The influence of inorganic matter on the pyrolysis of Canadian lignite in a fluidized bed

Fuel Processing Technology, 25 (1990) 201-214

201

Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

The I n f l u e n c e of I n o r g a n i c Matter on the P y r o l y s i s of Canadian Lignite in a Fluidized Bed A.J. ROYCE, S. MIYAWAKI*, J. PISKORZ, D.S. SCOTT and S. FOUDA**

Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario N2L 3G1 (Canada) (Received June 7th, 1989; accepted in revised form February 2nd, 1990)

ABSTRACT In low rank coals much of the inorganic matter is present as cations associated with organic carboxyl groups in the coal rather than as discrete mineral phases. By treating the coal with acid the inorganic content is reduced by cation exchange, as well as by acid leaching of discreteminerals. Whole and acid treated samples of pulverized lignitewere pyrolyzed in pilot scale (1 to 3 kg coal/h) and bench scale (60 to 100 g coal/h) fluidizedbed reactors at atmospheric pressure, 0.45 second vapour-residence time, and temperatures ranging from 500 to 730°C. Yields of char, tar, water and lightgases were determined. Removal of inorganic matter from a Saskatchewan lignite resulted in increased yields of tar and total volatilematter, with littleeffecton the yields of light gases. Increased yields of tar are largely a result of an increased asphaltene fraction in general and specifically of an increased catechol content. Char from acid-washed coal is less reactive to carbon dioxide than is char obtained from raw lignite.

INTRODUCTION

Work on coal pyrolysis was initiated at the University of Waterloo in 1979 using a small bench scale fluidized bed apparatus. The focus of the preliminary work was to determine the optimum pyrolysis conditions for the maximum yield of organic liquids for a variety of coals ranging in rank from lignite to bituminous [ 1 ]. This present study is concerned with the i~fiuence of the inherent inorganic matter in lignites on their pyrolysis behaviour and specifically on the yield of products, the composition of the tar, and on the reactivity of the char. Lignites represent an early stage of the coalification process and are thus relatively rich in humic substances compared to higher rank coals. The distinguishing fea*Present address: Nippon Mining Co., Tokyo, Japan. **Energy Research Laboratories, CANMET, Energy, Mines and Resources Canada, Ottawa, Canada.

0378-3820/90/$03.50

© 1990 Elsevier Science Publishers B.V.

202

tures of this organic matter of lignite are the lower aromaticity compared to bituminous coals and the prevalence of oxygen-containing functional groups consisting of carboxylates, phenols and ethers. The inorganic elements in lignite are found in three principal forms: as discrete mineral phases, as salts dissolved in the pore water or as cations combined chemically with the organic materials. The most significant feature of the inorganic matter in lignites is the large amount of alkali and alkaline-earth metal cations thought to be associated with the carboxyl groups in the coal. One of the important differences between low rank coals, such as lignites, and high rank coals is the higher carboxyl group content of lignites. Carboxyl groups impart ion exchange properties to the coal and a significant portion of them are associated with inorganic cations. A study of Australian brown coals [ 2 ] reported that from 17 to 30 percent of the total carboxyl groups were present as carboxylate anions. Direct determination of exchangeable calcium, magnesium, sodium, potassium and iron cations was found to account for 90 percent or more of the carboxylate determined by titration. Two investigations of American lignites [3,4 ] reported that 40 to 60 percent of the carboxyl groups in the lignites were associated with metal cations of which calcium and magnesium were the most abundant while sodium, potassium, strontium, and barium varied in both absolute and relative concentrations. Mallya and Zingaro [5 ] reported that as much as 80 percent of the total carboxyl groups present in a Texas lignite were present as carboxylates and that 40 percent of the phenolic groups were present as phenolates. It is important to differentiate between these exchangeable inorganic species and the discrete mineral components. The most important difference is their dispersion throughout the coal matrix. The exchangeable cations are dispersed on an atomic basis; that is, each metal cation is associated with one or two carboxyl groups, whereas it is well established that the great majority of minerals exist in size ranges greater than 1/~m. Thus, they are several orders of magnitude greater in size than the cations. A number of investigators have studied the effect of exchangeable cations in coal on the pyrolysis process. Australian workers have been involved in several studies involving the pyrolysis of low rank coals. Tyler and Schafer [ 2 ] studied the effect of exchangeable cations on the pyrolysis behaviour of brown coals under flash heating conditions using a small fluidized bed reactor. Coal pretreatment consisted of acid washing samples, by which means the inherent cations were removed from the coal, and then preparing from the acid treated samples a series of coals with a range of calcium contents by exchanging calcium onto the protonated coal. With all coals the maximum tar yield was obtained at 600 ° C. Removal of cations from the coals resulted in increased yields of tar (by factors of 1.2 to 2) and total volatile matter (by factors of 1.1 to 1.4 ) with little change in yields of C1 to C3 hydrocarbon gases. Morgan and Jenkins [6,7] studied the effects of ion exchangeable cations

203 on the flash pyrolysis of a Montana lignite using an entrained flow reactor. Removal of cations resulted in increased yields of tar and total volatile matter and increased the rate of volatile release. The tars produced from the raw coal were black and gummy and contained three times as much aliphatic hydrogen as tars from the acid washed coal, which were brown in colour. Morgan suggested that the cations may catalyse the cracking of the tar resulting in the simultaneous deposition of highly aromatic material and release of aliphatic material. The effect of cations on the rapid pyrolysis of a subbituminous coal was studied by Franklin et al. [8], who concluded that all alkali cations (Ca 2+, Na + and K + ) reduce tar yields at 727°C. In all cases sodium and potassium ions were more effective than calcium ions. Carbon dioxide yields from acid washed coal at high temperatures corresponded almost exactly with carboxyl group content and therefore it was concluded that any excess carbon dioxide from cation-form coals must originate from non-carboxyl oxygen. Similar results were also reported by Schafer [9-12 ], who studied the effects of exchangeable cations on the decomposition of acidic groups during the slow pyrolysis of brown coals. He showed, as did Franklin [8 ] and Otake [ 13 ], that carboxyl groups in coal decompose to give one mole of carbon dioxide for each equivalent of carboxyl present. However, cation-rich coals yield more carbon dioxide than can be accounted for by the carboxyl groups present. It was proposed that some other oxygen-containing functional groups are associated with the carboxyl groups and their decomposition mode is altered when the carboxyl groups are associated with cations. In a protonated brown coal, however, the carboxyl groups yield carbon dioxide, and phenolic groups yield carbon monoxide. To investigate these phenomena during fast pyrolysis, a Saskatchewan (Estevan, Canada) lignite was acid treated to remove varying amounts of inherent cations and subsequently pyrolyzed in a bench scale fluidized bed pyrolyzer to determine the effects of exchangeable cation content on the yields of char, tar, water, carbon oxides, hydrogen and light hydrocarbon gases. In addition, a second sample of a different Saskatchewan (Coronach) lignite was acid washed and pyrolyzed in a pilot scale fluidized bed pyrolyzer to confirm the results from the bench scale unit and to prepare sufficient amounts of tar for detailed analysis. EXPERIMENTAL "As received" coals were ground and sieved to provide samples with a particle size range of 74 to 250 Hm for the bench scale unit and 62 to 600 p m for the pilot plant. Acid washed coals with a range of cation contents were prepared from the raw coal by extraction with HC1 at five different constant acidities (pH = 4.0,

204

3.5, 2.5, 1.5, 0.8 ). Another acid washed sample was also prepared by extraction with sulphuric acid at a pH of 3.0. During the extractions the pH of the coalacid slurry was maintained at the desired acidity by incrementally adding acid until the exchange was complete. The bench scale fluidized bed pyrolyzer has been described in detail by Scott et al. [1 ]. In this work, two sizes of bench scale fluidized bed were used, having volumes of 24 ml and 155 ml. Within our experimental accuracy, results were the same from both reactors. A schematic of the small pilot plant unit is shown in Fig. 1. Briefly, in each case an entrained flow feeder injected a stream of coal particles directly into an electrically heated fluidized sand bed. Coal was fed to the bench scale fluidized bed at rates of 60 to 100 g / h in nitrogen for an average run time of 30 minutes. In the pilot scale unit, circulating pyrolysis gas fluidized the sand bed and carried coal to the reactor at rates of 1 to 3 kg/h, approximately 2 kg of coal being pyrolyzed in each experiment. Both char and volatile matter were entrained out the top of the reactor and collected separately downstream. Liquid product collection consisted of a system of two conPYROLYSIS

PILOT

PLANT

HOT WATER CONDENS SCREW FEEDER

D

ICE WATER DNDENSER

I IC'AR " W ~ H~T

r

I

ROTAMETER

!

FLOW INDICATING CONTROLLER

"URNACEI

LIGHT ORGANICS

I~ Zz-----BALLAST TANK

I I I I

PRESSURE IJ--"-i ORIFICE INDICATING I L METER CONTROLLERL.........J . . . . . . .

,FILTER

TAR PRODUCTS

BACK PRESSURE

> TO VENT GAS

ANALYZER

Fig. 1. Waterloo flash pyrolysis pilot plant unit.

-

,,1

I ~GAS

METER

205 densers and a cotton wool filter connected in series. Additional tar products were washed from the condensers by using tetrahydrofuran (THF). The solvent was evaporated at about 65 ° C and 50 m m H g vacuum, and the liquid product was then weighed and stored in a refrigerator prior to further analysis. Char was collected in a cyclone, and also recovered by filtration of solvent washings. Raw and acid washed coals were characterized by proximate and ultimate analysis, as well as by the analysis of their inorganic cations (Ca e+, Mg z+, Na + and K + ). Gaseous pyrolysis products were analyzed by gas chromatography (GC). Char and tar products were analyzed for their carbon, hydrogen and nitrogen contents. The tar was further characterized by solvent fractionation and the oil and asphaltene fractions were analyzed by GC-MS. Finally, the reactivity of the char to carbon dioxide was determined by thermogravimetric analyses (TGA). Two methods of analysis were attempted to determine the total amounts of specific inorganic elements in Estevan lignite (i.e. the combined amount in the mineral matter and in the organic matrix). The inorganic content of the raw coal was determined by first demineralizing a coal sample by using the method of Bishop and Ward [ 14 ] and then analyzing the extracted acidic solution for the desired elements by atomic absorption spectrophotometry (AAS). A second method tested was the ash digestion procedure described by Perkin-Elmer Corp. [ 15 ] followed by analysis of the extract solution. Thirdly, a procedure for the determination of exchangeable cations was adapted from the study of Morgan, Jenkins and Walker [4] in which the coal was subjected to ion exchange with a m m o n i u m acetate in order to selectively remove only the exchangeable cations and leave essentially intact the discrete mineral matter. The method of solvent fractionation described by Mima et al. [ 16 ] was used as the basic procedure for characterization of the liquid tar products. In using this method, however, it was found that some insoluble matter was formed in the toluene soluble fraction when evaporating toluene under heating conditions. This insoluble matter was highly viscous and was apparently formed by polymerization or condensation between toluene soluble materials. Since heating conditions are likely to accelerate these reactions, every heating step was eliminated from the procedure of solvent removal. Thereafter, the solvent removal was performed at room temperature in a draught chamber under an air stream. This solvent fractionation procedure is designed particularly for hydrocarbon systems, and the residual insoluble materials, after tetrahydrofuran and toluene extraction are classed as preasphaltenes. However, the lignites used contained 22 to 30 percent oxygen, and on pyrolysis could be expected to yield significant amounts of oxygenated polar compounds which would appear in the preasphaltene fraction, but would not necessarily be in the same molecular weight range. Therefore, an additional step was added to the solvent fractionation procedure in which the "preasphaltene" residue was divided into meth-

206

anol-soluble and methanol-insoluble fractions. In general, 50 percent or more of this preasphaltene material was soluble in methanol. Probably only the methanol-insoluble fraction should be considered as related to materials commonly classed as preasphaltenes. RESULTS

Materials Proximate and ultimate analyses of the raw and acid washed lignite samples are given in Table 1. Acid washing at different levels of severity does not appear to cause any statistically significant trends in the elemental composition of the organic portion of the lignites. It is apparent, however, that the inorganic matter, as indicated by the ash content, is reduced with increasing severity of acid treatment. Table 2 gives the inorganic content of the raw and acid washed coals. The Ca and Mg contents of raw Coronach lignite are nearly twice those of Estevan lignite, however, the sodium content of Estevan lignite is about 13 times that of the Coronach sample. The cation contents of the two coals, when both are washed in HC1 at a pH of 2.5, are very similar, with the exception of sodium. For the acid washed Estevan lignite sample it is apparent that the exchangeable cation contents decrease with increasing severity of acid treatment. Pyrolysis of raw and acid washed coals Earlier studies established that the maximum tar yield for low rank coals occurred on flash pyrolysis at about 600 to 650 ° C and short residence times of TABLE 1

Ultimate and proximate analysis of raw and acid washed Saskatchewan lignites As/pyrolyzed

Moisture and ash free basis

Atomic

moisture

volatile matter

fixed carbon

C

50.22 45.99 47.83

49.78 54.01 52.15

66.54

4.38

0.83

28.25

0.79

49.30

50.56

68.75 70.99 69.22 70.96 65.07 64.53 68.22

4.36 4.54 4.67 4.38 4.68 4.44 4.63

1.80 1.84 1.58 1.41 1.65 0.95 1.44

25.10 22.63 24.53 23.25 28.60 30.10 25.70

0.76 0.77 0.81 0.74 0.86 0.83 0.81

H/C

Coronach, raw Coronach, p H = 2.5 Estevan, raw Estevan, pH = 4.0 Estevan, p H = 3 . 5 Estevan, p H - - 3.0 a Estevan, p H = 2.5 Estevan, p H -- 1.5 Estevan, pH = 0.8

aAcid washed in H2SO4.

27.77 20.91 6.66 12.64 10.41 0.84 0.0 1.34 0.17

ash

12.52 6.98 6.24 5.25 4.59 5.24 2.63 2.74 2.58

H

N

Diff. ( 0 + S)

207 TABLE 2 Inorganic contents of raw and acid washed lignites

Coronach, raw Coronach, pH = 2.5 Estevan, raw Estevan, pH = 4.0 Estevan, pH-- 3.5 Estevan, pH = 3.0a Estevan, pH = 2.5 Estevan, pH = 1.5 Estevan, pH = 0.8

gram equivalents/kg dry coal Na K

Ca

Mg

1.01 0.16 0.53 0.47 0.40 0.37 0.16 0.07 0.04

0.40 0.01 0.21 0.16 0.12 0.06 0.03 0.02 0.02

0.04 0.01 0.53 0.30 0.28 0.19 0.20 0.19 0.19

0.03 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.003

Total 1.48 0.20 1.29 0.94 0.81 0.63 0.40 0.29 0.25

aAcid washed in H2SO4. about 0.2 to 0.5 seconds [ 1 ]. Since the main objective of the present investigation was to determine the influence of inorganic matter on fast pyrolysis yields, and particularly the optimum tar yield, tests were confined to temperatures in the range 550 to 730 °C and vapour residence times of about 0.4 to 0.5 s. The raw and acid washed Estevan lignite samples were pyrolyzed in the bench scale pyrolyzer at 650°C and 0.55 s vapour residence time. The major effect of acid washing was a higher tar yield with a corresponding reduction in gas yield. The tar yield for an Estevan sample acid washed at a p H of 0.8 was 150 percent of that from the raw Estevan lignite, while the gas yield was only 85 percent as great. The trend, as shown in Fig. 2, is to higher tar yields and lower gas yields with more severe acid treatment, that is, with increasing removal of ion-exchanged alkali cations in the coal matrix. Char yields are very slightly reduced. A lower CO2 yield accounts for most of the reduction in gas yield observed with more severe acid washing. Hydrogen and methane also decrease slightly, whereas carbon monoxide increases slightly. Table 3 gives results of pyrolysis runs of raw and acid washed Coronach lignite carried out with the small pilot scale pyrolyzer. Results from the pyrolysis of raw Coronach at 650 °C indicate that this lignite decomposes to yield more tar than does the Estevan lignite. Gas and water yields are lower and char yields higher than those from Estevan lignite pyrolyzed in the bench scale unit. However, some of these differences can be attributed to the fact that the char pot of the bench scale pyrolyzer was at a higher temperature which promoted more complete devolatilization of the char, resulting in slightly higher gas yields and lower char yields compared to the pilot scale pyrolyzer. Results shown in Table 3 indicate that acid washing produces higher tar yields. Also, somewhat

208

O 14 •

24

O CONVERSIONTOORGANICLIQUID /', CONVERSIONTO GAS R G A N I C ~ ~'

w ,< 22 '-.~ O oZ

LIQUIDS

~_~ (.) £9

t~

GAS

> Z 0 u

50 Q-O 8

i

0.0

I

0.4

i

I

i

I

0.8

1.2

J

I

J

I

1.6

2.0

ACID CONSUMPTION, grn eg./kg DRY COAL

Fig. 2. Tar and gas yields from Estevan lignite as a function of severity of acid treatment. TABLE 3 Effect of acid washing on product yields, pilot plant pyrolysis, Coronach lignite #3, pH = 2.5, HC1 Parameter

Temperature (°C) Residence time (s) Feed rate {kg/h) Moisture (wt.%) Ash (dry wt.%)

Acid washed

Original

P-84

P-82

P-83

P-85

P-80

553 0.50 1.657 20.91 8.83

652 0.46 1.535 20.91 8.83

732 0.41 1.577 20.91 8.83

581 0.46 1.635 22.7 16.21

652 0.45 1.590 22.77 16.21

13.85 15.20 63.18 9.25

16.55 19.95 57.31 7.41

21.07 17.53 55.69 12.22

13.87 12.47 67.99 6.37

17.06 17.32 58.97 8.50

101.48

101.22

106.51

100.70

101.85

Product yields, wt. % mar coal Gas Tar Char Water Total

Gaseous products, wt. % ma[ coal H2 CO CO2 CH4 C2H4 C2H~ C3 C4

0.07 4.14 7.39 1.31 0.54 0.17 0.14 0.09

0.11 5.93 8.38 1.59 0.33 0.08 0.07 0.06

0.17 9.69 8.29 2.05 0.57 0.12 0.09 0.09

0.08 3.26 8.80 1.25 0.23 0.12 0.08 0.06

0.14 5.81 8.83 1.77 0.32 0.09 0.06 0.04

Total

13.85

16.55

21.07

13.87

17.06

209

ACID W A S ~ D CORONACH $ 3

E 15 >~20~° C0RONa~CH Z 2.

* , MOISTURE

da po >O2

5

9.59 % MOISTURE [

40o

I

I

5oo 600 TEMPERATURE

7oo

800

("C) Fig. 3. Tar yieldsfromCoronachlignites,pilotplantunit. 3O

"6 E

2O

j.-'o

/

W rr

Io

O/

~ORORAC. LIGNITE • 2

JfEs,E .......... "s ( Scott, Plskorz, and gouda, I°J84 )

o

0.6

I

I

0.7

0,8

0,9

COAL H / C ATOMIC RATIO

Predictionof maximumtar yieldsfromrawor acidwashedlignites (afterSmith [ 18] and Scott et al. [ 1 ] ). Fig. 4.

lower gas yields are obtained, principally due to lower CO2 yields, in agreement with results from the bench scale unit. Figure 3 shows the tar yields from raw and acid washed Coronach lignite over the temperature range 550 ° to 730 ° C. Maximum tar yields are obtained at about 650 °C in all cases, which agrees with results from earlier work using the bench scale unit [1]. Tar yields from acid washed Coronach were higher than those from raw Coronach at the temperatures tested. From the results in Table 3 it is obvious that char yields are reduced by acid washing. This reduction is greater at lower temperatures and this trend coincides with the increased tar yield resulting from acid washing which is also greater at lower temperatures. The decrease of gas yields due to acid washing Coronach lignite, shown in Table 3, are not as great as those due to acid washing Estevan lignite. Gas yields from the Coronach sample are slightly reduced with acid washing for pyrolysis

210

at 650 ° C, but are very similar at lower temperatures for raw and acid washed coal. Tyler [ 2 ] and Smith [ 18 ] reported an empirical correlation for acid washed coals in which the maximum tar yields were related to the atomic H / C ratios of the parent coals. This correlation is illustrated in Fig. 4 together with the results from Coronach lignite. Figure 4 also shows the higher tar yields of raw Coronach compared to those of raw Estevan lignite. It appears that the maximum tar yields, obtainable for these Canadian liguites from both the raw and acid washed coals, can be approximately predicted from this correlation.

Composition of tar The results of solvent fractionation of tars from raw and acid washed Coronach lignite are illustrated in Fig. 5. The yield of a methanol insoluble frac20

8 o

,~ ~ ~~"~ASPHALTENE

E I0

¢--, --J hi >-

.,

,

METHANOL

INSOLUBLE TEMPERATURE(°C) RUN

553 P-84

COAL

652 P-B2

732 P-83

ACID WASHED

581 P-85

652

P-80 ORIGINAL

Fig. 5. Comparison of solvent fractionation of tars from raw and acid washed Coronach lignite #3. - - I PHENOL _ 2 O-CRESOL ---

3 m/p-CRESOL

ouO 0

- 4 XYLENOL 5 ETHYLPHENOL - 6 CATECHOL 7~ METHYLRESORCIHOL - S METHYLCATECHOL

ACIDWASHED CORONACH LIGNITE#3

0.5123

~J LLI

ORIGINAL CORONACH LIGNITE#3 FORESTBURG SUBBITUMINOUS COAL

I

I

I

I

I

I

~

i

I 2 3 4 5 6 7 8 COMPOUND CODE NUMBER

Fig. 6. Yields of phenolic compounds in asphaltene fractions of tars from raw and acid washed lignite.

211

tion of the preasphaltenes from acid washed Coronach lignite are slightly lower than those from the raw lignite. The methanol insoluble fractions are black agglomerates possibly consisting of oxygenated polymer-like matter that could be formed by the direct decomposition of the humic acid content of the lignite and/or by the repolymerization of highly reactive volatile oxygenated fragments. Generally, these highly polymerized compounds are undesirable in practical use, since they may cause some difficulty to potential users. From this point of view, it is indicated that the acid washing of Coronach lignite might be effective in reducing the amount of these methanol insoluble fractions of the tar. Although Runs P-83 and P-80 have very similar tar yields, the tar from acid washed Coronach (P-83) contains a larger asphaltene fraction than the tar produced from the raw lignite (P-80). Even the amount of asphaltene fraction obtained from acid washed coal at the lowest temperature of 553 ° C is slightly higher than that from the raw coal pyrolyzed at 652 ° C, indicating that the acid washing resulted in higher asphaltene yields at all temperatures. The preasphaltene fraction of the acid washed lignite pyrolyzed at 553 ° C is larger than that of the raw coal pyrolyzed at 581 ° C, however, this difference disappeared for the runs at 650 ° C. There was no observed effect on the oil yield due to acid washing. Consequently, it can be concluded that the increased tar yields by acid washing treatment are primarily due to the increase in the yields of asphaltene fractions. Figure 6 shows some results from the GC-MS analyses of the asphaltene fractions of a raw and of an acid washed Coronach lignite tar sample. The results show that the amounts of catechol and of methyl catechol increase remarkably in the tar product from pyrolysis of the acid washed lignite, while the concentrations of phenol and related compounds are less affected. It appears that the removal of alkaline cations from the lignite in some way affects the catechol precursors such that the production and evolution of catechol is enhanced. Vapour-phase reactions are not assumed to be responsible for the significant increase in catechol content in the asphaltene fraction of pyrolysis tar after protonation.

Char reactivity The effect of the cation content of the parent coal on char reactivity in carbon dioxide was also studied using thermogravimetric analysis. Figure 7 shows the results of four separate TGA runs with samples of Estevan lignite given different cation removal pretreatments but with identical devolatilization steps which conform to the TGA proximate analysis method [15]. This procedure ensures that devolatilization is complete at the reaction temperature and that the reaction of CO2 with the fixed carbon is the only significant weight-loss mechanism occurring.

212

100 950°C :E

I00 ° CI rnin bE MINERALIZED

50 (~

12(g/gh) roof

ACID WASHED

to~

0 0

I I0

~

R'~R~

~ 21C02 ~ 20 30

I 40

84S (g/ghlmaf

ei 50

I 60

TIME ( minutes )

Fig. 7. Effect of exchangeable cations on the reactivity of Estevan lignite char with C02 at 950°C.

The results shown in Fig. 7 clearly illustrate the differing char reactivities for different coal pretreatment. The numerical values given in Fig. 7 are the initial rates of reaction of the Estevan lignite chars in CO2 at 950 oC. Clearly the initial reaction rates varied directly with the inorganic content of the parent coal. The demineralized sample was virtually unreactive with COz. An acid washed sample (pH--0.8) reacted at a rate of 0.8 g/g h (maf), while a char sample from raw lignite in which the exchangeable cations were replaced with ammonium ions reacted at 1.47 g/g h (maf) or 1.8 times as fast as the acid washed sample. The raw lignite char had the highest reactivity, as expected, at 8.46 g/g h (maf). Apparently, the inorganic cations associated with the organic part of the coal are able to catalyze the char-CO2 reaction and therefore removing these cations reduces the char reactivity. These results confirm those reported by others, for example, Radovic et al. [17] on the effects of various inorganic constituents on coal char gasification rates. However, it is difficult to explain why the ammonium-form coal, in which it is assumed that all the exchangeable inorganic cations were removed, and in which the ammonium ion must have decomposed, reacted faster than the acid washed coal. Possibly the discrete mineral phase in the lignite might have been affected by the ammonia treatment and contributed a small catalytic effect, or the exchangeable calcium and sodium cations may not have been completely replaced by ammonium ion. SUMMARY About one half of the ash content of Canadian (Saskatchewan) lignite is associated in ion exchangeable form with the carboxyl groups of the coal matrix. This part of the ash, consisting largely of Ca, Mg, K and Na cations, can be over 80% removed by treatment with cold acid to a pH of about 1.0. Sodium is the most difficult to remove, although 65% removal is possible even with a

213

high sodium ash. The reduction in this part of the ash would considerably ease boiler fouling problems when burning high sodium lignites. In addition to ash reduction, the cold acid washing treatment increases the yield of liquids in fast fluidized bed pyrolysis. Maximum organic liquid yields can be increased by 50% by acid washing, at the expense of gas yield. Maximum yield occurs at about 650 ° C for both raw and treated lignites. The maximum yield of tars for raw or acid treated lignites can be approximately predicted from the correlation given by Smith [ 18 ] or by Scott et al. [ 1 ]. The increased yields of tars from acid washed lignites appears to be due mainly to an increase in asphaltene content. In particular, yields of catechols are much greater in tars produced by the fast pyrolysis of acid washed lignite as compared to raw lignite. ACKNOWLEDGEMENTS

The authors wish to express their thanks to the Natural Sciences and Engineering Research Council of Canada and to the Energy Research Laboratories of CANMET, Energy, Mines and Resources Canada for the financial support of this work. Thanks are also extended to Peter Majerski, operator of our pilot plant pyrolysis unit. One of the authors (S. Miyawaki) also wishes to express his appreciation to the Nippon Mining Co. for the award of a scholarship.

REFERENCES 1 2 3 4 5

6 7 8

9

Scott, D.S., Piskorz, J. and Fouda, S., 1986. Pyrolysis of low rank Canadian coals. Fuel Processing Technology, 13: 157-186. Tyler, R.G. and Schafer, H.N.S., 1980. Flash pyrolysis of coals: influence of cations on the devolatilization behaviour of brown coals. Fuel, 59: 487-493. Bobman, M.H., Golden, T.C. and Jenkins, R.G., 1983. Ion exchange in selected low rank coals. Part I: Equilibrium; Part II: Kinetics, Solvent Extraction and Ion Exchange, 1 (4): 791. Morgan, M.E., Jenkins, R.G. and Walker, P.L., Jr., 1981. Inorganic constituents in American lignites, Fuel, 60" 189. Mallya, N. and Zingaro, R.A., 1984. Some Structural Features of a Wilcox Lignite, in: Schobert, H.H. (Ed.), The Chemistry of Low Rank Coals. American Chemical Society, Washington, DC. Morgan, M.E. and Jenkins, R.G., 1983. Role of exchangeable cations on the rapid pyrolysis of lignites, Am. Chem. Soc., Div. Fuel Chem., Prepr., 28(4): 138. Morgan, M.E. and Jenkins, R.G., 1986. Pyrolysis of a lignite in an entrained flow reactor. Fuel, 65:757-763 and 65: 764-768. Franklin, H.D., Cosway, R.G., Peters, W.A. and Howard, J.B., 1983. Effects of cations on the rapid pyrolysis ofa Wyodak subbituminous coal. Ind. Eng. Chem. Process Des. Dev., 22: 3942. Schafer, H.N.S., 1979. Pyrolysis of brown coals: 1. Decomposition of acidic groups in coals containing carboxyl groups in the acid and cation forms. Fuel, 58: 667.

214 10 11 12 13

14 15 16 17

18

Schafer, H.N.S., 1979. Pyrolysis of brown coals: 2. Decomposition of acidic groups on heating in the range 100 to 900 ° C. Fuel, 58: 674. Schafer, H.N.S., 1980. Pyrolysis of brown coals: 3. Effect of cation content on the gaseous products containing oxygen from Yallourn coal. Fuel, 59: 296. Schafer, H.N.S., 1980. Pyrolysis of brown coals: 4. Relation between acidic groups and evolved carbon oxides. Fuel, 59: 302. Otake, Y. 1982. Volatile Matter Release During Pyrolysis of Demineralized and Metal Cation Loaded Lignites. M.Sc. Thesis, Dept. Materials Science and Engineering, Pennsylvania State University, University Park, PA. Bishop, M. and Ward, D.L., 1958. The direct determination of mineral matter in coal. Fuel, 37: 191. Perkin-Elmer Corp., 1978. Analytical Methods for Atomic Absorption Spectrophotometry. Perkin-Elmer, Norwalk, CT. Mima, M.J., Schultz, H. and McKinstry, W.E., 1978. Analytical Methods for Coal and Coal Products, Vol. 1, Karr, C. Jr., (Ed.). Academic Press, New York, NY, p. 557. Radovic, L.R., Steczko, K., Walker, P.L., Jr. and Jenkins, R.G., 1985. Combined effects of inorganic constituents and pyrolysis conditions on the gasification reactivity of coal chars. Fuel Processing Technology, 10:311. Smith, I.W., 1981. New approaches to coal pyrolysis at CSIRO. Paper presented at the EPRI Coal Pyrolysis Workshop, Palo Alto, CA, February.