- Email: [email protected]
Enhancing low severity coal liquefaction reactivity using mild chemical pretreatment
Enhancing low severity coal liquefaction reactivity using mild chemical pretreatment KayGhobad
Shams,
Ronald
L. Miller
and Robert
Chemical Engineering and Petroleum Refining Golden, CO 80401, USA (Received 2 August 1991; revised 7 February
Department,
M. Baldwin Colorado
School
of Mines.
7992)
We describe results from a study in which mild chemical pretreatment of coal has been used to enhance low severity liquefaction reactivity. We have found that ambient pretreatment of eight Argonne coals using methanol and a trace amount of hydrochloric acid improves tetrahydrofuran (THF)-soluble conversions by 24.5 wt% (daf basis) for Wyodak subbituminous coal and 28.4wt% for Beulah-Zap lignite with an average increase of 14.9 wt% for liquefaction of the eight coals at 623 K (350°C) reaction temperature and 30min reaction time. Similar enhancement results occurred using hexane or acetone in place of methanol. Pretreatment with methanol and HCl separately indicated that both reagents were necessary to achieve maximum liquefaction improvement. Acid concentration was the most important pretreatment variable studied; liquefaction reactivity increased with increasing acid concentration up to 2 ~01%. No appreciable effect on reactivity was observed at higher acid concentrations. Although vapour phase alcohol/HCi mixtures have been shown to partially alkylate bituminous coals, analysis of Wyodak and Illinois no. 6 coal samples indicated that no organic phase alteration occurred during pretreatment; however, over 90wt% of the calcium was removed from each coal. Calcium is thought to catalyse retrogressive reactions during coal pyrolysis, and thus calcium removal prior to low severity liquefaction minimizes the rate of THF-insoluble product formation. (Keywords: coal liquefaction; reactivity; chemical pretreatment)
Much of the recent research in direct coal liquefaction seeks to develop methods for dissolving coal at lower reaction severity [often defined as temperatures below 623 K (350@(Z)and pressures in the range of 6.9-10.3 MPa (lOOO-t SOOpsi)]. The incentives for developing a viable low severity liquefaction process are numerous and include : reduced hydrocarbon gas production resulting in reduced feed gas consumption and enhanced hydrogen utilization efficiency; suppressed retrogression of primary coal dissolution products resulting in enhanced distillate and residuum product quality; production of high boiling residuum which is less refractory and thus more amenable to catalytic upgrading in a conventional second-stage hydrocracker; substitution of less expensive off-the-shelf vessels, piping, valves, pumps, etc. in place of expensive, custom-designed units; less severe slurry handling and materials of construction problems as a result of lower operating temperatures and pressures. However, as shown schematically in Fiyure 1, lowering the reaction severity reduces coal conversion reactlon rates and liquid product yields unless the intrinsic coal reactivity can be sufficiently enhanced using some method of physical or chemical pretreatment prior to dissolution. Presented at the 2Olst ACS National 1419 April 1991
0016-2361,!92/09101549 c. 1992 Butterworth-Heinemann
Meeting,
Ltd
Atlanta,
Georgia,
USA,
Possible methods for reactivity enhancement include: dispersed homogeneous or heterogeneous catalysts; promoters such as basic nitrogen compounds; physical pretreatment of the coal structure; or chemical pretreatment of the coal’s inorganic and organic fractions. Generally these methods all improve low severity coal liquefaction reactivity, but for various reasons (use of exotic, expensive and sometimes hazardous chemical feedstocks, long pretreatment times and the potential for incorporating undesirable chemical constituents into the coal), none have been seriously considered as a process step in coal liquefaction. The objective of the research described here was to develop a simple, inexpensive coal pretreatment method using readily available commodity chemicals to enhance low severity liquefaction reactivity of lignites, subbituminous and bituminous coals. BACKGROUND Los
severity
coal liqwfucrion
The possibility of dissolving coal at low severity reaction conditions has intrigued researchers for many years. As early as 1921, Fischer and Schrader’ reported production of an ether-soluble material by liquefying coal at 623 K (350°C) using carbon monoxide and water as the reducing agent. More recently other groups, including the Pittsburgh Energy Technology Center2m5, the North Dakota Energy Research Center6-10, Stanford Research Institute”-14, Carbon Resources, Inc.‘s.‘h and the Colorado School of Mines’ 7-2 ‘, have investigated various methods of low severity liquefaction and have shown
FUEL, 1992, Vol 71, September
1015
Enhancing
coal liquefaction
reactivity:
K. Shams et al.
TreatedlLS
UntreatedlHS
UntreatedlLS
Reaction Figure 1
Schematic
representation
of reactivity
enhancement
using coal pretreatment
that, under certain conditions, high levels of coal dissolution can be achieved at temperatures in the range of 573-623 K (300-350°C). Conventional wisdom has held that coal must be heated to temperatures above 673-723 K (400-45O”C) before significant rates and extents of dissolution can be achieved”. Such temperatures are required to initiate thermal rupture of various labile cross-linked bonds within the three-dimensional coal structure and to form free radical intermediates. Stabilizing the free radicals with hydrogen results in products of lower molecular weight than the parent coal. This type of reaction scheme relies on excessive temperatures and pressures and long reaction times to ultimately derive a quantity of distillable products from the coal. However, high liquefaction temperatures are not selective towards distillable liquid production, and the resulting yield structure is far from optimal. Work by the groups cited above and others has demonstrated that: lignites, subbituminous and bituminous coals are inherently more reactive than previously thought; and chemical and physical coal structures can be selectively and efficiently weakened, and thus liquefaction reactivity improved, via selective chemical or physical action rather than high temperature thermal homolysis. Possible depolymerization techniques include reductive and non-reductive alkylation, acylation, partial oxidation and alkali hydrolysis. Treatments such as solvent swelling or selective demineralization can also be employed to modify coal structure. Ross and co-workers reported results of a low severity liquefaction study at temperatures below 673 K (400°C) using CO/H,0 but no organic solvent’ ‘-14. They found that CO/H,0 was superior to tetralin/H, as a conversion medium at low severity conditions. This result was attributed to the action of formate (an aqueous phase water-gas shift reaction intermediate) promoting various hydrolysis reactions involving ester groups in the coal. Porter and Kaesz15.r6 reported tetrahydrofuran (THF)soluble coal conversion levels of greater than 95 wt%
1016
FUEL, 1992, Vol 71, September
Time (LS = low reaction
severity,
HS = high reaction
severity)
(dmmf basis) at 608 K (335°C) and 3.45 MPa (500psi) CO pressure for low rank coals. Slightly lower conversion levels were noted with bituminous coals. Their process, known as ChemCoal, utilized phenolic solvent, alkali and water to chemically solubilize coal via ionic attack of selected bonds (predominantly oxygen functionalities) within the coal matrix. Results of this work are especially significant since they demonstrate that essentially complete coal dissolution can be achieved at low severity conditions. Numerous PETC researchers have conducted low temperature liquefaction studies for many yearszp5. Recent results using CO/H,0 and organic solvents suggest that higher levels of coal conversion can be achieved by using coal-derived solvent in combination with water than by using either solvent or water separately. These researchers also studied low severity liquefaction using H, gas rather than CO/H,0 or CO/HZ/H,0 and achieved THF-soluble coal conversion levels for Illinois no. 6 coal in excess of 88 wt% (daf basis) at 598 K (325°C) reaction temperature. Miller and co-workers20.2’ used basic nitrogen promoters such as 1,2,3,4-tetrahydroquinoline (THQ) and dipropylamine to enhance coal dissolution rates at low severity reaction conditions. Conversions of nearly 80wt% (daf basis) were obtained in single-stage runs using Kentucky no. 9 bituminous coal at 623 K (350°C) reaction temperature with nitrogen compound losses of less than 3 wt%. Decreased reaction severity lowered the extent of nitrogen incorporation, presumably by reducing both chemical adduction and physical entrapment. Two-stage experiments were completed in which coal was dissolved at low reaction severity in the first stage using THQ as a promoter, followed by high severity catalytic upgrading of the first stage coal-derived products to distillate. Coal conversion levels and distillate yields from this process scheme were comparable to conventional two-stage liquefaction (high severity thermal first stage/ high severity catalytic second stage) but with much lower hydrogen consumption and hydrocarbon gas
Enhancing
yields. These data demonstrated the potential for developing a viable direct liquefaction process in which coal was dissolved at milder conditions than traditionally envisaged. Coal pretreatment
to enhance liquefaction
reactivity
Modern coal structure theories suggest that coal consists of a complex, loosely bound three-dimensional structure containing numerous strong and weak crosslinkages, hydrogen bonds, guest molecules and other interactions among various functional groupsz3. Studies have shown that chemical treatments such as reductive and non-reductive alkylation24p29, acylation3’, partial oxidation31p33, alkali hydrolysis34p37 and cleavage with potassium crown ethers38, as well as physical treatments such as solvent swelling39p43, can render coal nearly 100 wt% soluble in simple organic solvents such as toluene or THF. This observation suggests that coal is a much less condensed and refractory material than once thought. Thus, the apparent low reactivity of many low rank and bituminous coals is, in fact, an artefact caused by the excessive thermal treatment and lack of available hydrogen during the initial stages of conventional thermal liquefaction. This suggests an opportunity for dissolving coal at relatively mild reaction conditions if conversion can be accomplished without irreversibly damaging the coal. We can then envisage an integrated pretreatment/liquefaction process in which: (1) the coal structure is partially disrupted via chemical or physical pretreatment to enhance reactivity; (2) the treated coal is dissolved at low severity conditions chosen to maximize coal conversion but minimize undesirable retrogressive reactions; and (3) the resulting residual products are catalytically hydrocracked to produce an additional yield of liquid products. Sternberg and Delle Donnez5 were perhaps first to demonstrate an ability to partially depolymerize subbituminous and bituminous coals at ambient reaction conditions. In this work, coal was reductively alkylated for 72-360 h using potassium metal, naphthalene, ethyl iodide and THF. Pyridine and benzene solubilities of the treated coals were found to increase with the extent of alkylation. Anthracite was found to be essentially unreactive at the reaction conditions studied. In addition to partially disrupting the coal structure by adding ethyl groups, a significant number of etheric and phenolic hydroxyl groups were also cleaved. Benzene solubilities as high as 95 wt’!! were achieved for the treated coals. Chow37 used alkali hydrolysis at 373-573 K (loo300°C) to pretreat bituminous coal, subbituminous coal and lignite. Results of high severity liquefaction experiments with tetralin solvent [683 K (41O”C), 6.9MPa (1000 psi) H,, 60 min] showed increased coal conversion to asphaltenes and oils. Schlosberg et a1.44 alkylated Wyodak subbituminous coal and Illinois no. 6 bituminous coal using aluminium chloride and methyl chloride at 373-423 K (100&l 50°C) for 1 h treatment time. The treated coal was water washed, filtered and vacuum dried prior to liquefaction in tubing bomb reactor runs. Results from high severity liquefaction runs with tetralin [700 K (427”(Z), 10.3 MPa (1500 psi) H,, 130min] showed a l&21 wt% increase (alkyl group-free basis) in cyclohexane soluble conversion for the alkylated coals. Neither of these studies measured liquefaction reactivity of the treated coals at lower reaction severity. Thus, any possible benefits due to increased coal
coal liquefaction
reactivity:
K. Shams
et al.
reactivity brought about by the pretreatment methods employed may not have been fully realized. Selected coal demineralization has also been studied as a method for enhancing liquefaction reactivity. Mochida45 reported that hydrochloric acid can be used to destroy cationic bridges present in low rank coals, thereby reducing coordination between oxygen-containing functional groups and allowing better contacting between coal and solvent during the initial stages of dissolution. Joseph and Forrai46 used ion exchange techniques to remove various cations from Wyodak subbituminous coal and North Dakota lignite. They found that removal of calcium, magnesium, sodium and potassium from each low rank coal improved high severity [673 K (4OO’C), 3.5 MPa (500 psig) H,, 60min] liquefaction conversion and product quality. This result was attributed to inhibited hydrogen transfer in the presence of alkaline and alkaline earth cations. Serio et a1.47 also found that ion-exchanged and demineralized low rank coals were more reactive at high reaction severity. They attributed these results to reduced cross-linking of the treated coals during the initial stages of dissolution. Brunson48 has shown that sulphurous acid can be used in place of hydrochloric acid to leach ion-exchangeable cations from carboxylate salts in coal. We have concluded that new methods of enhancing coal reactivity prior to liquefaction are required to obtain satisfactory coal dissolution rates and liquid yields at low severity reaction conditions. In this study, we examined the effect of pretreating coals of various ranks using the alkyl alcohol/hydrochloric acid system reported by Sharma et al. 29 because: (1) Sharma’s pretreatment method should attack ion-exchangeable sites in the coal; (2) Sharma’s chemistry can potentially alkylate selected sites in the coal (primarily phenolic and carboxylic oxygen), a prime source for retrogressive reactions during coal dissolution; and (3) small quantities of inexpensive chemical reagents are consumed during pretreatment. EXPERIMENTAL The entire suite of eight coals from the Argonne premium coal sample bank was used as the source of feed coals for this study. Ultimate analysis49 and calcium content5’ data for each coal are listed in Table 1. Additional chemical and physical properties of the Argonne coals have been reported5’. Pretreatment experiments were performed using a liquid phase technique we developed based on the gas phase alkylation chemistry reported by Sharma et a1.29. Coal was pretreated by suspending 5 g of undried coal and 0-2cm3 (O-5 ~01%) of in 40cm3 of methanol concentrated hydrochloric acid in a 100cm3 roundbottomed flask and continuously stirring the coal/ methanol slurry on a magnetic stirring plate for the desired pretreatment time (usually 3 h). The flask was connected to a cooling water condenser to reduce solvent losses by evaporation. Several experiments were completed in which dry nitrogen was used to blanket the coal/methanol slurry; elemental analyses of the treated coals showed no difference in the extent of oxidation when the system was purged with nitrogen and when it was vented to the atmosphere. Reactivity results for coal samples pretreated in each way (nitrogen blanket versus atmospheric venting) were not statistically different. Several pretreatment experiments using hexane or acetone
FUEL, 1992, Vol 71, September
1017
Enhancing Table 1
coal liquefaction
Analysis
of Argonne
reactivity:
K. Shams et al.
feed coals ~~____
Ultimate analysis (wt daf coal)
Wyodak
Beulah-Zap ~______~
Carbon
Illinois no. 6
_
Pittsburgh no. 8
Blind Canyon
~
LewistonStockton
Upper Freeport
Pocahontas
~~~~~
68.4
65.9
65.7
75.5
76.9
66.2
14.2
86.7
Hydrogen
4.9
4.4
4.2
4.8
5.5
4.2
4.1
4.2
Nitrogen
I.0
I.0
1.2
1.5
I.5
1.3
1.4
1.3
Sulphur
0.6
0.8
4.8
2.2
0.6
0.7
2.3
0.7
Oxygen
16.3
18.2
8.6
6.7
10.8
7.8
4.8
2.3
8.8
9.7
15.5
9.3
4.7
19.8
13.2
4.8
1.20
1.54
0.96
0.21
0.41
0.06
0.45
0.45
Coal rank
Subbit.
Lignite
HVB
HVB
HVB
HVB
MVB
LVB
Symbol ~_
WY
ND
IL
PITT
UT
WV
UF
POC
Ash
Calcium
content
(wt% daf coal)
in place of methanol were also completed. While a majority of the pretreatment work was performed at ambient temperature, we performed several experiments at 328 K (55°C) by placing the round-bottomed flask on a heated stirring plate. Two methods were used to recover the treated coal from the pretreatment slurry. Most of the samples were washed with 500cm3 aliquots of methanol and centrifuged to remove excess acid and soluble mineral species. Selected coal samples were rotoevaporated to remove excess solvent; residual acid and soluble mineral species were redeposited on the coal surface in this procedure. Coal samples from each separation method were then vacuum dried at 298 K (25°C) under 1.3-2.6 Pa pressure for 24 h. Untreated coal samples were vacuum dried at the same conditions before liquefaction. After drying, all treated and untreated coal samples were stored at room temperature in a vacuum desiccator at 13.3 Pa before analysis or liquefaction. Reactor runs were scheduled so that each coal sample was stored for less than 12 h before use. Portions of each untreated coal and pretreated coal were subjected to elemental analysis and ash analysis, as well as ‘H CRAMPS n.m.r., 13C CP/MAS n.m.r., FT-i.r., Mossbauer and XRD spectroscopy. Liquefaction experiments were conducted in the tubing bomb reactor system shown in Figure 2. This apparatus consisted of four main components: a matched pair of 20 cm3 tubing bomb reactors, a reciprocating arm which served to agitate each reactor and its contents, a fluidized sandbath to heat the reactors, and a high pressure gas delivery system. Prior to a liquefaction experiment, each reactor was charged with 0.5 g of coal and 1.Og of solvent; the reactor was sealed and flushed with hydrogen three times to remove any trace of oxygen from the system. Two reactors were then connected to individual gas delivery lines and pressurized to the desired initial hydrogen pressure. In most experiments, one reactor was charged with a sample of untreated coal and the second reactor was charged with a sample of the corresponding treated coal. This procedure allowed us to measure changes in liquefaction reactivity without bias from slight variations in coal composition or experimental procedures. To begin a run, the fluidized sandbath (Tecan model SBL-20) was preheated to approximately 10 K above the
1018
FUEL,
1992,
Vol 71, September
desired reaction temperature and then raised by means of an automated hoist system to immerse the shaking reactors. The temperature and pressure of the reactor contents were continuously monitored using thermocouples and pressure transducers, respectively. At the end of the desired reaction time, the sandbath was lowered and iceewater was used to quench each reactor. Reactor heat-up time to 623 K (350°C) was achieved in approximately 80 s and cool-down to temperatures below 473 K (200°C) occurred in less than 15 s. Coal conversion was monitored using THF extraction data corrected for the intrinsic THF solubilities of treated and untreated coals. Solubility measurements were conducted at ambient conditions and consisted of three steps: (1) sonicating the liquid products from the tubing bomb reactor (or feed coal sample) in excess THF for 1Omin; (2) centrifuging the mixture at 2000 rev min- ’ for 20 min; and (3) decanting THF-soluble products and excess THF from the THF-insoluble residuum. This procedure was repeated at least twice or until no additional THF-soluble products were recovered. Remaining THF-insolubles were dried at 373 K (1OOC) for 24 h to remove residual THF, weighed and finally ashed. Coal conversion to THF-soluble products was computed using the following formula: Coal conversion
M,-M = ~~~ Lx M,
100
(wt % daf basis) where M,=mass of daf THF-insoluble coal charged, M, = mass of daf THF-insoluble products recovered (corrected to satisfy the ash balance). It is important to note that this definition of coal conversion accounts for variations in the intrinsic THF solubility of treated coals. Thus reported changes in coal conversion levels can be attributed solely to inherent differences in low severity coal liquefaction reactivity and not to changes in intrinsic solubility. Unless otherwise noted, low severity reaction conditions were set at 623 K (35OC) reaction temperature, 6.9 MPa (1OOOpsig) initial cold hydrogen pressure, and 30min reaction time throughout this study. Dihydrophenanthrene (DHP) was used as hydrogen donor solvent (2:l solvent:coal wt ratio) in each liquefaction run.
Enhancing
Pressure Reactor
coal liquefaction
reactivity:
K. Shams
et al.
Reciprocating A?Xl
W
Compressor Reactors
I
I
Valve
Hydrogen Tank
Compressed Air Figure 2
RESULTS
Y Schematic
diagram
-
Arm Fluidized
-
of 20cm3
Sand Bath
microautoclave
Effect of mild chemical pretreatment on low severity liquefaction reactivity Baseline low severity liquefaction reactivity data for the untreated Argonne coals are summarized in Figures 3 and 4. At the conditions studied, three of the high volatile bituminous coals [Illinois no. 6 (75.0 wt%), Blind Canyon (69.9 wt%) and Pittsburgh no. 8 (57.0 wt%)] gave the highest THF conversions. Wyodak subbituminous coal was the next most reactive coal (42.0wt%), while
liquefaction
THF Conversion
AND DISCUSSION
Several sets of experiments were completed to evaluate the effectiveness of our pretreatment method in enhancing low severity liquefaction reactivity. In each set of runs, individual reactor experiments were duplicated and in some cases triplicated. Results shown in this paper represent average values of replicated runs; conversion differences of 2.1 wt% or greater (daf basis) represent statistically significant differences in liquefaction reactivity at the 95% confidence level.
Lift
100
reactor
system
(wt%,
daf basis)
q Untreated
q HexrnelHCI
--. ......... j..... ........ /I.......... H MeOHIHCI
”
WY
IL
ND
PITT
Feed Coal Figure 3 Effect of pretreatment with methanoliHC1 and hexane/HCl on low severity liquefaction reactivity of Argonne coals (pretreatment conditions: 1.5 ~01% HCI in methanol or hexane. 298 K, 3 h)
FUEL, 1992, Vol 71, September
1019
Enhancing coal liquefaction reactivity: K. Shams THF Conversion
(wt%,
et al.
daf basis)
Table 2
THF solubility
data for untreated
and treated Argonne
THF solubles
coals
(wt% daf basis)
80
80
40
20
Feed coal
Untreated
WY ND IL PITT UT WV UF POC
3.6 1.8 7.8 3.1 5.0
UT
WV
UF
POC
Treated
1.6
conditions:
I .5 ~01% HCI in methanol,
Table 3 Intrinsic solubility and extent of methylation pretreated by three methods
Feed Coal Figure 4 Effect of pretreatment with methanol/HCl and hexane/HCl on low severity liquefaction reactivity of Argonne coals (pretreatment conditions: I .5 ~01% HCI in methanol or hexane, 298 K, 3 h)
Pocahontas low volatile bituminous coal was the least reactive sample studied (15.6 wt%). These reactivity data follow the generally accepted trends reported for thermal conversion of the Argonne coals5’. Pretreatment with methanol and 1.5 ~01% HCl for 3 h at ambient conditions using the procedure described earlier enhanced low severity liquefaction reactivity for all eight Argonne coals. The absolute increase ranged from 24.5 wt% for Wyodak coal and 28.4wt% for Beulah-Zap lignite to 5.2 wt% for Blind Canyon coal, and averaged 14.9 wt% for the eight coals. No simple trends in reactivity improvement with chemical or physical properties of the coals were obvious, although reactivity of the pretreated low rank coals (Wyodak and Beulah-Zap) increased much more than reactivity of the six bituminous coals. As shown in Table 2, pretreatment with methanol and HCl improved the intrinsic THF solubility of each Argonne coal although, with the exception of Wyodak and Illinois no. 6 coals, the effect was minor. Once again, no simple trends in solubility improvement with physical or chemical properties of the coals were observed. As noted in the Experimental section, increased values of THF intrinsic solubility were accounted for in the THF conversion results shown in Figures 3 and 4. Although Sharma et al. 29 showed that vapour phase methanol/HCl mixtures would partially alkylate bituminous coals, elemental analyses of the treated coals indicated that little methylation (CO. 1 methyl groups/ 100 C atoms) had occurred during liquid phase methanol/ HCl pretreatment; no evidence of methylation was observed from n.m.r. and FT-i.r. measurements on untreated and treated coal samples. This result was confirmed by replacing methanol with hexane during coal pretreatment; as shown in Figures 3 and 4, hexane/HCl pretreatment also enhanced low severity liquefaction reactivity. Since hexane cannot participate in the alkylation chemistry, other effects must also contribute to the observed reactivity enhancement. Several pretreatment experiments were conducted using acetone in place of methanol or hexane and the reactivity results obtained were similar to those shown in Figures 3 and 4. Throughout this portion of the study we found no
1020
FUEL, 1992, Vol 71, September
Treatment None Methanol/HCl Liotta alkylation Sternberg
alkylation
coal”
15.2 2.1 18.2 5.6 6.2 3.1 I .4 4.0
1.0 2.7
a Pretreatment
0
coal
298 K, 3 h
for Wyodak
coal
THF solubles (wt% daf basis)
Extent of methylation (methyl groups/ 100 C atoms)
3.6 15.2 9.2 18.8 20.5
obvious trends relating reactivity enhancement, pretreatment solvent properties and coal properties. Our general conclusion is that any simple organic solvent may be used to conduct the liquid phase pretreatment step. To study the effect of methanol/HCl pretreatment versus standard alkylation pretreatment on low severity liquefaction reactivity, we methylated Wyodak coal samples using the Liotta28 and Sternbergz5 techniques. The extent of methylation and THF intrinsic solubility of each treated coal sample are summarized in Table 3. As shown, all three types of pretreatment improved THF solubility, although the extent of alkylation varied widely. Figure 5 compares the reactivity behaviour of these coal samples at low liquefaction severity. Our method provided a significant reactivity improvement even though the extent of alkylation was very low, again suggesting that effects in addition to alkylation of the
Extent
of
Alkylation
15
1”
3
”
(methyl
groups/l00
C
atoms)
Figure 5 Low severity liquefaction reactivity enhancement of Wyodak coal as a function of alkylation extent for three pretreatment methods: 0, present study; n , Liotta”. A, Sternberg”
Enhancing
coal’s organic phase contribute to the observed improvements in low severity coal liquefaction reactivity. To separate the influences of methanol and hydrochloric acid on coal pretreatment we performed a series of experiments in which Wyodak subbituminous coal and Illinois no. 6 bituminous coal were pretreated using methanol only (no HCl addition) and hydrochloric acid only (1.5~01% HCl in water, no methanol addition). Results of low severity reactivity experiments using these treated coals are summarized in Figure 6. As expected, no reactivity enhancement occurred when coal samples were pretreated with only methanol. However, coal samples treated with HCl/water exhibited significant reactivity improvement, although less than observed using methanol/HCl pretreatment. Blank pretreatment using only distilled water did not affect low severity reactivity of either coal. Thus, we can conclude that, while the presence of a small concentration of HCl is essential for successful pretreatment, the addition of methanol or other simple organic solvent enhances pretreatment effectiveness. Mochida et a1.45 attributed this effect to improved wettability of the coal surface by the organic solvent and thus better contacting between acid and coal. l?flect of pretreatment
on coal composition
To begin elucidating the mechanism of reactivity enhancement with methanol/HCl, we used several analytical techniques to study changes in the organic and inorganic phases of Wyodak and Illinois no. 6 coals after pretreatment. As mentioned earlier, n.m.r. and FT-i.r. analyses indicated no measurable organic phase alterations indicative of alkylation. However, the Wyodak FT-i.r. spectra indicated formation of carboxylic functional groups during pretreatment, probably as a result of divalent (Ca, Mg) cationic bridge destruction45346. Mijssbauer spectroscopy results demonstrated that pyrite was largely unaffected by methanol/HCl treatment, eliminating the possibility that FeCl,, a known coal dissolution catalyst, was formed in the treated coals. X-ray diffraction measurements were conducted on the low temperature ash (LTA) from treated and untreated Wyodak and Illinois no. 6 coals. These results indicated that over 90 wt% of the calcium was leached from each
THF
Conversion
(wt%,
daf
basis)
q Wyodak n Illinois
Untreated
MeOHIHCI
MeOH
coal liquefaction
reactivity:
K. Shams
et al.
coal during pretreatment. In these spectra, calcium was observed as CaCO,, with no CaO or CaSO, present. This observation agrees with a report by Miller and Givens53 that organically bound calcium converts to CaCO, rather than CaO during low temperature ashing. Elemental analyses of treated coal samples confirmed the extent of calcium loss during pretreatment. Other ion-exchangeable species such as Mg, K and Na were also leached, but were present in such small concentrations that they were not studied further in this work. Eflect of calcium reactivity
content on low secerity
liqwfaction
To study further the effect of calcium content on liquefaction reactivity, we prepared Wyodak and Illinois no. 6 coal samples with different calcium contents by varying the amount of acid used during pretreatment. Other pretreatment conditions were the same as described earlier. Results of low severity liquefaction experiments using these coals are summarized in Figure 7. The reactivity of both coals was enhanced as calcium was removed; the effect was more pronounced for Wyodak coal, particularly at a calcium content of less than about 0.2 wt%. Several low severity liquefaction experiments were completed in which calcium (as CaCO,) was added back to the reactor prior to liquefaction. In each experiment, the amount of calcium added was equivalent to the amount extracted during pretreatment. Results of these experiments are shown in Figure 8. As these data show, the beneficial effect of methanol/HCl pretreatment was almost completely negated by adding CaCO, to the reaction system. A similar effect was observed when CaO was added during low severity liquefaction of pretreated Wyodak and Illinois no. 6 coals. The mechanistic role of calcium during coal dissolution is not completely understood. Mochida et a1.45 attributed accelerated rates of low rank coal dissolution to the destruction of calcium dicarboxylate bridging structures by HCI and, therefore, less coordination of oxygencontaining functional groups. Joseph and Forrai46 speculated that calcium and other exchangeable alkaline and alkaline earth cations impeded hydrogen transfer during coal dissolution; removal of these cations would improve hydrogen transfer to coal free radicals as they form and thus improve the extent of coal conversion.
+6
UC/ wt%
Pretreatment Figure 6 Effect of pretreatment with methanol/HCl, methanol, and HCI on low severity liquefaction reactivity of Wyodak and Illinois no. 6 coals (pretreatment conditions: 298 K, 3 h)
Calcium
Content
Figure 7 Effect of calcium content on low severity liquefaction reactivity of Wyodak (0) and Illinois no. 6 (0) coals (pretreatment conditions: O-5 ~01% HCI in methanol, 298 K, 3 h)
FUEL,
1992,
Vol 71, September
1021
Enhancing
coal liquefaction
THF Conversion
(wt%,
reactivity:
K. Shams
et al.
daf basis)
THF Conversion
---I I #6
100
100
80
80
60
80
40
40
20
20
Untreated
Treated
q Untreated
/
TreatedlCaC03
588
Joseph and Forrai also cited the propensity of calcium dicarboxylate structures to undergo cross-linking reactions during the initial stages of coal dissolution. However, the fact that bulk addition of calcium (as CaCO,) inhibits low severity coal dissolution, as shown in Figure 8, suggests that calcium actively participates in primary coal dissolution reactions. We hypothesize that, in addition to the effects cited above, calcium can directly catalyse retrogressive reactions involving coal-derived free radical species during low severity liquefaction. Numerous studies have cited the role of calcium (as CaCO, or CaO) in increasing char yields and reducing tar yields during coal pyrolysis. The char yield enhancement has been attributed to catalysis of metaphast recombination prior to devolatilization54 and catalysis of repolymerization and secondary cracking reactions55. We are presently conducting a series of model compound studies to further elucidate the mechanistic effects of calcium during low severity liquefaction. parameters
on liquefaction
Pretreatment experiments were performed to study the effects of varying acid concentration, pretreatment tem-
/
FUEL, 1992, Vol 71, September
fx)
the effectiveness of coal pretreatment with methanol/HCl over a range of THF conversion levels. As shown, treated Wyodak coal was significantly more reactive at every reaction temperature studied and at every conversion level obtained. Nearly 93 wt% (daf basis) of treated
Untreated 298
Acid
TemperaWe
Eflect of reaction temperature on liquefaction reactivity of pretreated coals Figure 10 summarizes the results of a study to measure
(K)
Figure 9 Effect of HCl concentration on low severity liquefaction reactivity of Wyodak (0) and Illinois no. 6 (0) coals (pretreatment conditions: HCI in methanol, 298 K, 3 h)
673
perature and time, and separation method on low severity liquefaction reactivity of Wyodak and Illinois no. 6 coals. Results from these runs are summarized in Figure 9 and Table 4. An increased concentration of hydrochloric acid enhanced liquefaction reactivity for both coals up to 2 ~01% acid, but higher acid concentrations did not affect coal reactivity appreciably. As shown in Table 4, none of the other pretreatment parameters studied significantly influenced low severity reactivity enhancement of either coal. These data demonstrate that our pretreatment method is relatively insensitive to changes in the parameters studied. Thus, ambient temperature, a moderately short pretreatment time (approximately 3 h) and an acid concentration of about 2~01% provides near optimal pretreatment conditions for low severity coal liquefaction enhancement.
Temperature
1022
coal
Figure 10 Liquefaction reactivity of untreated and treated Wyodak coal as a function of reaction temperature (pretreatment conditions: 1.5~01% HCI in methanol, 298 K, 3 h; liquefaction conditions: 6.9MPa (1OOOpsig) H,, 30min, DHP solvent)
Table 4 Effect of pretreatment reactivity”
Vol%
n Water/
623 Reactfon
Effect of calcium carbonate addition on low severity reactivity of Wyodak and Illinois no. 6 coals (pretreatment 1.5~01%HCI in methanol, 298 K, 3 h)
Effect of pretreatment reactivity
coal
.’
Pretreatment Figure 8 liquefaction conditions:
daf basis)
0,t
1
0
(wt%,
Pretreatment time (h)
parameters
Separation method
3
328
3
298 298
3 24
filtration/ centrifugation filtration/ centrifugation rotoevaporation filtration: centrifugation
on low severity liquefaction
Wyodak conversion (wt%)
Illinois no. 6 conversion (wt%)
42.0 66.5
75.0 82.5
62.7
81.2
58.2 68.5
84.7 84.6
“Pretreatment solvent: 1.5 ~01% HCI in methanol; conditions: 623 K, 6.9 MPa H,, 30min, DHP solvent
liquefaction
Enhancing
Wyodak coal was converted to THF-soluble products at 673 K (400°C) in 30min reaction time, a conversion enhancement of 14 wt% compared to untreated Wyodak coal. Thus, our pretreatment method improved liquefaction reactivity over a wide range of conversion levels. This result is significant since it shows that reactivity enhancement can be achieved at conversion levels of commercial interest. CONCLUSIONS We have presented data which demonstrate the beneficial use of ambient, liquid phase coal pretreatment using methanol/HCl to significantly improve low severity liquefaction reactivity. The pretreatment enhances reactivity for coals over a wide range of rank, but works best with lignite and subbituminous coals. Other organic solvents such as hexane and acetone can be used to replace methanol with equally effective results. Reactivity enhancement appears to be caused by calcium removal from the pretreated coal, although the specific mechanisms involved are not completely clear. Of the pretreatment parameters studied, only acid concentration affects liquefaction reactivity; results are not influenced by changes in pretreatment times between 3 and 24 h and pretreatment temperatures between 25 and 55°C. Further work must be completed to study the impact of our mild chemical pretreatment on the conversions and yields obtained in an integrated twostage liquefaction process.
16
17 18 19
20 21 22 23 24
25 26 21 28 29 30 31
32
33
34
ACKNOWLEDGEMENTS We acknowledge financial support from the US Department of Energy under contract nos DE-AC22-88PC88812 and DE-FG22-90PC90289. Jenefer R. Olds performed the intrinsic solubility analyses for untreated and treated coal samples.
35 36
37 38
REFERENCES
11 12 13
14 15
Fischer, F. and Schrader, H. Brennstoff-Chem. 1921, 2, 257 Appell, H. R., Moroni, E. C. and Miller, R. D. Am. Chem. Sot. Diti. Fuel Chem. Prepr. 1975, 20(l), 58 Fu, Y. C. and Illig, E. G. Ind. Eng. Chem. Proc. Des. Dec. 1976. 15, 392 Bockrath, B. C., Illig, E. G., Finseth, D. H. and Sprecher, R. F. Am. Chem. Sot. Die. Fuel Gem. Prepr. 1984, 29(5), 76 Blaustein, B. D., Bockrath, B. C., Davis, H. M., Friedman, S., Illig, E. G. and Mikita, M. A. Am. Chem. Sot. Dia. Fuel Chem. Prepr. 1985, 30(2), 359 Severson, D. E., Souby, A. M. and Baker, G. G. Am. Chrm. SM. Div. Fuel Chem. Prepr. 1977, 22(6), 161 Severson, D. E., Souby, A. M. and Owens, T. C. Energy Sources 1982, 6, 173 Sondreal. E. A., Wilson, W. G. and Stenberg, V. I. Fuel 1982, 61, 925 Knudson. C. L., Willson, W. G. and Baker, G. G. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1981, 26(l), 132 Farnam, S. A., Wolfson, A. C., Miller, D. J., Gaides, G. E. and Merrick, D. D. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1985. 30(2), 354 Ross, D. S., Blessing, J. E., Nguyen, Q. C. and Hum, G. P. Fuel 1984. 63. 1206 Ross, D. S., Hum, G. P., Miin, T. C. and Green, T. K. Am. Chem. Sot,. Dit-. Fuel Chem. Prepr. 1985, 30(3), 94 Ross. D. S. ‘Exploratory Study of Coal Conversion Chemistry’, final report for US DOE contract no. DE-AC22-81PC40785, Washington, DC, December 1984 Ross, D. S., Hum, G. P. and Green, T. K. Am. Chem. Sot. Dio. Fuel Chem. Prepr. 1985, 30(4), 339 Porter, C. R. and Kaesz, H. D. Proceedings of the Thirteenth
39 40
41 42 43 44 45 46 47
48 49 50 51
52 53
54 55
coal liquefaction
reactivity:
K. Shams
et al.
Biennial Lignite Symposium, Bismarck. Grand Forks, ND, 21-23 May 1985, US DOE. pp. 357-365 Porter. C. R. and Kaesz. H. D. ‘Coal Processing Technolouv’. vol. VII, American Institute of Chemical Engmeering, New York, 1981, pp. 135-140 Miller, R. L. and Baldwin, R. M. .4m. Chem. SW. Div. Fuel Chem. Prepr. 1986, 31(l), 301 Miller, R. L. and Baldwin, R. M. .4m. Chem. Sot. Div. Fuel Chem. Prepr. 1986, 31(4), 152 Miller, R. L. ‘Liquefaction Co-Processing: A Study of Critical Variables’, Electric Power Research Institute final report for project no. 2383-1, Palo Alto, October 1987 Miller, R. L., Giacomelli, G. F., McHugh, K. J. and Baldwin, R. M. Energy & Fuels 1989,3. 127 Miller. R. L.. Baldwin. R. M. and Kennar. D. R. Am. Chem. Sot. Dir. Fuel Chem. Preor. 1990, 35(l), 9 Neavel, R. C. Fuel 1976,‘55, 237 Given. P. H.. Marzec. A.. Banton. W. A.. Lvnch. L. J. and Gerstein, B. C. Fuel 1986, 65, 155 Wender, I., Herdy, L. A., Heuworth, M. B. and Dryden, I. G. C. ‘Chemistry of Coal Utilization 2nd suppl. vol. (Ed. M. A. Elliott), John Wiley, New York, 1981 Sternberg, H. and Delle Donne, C. L. Fuel 1974, 53, 172 Larsen, J. W. and Kuemmerle, E. W. Fuel 1976, 55, 162 Flores, R. A.. Geigel, M. A. and Mayo. F. R. Fuel 1978,57,697 Liotta, R. Fuel 1979, 58, 724 Sharma, D. K., Sarkar, M. K. and Mirza, Z. B. Fuel 1985,64,449 Hodek, W. and Kolling, G. Fuel 1973, 52, 220 Hessley, R. K. ‘Co-Oxidative Depolymerization of Coal’, Electric Power Research Institute final report for project no. 2383-2, Palo Alto, June 1985 Deno, N. C., Curry, K. W., Greigger, B. A.. Jones, A. D., Rakitsky, W. G., Smith, K. A., Wagner. K. and Minard, R. D. Fuel 1980, 59, 694 Deno, N. C., Jones, A. D., Koch, C. C., Minnard, R. D., Potter, T., Sherrard, R. S., Stroh, J. G. and Yevak, R. J. Fuel 1982, 61, 490 Mirza, 2. B., Sarkar, M. K. and Sharma, D. K. Fuel Process. Technol. 1984, 9, 149 Shreve, R. N. and Brink, J. A. ‘Chemical Process Industries’ 4th edn., McGraw-Hill, New York, 1977, Ch. 33 Fischer, P. E., Ouyang, W. and Shah, R. in Proceedings of the Thirteenth Biennial Lignite Symposium, Bismarck, ND, 21-23 May 1985, US DOE, pp.421434 Chow, C. K. Fuel 1983, 62, 317 Narayan, R. and Huang, S. in Proceedings of the Tenth Electric Power Research Institute Contractors’ Conference on Clean Liquid and Solid Fuels, Palo Alto, CA, 23-25 April 1985, Electric Power Research Institute, pp. 7-l-7-25 Sanada, Y. and Honda, H. Fuel 1986.65, 295 Gorbaty, M. L. in ‘Magnetic Resonance, Advanced Topics and Applications to Fossil Energy’ (Eds L. Petrakis and J. P. Fraissard), D. Reidel, Dordrecht, 1984 Marzec, A. Fuel Process Technol. 1979. 2, 35 Bockrath, B. C., Illig. E. G. and Wassell-Bridger, W. D. Energ) & Fuels 1987. 1, 226 Larsen, J. W. and Kovac, J. Am. Chem. Sot. S.vmp. Ser. 1978, 71, 36 Schlosberg, R. H., Neavel, R. C., Maa, P. S. and Gorbaty, M. L. Fuel 1980, 59,45 Mochida, I., Shimohara, T., Korai, Y.. Fujitsu, H. and Takeshita, K. Fuel 1983, 62, 659 Joseph, J. T. and Forrai, T. R. Fuel 1992, 71, 75 Serio, M. A., Solomon, P. R., Kroo, E., Bassilakis, R., Malhotra, R., and McMillen, D. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1990, 35(l), 61 Brunson, R. J. US Patent 4 161 440, 1979 Vorres, K. S. Energy & Fue1.r 1990, 4, 420 Doughten, M. W. and Gillison, J. R. Energy & Fuels 1990,4,426 Vorres, K. S. ‘Users Handbook for the Argonne Premium Coal Sample Program’, Argonne National Laboratory, Argonne, IL. 1989 Shin, S. C., Baldwin. R. M. and Miller, R. L. Energ.v & Fuels 1989,3, 193 Miller, R. N. and Givens, P. H. ‘A Geochemical Study of the Inorganic Constituents in Some Low Rank Coals’, Technical Report for DOE contract no. EX-76-C-01-2494, Washington, DC, February 1978 Tyler, R. J. and Schafer, H. N. S. Fuel 1980, 59, 487 Franklin, H. D., Peters, W. A. and Howard, J. B. Am. Chem. Sot. Dir. Fuel Chem. Prepr. 1981, 26(2), 121 I.
FUEL,
1992,
Vol 71, September
1023
Shams,
Ronald
L. Miller
and Robert
Chemical Engineering and Petroleum Refining Golden, CO 80401, USA (Received 2 August 1991; revised 7 February
Department,
M. Baldwin Colorado
School
of Mines.
7992)
We describe results from a study in which mild chemical pretreatment of coal has been used to enhance low severity liquefaction reactivity. We have found that ambient pretreatment of eight Argonne coals using methanol and a trace amount of hydrochloric acid improves tetrahydrofuran (THF)-soluble conversions by 24.5 wt% (daf basis) for Wyodak subbituminous coal and 28.4wt% for Beulah-Zap lignite with an average increase of 14.9 wt% for liquefaction of the eight coals at 623 K (350°C) reaction temperature and 30min reaction time. Similar enhancement results occurred using hexane or acetone in place of methanol. Pretreatment with methanol and HCl separately indicated that both reagents were necessary to achieve maximum liquefaction improvement. Acid concentration was the most important pretreatment variable studied; liquefaction reactivity increased with increasing acid concentration up to 2 ~01%. No appreciable effect on reactivity was observed at higher acid concentrations. Although vapour phase alcohol/HCi mixtures have been shown to partially alkylate bituminous coals, analysis of Wyodak and Illinois no. 6 coal samples indicated that no organic phase alteration occurred during pretreatment; however, over 90wt% of the calcium was removed from each coal. Calcium is thought to catalyse retrogressive reactions during coal pyrolysis, and thus calcium removal prior to low severity liquefaction minimizes the rate of THF-insoluble product formation. (Keywords: coal liquefaction; reactivity; chemical pretreatment)
Much of the recent research in direct coal liquefaction seeks to develop methods for dissolving coal at lower reaction severity [often defined as temperatures below 623 K (350@(Z)and pressures in the range of 6.9-10.3 MPa (lOOO-t SOOpsi)]. The incentives for developing a viable low severity liquefaction process are numerous and include : reduced hydrocarbon gas production resulting in reduced feed gas consumption and enhanced hydrogen utilization efficiency; suppressed retrogression of primary coal dissolution products resulting in enhanced distillate and residuum product quality; production of high boiling residuum which is less refractory and thus more amenable to catalytic upgrading in a conventional second-stage hydrocracker; substitution of less expensive off-the-shelf vessels, piping, valves, pumps, etc. in place of expensive, custom-designed units; less severe slurry handling and materials of construction problems as a result of lower operating temperatures and pressures. However, as shown schematically in Fiyure 1, lowering the reaction severity reduces coal conversion reactlon rates and liquid product yields unless the intrinsic coal reactivity can be sufficiently enhanced using some method of physical or chemical pretreatment prior to dissolution. Presented at the 2Olst ACS National 1419 April 1991
0016-2361,!92/09101549 c. 1992 Butterworth-Heinemann
Meeting,
Ltd
Atlanta,
Georgia,
USA,
Possible methods for reactivity enhancement include: dispersed homogeneous or heterogeneous catalysts; promoters such as basic nitrogen compounds; physical pretreatment of the coal structure; or chemical pretreatment of the coal’s inorganic and organic fractions. Generally these methods all improve low severity coal liquefaction reactivity, but for various reasons (use of exotic, expensive and sometimes hazardous chemical feedstocks, long pretreatment times and the potential for incorporating undesirable chemical constituents into the coal), none have been seriously considered as a process step in coal liquefaction. The objective of the research described here was to develop a simple, inexpensive coal pretreatment method using readily available commodity chemicals to enhance low severity liquefaction reactivity of lignites, subbituminous and bituminous coals. BACKGROUND Los
severity
coal liqwfucrion
The possibility of dissolving coal at low severity reaction conditions has intrigued researchers for many years. As early as 1921, Fischer and Schrader’ reported production of an ether-soluble material by liquefying coal at 623 K (350°C) using carbon monoxide and water as the reducing agent. More recently other groups, including the Pittsburgh Energy Technology Center2m5, the North Dakota Energy Research Center6-10, Stanford Research Institute”-14, Carbon Resources, Inc.‘s.‘h and the Colorado School of Mines’ 7-2 ‘, have investigated various methods of low severity liquefaction and have shown
FUEL, 1992, Vol 71, September
1015
Enhancing
coal liquefaction
reactivity:
K. Shams et al.
TreatedlLS
UntreatedlHS
UntreatedlLS
Reaction Figure 1
Schematic
representation
of reactivity
enhancement
using coal pretreatment
that, under certain conditions, high levels of coal dissolution can be achieved at temperatures in the range of 573-623 K (300-350°C). Conventional wisdom has held that coal must be heated to temperatures above 673-723 K (400-45O”C) before significant rates and extents of dissolution can be achieved”. Such temperatures are required to initiate thermal rupture of various labile cross-linked bonds within the three-dimensional coal structure and to form free radical intermediates. Stabilizing the free radicals with hydrogen results in products of lower molecular weight than the parent coal. This type of reaction scheme relies on excessive temperatures and pressures and long reaction times to ultimately derive a quantity of distillable products from the coal. However, high liquefaction temperatures are not selective towards distillable liquid production, and the resulting yield structure is far from optimal. Work by the groups cited above and others has demonstrated that: lignites, subbituminous and bituminous coals are inherently more reactive than previously thought; and chemical and physical coal structures can be selectively and efficiently weakened, and thus liquefaction reactivity improved, via selective chemical or physical action rather than high temperature thermal homolysis. Possible depolymerization techniques include reductive and non-reductive alkylation, acylation, partial oxidation and alkali hydrolysis. Treatments such as solvent swelling or selective demineralization can also be employed to modify coal structure. Ross and co-workers reported results of a low severity liquefaction study at temperatures below 673 K (400°C) using CO/H,0 but no organic solvent’ ‘-14. They found that CO/H,0 was superior to tetralin/H, as a conversion medium at low severity conditions. This result was attributed to the action of formate (an aqueous phase water-gas shift reaction intermediate) promoting various hydrolysis reactions involving ester groups in the coal. Porter and Kaesz15.r6 reported tetrahydrofuran (THF)soluble coal conversion levels of greater than 95 wt%
1016
FUEL, 1992, Vol 71, September
Time (LS = low reaction
severity,
HS = high reaction
severity)
(dmmf basis) at 608 K (335°C) and 3.45 MPa (500psi) CO pressure for low rank coals. Slightly lower conversion levels were noted with bituminous coals. Their process, known as ChemCoal, utilized phenolic solvent, alkali and water to chemically solubilize coal via ionic attack of selected bonds (predominantly oxygen functionalities) within the coal matrix. Results of this work are especially significant since they demonstrate that essentially complete coal dissolution can be achieved at low severity conditions. Numerous PETC researchers have conducted low temperature liquefaction studies for many yearszp5. Recent results using CO/H,0 and organic solvents suggest that higher levels of coal conversion can be achieved by using coal-derived solvent in combination with water than by using either solvent or water separately. These researchers also studied low severity liquefaction using H, gas rather than CO/H,0 or CO/HZ/H,0 and achieved THF-soluble coal conversion levels for Illinois no. 6 coal in excess of 88 wt% (daf basis) at 598 K (325°C) reaction temperature. Miller and co-workers20.2’ used basic nitrogen promoters such as 1,2,3,4-tetrahydroquinoline (THQ) and dipropylamine to enhance coal dissolution rates at low severity reaction conditions. Conversions of nearly 80wt% (daf basis) were obtained in single-stage runs using Kentucky no. 9 bituminous coal at 623 K (350°C) reaction temperature with nitrogen compound losses of less than 3 wt%. Decreased reaction severity lowered the extent of nitrogen incorporation, presumably by reducing both chemical adduction and physical entrapment. Two-stage experiments were completed in which coal was dissolved at low reaction severity in the first stage using THQ as a promoter, followed by high severity catalytic upgrading of the first stage coal-derived products to distillate. Coal conversion levels and distillate yields from this process scheme were comparable to conventional two-stage liquefaction (high severity thermal first stage/ high severity catalytic second stage) but with much lower hydrogen consumption and hydrocarbon gas
Enhancing
yields. These data demonstrated the potential for developing a viable direct liquefaction process in which coal was dissolved at milder conditions than traditionally envisaged. Coal pretreatment
to enhance liquefaction
reactivity
Modern coal structure theories suggest that coal consists of a complex, loosely bound three-dimensional structure containing numerous strong and weak crosslinkages, hydrogen bonds, guest molecules and other interactions among various functional groupsz3. Studies have shown that chemical treatments such as reductive and non-reductive alkylation24p29, acylation3’, partial oxidation31p33, alkali hydrolysis34p37 and cleavage with potassium crown ethers38, as well as physical treatments such as solvent swelling39p43, can render coal nearly 100 wt% soluble in simple organic solvents such as toluene or THF. This observation suggests that coal is a much less condensed and refractory material than once thought. Thus, the apparent low reactivity of many low rank and bituminous coals is, in fact, an artefact caused by the excessive thermal treatment and lack of available hydrogen during the initial stages of conventional thermal liquefaction. This suggests an opportunity for dissolving coal at relatively mild reaction conditions if conversion can be accomplished without irreversibly damaging the coal. We can then envisage an integrated pretreatment/liquefaction process in which: (1) the coal structure is partially disrupted via chemical or physical pretreatment to enhance reactivity; (2) the treated coal is dissolved at low severity conditions chosen to maximize coal conversion but minimize undesirable retrogressive reactions; and (3) the resulting residual products are catalytically hydrocracked to produce an additional yield of liquid products. Sternberg and Delle Donnez5 were perhaps first to demonstrate an ability to partially depolymerize subbituminous and bituminous coals at ambient reaction conditions. In this work, coal was reductively alkylated for 72-360 h using potassium metal, naphthalene, ethyl iodide and THF. Pyridine and benzene solubilities of the treated coals were found to increase with the extent of alkylation. Anthracite was found to be essentially unreactive at the reaction conditions studied. In addition to partially disrupting the coal structure by adding ethyl groups, a significant number of etheric and phenolic hydroxyl groups were also cleaved. Benzene solubilities as high as 95 wt’!! were achieved for the treated coals. Chow37 used alkali hydrolysis at 373-573 K (loo300°C) to pretreat bituminous coal, subbituminous coal and lignite. Results of high severity liquefaction experiments with tetralin solvent [683 K (41O”C), 6.9MPa (1000 psi) H,, 60 min] showed increased coal conversion to asphaltenes and oils. Schlosberg et a1.44 alkylated Wyodak subbituminous coal and Illinois no. 6 bituminous coal using aluminium chloride and methyl chloride at 373-423 K (100&l 50°C) for 1 h treatment time. The treated coal was water washed, filtered and vacuum dried prior to liquefaction in tubing bomb reactor runs. Results from high severity liquefaction runs with tetralin [700 K (427”(Z), 10.3 MPa (1500 psi) H,, 130min] showed a l&21 wt% increase (alkyl group-free basis) in cyclohexane soluble conversion for the alkylated coals. Neither of these studies measured liquefaction reactivity of the treated coals at lower reaction severity. Thus, any possible benefits due to increased coal
coal liquefaction
reactivity:
K. Shams
et al.
reactivity brought about by the pretreatment methods employed may not have been fully realized. Selected coal demineralization has also been studied as a method for enhancing liquefaction reactivity. Mochida45 reported that hydrochloric acid can be used to destroy cationic bridges present in low rank coals, thereby reducing coordination between oxygen-containing functional groups and allowing better contacting between coal and solvent during the initial stages of dissolution. Joseph and Forrai46 used ion exchange techniques to remove various cations from Wyodak subbituminous coal and North Dakota lignite. They found that removal of calcium, magnesium, sodium and potassium from each low rank coal improved high severity [673 K (4OO’C), 3.5 MPa (500 psig) H,, 60min] liquefaction conversion and product quality. This result was attributed to inhibited hydrogen transfer in the presence of alkaline and alkaline earth cations. Serio et a1.47 also found that ion-exchanged and demineralized low rank coals were more reactive at high reaction severity. They attributed these results to reduced cross-linking of the treated coals during the initial stages of dissolution. Brunson48 has shown that sulphurous acid can be used in place of hydrochloric acid to leach ion-exchangeable cations from carboxylate salts in coal. We have concluded that new methods of enhancing coal reactivity prior to liquefaction are required to obtain satisfactory coal dissolution rates and liquid yields at low severity reaction conditions. In this study, we examined the effect of pretreating coals of various ranks using the alkyl alcohol/hydrochloric acid system reported by Sharma et al. 29 because: (1) Sharma’s pretreatment method should attack ion-exchangeable sites in the coal; (2) Sharma’s chemistry can potentially alkylate selected sites in the coal (primarily phenolic and carboxylic oxygen), a prime source for retrogressive reactions during coal dissolution; and (3) small quantities of inexpensive chemical reagents are consumed during pretreatment. EXPERIMENTAL The entire suite of eight coals from the Argonne premium coal sample bank was used as the source of feed coals for this study. Ultimate analysis49 and calcium content5’ data for each coal are listed in Table 1. Additional chemical and physical properties of the Argonne coals have been reported5’. Pretreatment experiments were performed using a liquid phase technique we developed based on the gas phase alkylation chemistry reported by Sharma et a1.29. Coal was pretreated by suspending 5 g of undried coal and 0-2cm3 (O-5 ~01%) of in 40cm3 of methanol concentrated hydrochloric acid in a 100cm3 roundbottomed flask and continuously stirring the coal/ methanol slurry on a magnetic stirring plate for the desired pretreatment time (usually 3 h). The flask was connected to a cooling water condenser to reduce solvent losses by evaporation. Several experiments were completed in which dry nitrogen was used to blanket the coal/methanol slurry; elemental analyses of the treated coals showed no difference in the extent of oxidation when the system was purged with nitrogen and when it was vented to the atmosphere. Reactivity results for coal samples pretreated in each way (nitrogen blanket versus atmospheric venting) were not statistically different. Several pretreatment experiments using hexane or acetone
FUEL, 1992, Vol 71, September
1017
Enhancing Table 1
coal liquefaction
Analysis
of Argonne
reactivity:
K. Shams et al.
feed coals ~~____
Ultimate analysis (wt daf coal)
Wyodak
Beulah-Zap ~______~
Carbon
Illinois no. 6
_
Pittsburgh no. 8
Blind Canyon
~
LewistonStockton
Upper Freeport
Pocahontas
~~~~~
68.4
65.9
65.7
75.5
76.9
66.2
14.2
86.7
Hydrogen
4.9
4.4
4.2
4.8
5.5
4.2
4.1
4.2
Nitrogen
I.0
I.0
1.2
1.5
I.5
1.3
1.4
1.3
Sulphur
0.6
0.8
4.8
2.2
0.6
0.7
2.3
0.7
Oxygen
16.3
18.2
8.6
6.7
10.8
7.8
4.8
2.3
8.8
9.7
15.5
9.3
4.7
19.8
13.2
4.8
1.20
1.54
0.96
0.21
0.41
0.06
0.45
0.45
Coal rank
Subbit.
Lignite
HVB
HVB
HVB
HVB
MVB
LVB
Symbol ~_
WY
ND
IL
PITT
UT
WV
UF
POC
Ash
Calcium
content
(wt% daf coal)
in place of methanol were also completed. While a majority of the pretreatment work was performed at ambient temperature, we performed several experiments at 328 K (55°C) by placing the round-bottomed flask on a heated stirring plate. Two methods were used to recover the treated coal from the pretreatment slurry. Most of the samples were washed with 500cm3 aliquots of methanol and centrifuged to remove excess acid and soluble mineral species. Selected coal samples were rotoevaporated to remove excess solvent; residual acid and soluble mineral species were redeposited on the coal surface in this procedure. Coal samples from each separation method were then vacuum dried at 298 K (25°C) under 1.3-2.6 Pa pressure for 24 h. Untreated coal samples were vacuum dried at the same conditions before liquefaction. After drying, all treated and untreated coal samples were stored at room temperature in a vacuum desiccator at 13.3 Pa before analysis or liquefaction. Reactor runs were scheduled so that each coal sample was stored for less than 12 h before use. Portions of each untreated coal and pretreated coal were subjected to elemental analysis and ash analysis, as well as ‘H CRAMPS n.m.r., 13C CP/MAS n.m.r., FT-i.r., Mossbauer and XRD spectroscopy. Liquefaction experiments were conducted in the tubing bomb reactor system shown in Figure 2. This apparatus consisted of four main components: a matched pair of 20 cm3 tubing bomb reactors, a reciprocating arm which served to agitate each reactor and its contents, a fluidized sandbath to heat the reactors, and a high pressure gas delivery system. Prior to a liquefaction experiment, each reactor was charged with 0.5 g of coal and 1.Og of solvent; the reactor was sealed and flushed with hydrogen three times to remove any trace of oxygen from the system. Two reactors were then connected to individual gas delivery lines and pressurized to the desired initial hydrogen pressure. In most experiments, one reactor was charged with a sample of untreated coal and the second reactor was charged with a sample of the corresponding treated coal. This procedure allowed us to measure changes in liquefaction reactivity without bias from slight variations in coal composition or experimental procedures. To begin a run, the fluidized sandbath (Tecan model SBL-20) was preheated to approximately 10 K above the
1018
FUEL,
1992,
Vol 71, September
desired reaction temperature and then raised by means of an automated hoist system to immerse the shaking reactors. The temperature and pressure of the reactor contents were continuously monitored using thermocouples and pressure transducers, respectively. At the end of the desired reaction time, the sandbath was lowered and iceewater was used to quench each reactor. Reactor heat-up time to 623 K (350°C) was achieved in approximately 80 s and cool-down to temperatures below 473 K (200°C) occurred in less than 15 s. Coal conversion was monitored using THF extraction data corrected for the intrinsic THF solubilities of treated and untreated coals. Solubility measurements were conducted at ambient conditions and consisted of three steps: (1) sonicating the liquid products from the tubing bomb reactor (or feed coal sample) in excess THF for 1Omin; (2) centrifuging the mixture at 2000 rev min- ’ for 20 min; and (3) decanting THF-soluble products and excess THF from the THF-insoluble residuum. This procedure was repeated at least twice or until no additional THF-soluble products were recovered. Remaining THF-insolubles were dried at 373 K (1OOC) for 24 h to remove residual THF, weighed and finally ashed. Coal conversion to THF-soluble products was computed using the following formula: Coal conversion
M,-M = ~~~ Lx M,
100
(wt % daf basis) where M,=mass of daf THF-insoluble coal charged, M, = mass of daf THF-insoluble products recovered (corrected to satisfy the ash balance). It is important to note that this definition of coal conversion accounts for variations in the intrinsic THF solubility of treated coals. Thus reported changes in coal conversion levels can be attributed solely to inherent differences in low severity coal liquefaction reactivity and not to changes in intrinsic solubility. Unless otherwise noted, low severity reaction conditions were set at 623 K (35OC) reaction temperature, 6.9 MPa (1OOOpsig) initial cold hydrogen pressure, and 30min reaction time throughout this study. Dihydrophenanthrene (DHP) was used as hydrogen donor solvent (2:l solvent:coal wt ratio) in each liquefaction run.
Enhancing
Pressure Reactor
coal liquefaction
reactivity:
K. Shams
et al.
Reciprocating A?Xl
W
Compressor Reactors
I
I
Valve
Hydrogen Tank
Compressed Air Figure 2
RESULTS
Y Schematic
diagram
-
Arm Fluidized
-
of 20cm3
Sand Bath
microautoclave
Effect of mild chemical pretreatment on low severity liquefaction reactivity Baseline low severity liquefaction reactivity data for the untreated Argonne coals are summarized in Figures 3 and 4. At the conditions studied, three of the high volatile bituminous coals [Illinois no. 6 (75.0 wt%), Blind Canyon (69.9 wt%) and Pittsburgh no. 8 (57.0 wt%)] gave the highest THF conversions. Wyodak subbituminous coal was the next most reactive coal (42.0wt%), while
liquefaction
THF Conversion
AND DISCUSSION
Several sets of experiments were completed to evaluate the effectiveness of our pretreatment method in enhancing low severity liquefaction reactivity. In each set of runs, individual reactor experiments were duplicated and in some cases triplicated. Results shown in this paper represent average values of replicated runs; conversion differences of 2.1 wt% or greater (daf basis) represent statistically significant differences in liquefaction reactivity at the 95% confidence level.
Lift
100
reactor
system
(wt%,
daf basis)
q Untreated
q HexrnelHCI
--. ......... j..... ........ /I.......... H MeOHIHCI
”
WY
IL
ND
PITT
Feed Coal Figure 3 Effect of pretreatment with methanoliHC1 and hexane/HCl on low severity liquefaction reactivity of Argonne coals (pretreatment conditions: 1.5 ~01% HCI in methanol or hexane. 298 K, 3 h)
FUEL, 1992, Vol 71, September
1019
Enhancing coal liquefaction reactivity: K. Shams THF Conversion
(wt%,
et al.
daf basis)
Table 2
THF solubility
data for untreated
and treated Argonne
THF solubles
coals
(wt% daf basis)
80
80
40
20
Feed coal
Untreated
WY ND IL PITT UT WV UF POC
3.6 1.8 7.8 3.1 5.0
UT
WV
UF
POC
Treated
1.6
conditions:
I .5 ~01% HCI in methanol,
Table 3 Intrinsic solubility and extent of methylation pretreated by three methods
Feed Coal Figure 4 Effect of pretreatment with methanol/HCl and hexane/HCl on low severity liquefaction reactivity of Argonne coals (pretreatment conditions: I .5 ~01% HCI in methanol or hexane, 298 K, 3 h)
Pocahontas low volatile bituminous coal was the least reactive sample studied (15.6 wt%). These reactivity data follow the generally accepted trends reported for thermal conversion of the Argonne coals5’. Pretreatment with methanol and 1.5 ~01% HCl for 3 h at ambient conditions using the procedure described earlier enhanced low severity liquefaction reactivity for all eight Argonne coals. The absolute increase ranged from 24.5 wt% for Wyodak coal and 28.4wt% for Beulah-Zap lignite to 5.2 wt% for Blind Canyon coal, and averaged 14.9 wt% for the eight coals. No simple trends in reactivity improvement with chemical or physical properties of the coals were obvious, although reactivity of the pretreated low rank coals (Wyodak and Beulah-Zap) increased much more than reactivity of the six bituminous coals. As shown in Table 2, pretreatment with methanol and HCl improved the intrinsic THF solubility of each Argonne coal although, with the exception of Wyodak and Illinois no. 6 coals, the effect was minor. Once again, no simple trends in solubility improvement with physical or chemical properties of the coals were observed. As noted in the Experimental section, increased values of THF intrinsic solubility were accounted for in the THF conversion results shown in Figures 3 and 4. Although Sharma et al. 29 showed that vapour phase methanol/HCl mixtures would partially alkylate bituminous coals, elemental analyses of the treated coals indicated that little methylation (CO. 1 methyl groups/ 100 C atoms) had occurred during liquid phase methanol/ HCl pretreatment; no evidence of methylation was observed from n.m.r. and FT-i.r. measurements on untreated and treated coal samples. This result was confirmed by replacing methanol with hexane during coal pretreatment; as shown in Figures 3 and 4, hexane/HCl pretreatment also enhanced low severity liquefaction reactivity. Since hexane cannot participate in the alkylation chemistry, other effects must also contribute to the observed reactivity enhancement. Several pretreatment experiments were conducted using acetone in place of methanol or hexane and the reactivity results obtained were similar to those shown in Figures 3 and 4. Throughout this portion of the study we found no
1020
FUEL, 1992, Vol 71, September
Treatment None Methanol/HCl Liotta alkylation Sternberg
alkylation
coal”
15.2 2.1 18.2 5.6 6.2 3.1 I .4 4.0
1.0 2.7
a Pretreatment
0
coal
298 K, 3 h
for Wyodak
coal
THF solubles (wt% daf basis)
Extent of methylation (methyl groups/ 100 C atoms)
3.6 15.2 9.2 18.8 20.5
obvious trends relating reactivity enhancement, pretreatment solvent properties and coal properties. Our general conclusion is that any simple organic solvent may be used to conduct the liquid phase pretreatment step. To study the effect of methanol/HCl pretreatment versus standard alkylation pretreatment on low severity liquefaction reactivity, we methylated Wyodak coal samples using the Liotta28 and Sternbergz5 techniques. The extent of methylation and THF intrinsic solubility of each treated coal sample are summarized in Table 3. As shown, all three types of pretreatment improved THF solubility, although the extent of alkylation varied widely. Figure 5 compares the reactivity behaviour of these coal samples at low liquefaction severity. Our method provided a significant reactivity improvement even though the extent of alkylation was very low, again suggesting that effects in addition to alkylation of the
Extent
of
Alkylation
15
1”
3
”
(methyl
groups/l00
C
atoms)
Figure 5 Low severity liquefaction reactivity enhancement of Wyodak coal as a function of alkylation extent for three pretreatment methods: 0, present study; n , Liotta”. A, Sternberg”
Enhancing
coal’s organic phase contribute to the observed improvements in low severity coal liquefaction reactivity. To separate the influences of methanol and hydrochloric acid on coal pretreatment we performed a series of experiments in which Wyodak subbituminous coal and Illinois no. 6 bituminous coal were pretreated using methanol only (no HCl addition) and hydrochloric acid only (1.5~01% HCl in water, no methanol addition). Results of low severity reactivity experiments using these treated coals are summarized in Figure 6. As expected, no reactivity enhancement occurred when coal samples were pretreated with only methanol. However, coal samples treated with HCl/water exhibited significant reactivity improvement, although less than observed using methanol/HCl pretreatment. Blank pretreatment using only distilled water did not affect low severity reactivity of either coal. Thus, we can conclude that, while the presence of a small concentration of HCl is essential for successful pretreatment, the addition of methanol or other simple organic solvent enhances pretreatment effectiveness. Mochida et a1.45 attributed this effect to improved wettability of the coal surface by the organic solvent and thus better contacting between acid and coal. l?flect of pretreatment
on coal composition
To begin elucidating the mechanism of reactivity enhancement with methanol/HCl, we used several analytical techniques to study changes in the organic and inorganic phases of Wyodak and Illinois no. 6 coals after pretreatment. As mentioned earlier, n.m.r. and FT-i.r. analyses indicated no measurable organic phase alterations indicative of alkylation. However, the Wyodak FT-i.r. spectra indicated formation of carboxylic functional groups during pretreatment, probably as a result of divalent (Ca, Mg) cationic bridge destruction45346. Mijssbauer spectroscopy results demonstrated that pyrite was largely unaffected by methanol/HCl treatment, eliminating the possibility that FeCl,, a known coal dissolution catalyst, was formed in the treated coals. X-ray diffraction measurements were conducted on the low temperature ash (LTA) from treated and untreated Wyodak and Illinois no. 6 coals. These results indicated that over 90 wt% of the calcium was leached from each
THF
Conversion
(wt%,
daf
basis)
q Wyodak n Illinois
Untreated
MeOHIHCI
MeOH
coal liquefaction
reactivity:
K. Shams
et al.
coal during pretreatment. In these spectra, calcium was observed as CaCO,, with no CaO or CaSO, present. This observation agrees with a report by Miller and Givens53 that organically bound calcium converts to CaCO, rather than CaO during low temperature ashing. Elemental analyses of treated coal samples confirmed the extent of calcium loss during pretreatment. Other ion-exchangeable species such as Mg, K and Na were also leached, but were present in such small concentrations that they were not studied further in this work. Eflect of calcium reactivity
content on low secerity
liqwfaction
To study further the effect of calcium content on liquefaction reactivity, we prepared Wyodak and Illinois no. 6 coal samples with different calcium contents by varying the amount of acid used during pretreatment. Other pretreatment conditions were the same as described earlier. Results of low severity liquefaction experiments using these coals are summarized in Figure 7. The reactivity of both coals was enhanced as calcium was removed; the effect was more pronounced for Wyodak coal, particularly at a calcium content of less than about 0.2 wt%. Several low severity liquefaction experiments were completed in which calcium (as CaCO,) was added back to the reactor prior to liquefaction. In each experiment, the amount of calcium added was equivalent to the amount extracted during pretreatment. Results of these experiments are shown in Figure 8. As these data show, the beneficial effect of methanol/HCl pretreatment was almost completely negated by adding CaCO, to the reaction system. A similar effect was observed when CaO was added during low severity liquefaction of pretreated Wyodak and Illinois no. 6 coals. The mechanistic role of calcium during coal dissolution is not completely understood. Mochida et a1.45 attributed accelerated rates of low rank coal dissolution to the destruction of calcium dicarboxylate bridging structures by HCI and, therefore, less coordination of oxygencontaining functional groups. Joseph and Forrai46 speculated that calcium and other exchangeable alkaline and alkaline earth cations impeded hydrogen transfer during coal dissolution; removal of these cations would improve hydrogen transfer to coal free radicals as they form and thus improve the extent of coal conversion.
+6
UC/ wt%
Pretreatment Figure 6 Effect of pretreatment with methanol/HCl, methanol, and HCI on low severity liquefaction reactivity of Wyodak and Illinois no. 6 coals (pretreatment conditions: 298 K, 3 h)
Calcium
Content
Figure 7 Effect of calcium content on low severity liquefaction reactivity of Wyodak (0) and Illinois no. 6 (0) coals (pretreatment conditions: O-5 ~01% HCI in methanol, 298 K, 3 h)
FUEL,
1992,
Vol 71, September
1021
Enhancing
coal liquefaction
THF Conversion
(wt%,
reactivity:
K. Shams
et al.
daf basis)
THF Conversion
---I I #6
100
100
80
80
60
80
40
40
20
20
Untreated
Treated
q Untreated
/
TreatedlCaC03
588
Joseph and Forrai also cited the propensity of calcium dicarboxylate structures to undergo cross-linking reactions during the initial stages of coal dissolution. However, the fact that bulk addition of calcium (as CaCO,) inhibits low severity coal dissolution, as shown in Figure 8, suggests that calcium actively participates in primary coal dissolution reactions. We hypothesize that, in addition to the effects cited above, calcium can directly catalyse retrogressive reactions involving coal-derived free radical species during low severity liquefaction. Numerous studies have cited the role of calcium (as CaCO, or CaO) in increasing char yields and reducing tar yields during coal pyrolysis. The char yield enhancement has been attributed to catalysis of metaphast recombination prior to devolatilization54 and catalysis of repolymerization and secondary cracking reactions55. We are presently conducting a series of model compound studies to further elucidate the mechanistic effects of calcium during low severity liquefaction. parameters
on liquefaction
Pretreatment experiments were performed to study the effects of varying acid concentration, pretreatment tem-
/
FUEL, 1992, Vol 71, September
fx)
the effectiveness of coal pretreatment with methanol/HCl over a range of THF conversion levels. As shown, treated Wyodak coal was significantly more reactive at every reaction temperature studied and at every conversion level obtained. Nearly 93 wt% (daf basis) of treated
Untreated 298
Acid
TemperaWe
Eflect of reaction temperature on liquefaction reactivity of pretreated coals Figure 10 summarizes the results of a study to measure
(K)
Figure 9 Effect of HCl concentration on low severity liquefaction reactivity of Wyodak (0) and Illinois no. 6 (0) coals (pretreatment conditions: HCI in methanol, 298 K, 3 h)
673
perature and time, and separation method on low severity liquefaction reactivity of Wyodak and Illinois no. 6 coals. Results from these runs are summarized in Figure 9 and Table 4. An increased concentration of hydrochloric acid enhanced liquefaction reactivity for both coals up to 2 ~01% acid, but higher acid concentrations did not affect coal reactivity appreciably. As shown in Table 4, none of the other pretreatment parameters studied significantly influenced low severity reactivity enhancement of either coal. These data demonstrate that our pretreatment method is relatively insensitive to changes in the parameters studied. Thus, ambient temperature, a moderately short pretreatment time (approximately 3 h) and an acid concentration of about 2~01% provides near optimal pretreatment conditions for low severity coal liquefaction enhancement.
Temperature
1022
coal
Figure 10 Liquefaction reactivity of untreated and treated Wyodak coal as a function of reaction temperature (pretreatment conditions: 1.5~01% HCI in methanol, 298 K, 3 h; liquefaction conditions: 6.9MPa (1OOOpsig) H,, 30min, DHP solvent)
Table 4 Effect of pretreatment reactivity”
Vol%
n Water/
623 Reactfon
Effect of calcium carbonate addition on low severity reactivity of Wyodak and Illinois no. 6 coals (pretreatment 1.5~01%HCI in methanol, 298 K, 3 h)
Effect of pretreatment reactivity
coal
.’
Pretreatment Figure 8 liquefaction conditions:
daf basis)
0,t
1
0
(wt%,
Pretreatment time (h)
parameters
Separation method
3
328
3
298 298
3 24
filtration/ centrifugation filtration/ centrifugation rotoevaporation filtration: centrifugation
on low severity liquefaction
Wyodak conversion (wt%)
Illinois no. 6 conversion (wt%)
42.0 66.5
75.0 82.5
62.7
81.2
58.2 68.5
84.7 84.6
“Pretreatment solvent: 1.5 ~01% HCI in methanol; conditions: 623 K, 6.9 MPa H,, 30min, DHP solvent
liquefaction
Enhancing
Wyodak coal was converted to THF-soluble products at 673 K (400°C) in 30min reaction time, a conversion enhancement of 14 wt% compared to untreated Wyodak coal. Thus, our pretreatment method improved liquefaction reactivity over a wide range of conversion levels. This result is significant since it shows that reactivity enhancement can be achieved at conversion levels of commercial interest. CONCLUSIONS We have presented data which demonstrate the beneficial use of ambient, liquid phase coal pretreatment using methanol/HCl to significantly improve low severity liquefaction reactivity. The pretreatment enhances reactivity for coals over a wide range of rank, but works best with lignite and subbituminous coals. Other organic solvents such as hexane and acetone can be used to replace methanol with equally effective results. Reactivity enhancement appears to be caused by calcium removal from the pretreated coal, although the specific mechanisms involved are not completely clear. Of the pretreatment parameters studied, only acid concentration affects liquefaction reactivity; results are not influenced by changes in pretreatment times between 3 and 24 h and pretreatment temperatures between 25 and 55°C. Further work must be completed to study the impact of our mild chemical pretreatment on the conversions and yields obtained in an integrated twostage liquefaction process.
16
17 18 19
20 21 22 23 24
25 26 21 28 29 30 31
32
33
34
ACKNOWLEDGEMENTS We acknowledge financial support from the US Department of Energy under contract nos DE-AC22-88PC88812 and DE-FG22-90PC90289. Jenefer R. Olds performed the intrinsic solubility analyses for untreated and treated coal samples.
35 36
37 38
REFERENCES
11 12 13
14 15
Fischer, F. and Schrader, H. Brennstoff-Chem. 1921, 2, 257 Appell, H. R., Moroni, E. C. and Miller, R. D. Am. Chem. Sot. Diti. Fuel Chem. Prepr. 1975, 20(l), 58 Fu, Y. C. and Illig, E. G. Ind. Eng. Chem. Proc. Des. Dec. 1976. 15, 392 Bockrath, B. C., Illig, E. G., Finseth, D. H. and Sprecher, R. F. Am. Chem. Sot. Die. Fuel Gem. Prepr. 1984, 29(5), 76 Blaustein, B. D., Bockrath, B. C., Davis, H. M., Friedman, S., Illig, E. G. and Mikita, M. A. Am. Chem. Sot. Dia. Fuel Chem. Prepr. 1985, 30(2), 359 Severson, D. E., Souby, A. M. and Baker, G. G. Am. Chrm. SM. Div. Fuel Chem. Prepr. 1977, 22(6), 161 Severson, D. E., Souby, A. M. and Owens, T. C. Energy Sources 1982, 6, 173 Sondreal. E. A., Wilson, W. G. and Stenberg, V. I. Fuel 1982, 61, 925 Knudson. C. L., Willson, W. G. and Baker, G. G. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1981, 26(l), 132 Farnam, S. A., Wolfson, A. C., Miller, D. J., Gaides, G. E. and Merrick, D. D. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1985. 30(2), 354 Ross, D. S., Blessing, J. E., Nguyen, Q. C. and Hum, G. P. Fuel 1984. 63. 1206 Ross, D. S., Hum, G. P., Miin, T. C. and Green, T. K. Am. Chem. Sot,. Dit-. Fuel Chem. Prepr. 1985, 30(3), 94 Ross. D. S. ‘Exploratory Study of Coal Conversion Chemistry’, final report for US DOE contract no. DE-AC22-81PC40785, Washington, DC, December 1984 Ross, D. S., Hum, G. P. and Green, T. K. Am. Chem. Sot. Dio. Fuel Chem. Prepr. 1985, 30(4), 339 Porter, C. R. and Kaesz, H. D. Proceedings of the Thirteenth
39 40
41 42 43 44 45 46 47
48 49 50 51
52 53
54 55
coal liquefaction
reactivity:
K. Shams
et al.
Biennial Lignite Symposium, Bismarck. Grand Forks, ND, 21-23 May 1985, US DOE. pp. 357-365 Porter. C. R. and Kaesz. H. D. ‘Coal Processing Technolouv’. vol. VII, American Institute of Chemical Engmeering, New York, 1981, pp. 135-140 Miller, R. L. and Baldwin, R. M. .4m. Chem. SW. Div. Fuel Chem. Prepr. 1986, 31(l), 301 Miller, R. L. and Baldwin, R. M. .4m. Chem. Sot. Div. Fuel Chem. Prepr. 1986, 31(4), 152 Miller, R. L. ‘Liquefaction Co-Processing: A Study of Critical Variables’, Electric Power Research Institute final report for project no. 2383-1, Palo Alto, October 1987 Miller, R. L., Giacomelli, G. F., McHugh, K. J. and Baldwin, R. M. Energy & Fuels 1989,3. 127 Miller. R. L.. Baldwin. R. M. and Kennar. D. R. Am. Chem. Sot. Dir. Fuel Chem. Preor. 1990, 35(l), 9 Neavel, R. C. Fuel 1976,‘55, 237 Given. P. H.. Marzec. A.. Banton. W. A.. Lvnch. L. J. and Gerstein, B. C. Fuel 1986, 65, 155 Wender, I., Herdy, L. A., Heuworth, M. B. and Dryden, I. G. C. ‘Chemistry of Coal Utilization 2nd suppl. vol. (Ed. M. A. Elliott), John Wiley, New York, 1981 Sternberg, H. and Delle Donne, C. L. Fuel 1974, 53, 172 Larsen, J. W. and Kuemmerle, E. W. Fuel 1976, 55, 162 Flores, R. A.. Geigel, M. A. and Mayo. F. R. Fuel 1978,57,697 Liotta, R. Fuel 1979, 58, 724 Sharma, D. K., Sarkar, M. K. and Mirza, Z. B. Fuel 1985,64,449 Hodek, W. and Kolling, G. Fuel 1973, 52, 220 Hessley, R. K. ‘Co-Oxidative Depolymerization of Coal’, Electric Power Research Institute final report for project no. 2383-2, Palo Alto, June 1985 Deno, N. C., Curry, K. W., Greigger, B. A.. Jones, A. D., Rakitsky, W. G., Smith, K. A., Wagner. K. and Minard, R. D. Fuel 1980, 59, 694 Deno, N. C., Jones, A. D., Koch, C. C., Minnard, R. D., Potter, T., Sherrard, R. S., Stroh, J. G. and Yevak, R. J. Fuel 1982, 61, 490 Mirza, 2. B., Sarkar, M. K. and Sharma, D. K. Fuel Process. Technol. 1984, 9, 149 Shreve, R. N. and Brink, J. A. ‘Chemical Process Industries’ 4th edn., McGraw-Hill, New York, 1977, Ch. 33 Fischer, P. E., Ouyang, W. and Shah, R. in Proceedings of the Thirteenth Biennial Lignite Symposium, Bismarck, ND, 21-23 May 1985, US DOE, pp.421434 Chow, C. K. Fuel 1983, 62, 317 Narayan, R. and Huang, S. in Proceedings of the Tenth Electric Power Research Institute Contractors’ Conference on Clean Liquid and Solid Fuels, Palo Alto, CA, 23-25 April 1985, Electric Power Research Institute, pp. 7-l-7-25 Sanada, Y. and Honda, H. Fuel 1986.65, 295 Gorbaty, M. L. in ‘Magnetic Resonance, Advanced Topics and Applications to Fossil Energy’ (Eds L. Petrakis and J. P. Fraissard), D. Reidel, Dordrecht, 1984 Marzec, A. Fuel Process Technol. 1979. 2, 35 Bockrath, B. C., Illig. E. G. and Wassell-Bridger, W. D. Energ) & Fuels 1987. 1, 226 Larsen, J. W. and Kovac, J. Am. Chem. Sot. S.vmp. Ser. 1978, 71, 36 Schlosberg, R. H., Neavel, R. C., Maa, P. S. and Gorbaty, M. L. Fuel 1980, 59,45 Mochida, I., Shimohara, T., Korai, Y.. Fujitsu, H. and Takeshita, K. Fuel 1983, 62, 659 Joseph, J. T. and Forrai, T. R. Fuel 1992, 71, 75 Serio, M. A., Solomon, P. R., Kroo, E., Bassilakis, R., Malhotra, R., and McMillen, D. Am. Chem. Sot. Div. Fuel Chem. Prepr. 1990, 35(l), 61 Brunson, R. J. US Patent 4 161 440, 1979 Vorres, K. S. Energy & Fue1.r 1990, 4, 420 Doughten, M. W. and Gillison, J. R. Energy & Fuels 1990,4,426 Vorres, K. S. ‘Users Handbook for the Argonne Premium Coal Sample Program’, Argonne National Laboratory, Argonne, IL. 1989 Shin, S. C., Baldwin. R. M. and Miller, R. L. Energ.v & Fuels 1989,3, 193 Miller, R. N. and Givens, P. H. ‘A Geochemical Study of the Inorganic Constituents in Some Low Rank Coals’, Technical Report for DOE contract no. EX-76-C-01-2494, Washington, DC, February 1978 Tyler, R. J. and Schafer, H. N. S. Fuel 1980, 59, 487 Franklin, H. D., Peters, W. A. and Howard, J. B. Am. Chem. Sot. Dir. Fuel Chem. Prepr. 1981, 26(2), 121 I.
FUEL,
1992,
Vol 71, September
1023