite allowed substantial reduction inthe gaseous sulphur release. The lowest concentrations inthe condensation system, as SO2�

4 , were obtained and thesulphur transported by the gas, as H2S was also found tobe low. Consequently, it was in the presence of dolomitethat the highest sulphur retention in the solid phase wasobtained, being the percentage of gaseous sulphur lowerthan 10% for all the feedstocks studied.

In Fig. 5 it is observed that when using dolomite, a lowerdecrease of H2S gas than those of Figs. 3 and 4 wasobtained. Fig. 5 refers to a blend with petcoke, it isexpected that much sulphur in petcoke is in an organicform, as it is a residue from petroleum industry, while incoal and due to the presence of pyrite, some of it is in aninorganic form. Therefore, different interactions with dolo-mite are expected.

The dolomite was added to the gasification bed alreadycalcined, therefore, its main component was CaO, whichcould have reacted with H2S to form CaS that remainedin the gasification bed, according to reaction (1), M beingCa

MOþH2S ¡ MSþH2O ð1Þ

When Ni–dolomite was used, low sulphur releases into thegas phase were obtained, though higher than thoseachieved when only dolomite was added to the gasificationmedium, probably due to the lower CaO amount in the Ni–dolomite. The reduction of H2S in presence of dolomitevaried between 60% and 80%, depending on sulphur con-tent in the fuel, whilst those obtained in presence of Ni–dolomite was around 40%. The presence of this sorbentled to gaseous sulphur releases of about 15% (of sulphurin the feed).

In Fig. 5, Ni–dolomite achieves higher reduction in H2Sgas than dolomite, probably due to the presence of nickel,tars and other hydrocarbons were easily converted intosmaller molecules, which might have facilitated the releaseof sulphur and the formation of H2S. Therefore, the reac-tion between H2S and CaO from dolomite would be alsoeasier. However, this is only a hypothesis that needs tobe further confirmed.

On the other hand, the addition of Ni–Mg to the gasifi-cation bed did not affect the H2S content in the gas, nor theSO2�

4 concentration and hence sulphur retention in thesolid phase was similar to those observed in the absenceof catalyst. As the Ni–Mg was placed inside the gasificationbed and due to the high sulphur content of same fuels, itwas also possible that there was some sulphur poisoning.On the other hand, as reported by several authors, nickelcatalysts are better suited for reducing tars and nitrogencompounds than for sulphur compounds reduction [3].

The presence of natural olivine in the gasification bed didnot lead to great reductions in H2S content in the fuel gas, as

only very small reductions were observed, for all fuels stud-ied. Olivine is usually represented as (Mg, Fe) Æ 2SiO4, there-fore, it could be expected that the presence of olivine couldhave increased sulphur retention in the solid phase. In fact,Najjar et al. [21] reported that sulphur capture was signifi-cantly increased when iron compounds were added to coalash, due to the formation of a new major oxysulphide phasein addition to the silicate one. However, the presence of nat-ural olivine did not lead to significant sulphur retentions inthe solid phase, probably, this process did not occur to a sig-nificant extent and/or the iron content in olivine was nothigh enough to achieve further sulphur capture.

On the other hand, the use of calcined olivine was foundto bring about lower H2S concentrations in the gasobtained and greater sulphur retentions in the char thanthose achieved in presence of natural olivine, probablydue to the calcination process that converted differentmetallic compounds into their oxides. These oxides seemto have favoured sulphur retention, as metallic sulphides,MS, according to reaction (1), M being mainly Ca, Feand Mg. The presence of calcined olivine led to sulphurreleases to the gas phase of about 25%, as may be observedin Figs. 3–5.

The addition of C49 TRX to the gasification medium ledto small reductions in gaseous sulphur release. The reduc-tion in H2S content was around 15%, probably, becausepotential participation of cobalt and molybdenum oxidesin reaction (1) was not as effective as those observed forother metals, such as Zn, present in the G 72D catalyst.

The presence of G 72D catalyst was found to give rise toa reduction of H2S in the fuel gases of around 30%, in com-parison with to the absence of catalyst. Therefore, the sul-phur content in the gas phase was higher than thoseobtained when dolomite was added to the gasification med-ium, either as calcined dolomite, or as Ni–dolomite. Thepercentage of gaseous sulphur release was around 20%,probably due to sulphur retention in the solid phase byreaction (1), M being now Zn.

Another important issue is that, both C49 TRX and G72D were used as pellets, while all the other catalysts wereused as small particles. Therefore, it is possible that due tocoke deposition in the catalyst pellets, they were deacti-vated faster than other catalysts.

As reported by Tijmensen et al. [22], further sulphurreductions could be achieved by using adsorbents contain-ing Fe, Zn and Mn. These metals or their oxides have thestrongest affinity for H2S, but oxides might be less resistantto the reaction by means of H2 or CO as compared withalkaline earth oxides. The removal of H2S by means ofFe2O3, or Fe3O4 may be described as follows:

Fe3O4 þH2S ¡ FeSþ Fe2O3 þH2O ð2Þ2Fe2O3 þH2Sþ 2H2 ¡ FeSþ Fe3O4 þ 2H2O ð3Þ2Fe2O3 þH2Sþ 2CO ¡ FeSþ Fe3O4 þ 2CO2 ð4Þ

Reactions (2)–(4) may be repeated until all the iron oxidesare consumed. The presence of steam and CO2 in hot gases

0

25

50

75

100

100%Coal

80%Coal 20%Pine

80%Coal 20%PE

80%Coal 10%Pine 10%PE

75%Coal 10%Pine 10%PE 5%Petc

60%Coal 20%Pine 20%PE

55%Coal 20%Pine 20%PE 5%Petc0

20

40

60

80

100

Cl- (ppmv) F- (ppmv) Cl- (%)

Cl- co

nden

s.,

F- c

onde

ns. (

ppm

v)

Cl- ,

F-(%

)

Fig. 6. Effect of waste type and content on Cl� and F� formation duringco-gasification of coal mixed with wastes at 850 �C.

0

25

50

75

100

No Catalyst

DolomiteOlivine

Olivine Calcin.

G-72 D (ZnO)

C49 TRX (CoO+MoO3)0

20

40

60

80

100

Cl- (ppmv) F- (ppmv) Cl- (%)

Cl- co

nden

s.,

F- con

dens

. (pp

mv)

Cl- ,

F-(%

)

Fig. 7. Effect of catalyst or sorbent type on Cl� and F� formation duringgasification of coal at 850 �C. Other experimental conditions as shown inFig. 2.

0

25

50

75

100

NoCatalyst

DolomiteOlivinea

Olivine Clcin.

Ni-Mg

Ni-Dolomite

G-72 D (ZnO)

C49 TRX (CoO+MoO3)

0

20

40

60

80

100

Cl- (ppmv) F- (ppmv) Cl- (%)

Cl- co

nden

s.,

F- con

dens

. (pp

mv)

Cl- ,

F-(%

)

Fig. 8. Effect of catalyst or sorbent type on Cl� and F� formation duringco-gasification at 850 �C of 80% of coal mixed with 20% of PE. Otherexperimental conditions as shown in Fig. 2.

1056 F. Pinto et al. / Fuel 87 (2008) 1050–1062

affects sulphur retention, as they might reverse the previousreactions. It is due to this fact that zinc ferrites (ZnO Æ x-

Fe2O3) and zinc titanates (ZnO Æ xTiO2) have shown tobe more strong sorbents for desulphurisation [23].

Some experiments were done at a higher gasificationtemperature, 900 �C, both without and in presence of dolo-mite using a blend of coal with 10% of pine, 10% of PE and5% of (w/w) of petcoke. The increase of temperature in theabsence of dolomite was disadvantageous, as it led to anincrease in H2S formation of about 15%. The increase oftemperature also led to a small decrease in SO2�

4 concentra-tion. Therefore, in the absence of dolomite the rise in tem-perature also allowed a decrease in sulphur retention inthe solid phase, these results agree with others in the litera-ture [5]. On the other hand, the increase in temperature inpresence of dolomite led to a reduction in H2S content ofaround 20%. Probably, the increase of temperaturefavoured reaction (1) and sulphur capture in the solidphase, due to the formation of higher amounts of CaS.Due to the small variation on SO2�

4 concentration, total sul-phur released to the gas phase decreased from 16% to 7%.

The effect of increasing steam flow rate on the formationof H2S was also studied. Steam/fuel ratio rise from 0.8 to1.8 led to an increase in H2S content in the gas leavingthe condensation system of around 37%, which was accom-panied by a decrease in SO2�

4 concentration of around38%. When studying the effect of steam flow rate it wasobserved that H2S concentration in the syngas was inver-sely related with the sulphur retention in the form ofSO2�

4 in the condensation system, which agrees with theresults reported by Khan et al. [6]. Due the opposite varia-tions in H2S and SO2�

4 concentrations, no significantchanges were detected in sulphur retention in char. In theabsence of catalyst or sorbent the percentage of sulphurreleased to the gaseous phase was around 25% over theexperimental conditions used.

3.2. Halogen compounds abatement

As may be observed in Table 1, coal and pine had thehighest chlorine contents, being followed by petcoke,whose chlorine content, being 0.02 (%daf), was 70% lower.In Fig. 6, it is shown that the chlorine concentrationreleased into the gas phase, presented in ppmv, as well asthe percentage of the chlorine released to the gas in relationto those in the fuel, referred to as Cl� (%). It may beobserved that fuels with higher chlorine contents also ledto its greater release into the gas phase. Therefore, the low-est chlorine concentration was obtained when 20% (w/w)of PE was mixed with coal, as no chlorine content wasdetected in this waste. As shown in Fig. 6, most chlorinepresented in the fuel was released to the gas phase andretained in the condensation system, as the percentage ofchlorine released with regards to chlorine input in the fuelwas about 100% for most experiments.

In Figs. 7–9 it may be observed that the effect of thepresence of catalysts or sorbents in gaseous HCl formation.

0

25

50

75

100

No Catalyst

DolomiteOlivine

Olivine Calcin.Ni-M

g

Ni-Dolomite

G-72 D (ZnO)

C49 TRX (CoO+MoO3)

0

20

40

60

80

100

Cl- (ppmv) F- (ppmv) Cl- (%)

Cl- co

nden

s.,

F- con

dens

. (pp

mv)

Cl- ,

F-(%

)

Fig. 9. Effect of catalyst or sorbent type on Cl� and F� formation duringco-gasification at 850 �C of 75% of coal mixed with 10% of pine, 10% ofPE and 5% of (w/w) of petcoke.

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1057

The lowest HCl release to the gas phase was obtained withdolomite, as reductions varied between 54% and 64%,depending on fuel composition. Therefore, the use of thissorbent allowed higher chlorine retentions in the solidphase, as the percentage of chlorine released to the gaseousphase varied between 30% and 35%.

The use of Ni–dolomite during co-gasification of coalmixed with 20% (w/w) of PE or with 20% of PE, 20% ofpine and 5% (w/w) of petcoke led to reductions in HCl con-centrations of about 30%, which corresponded to higherchlorine retentions in the solid phase. The percentage ofchlorine released to the gaseous phase was about 53%,against the 30% obtained when only dolomite was used,or the almost 100% when Ni–Mg was added to the gasifi-cation medium. In fact, the presence of Ni–Mg did not leadto any significant change in HCl content released to the gasphase, in comparison with the values obtained when nocatalyst or sorbent was used. Therefore, the changesobserved in chlorine capture when Ni–dolomite was used,were probably due to dolomite.

The presence of CaO and Ca(OH)2 in dolomite, asshown in Table 2, could have reacted with HCl, accordingto reactions (5) and (6), thus retaining it in the solids, asCaCl2. On the other hand, CaCl2 starts to decompose at740 �C, however, as gasification experiments were carriedout at 850 �C, some of the CaCl2 formed by reactions (5)and (6) might have released some HCl, which could thenhave caused a decrease in chlorine capture in the solidphase

CaOþ 2HCl ¡ CaCl2 þH2O ð5ÞCaðOHÞ2 þ 2HCl ¡ CaCl2 þ 2H2O ð6Þ

At the gasification temperature used, the utilisation ofsteam as gasification medium might also have reversedreactions (5) and (6), which could explain why no higherchlorine retentions were obtained.

HCl formation and the direction of reaction (5) could beinfluenced by the content of other elements such as: Si, Al,

alkaline earth metals and heavy metals, as they might par-ticipate in reactions to retain Cl� in the solid phase andcould consequently compete with CaO to react with chlo-rine ion. On the other hand, at gasification temperatures,the presence of high K content in the feedstock could pro-duce gaseous KCl instead of HCl, but this depended on thecontent of other elements and the competition among theseto react with chlorine ion [24].

CaC12 has a melting point at 772 �C, and the lowestliquid temperature for the CaCl2–CaO system is 750 �Cwith a eutectic composition of about 6% (mol) CaO, asreported by Weinell et al. [24]. Therefore, a liquid phaseof CaCl2 saturated with CaO might be formed at the gasi-fication temperature of 850 �C, which could have reducedthe sorption capacity of CaO. Also, the sorption capacityof CaO might also have been diminished by the reversibil-ity of reaction (5) at high temperature, causing the releaseof HCl.

Apart from chemical reactions influencing, chlorineretention in the solid phase, physical properties of solids,namely particle size and specific area could also contributeto the level of chlorine retention. According to Weinellet al. [24], the surface area only exerted a minor influenceon the final conversion of HCl at high temperatures andfor SBET > 12 m2/g, whilst chlorine retention in the solidphase was independent on particle size. In fact, there aremany different phenomena occurring during gasificationand some of them act in opposing directions.

Increasing the gasification temperature from 850 to900 �C, led to higher chlorine release into the gaseousphase, independently of the feedstock blend, probablydue to CaCl2 decomposition, which starts at 740 �C. Thistendency was also predicted by Wei et al. [10] for somefeedstocks. However, during sewage sludge gasification,HCl did not vary much with temperature, when sodiumcontent was greater than that of potassium. Due to highcontents of silicon and aluminium in sludge, sodium wasretained in ash as aluminosilicate and was only likely tobe released at high temperatures, higher than 1027 �C [10].

The presence of olivine, either natural or calcined, didnot lead to great changes in HCl release into the gas phasefor all the fuels tested, as may be observed in Figs. 7–9.Only when natural olivine was added to the gasificationmedium a slight decrease in HCl content was observed,which varied between 7% and 13%, these values were onlyslightly higher than the deviation commonly observed inexperimental results. This reduction in HCl content wasprobably due to the reaction between FeO(OH) and HCl,according to reaction (7). As may be observed in Table 2,the presence of FeO(OH) was detected in natural olivine,but not in calcined olivine, probably due to its thermaldestruction during the calcination process and its conver-sion into iron oxide and further formation of silicates

FeOðOHÞ þ 3HCl ¡ FeCl3 þ 2H2O ð7Þ

Consequently, the presence of natural olivine only led tosmall chlorine capture in the solid phase. The percentage

1058 F. Pinto et al. / Fuel 87 (2008) 1050–1062

of chlorine released to the gaseous phase was found to bearound 91%.

The presence of G-72 D in the bed material of the flui-dised bed gasifier was observed to promote HCl reductionof around 37% in the gas phase, varying HCl concentra-tions between 44 and 25 ppmv, depending on chlorine con-tent in the fuel. The lower HCl release meant higherchlorine retentions in the solid phase. When coal was co-gasified both with 20% (w/w) of PE and with 20% of PE,20% of pine and 5% (w/w) of petcoke, 53% of chlorinewas found to be released to the gas phase. During gasifica-tion of only coal, chlorine content in the gas phase washigher, about 65%, probably due to the higher chlorinecontent in coal and consequently leading to higher forma-tion of HCl.

ZnO is present in G-72 D, therefore it might havereacted with HCl, according to reaction (8). However, atthe gasification temperature ZnCl2 was liquid and due toits significant vapour pressure, it could have escaped fromthe sorbent. On the other hand, the presence of steam in thegasification medium could also have affected the equilib-rium of reaction (8) and chlorine retention

ZnOþ 2HCl ¡ ZnCl2 þH2O ð8Þ

On the other hand, due to the presence of H2S in the reac-tion medium, it was also possible that some of the ZnCl2formed in accordance with reaction (8) might also have re-acted with H2S, according to reaction (9). However, Guptaet al. [25] calculated the Gibbs free energy and equilibriumconstant values for reaction (9) at various temperaturesand found that the relatively high equilibrium constant val-ues, strongly favoured ZnCl2 reduction by reaction withH2S to form ZnS, allowing the formation of more ZnCl2.Gupta et al. [25] reported that the formation of ZnCl2was favoured at a temperature lower than 650 �C.

ZnCl2 þH2S ¡ ZnSþ 2HCl ð9ÞHowever, at gasification temperature range used in thisstudy, ZnCl2 is most likely in a liquid phase, which couldmake it more accessible for reactions with H2S than ZnO,which is a solid. Therefore, the presence of HCl in gasifica-tion reaction medium may lead to higher ZnCl2 concentra-tions and hence to more efficient removal of the H2S byreaction (9). According to Park et al. [26] ZnO could alsoreact with fuel gas produced through reactions (10) and(11), consequently the metallic zinc gradually diffused tothe surface and vaporized during the sulfidation/regenera-tion cycles, for this reason a catalyst composed of nearlypure ZnO, exhibited excessive zinc vaporization. Thesereactions can be inhibited by the addition of a natural zeo-lite to ZnO as a structural stabilizer. To improve ZnO per-formance in multiple tests some authors have added someinorganic oxides, like TiO2 to increase ZnO stability,mechanical strength and capacity of retention [27].

ZnOþH2 ! ZnþH2O ð10ÞZnOþ CO! ZnCþO2 ð11Þ

As may be observed in Figs. 7–9, no significant changeswere detected in both gaseous HCl content and in chlorineretention in the solid phase when C49 TRX was used ascatalyst. The main compounds present in this catalyst,CoO and MoO did not seem to play any major role inretaining HCl in the solid phase. It should be pointed outthat C49 TRX was added directly to the gasifier bed andhence could have suffered from deactivation due to thedeposition of char particles on its surface. In future work,this and other catalysts or sorbents will be tested in a sec-ondary reactor, following the gasifier, for syngas treatment,which is expected to improve catalyst performance.

Although fluorine concentrations in the feedstock werevery low, below the detection limit of the method used,the fluorine content in the condensate was analysed usingCapillary Ion Electrophoresis, as described in the experi-mental part. Fluorine was detected in low contents in thecondensate, probably due to the gradual accumulation offluorine over the sampling period. The amount of fluorineretained increased with the length of sampling time. InFig. 6, it is shown that the fluorine concentration releasedinto the gas phase, presented in ppmv. It was not possibleto calculate the percentage of the fluorine released into thegas phase, because the corresponding values in the feed-stock were below the detection limit of the method used.

Due to the very low fluorine contents in both coal andwastes, fluorine concentration in the condensation systemwas usually very low, varying between 22 and 43 ppmv,depending on fuel fluorine content. As experienced withchlorine, most fluorine was also released to the gas phase,hence, fluorine capture in the solid phase was usually verylow, probably almost zero.

Fig. 6 shows the fluorine levels obtained with differentcoal and waste blends. Although, all feedstocks presentedfluorine contents below the detection limit, a close examina-tion of Fig. 6 shows that when higher contents of either coalor pine wastes were used, specially the last one, higher fluo-rine contents were detected. This suggests that coal and pinemight have significant fluorine contents. For the mixturewith 10% of pine, 10% of PE and 5% (w/w) of petcoke itwas also detected an increase in fluorine release, this wouldsuggest that petcoke also had a high fluorine content, whichalso agrees with the results obtained when the blend with20% of pine, 20% of PE and 5% of petcoke was used.

The results presented in Figs. 7–9 show that for all thefeedstocks studied, fluorine was never detected in the con-densation system when dolomite or Ni–dolomite wereadded to the gasification medium, probably due to HFreacting with CaO through reaction (12) and (13), leadingto the formation of CaF2, which was probably retainedin the char

CaOþ 2HF ¡ CaF2 þH2O ð12ÞCaðOHÞ2 þ 2HF ¡ CaF2 þ 2H2O ð13Þ

These results agree with those reported by Weinell et al.[24], who stated that HF, HBr and HI could participate

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1059

in the same kind of reactions as HCl, being CaO,(Ca(OH)2), together with limestone (CaCO3), nahcolite(NaHCO3), soda (Na2CO3) or trona ðNa2CO�3NaHCO3Þthe most suitable sorbents for the control and retentionof halogens.

3.3. The solid residue

The gasification process produced solid residue, accu-mulated in the bed, mixed with the fluidizing sand and cat-alysts or sorbents used. In addition, smaller amounts ofparticulate matter were collected in the hot cyclone, inwhich the temperature was about 600 �C. Due to the highash content of coal, the fraction of solid material remainingin the bed varied from 30% to 45%, excluding sand, of theinlet fuels. For the runs without catalyst or sorbent the finematerial collected in the cyclone represented 10–15% of thetotal solid residue produced. These solids contained signif-icant amounts of unconverted carbon, varying over a rangeof 20–47% (w/w), and sulphur which varied from 0.5% to2.4%. The higher values of sulphur corresponded to thoseobtained when petcoke was used. Global retention of sul-phur in solid residues increased with the substitution ofcoal from 24% to 63% at 850 �C. For the co-gasificationtest at 900 �C, sulphur retention decreased to 29%. Heavymetals were also partially retained in the solid residues.Coal had the highest levels of heavy metals, except for Ni

Ni

01 00 200 300 400

bedcycbedcycbedcycbedcycbedcycbedcyc

AE

mg/kg

Pb

0 200 400 600 800

bedcycbedcycbedcycbedcycbedcycbedcyc

mg/kg

A(100%Coal) B(80%Coal20%Pine) C(80%Coal20E(55%Coal20%Pine20%PE5% Petcoke) F(75%

DF

CB

AE

DF

CB

Fig. 10. Concentrations of Ni, Zn, Pb and Hg in residu

that was greater in petcoke. In Fig. 10, the analysis of somemetals in bed and cyclone collections is presented for thecase of tests without catalyst. Tests with catalysts or sor-bents were affected by metal concentration originatingfrom the catalyst; therefore, it was difficult to identify theirsource. In the case of tests with only coal, metals concen-trations in the bed and in the cyclone were found to be sim-ilar, except for Hg.

When PE, pine or petcoke were added there were foundconsiderable increases in Ni and Pb concentrations in thecyclones. For the case of tests with petcoke, the increaseof Ni may be due to its higher content in the fuel. Althougha different partitioning of these metals between residuescollected in the bed and in the cyclone was observed, theglobal amounts retained in the solid phase appeared tobe similar in all tests (without catalyst), except at the highertemperature (900 �C). Metal retentions in solid residues at850 �C were as follows: 72 ± 13% for Cr, 74 ± 17% for Zn,51 ± 4% for Ni, 43 ± 14% for Mn and 42 ± 9% for Pb.

The effect of temperature may be analyzed through thecomparison of the tests carried out with the same fuelblend; D, for 850 �C and F, at 900 �C (white bars in thegraphics). For a relatively small increase of temperature amore pronounced partitioning of Hg, Pb, Zn, Ni, Mnand Cu was observed. Except for Hg, concentrations ofall metals decreased in the bed and increased in the cycloneat the higher temperature. It was verified that this was not a

Zn

0 1000 2000 3000 4000

bedcycbedcycbedcycbedcycbedcycbedcyc

mg/kg

Hg

0.0 0.2 0.4 0.6 0.8

bedcycbedcycbedcycbedcycbedcycbedcyc

mg/kg

%PE) D(75%Coal10%Pine10%PE5%Petcoke)Coal10%Pine10%PE5%Petcoke (900ºC))

AE

DF

CB

AE

DF

CB

al solids (bed and cyclone) of the gasification tests.

1060 F. Pinto et al. / Fuel 87 (2008) 1050–1062

dilution effect of concentrations, as the level of unburnedcarbon was similar in solids of both tests. The total metalfraction retained in all solid streams was found to decreaseat the higher temperature. The retention of Mn and Nidecreased 6%, Pb decreased 19% and Zn 39%, comparingwith the results obtained at 850 �C.

In the case of Hg, much smaller concentrations werefound in the bed than in the cyclone. Probably due to thehigher volatility of this metal, it was volatilized from thebed char and captured by particles collected in the cyclonecontaining unburned carbon (20–47%) and sulphur (0.5–

Table 3Leachability of residual solids (bed and cyclone) of the gasification tests

pH Cond.mS/cm

Cl�

(mg/kg)SO2�

4

(mg/kg)Cd(mg/k

A-Bed 11.3 0.54 37 466 <0.2A-Cyc 11.0 0.41 91 772 <0.2B-Bed 11.3 0.50 17 434 <0.2B-Cyc 9.9 0.28 101 1066 <0.2C-Bed 11.5 0.62 18 257 <0.2C-Cyc 9.5 0.27 77 1019 <0.2D-Bed 10.8 0.35 <10 700 <0.2D-Cyc 11.0 0.36 61 622 <0.2E-Bed 11.3 0.52 16 466 <0.2E-Cyc 11.0 0.27 24 446 <0.2F-Bed 12.8 8.15 43 417 <0.2F-Cyc 12.0 1.55 163 706 <0.2

A (100%Coal) B (80%Coal20%Pine) C (80%Coal20%PE) D (75%Coa(75%Coal10%Pine10%PE5%Petcoke (900 �C)).

Table 4Leachability of residual solids (bed and cyclone) of the gasification tests with

pH Cond.mS/cm

Cl�

(mg/kg)SO2�

4

(mg/kg)Cd(mg/k

Ad-Bed 12.8 8.39 32 270 <0.2Ad-Cyc 12.5 8.73 1308 2028 <0.2Bd-Bed 12.6 7.90 50 748 <0.2Bd-Cyc 12.7 8.91 628 934 <0.2Cd-Bed 12.8 8.10 46 228 <0.2Cd-Cyc 12.8 8.81 475 934 <0.2Dd-Bed 12.7 8.33 47 358 <0.2Dd-Cyc 12.5 8.64 1181 1090 <0.2Fd-Bed 12.1 1.88 9.8 160 <0.2Fd-Cyc 12.4 6.00 140 870 <0.2

Ad (100%Coal) Bd (80%Coal20%Pine) Cd (80%Coal20%PE) Dd (75%Coal10%

Table 5Leachability of residual solids (bed and cyclone) of the gasification tests with

pH Cond.mS/cm

Cl�

(mg/kg)SO2�

4

(mg/kg)Cd(mg

Ao-Bed 11.5 0.89 22 404 <0.2Ao-Cyc 11.1 0.29 158 485 <0.2Bo-Bed 11.5 1.04 10 1037 <0.2Bo-Cyc 11.0 0.64 163 1164 <0.2Co-Bed 11.8 1.05 12 493 <0.2Co-Cyc 11.0 0.34 59 280 <0.2Do-Bed 11.1 0.56 <10 573 <0.2Do-Cyc 9.6 0.32 406 319 <0.2

Ao (100%Coal) Bo (80%Coal20%Pine) Co (80%Coal20%PE) Do (75%Coal10%

2.4%). The concentrations of Hg decreased with thedecrease of coal in the blends. The retention of Hg variedfrom 15% to 23% at 850 �C (16 ± 7%). For the test F, per-formed at higher temperature, the retention of Hgdecreased to 4.2%, although the higher level of sulphur inthe cyclone compared with test D (2.4% and 1.5%,respectively) was not observed to enhance Hg retained,most likely, because cyclone temperature was high(600 �C).

Due to the necessity of finding suitable options forsolid residues produced during gasification, their leach-

g)Cr(mg/kg)

Cu(mg/kg)

Ni(mg/kg)

Pb(mg/kg)

Zn(mg/kg)

<0.5 <0.5 <0.5 <0.5 0.2<0.5 <0.5 0.5 < 0.5 0.2<0.5 <0.5 <0.5 <0.5 <0.2<0.5 <0.5 <0.5 <0.5 0.2<0.5 <0.5 <0.5 <0.5 <0.2<0.5 <0.5 <0.5 <0.5 <0.2<0.5 <0.5 < 0.5 <0.5 <0.2<0.5 <0.5 <0.5 <0.5 0.2<0.5 <0.5 <0.5 <0.5 <0.2<0.5 <0.5 <0.5 <0.5 <0.2<0.5 <0.5 0.5 0.7 <0.2

0.5 <0.5 <0.5 3.4 11.0

l10%Pine10%PE5%Petcoke) E (55%Coal20%Pine20%PE5%Petcoke) F

dolomite

g)Cr(mg/kg)

Cu(mg/kg)

Ni(mg/kg)

Pb(mg/kg)

Zn(mg/kg)

2.7 <0.5 <0.5 <0.5 <0.212.7 <0.5 0.1 0.8 12.66.0 <0.5 <0.5 <0.5 <0.25.6 <0.5 <0.5 0.9 2.4

10.8 <0.5 <0.5 <0.5 <0.211.0 <0.5 <0.5 1.6 2.31.6 <0.5 <0.5 <0.5 <0.2

10.0 <0.5 0.7 0.7 0.2<0.5 < 0.5 <0.5 <0.5 <0.2

2.6 <0.5 1.4 <0.5 2.2

Pine10%PE5%Petcoke) Fd (75%Coal10%Pine10%PE5%Petcoke (900 �C)).

olivine

/kg)Cr(mg/kg)

Cu(mg/kg)

Ni(mg/kg)

Pb(mg/kg)

Zn(mg/kg)

4.0 <0.5 <0.5 <0.5 <0.20.5 <0.5 <0.5 0.6 0.73.2 <0.5 <0.5 <0.5 <0.20.7 <0.5 2.5 <0.5 0.65.5 <0.5 <0.5 <0.5 <0.2

<0.5 < 0.5 <0.5 <0.5 1.312.4 < 0.5 <0.5 <0.5 <0.22.1 <0.5 0.7 0.5 0.2

Pine10%PE5%Petcoke).

F. Pinto et al. / Fuel 87 (2008) 1050–1062 1061

ability was evaluated according to the European Standard(EN 12457). In Tables 3–5 are presented the valuesobtained. Leachates produced were all alkaline. The solidsobtained in the test at higher temperature had pH valuesabove 12 and the same happened for tests performed withdolomite. The use of olivine did not increase pH of leach-ates of the residues. For these tests the solubility wasrather low, having conductivity below 1, except for thecase of the test at higher temperature (F). For tests withdolomite, the conductivity was found to be higher, themaximum was 9 mS cm�1. The levels of solubilized Cl�

and SO2�4 of solids captured in cyclones were greater com-

pared with those of the bed. Usually Cl� in bed residuesattained a maximum of 50 mg/kg and for cyclone materialthe maximum was 163 mg/kg. Higher values wereobtained for cyclone material collected in the tests withdolomite, varying from 140 to 1300 mg/kg, as shown inTable 4. The same happened for sulphur. Although thesulphur content was higher in coal and petcoke than inpine and PE, no significant differences were obtainedin SO2�

4 solubility. But the use of dolomite increasedSO2�

4 solubility. These results showed the scavengingeffect of dolomite for chlorine and sulphur compoundsin the cyclones. Leachability of metals was insignificantfor most of the residues, although the higher pH valuesallowed a small leachability of Pb, Zn and Ni. This wasseen for test F and also for the tests with dolomite, as pre-sented in Table 4. In those tests with dolomite or olivine,some leachability of Cr is also found, reaching 14 mg/kgfor the tests with dolomite.

4. Conclusions

Due to the energy value of petcoke, pine and PE, theirco-gasification with coal can be advantageous, as itincrease process efficiency, without significant changes inexisting coal gasification plant. However, due to the highsulphur content of petcoke and chlorine amount of pine,their increased amount in blends with coal could lead tothe release of pollutants, such as H2S and HCl. This issueneeds to be addressed. One way is the addition of catalystsor sorbents to the gasifier to promote destruction of thesecompounds and the retention of sulphur and chlorine inthe bed solids.

The addition of dolomite to the gasification bed led tothe highest H2S reduction and consequently to the greatestsulphur retention in the solid phase, with values higherthan 90% for all the fuels studied. It was in the presenceof Ni–dolomite that the second lowest sulphur release tothe gas phase was obtained, probably due to the dolomitepresence even in lower amounts. The observed decreases insulphur releases into the gas phase resulted from CaO reac-tion with H2S to form CaS that remained in the gasificationbed.

The addition of G-72-D to the gasification bed led tosmaller reductions in chlorine released to the gas phasethan those obtained with dolomite. However, as both G

72-D and C49 TRX were used as pellets, while the otherswere as small particles, the pelletised catalysts might havedeactivated faster. The action of different catalysts or sor-bents during coal gasification and when coal was mixedwith wastes produced quite clear tendencies, which alloweda classification of theirs activity as given below:

dolomite > Ni–dolomite > G 72-D > C49TRX � olivine

> Ni–Mg

It was also in presence of dolomite that the highest reduc-tions in gaseous chlorine were achieved, probably due tothe reaction between CaO and HCl to form CaCl2. AsCaC12 has a melting point at 772 �C, it might have re-mained inside the reactor, probably impregnated in thechar. It was probably due to this reason, that the releaseof HCl by the reversion of reaction (5), caused by steampresence was prevented in some extent.

The presence of ZnO in G 72-D might have retainedmore chlorine in the gas phase, due to ZnO reaction withHCl to form ZnCl2. However, some of the H2S presentin the reaction medium could have reacted with ZnCl2to form ZnS thus causing the release of HCl, whichwould explain why no greater chlorine retentions wereobtained.

Coal gasification and its co-gasification blended withwastes showed that the action of catalysts or sorbentstested in chlorine retention in the solid phase could be inthe following order:

dolomite > Ni–dolomite > G72-D:

Olivine and C49 TRX, led to chlorine releases into the gas-eous phase only slightly lower than those obtained when nocatalyst was used, whilst the values obtained in presence ofNi–Mg were even similar. The low cost of dolomite and itssatisfactory capability to capture sulphur and chlorine,make it very suitable to be used in the gasifier bed.

Heavy metals present in the fuels or introduced with thecatalysts or sorbents were partially retained in solids col-lected in the bed and hot cyclone. Apparently the additionof residues to the coal increased the volatility of some met-als, especially in the case of Ni and Pb. However, metalswere further retained in the solids collected in the cyclones,as metal were not detected in the condensates of the gascooling system. The increase of gasification temperatureis advantageous to increase syngas quality, however itincreased metal volatility especially that of Hg. However,leachability of solid residues was found to be insignificant,although the addition of catalysts or sorbents containingalkalis to the process slightly increased metals leachabilitydue to the increase in pH.

Acknowledgements

Authors are grateful to the European Commission forthe financial support to do this work through the ProjectRFC-CR-04006.

1062 F. Pinto et al. / Fuel 87 (2008) 1050–1062

References

[1] Sage PW, Welford GB. Fluidized bed combustion, vol. 2. ASME;1997. p. 1295–302.

[2] Paterson N , Reed GP, Dugwell DR, Kandiyoti R. ASME TurboExpo 2002 – Land, Sea & Air (2002) Amesterdam, The Netherlands,June, GT2002-530013.

[3] Sutton David, Kelleher Brian, Ross Julian. Fuel Process Technol2001;73:155–73.

[4] van der Drift A, van Doorn J, Vermeulen JW. Biomass Bioenergy2001;20:45–56.

[5] Kuramochi H, Wu W, Kawamoto K. Fuel 2005;84:377–87.[6] Khan MR. Fuel 1989;68:1439–49.[7] Park N, Lee JD, Lee TJ, Ryu SO, Chang CH. Fuel 2005;84:2165–71.[8] Xu C, Tsubouchi N, Hashimoto H, Ohtsuka Y. Fuel

2005;84:1957–67.[9] Zhao Z, Li W, Qiu J, Li B. Fuel 2003;82:1839–44.

[10] Wei Xiaolin, Schnell Uwe, Hein Klaus RG. Fuel 2005;84:841–8.[11] Arton B RG, Clark WD, Seeker WR. Combust Sci Tech

1990;74:327–42.[12] Linak WP, Wendt JOL. Prog Energy Combust Sci 1993;19:145–85.[13] Linak WP, Wendt JOL. Fuel Process Technol 1994;39:173–98.[14] Frandsen F, Dam-Johansen K, Rasmussen P. Prog Energy Combust

Sci 1994;20:115–38.

[15] Helble JJ, Mojtahedi W, Lyranem J, Jokiemi J, Kauppinen E. Fuel1996;75:931–9.

[16] Elled A-L, Amand L-E, Leckner B, Andersson B-A. Fuel2007;86:843–52.

[17] Pinto Filomena, Franco Carlos, Lopes Helena, Andre Rui Neto,Gulyurtlu I, Cabrita I. Fuel 2005;84(17):2236–47.

[18] Garcia L, Benedicto A, Romeo E, Salvador ML, Arauzo J, Bilbao R.Energy Fuels 2002;16(5):1222–30.

[19] Wang Shaobin, Lu GQ. (Max) Energy Fuels 1998;12:248–56.[20] Sciazko M, Kubica K. Fuel Process Technol 2002;77–78:95–102.[21] Najjar Mitri S, Jung Dick Y. Fuel Process Technol 1995;44:173–80.[22] Tijmensen MJA. The production of Fischer Tropsch liquids and

power through biomass gasification. Springer; 2000. November.[23] Zevenhoven, R, Kilpinen, P. Control of pollutants in flue gases and

fuel gases (sulphur chapter), The Nordic Energy Research Pro-gramme, Helsinki University of Technology, February, 2004.

[24] Weinell Claus E, Jensen Peter I, Dam-Johansen Kim. Hans Livbjerg.Ind Eng Chem Res 1992;31:164–71.

[25] Gupta Raghubir P, O’Brien William S. Ind Eng Chem Res2000;39:610–9.

[26] Park No-Kuk, Lee Dong-Hwal, Jun Jin Hyuk, Lee Jong Dae, Ryu SiOk, Lee Tae Jin, et al. Fuel 2006;85:227–34.

[27] Pineda M, Palacios JM, Alonso L, Garcia E, Moliner R. Fuel2000;79:885–95.