Action of Solvents on Coal: Extraction of a Pittsburgh Seam Coal with

Action of Solvents on Coal: Extraction of a Pittsburgh Seam Coal with Aniline, Tetralin, and Phenol at Elevated Temperatures. R. S. Asbury. Ind. Eng. ...
1 downloads 0 Views 705KB Size
Action of Solvents on Coal Extraction of a

..

Soxhlet extraction studies of coal were made with three aromatic solvents. With aniline a t 225" C., tetralin up t o 400" C., and phenol up to 300" C., yields of 47, 85, and 67 per cent, respectively, were obtained. Comparison of yields a t comparable temperatures obtained by the two polar solvents (aniline and phenol) with yields from the nonpolar solvents (tetralin and benzene) shows the greater solvent action of the polar liquids; since greater action indicates greater dissociating power, it is suggested that increased yields with aniline and phenol result from depolymerization of the coal structure. Higher yields with tetralin a t the higher temperatures may be the result of greater depolymerization of the coal and decomposition of the solvent to hydrogen and consequent reaction with the coal. Microscopic examination of the extracts seems t o show that the extracts were in true solution under the conditions of extraction. Comparison of molecular weights of the ether-insoluble fractions from the extracts with the ratios of ether insolubles t o ether solubles in the extract also shows that the polar solvents exert a stronger dissociating or depolymerizing action on the coal substance than do the nonpolar solvents. Further, comparison of these molecular weights with those of bitumens, pseudo bitumens, and humic acids derived from coal substantiates the thesis t h a t coal may be regarded as essentially a polymer of relatively small units.

Pittsburgh Seam Coal with Aniline, Tetralin, and Phenol at Elevated Temperatures R. S. ASBURY Coal Research Laboratory, Carnegie Institute of Technology, Pittsburgh, Pa.

T

HE action of heat on coal in the presence of

solvents has been shown to be the mildest type of pyrolysis which can be readily investigated. Solvent extraction provides, therefore, an important tool in a study of the mechanism of thermal decomposition of coal, an understanding of which is essential for an accurate description of carbonization, combustion, and hydrogenation of coal. In all these processes coal is subjected to elevated temperatures: in carbonization, in the absence of air; in combustion, in the presence of air; and in hydrogenation, in the presence of hydrogen, a dispersion medium, and a catalyst. Besides having general interest in these industrial processes, the work reported in this paper has a special bearing on the hydrogenation of coal, since previous investigators have shown that the results obtained depend on the dispersion medium, or solvent, used. To be effective, a dispersion medium must render the coal susceptible to attack by hydrogen. Certain solvents may aid in the process also by their capacity for adding hydrogen. Tetralin, in particular, has been shown to act in part as a hydrogen carrier, thereby effectively accelerating the rate of hydrogenation. A thorough knowledge of the action of solvents on coals is essential, therefore, for an understanding of the hydrogenation process. Although the action of many solvents on various coals has been studied by numerous investigators, the mechanism of the extraction process is still not known. The effects of the variables time, temperature, and coal particle size were studied in a previous investigation ( I ) in which benzene was used as solvent for the same coal as used in the present study, a Pittsburgh seam coal from the Edenborn Mine. This study of the extraction process has been extended to include three additional aromatic solvents, aniline, tetralin, and phenol, which, because of their higher critical temperatures, allow the extraction to be conducted at higher temperatures, and also permit partial evaluation of the effect of chemical nature of the solvent on the solvent process. The action of one or more of these solvents was investigated by Berl and Schildwachter (S), Brazer and Hoffman (6),Keppeler and

Borchers ( 6 ) ,Novak and Hubacek (?'), Parr and Hadley (8), Pertierra (9),Pott, Broche, and Scheer (IO), and others, but in most cases the examination of the extract was superficial. In the present study the chemical nature of the solvent was of primary interest. Of the remaining variables, the effects of time and temperature were also investigated, but particle size was not considered. I n order to investigate thoroughly the effect of nature of the solvent, yield determinations were made with each solvent for soluble products, fine coal, and undissolved residues; the soluble products were split into acidic, basic, phenolic, and ether-soluble and -insoluble bodies; elementary analyses were made of all products; molecular weight determinations were run on the ether-insoluble fractions of the dissolved materials; and examination of the 687

INDUSTRIAL AND ENGXNEERING CHEMISTRY

688

extracts in their respective solvents were made with the aid of the microscope.

Apparatus and Procedure The apparatus used in these experiments was the same as that in which Edenborn coal was extracted with benzene (1) except for the following items: (a) the water-cooled steel condenser was replaced by a coil of seamless steel tubing through which medium-grade lubricating oil was pumped by a small gear pump to a water-cooled heat exchanger; (a) a device for indicating the frequency of siphoning (B) was attached to the outlet from the siphon cup and made possible for the first time complete knowledge of siphoning in pressure extraction systems. Siphonings were kept between six and ten per hour by changing the speed of pumping, the temperature of inlet oil to condenser, and wattage input t o the furnace surrounding the extractor. Previous work with benzene as solvent showed that the yield of extract was independent of the particle sBe from 4 to 80 mesh. With all three solvents only one size of coal was studied-16 to 20 mesh Pittsburgh seam coal from the Edenborn Mine-since this size was most readily handled in the apparatus. Extractions were made in a manner similar to that described in the benzene extraction studies (Z), except for the following differences:

ANILINEL One liter of c. P. aniline was used in each stage. The yield of extract in each stage was determined by acidifying aliquot samples from each stage with hydrochlorio acid, filtering off the precipitated extract from the aqueous solution of aniline hydrochloride, washing, drying, and weighing.

TABLEI. SUMMARIZED DATAON ANILINE,TETRALIN, AND PH~NO PRESSURE L EXTRACTIONS Solvent

Aniline 25 18-20 225

Run. No.

Coalaiae, mesh Temp., O C. Time hours: 22& c. 250" C. 300" C. 350' C. 400' C. Total Total No. of stages Coal sample weight: Grama Loss in weight of coal: Grams

Tetralin 27 18-20

{ EOO:200: f

484

... ... ... ...

484 72

li0 128 108 4 374 79

125

150

57.3" 45.8" 53.4"

w-

W&ht of extract, grams Yield of extract, based on orig. wt. of 42.7" coal, % Yield of extract, cor. for fine coal % 47b Yield at infinite time, cor. for hne 67" cpal, % Weight of fine coal, grams 20.6" Weight of residue, grama 47.P Weight of extract fine ooal residue, g r a m 121.15 Balance, 98.8" Corrected for aniline assumed to be in products at nitrogen content. b Estimated. Equal to extract fine coal residue original weight of coal 'O0'

+

+

+

Phenol 31 16-20 250.300

...

44 21

... ...

86 23 150

96.5 63.7 117.1

87.3 58.2 97.5

78.1 85.5

85.0 86.7

> 90

88 10.9 51.8

26.2 28.3

171.6 180.2 114.3 107.0 end of run, based on

+

TABLE11. YIELDSOF TETRALIN EXTRACTS Temp.

c. 260

Yield Qtanaa 49.8

Yield Based on Original Coal

% RR 3

Yield Cor. for Fine Coal Extrapolated t o infinite Obsvd. time % % R? Q

VOL. 28, NO. 6

TETRALIN.One liter of tetralin, made by vacuum distillation of Eastman technical-grade tetralin, discarding the last 10 per cent, was used in each stage. The extract yield for each stage was determined by vacuum distillation of aliquot samples over a water bath. PHENOL. About 900 cc. of warmed c. P. phenol were used in each stage. On removal of the extract solution from the pressure extractor, sufficient water was added to make a 7 per cent solution of water in phenol, thereby keeping the mixture fluid a t room temperature and facilitating sampling and handling. Yield determinations were made by vacuum distillation of aliquot samples over a water bath. Corrections for the nonvolatile matter in phenol were applied to each stage yield. I n all cases it was necessary to correct for fine coal mechanically washed into the extract within the pressure extractor. The fine coal was determined by glass Soxhlet extraction of the extract-fine coal mixture with the solvent in question at atmospheric pressure. Undissolved residues remaining within the alundum thimble were weighed and recorded as fine coal. Corrections were applied to the yield of extract a t each stage for this fine coal removed. Consequently in all three cases the percentage yield of extract is somewhat higher than the corresponding value obtained on the basis of the original weight of coal. Determination of the yields with each solvent was followed by investigations of the chemical nature of the extraot and properties of the products.

Yields of Soluble Products The data on yields of soluble products obtained with all three solvents are summarized in Table I. Yields a t infinite time were obtained by extrapolation of the curves of per cent yield us. inverse time of extraction. Results of analyses of the products by micromethods (given later) in the case of run 25 showed such an excess of nitrogen that a correction of the yield was made on the assumption that aniline was present in the products in proportion to the excess nitrogen. From the hydrogen contents of the extracts, we must conclude that the aniline was not physically retained as such but reacted with the coal with loss of hydrogen which did not appear as gas. The yield figures uncorrected for retpined aniline are 59.3 per cent yield based on original weight of coal, and 65.0 per cent corrected for fine coal, with a material balance of over 125 per cent. In the tetralin run a high balance, 114.3 per cent, was obtained with high hydrogen and oxygen contents in the products. Excess oxygen may be attributed to air oxidation of the products, excess hydrogen probably to tetralin or its polymers. Unfortunately it was not possible to check this point since no element could serve as a basis as was the case with nitrogen in aniline. I n agreement with Pott, Broche, and Scheer (IO) and others, decomposition of tetralin was found to occur around 400" C. Consequently only a few stages at this temperature were attempted. Difficulties of a similar nature were encountered with phenol, but a t a lower temperature. At 300' C . much of the extract produced with this solvent had the odor of, and a refractive index close to that of, phenyl ether. The increase in extract a t this higher temperature was only 1 per cent, and therefore extraction a t 300" C. was abandoned after the seventh stage. Extract solutions with all solvents were dark brown in the early stages, lightening in color somewhat toward the end of the runs. The tetralin extracts were highly fluorescent and precipitated on the walls of the bottles slowly. The aniline and phenol extracts were not fluorescent, and precipitation on the walls was negligible. The yields of tetralin extract produced up to and including each temperature shown are included in Table 11.

JUNE, 1936

INDUSTRIAL AND ENGINEERING CHEMISTRY

The maximum yield of products soluble in benzene at 260' C., extrapolated to infinite time, amounted to about 30 per cent (1). The indicated yields for .aniline, tetralin, and phenol a t infinite time a t the lowest temperature of extraction used in this work are 57,40, and 67 per cent, respectively. With the two nonpolar solvents, benzene and tetralin, the yields are not greatly different but are distinctly lower than the yields with the polar solvents, aniline and phenol. The greater solvent action of these polar liquids indicates their greater dissociating power, and hence it is suggested that increased yields with aniline and phenol a t this lowest temperature result from depolymerization of the coal structure into smaller structural units. With tetralin as solvent, a temperature of about 300" C. is required to give the same yield as obtained with phenol a t 250". Extrapolation to infinite time of the yield of extract obtained with tetralin as solvent a t 350° and 400" C. is uncertain. The data do not exclude the possibility of complete solution of the coal in tetralin a t these temperatures, which may result from greater depolymerization of the coal a t these high temperatures and from decomposition of the solvent to give hydrogen which would probably react with the coal under these conditions. Biggs (4) has shown that hydrogenation of this Pittsburgh seam coal in the absence of a solvent a t temperatures of 350" to 380" C. greatly increases its solubility in benzene. He suggests that the coal is depolymerized a t these temperatures, and that the hydrogen reacts a t the bonds liberated by the depolymerization so that the smaller stabilized units become soluble in weakly dissociating solvents such as benzene and tetralin. The data reported in this paper indicate that this stabilization is unnecessary for solution in polar solvents of considerable dissociating power, such as aniline and phenol.

Chemical Separations Separations of the aniline, tetralin, and phenol extracts into acids, bases, phenols, neutral ether solubles, and neutral ether insolubles were made with the aid of caustic and mineral acid as described under the benzene studies ( I ) . Table HI summarizes these data. Comparison of the data in Table I11 shows the similarity in amounts of these products to those obtain%d with benzene a t the lower temperatures, in that acids, bases, and phenols are but a small part of the extract, the neutral bodies predominating. However, the ratio of ether-soluble to etherinsoluble bodies does not remain constant; the latter in the case of the tetralin rises to a maximum between 300" and 350" C. and decreases thereafter, as occurred in the work on benzene. Likewise, increase in phenols and bases occurs with increasing temperature, further substantiating the original idea that basic and phenolic materials are formed from etherinsoluble bodies.

Elementary Analyses of Products Elementary analyses of all products produced by the solvents under investigation were made and are shown in Table IV. Analyses of these results were made with the aid of material balances. As pointed out under the results on chemical separation of the tetralin products, some material remained in the products even after long vacuum treatment. Run 27 (Table IV) shows that hydrogen-carbon ratios for the extracts are higher than for ether insolubles from the same stages. For example, extract for 300" C. stages is 8.72; ether insolubles for 300" C. stages, 6.12. Hence some material high in hydrogen is retained by the ether solubles. Since the hydrogen-carbon ratio in tetralin is 9.09, in all probability polymers of tetralin are responsible for high ratios in the extract.

689

TABLE 111. COMPOSITION OF EXTRACTS Solvent Aniline Temp., C. 225 Total extract, yo 47 Compn. of extract, %: Acids 0.07 Bases 1.30 Phenols 0.48 Ether-sol. 12.78 Ether-insol. 9 5 , 4 0 Loss

250

300

Tetralin

350

400

.

Phenol

250

33.8

17.0

30.8

3.9

66.7

1.2 0.5 0.9 38.2 59.2

0.9 0.4 1.0 25.5 72.3

0.6 0.6 1.8 31.3 65.7

0.3 1.1 2.0 69.2 27.4

0.2 0.2 1.0 10.6 86.4 1.7 100.0

-10.03 On Oa 00 Oa Total 100.00 100.0 100.0 100.0 100.0 a In the case of tetralin extracts, when aliquot samples had been dried down in preparation for chemical separations, excess product (sometimes 300 per cent of that calculated to be present) was found. Since long highvacuum treatment did not reduce the weights of materials. separations were made. In every case ether-insoluble compounds remained relatively constant and the ether solubles made up the difference whether the yield had been 150 or 250 per cent of the amount calculated to be present from yleld data. Knowing that tetralin polymerizes readily above room temperature these excess products must have been tetralin polymers which could be discarded by use of the formula, ether solubles = (100 - (acids bases phenols ether insolubles) j resulting in the figures in this table. These figures on yields of acidic, phenolic, and basic constituents in the tetralin ex eriments are in substantial agreement with those found by Berl and Schdwschter (8).

+

+

TABLE IV.

+

ELEMENTARY ANALYSES OF PRODUCTS (DRYBASIS, IN PERCENT)

Coal

100 H C

c

H

N

S

0

79.06

5.13

1.65

1.00

5.61

7.55

6.48

7.62 16.23 6.79

1.08 8.54 14.23

5.62 6.70 6.54

Ash

Run 25, Aniline Extract Fine coal Residue

80.49 65.98 70.38

4.52 4.42 3.90

5.73 0.56 4.36 0.47 4.20 0.50 Run 27. Tetralin

83.80 87.08 87.45 87.80

6.81 7.78 7.62 7.12

0.44 0.56 0.38 0.77

0.14 0.49 0.27 0.23

8.81 3.57 4.17 3.98

0.0 0.52 0.11 0.10

8.13 8.92 8.72 8.12

1.16 1.46

0.43 0.70

10.00 9.51

1.14 0.95

6.16 6.12

1.67 0.81 1.62 0.81 1.50 0.82 0.71 1.28 31, Phenol 0.88 1.44 0.89 1.62 0.88 0 . 8 0

10.47 11.84 9.93 4.13

4.47 3.65 4.39 42.47

6.18 6.35 6.20 5.15

8.31 9.32 9.72

2.40 2.24 21.16

6.37 6.01 5.65

Extract: Composite 250° C. stages 300° C. stages 350° C.atages Ether-insol: Composite 30O0 C.stages Fine Coal: 250° C. stages 300° C. stages 350° C.stages Residue

82.21 82.35

5.06 5.03

77.78 77.18 78.50 48.89

4.80 4.90 4.86 2.52

Extract Ether-insol. Residue

81.77 81.15 64.77

5.20 4.88 3.66

Run

The sum of the oxygen in the extracts, fine coals, and residues was greater than that of the original coal in all three cases. In all probability this excess was due to atmospheric oxidation of products and to hydrolysis during extraction. Part of the excess oxygen from phenol extractions may be accounted for by phenol reacting with or remaining in the products. If the yield of extract is calculated from the ash content of the residue, assuming the extract to contain none of t h e mineral matter of the original coal, the yield obtained with aniline is 46.9 against 4 i per cent as given in Table I, and similarly for tetralin, 82.2 against 85.5 per cent, and for phenol 64.4 against 66.7 per cent. The ash observed in t h e analysis of the extracts can be explained by slight disintegration of the alundum thimbles used for fine coal determinations and filtration of aniline extracts. The checks on yield figures obtained above tend to indicate that the extract was in true solution in the solvent since peptization would probably carry mineral matter into the extract in proportion to t h e amount in the original coal. Evidence from microscopic examination, presented later, likewise indicates true solution of extract in the solvent.

Molecular Weights of Ether Insolubles Molecular weights of the ether insolubles were determined in catechol by the method described by Smith and Howard

INDUSTRIAL AND ENGINEERING CHEMISTRY

690

VOL. 28, NO. 6

(la)and are given in Table V. Previous work has shown this fraction of the extract to represent the compounds of higher molecular weight. TABLE V. MOLECULAR WEIGHTSOF Run No. 25 27

a

Solvent Aniline, 225' C. Tetralin 250' C. stages 27 Tetrelin' 300' C. stages Tetralin: 350' C.stages 27 27 Tetralin, 400' C. stages 31 Phenol, 250' C. Averages of two or more determinations.

and residues range from 200 to 300, as pointed out by Biggs (4); and they are of the same order of magnitude as aniline, tetralin, and phenol extracts, indicating that depolymerization by heat in the presence of a solvent yields a material similar in size to that obtained by chemical reaction and supETHERINSOLUBLESporting the thesis that coal may be regarded as essentially a Mol. Weight polymer of relatively small units ( 1 2 ) . i n Catechola 354 305 407 293 205 358

Molecular weights of tetralin extracts in catechol, like the percentages of ether insolubles in the extract, increase to a maximum a t 300' C. and decrease a t higher temperatures. Obviously a thermal dissociation occurs with this coal in the range 300" to 350' C., causing reduction in molecular size of the extract. Earlier work ( I ) indicated that the ether-insoluble substances represent that fraction of the extract which has suffered least thermal decomposition and that the ether solubles, phenols, etc., may result from a secondary thermal decomposition of the ether insolubles. A relatively large yield of high-average molecular weight, therefore, indicates a relatively small amount of secondary thermal decomposition. From this point of view an analysis of molecular weight data in Table V in conjunction with ratios of ether insolubles to ether solubles in Table I11 and yields of ether insolubles a t each temperature offers further support to a stronger dissociating or depolymerizing action of the polar substances on the coal substance. With both the polar solvents aniline and phenol, the ratios of ether insolubles to ether solubles are high and molecular weights are almost identical; with tetralin, the ratios are much lower and molecular weights increase and decrease together with these ratios as the temperature increases. Since the ether insolubles do not make up the entire extract, and since the molecular weights of ether solubles are lower than ether insolubles, the average molecular weights of the entire extract in catechol would be less than the values given in Table V. Molecular weights of bitumens, pseudo bitumens, and humic acids prepared from benzene extracts

Microscopic Examination of Products Examination with the dark-field microscope a t magnifications up to 980 diameters shows that with aniline and phenol extracts in their respective solvents, numerous colloidal particles below u , size are in evidence in Brownian movement, and with tetralin extracts only the aggregates of particles not in motion appear. Considering the fact that these solutions were examined a t room temperature. it is probable that most of the extract was in true solution under the conditions of pressure extraction, in agreement with the studies on benzene extracts.

Acknowledgment The author is indebted to N. W. Franke for assistance throughout this work, and to F. C . Silbert and T. B. Smith for determinations of ultimate composition of the products.

Literature Cited (1) Asbury, R. S., IND. EXQ.CHEM.,26, 1301 (1934). ENQ.CHEM.,Anal. Ed., 8, 152 (1936). (2) Asbury, R. S.,IND. (3) Berl, E., and Schildwachter, H., Brennstof-Chem., 9, 105-13 (1928). (4) Biggs, B. S.,Carnegie Inst. Tech., Coal Research Lab., unpub-

lished data. (5) Frazer, J. C. W., and Hoffman, E. J., Bur. Mines, Tech. Paper 5 (1912). (6) Keppeler, G.,and Borchers, H., Brennstof-Chem., 15, 241-5 (1934). (7) Novak, H., and Hubacek, J., Paliva a Topent, 10,28-33 (1928). (8) Parr, S. W., and Hadley, H. F., J. Gas Lighting, 129, 260 (1915); Univ. Ill. Eng. Expt. Sta., Bull. 76 (1914). (9) Pertierra, J. M.,Fuel, 13, 23-6 (1933). (IO) Pott, A,, Broche, H., and Scheer, W., Fuel, 13,91 (1934). (11) Smith, R. C.,and Howard, H. C., Carnegie Inst. Tech., Coal

Research Lab., unpublished data.

(12) Smith, R. C.,and Howard, H. C., J . Am. Chem. Soc., (1935).

57, 512

RECEIVED March 9, 1936. Presented before the Division of G w and Fuel Chemistry a t the 9 l s t Meeting of the American Chemical Society, Kansaa City, Mo., April 13 t o 17,1936.

Courtesy, Bromborough Port Est4