330
INDUSTRIAL AND ENGINEERING CHEMISTRY
needed. Many of the deposits vere extremely irregular and would have been objectionable if they had been applied to leaves for protection from fungi. Others were satisfactory as first laid down but became much more defective after exposure t o rain than the loss in their weight indicated. It was intended to record the characteristics of the deposits by suitable photographs, but time and facilities did not permit it. This important step together with a comparison with the deposits on leaves for a selected few of the materials should be taken. I n spite of the fact that little information was supplied about the composition of the samples, some conchaions can be drawn fiom the results of the tests even if nothing more than the weights of the deposits are considered. Thirty-one of the 110 samples rctained weighable residues through four artificial rains of 1.33 inches; 23 withstood eight of these rains or else became stabilized and apparently would have maintained themselves through an indefinite number of rains. From the little that is known about their chemical make-up the following conclusions can be drawn: Those materials that are chiefly notable for their detergency, or effect upon the wetting and spreading properties of water, are unsuited for the purposes of the present work. Examples are the Tergitols, Penetrol, and petroleum oil sulfonates (Acto materials). None of the bentonites and none of the starch products offered promise. There is a suggestive similarity between the performance and appearance of materials under code numbers 4001 and 1501. Both appear t o have several com-
Coal Hy
Vol. 42, No. 2
ponents, one of which probably confers the observed effectiveness and the rest are required to permit that one to be properly dispersed. That such auxiliaries are not always necessary is demonstrated by 2501, polyethylene glycol monooleate, which was dispersible-actually soluble-in Tvater jus? as received, and which gave a good performance. The parallel performance manifested by 2513, propylene glycol dioleate, and 2515, glycerol monooleate, prevents an interpretation of the effect of a angle as opposed to two esterifying fatty acid residues. The unmodified soaps were unsuccessful. SUMMARY
An apparatus for the testing of adhesives for horticultural mist dusting and the method of operation is described. One hundred and ten samples, mostly commercial offerings, werc tested by it, and the results are presented. LITER4TURE CITED Annual Report, Cornel1 University, 1945. Cupples, H. L., U . S. Dept. A g r . , Agr. Research Admin., Bur. Entornol. Plant Quarantine, E Series, Ciw. 426, 504, and 607. Glaves, J . Agr. Eng., 28, 551-2 (1947). Green, E. I,., IKD.Esc. CHEM.,19, 931 (1927). Green, E. L., unpublished d a t a . Green, E. L., a n d Goldsvorthy, M. C., Phytopathology, 27, 957-70 (1937). V a n h n t w e r p e n , F. J., IXD. ENG.CHEM.,35, 126 (1943).
RECEIVED January 12, 1949.
enation Ca
BATCH AUTOCLAVE TESTS SOL W’ELLER, R I . G. PELIPETZ, SAM FRIEDBIAN, AND H. H. STORCH B u r e a u of Mines, P i t t s b u r g h , P a . Comparative catalyst tests have been made for the hydrogenation of whole Bruceton coal (Pittsburgh bed), hand-picked Bruceton anthraxylon, and Rock Springs coal (Wyoming subbituminous). In all cases, ammonium chloride added by itself either had no action or decreased the liquefaction of coal. Tin added by itself showed moderate catalytic activity. The combination of tin plus halogen acid, however, shows a remarkable synergism and, with the possible exception of germanium plus halogen acid, constitutes perhaps the best known catalyst for coal hydrogenation. Ammonium chloride, hydrochloric acid, carbon tetrachloride, and the chloroacetic acids were all found to be essentially equivalent as promoters for tin. Sodium chloride is inert, while elemental chlorine exerts a harmful effect. RIolybdena, nickel on kieselguhr, and copper chromite were found to be relatively ineffective in all of these tests. Iron compounds were also found to be useless at initial hydrogen pressures of 1000 pounds per square inch. Zinc shows appreciable catalytic activity in the presence of ammonium chloride. I t was found possible to replace a t least 90% of the tin in a tin-ammonium chloride catalyst by zinc without appreciable loss of catalytic effectiveness. The physical distribution of tin was found to be important. However, if the physical distribution is good, tin may apparently be used i n almost any chemical form; tin, stannous sulfide, metastannic acid, and tin tetraphenyl were found to be equally effective in combination with ammonium chloride. Increase of metal (tin or zinc) concentration above 0.5% has little ef-
fect on the coal hydrogenation. The principal effect of an increase in ammonium chloride content (in the presence of tin or zinc) is to decrease the production of asphalt. High pressure tests (3700 pounds per square inch initial hydrogen pressure) on Rock Springs coal indicate that iron catalysts, such as ferrous sulfate, pyrite, and “red mud,” can be used to good effect though they fall considerably short of tin plus ammonium chloride or zinc plus ammonium chloride in promoting coal liquefaction.
A
TTEMPTS to improve ?he economic status of coal hydro genation has led the Bureau of Mines t o a continuing search for cheaper and more effective catalysts for the process. These studies have two aspects: one, a semiempirical trial of all reasonable catalysts; the other, an attempt to understand the mechanism by which the best catalysts operate, so that improvement of catalysts can be put on a rational basis. Only a few comparative studies of coal hydrogenation catalysts have been published ( 1 , 4 , 7 , 8, 10). The work reported here constitutes an extension of published results to additional catalysts and coals. APPARATUS AA’D PROCEDURE
A detailed description of the hydrogenation equipment has been published ( 6 ) . A Pyrex No. 774 glass liner was used in all of the experiments; this is vital in catalyst studies, for otherwise the autoclave will show “memory” effects. No vehicle was used in any of the experiments. The 1.2-liter autoclave was charged with 50 grams of dry, powdered coal and v i t h catalyst,
INDUSTRIAL AND ENGINEERING CHEMISTRY
February 1950
II
*
flushed and filled with hydrogen t o an initial pressure of 1000 pounds per square inch (or 3700 pounds per square inch for the high-pressure tests), heated to 450” C. during a eriod of about 80 minutes, held at 450” C. for 1 hour, and coofed. After the autoclave was cooled, the gases were bled and analyzed by mass spectrometer. The autoclave residue was extracted a t room temperature with 400 ml. of benzene; the slurry was centrifuged and the centrifuge residue was exhaustively extracted with benzene in a Soxhlet apparatus. The residue in the Soxhlet thimble, after drying, constituted the benzene-insoluble material. The total benzene extract was freed of benzene by distillation; the residue was then treated with about five times its own weight of n-hexane. After the mixture had stood 16 hours, the hexaneinsoluble material was filtered, washed with two 75-ml. portions of hexane, and dried. This benzene-soluble, hexane-insoluble material is defined as “asphalt.” The per cent liquefaction was based on the organic matter in the benzene-insoluble material; it was obtained by the formula - benzene insolubles x 100 yo liquefaction = original coaloriginal coal all quantities being expressed on a moisture- and ash-free (m.a.f.) basis. The quantity given in the following tables as “asphalt per unit liquefaction” is the ratio of grams of asphalt per grams of coal liquefied. I n the high liquefaction range (per cent liquefaction >SO) the values of per cent li uefaction found in duplicate experiments agree within about 2 2 ; when the liquefaction is low (< 60%) the agreement between duplicates is 4 to 5%. The per cent liquefaction is a measure of the coal converted to gases and liquids. The amount of coal converted to liquid products may be determined by subtracting the per cent gas from the per cent liquefaction. Since the gas production did not vary greatly in these experiments (Tables I t o V), differences in the liquefaction reflect (approximately) differences in the production of liquids. The figures for “asphalt per unit liquefaction” are of significance since they show the fraction of the coal converted which goes into the production of asphalt. The production of asphalt expressed as per cent of moisture- and ash-free coal can be obtained by multiplying the per cent liquefaction by the asphalt per unit liquefaction. OF COALS TABLE I. ULTIMATEANALYSES
Constituent
Whole Bruceton Coal, %
Bruceton Anthraxylon, %
Rock Springs Coal, %
TABLE 11. CATALYSTS FOR BRUCETON COAL (lOOO-lb./sq. inch initial Hz pressure: 1 hr. a t 450° C.)
%
Run No. Catalyst 311 None 516 1% Sn 518 ~~. 0.6% NHaCl 481 1% Sn 0.5% NHaCl 684 I 7 SnS 306 SnS 0.5% NH4C1
Lique. faction 55.0 66.8 40.8 86.1 70.0 84.5
16.2 12.9 13.8
Asphalt per Unit Liquefaction 0.28 0.47 0.21 0.28 0.53 0.16
433 434 441 435 996 1117 1118
70.0 66.3 65.8 55.2 78.5 51.4
13.6 13.3 13.1 13.5 13.5 13.4
0.38 0.32 0.28 0.31 0.33 0.43
+
3n4 . .
314 310 295 309 294 316 313 303
CATALYSTS FOR BRUCETON COAL
In Table I1 are summarized comparative tests of a number of catalysts on Bruceton coal. All of these experiments were carried out a t an initial pressure of 1000 pounds per square inch. The tin-ammonium chloride or tin sulfide-ammonium chloride combination was studied in the first group of experiments. Tin and tin sulfide used alone are about equally effective; each increases the liquefaction by a moderate amount but also increases the proportion of asphalt in the liquefied product. Ammonium chloride by itself is a negative liquefaction catalyst. The combination of tin and ammonium chloride or tin sulfide and ammonium chloride shows a remarkable synergism, resulting in a high liquefaction and low asphalt production. Molybdenum trioxide by itself is comparable in effectiveness to tin or tin sulfide. It seems, however, not to be susceptible to the
1% Moos 1% MOO3 0.5% NHaCl 1% Moos on F.E.a 17 MoOaon F.E.= 0.570 NH&1 MOO3 HI Ammonium molybdate (1% Mo) Ammonium molybdate (1% Mo) 0.5% NHiCl
+
+
ld
+
%
Gas 14.0 12.9
...
44.2
13.6
0.36
AS 0.5% NHaCl 0.5% NHICl 1% Zn 0.5% NHaCl 1% Sb Cd 0.5% NHaCl 1% 0.5% NH4Cl 1% Se Bi 0.5% NHaCl 1% 0.5% NHaCl 1 % As-Zn (50:50) Zn-Sb (50:50) 0.5% NH4Cl 1% 0.5% NHaCl 1% Zn-Sb (50:50)
62.5 59.6 30.0 33.9 33.3 31.0 63.0 35.7 87.1
12.9 14.6 14.6 13.4 14.2 12.3 14.2
0.43 0.20 0.23 0.72 0.31 0.73 0.62 0.27 0.19
1% fuller’s earth
13.6 11.7 12.2 14.0 12.6 12.7
0.54 0.23 0.18 0.64 0.21 0.14 0.34
~1%
++ + +++
+ ++
...
12.8
682 71 1 774 471 713 714 772
2%
5 % Dicalite 911 5 % Dicalite, speed flow 5 % Activated charcoal
53.6 39.5 29.9 65.0 40.0 38.8 28.7
896 918 972
Ni on kieselguhrb (1% Ni) 2% Cu ohromitec 2% Cu chromite (CIT)d
68.6 63.8 63.4
13.4 12.9 13.4
0.59 0.54 0.46
HC1 HI
40.2 66.6 46.6 41.8
12.6 11.7 11.4
0.48 0.39 0.49 0.26
30.0 44.8 51.0 37.0 44.3 77.1
12.1 13.6 11.9 11.4 11.9 11.9
0.75 0.34 0.65 0.15 0.35 0.58
83.8
13.8
0.51
85.8
...
0.42
86.2
14.0
0.26
77.6 88.3 80.3
14.7 14.1
...
0.29 0.18 0.28
39.0
...
0.22
85.2
14.5
0.53
66.1 73.3 76.7
12.6 13.5 13.6
0.56 0.54 0.53
79.8 83.4 70.2 80.0 88.8 86.4 84.6 79.9 85.2 87.8 88.2 87.5 70.0
13.5 12.8
0.52 0.61 0.57 0.44 0.54 0.32 0.37 0.53 0.32 0.44 0.52 0.21 0.69
862 850 962 1270
5 % fuller’s earth 0.5% NHaCl 5 % Bentonite
+ Thermofor cracking catalyst
0.5% NHaCle 0.5% NH4Cle HzsOa 2: 0.5% NHaCle HBr C 0.5% Cle
+
296 773 873 867 841 789
170Fe 0.6% NH4Cl FeSOa (1% Fe) Pyrite (1% Fe) FeCl3 on bentonitef (1% Fe) 1% stainless steel (303) filings 5 % driedredmudg 0.5% NHaCl
655
1% Zn 0.1% Sn (as NHaC1) 0 1 % Sn 1% Zn 1% (asZn NHaC1) 0.1% Sn
666
Three coal materials were used in the course of these studies. These were Pittsburgh-bed coal from the Bureau’s experimental mine at Bruceton, Pa., hand-picked anthraxylon from the same mine, and Rock Springs coal from the D. 0. Clarke mine, No. 9, Superior, Wyo. Ultimate analyses of these coals, on a moisturefree basis, are shown in Table I. With the exception of the stainless steel (303) filings, all of the solid catalysts were powdered before use. Hydrochloric, sulfuric, and hydriodic acids were added as aqueous solutions. Rotation of the autoclave during the heating-up period was relied on to provide mixing of coal and catalyst.
+ +
1.3
664
MATERIALS
331
616 526 575
+
+
+ 0.5%
C1
1% C1 ++ ++ 1.5% C1 0.1% (as NHaCl) Sn + 1.5% C1 (as NH4C1) 17’ Sn + 1 5 % C1 (as NHaC1) 1% Zn + 1:5% C1 (as NHaC1)
728 76 1 809 781 835 ~ 7 n SnClt (0.1% Sn) 806 SnCh (0.1% Sn), impregnated 611 0.1% Sn 0.1% Cl (as NHaCl) 1110 SnClz (0.5% Sn) 802 SnClz (0.6% Sn), impregnated 589 SnClz ( 1 7 Sn) 1147 SnClz (1 Sn) i impregnated 871 SnClz on F.E.L (0.1% Sn) GOS SnClz on F.E.h (1% Sn) Sn(CaHs)a (1%Sn) 0.570 ”acl 824 813 0 , 5 % NHaCl SnOz ( 1 7 Sn) 0.5% NHaCle 1% Sn HI 860 1% Sn HaSOa C 0 . 5 % NHaCle 818 1% Sn (single piece) 912 0.5% NHaCl 913 1% Sn (shot, ca. 8 mesh) 0.5% NHaCl
+
%
+ +
964 a
GeOt (1%Ge)
+ +
+ +
+ 0.5% NHaCl
5 % Moos on fuller’s earth.
11.6
...
...
14.0 13.7
ii.5
12.9 14.2 15.0 13.3 14.3 14.6
56.5
13.8
0.41
58.2
13.6
0.37
86.8
13.1
0.51
Catalyst contains 46%,Ni. Adkins’ catalyst. obtained from the Harshaw Chemical Co. Adkins’ catalyst(; obtained through courtesy of R. A. Glenn, Carnegie Institute of Technolog 6 The equivalence orAcids was on a molar basis. f Catalyst contains 27% Fe. 0 Residue from Bayer aluminum process, obtained by courtesy of Aluminum Company of America. h Catalyst contains 10% Sn. b
C
d
332
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 42, No. 2
the apparent negative catalysis by most of these high surfaceCONCEI~TRATION, BRUCETOW area materials is due to an increase of organic insolubles because TABLE111, EFFECTOF CATALYST COAL of an irreversible adsorption by these materials of normally soluble (lOOO-lb./sq. inch initial Hz pressure: 1 hr. a t 450' C.) coal hydrogenation products. Asphalt % Runs 896, 918, and 972 indicate that nickel-kieselguhr and per Unit Liquecopper chromite (Adkins' catalyst), which are very effective catafac% LiqueRun tion Gas faction Catalyst No. lysts in other systems, are only weakly effective for coal hy570 1% Zn + 0.1% Cl (as xHaC1) 62.3 13.1 0.53 drogenation; this may be due to a poisoning by sulfur compounds 79.6 1 3 . 7 0 . 4 8 1% Zn + 0.5% C1 (as NHaClj 571 79.7 13.0 0.32 1% Zn + 1.0% C1 (as NHaCU liberated during coal hydrogenation. 573 575 1% Zn + 1.5% C1 (as NHaCI) 80.3 13.6 0.28 Hydrochloric acid by itself is a negative catalyst (run 862), as 1% Zn + 2.0% C1 (as NHaCl) 73.5 13.6 0.25 627 628 1% Zn + 2.5% C1 (as NHaCl) 76.0 ,.. 0.20 are sulfuric acid (run 862) and hydrobromic acid (run 1270). 631 1% 75,3 13.7 0.23 . - Zn + 3.0% C1 (as "aC1) Hydriodic acid is unique among halogen acids in promoting 654 1.5% Zn + 2.5% C1 (as NHaC1) 76.8 15.8 0.28 656 2.0'7 Zn 4- 2.5% C1 (as NHaC1) 78.6 ... 0.24 liquefaction when added alone ( 2 ) . I n run 996 hydriodic acid 663 2.5d Zn + 2.5% C1 (as "&I) 18',.6 acted as a promoter in enhancing the effect of molybdenum 3.0% Zn + 2.5% C1 ( a s hHaC1) 665 trioxide. 523 1% Sn 4- 0.1% C1 (as XI-IaCl) 73.8 12.8 0.48 1% Sn + 0.5% C1 (as NH4Clj 86.7 14.5 0.52 524 With one exception, all of the iron compounds tested are non525 1% Sn + 1.0% C1 (as NHaClj 86.2 13.1 0.20 526 1% Sn + 1.5% C1 (as SHPC1) 88.3 14.1 0.18 catalytic for Bruceton coal. The exception is dried red mud, a 76.4 ... 0.59 by-product of the Bayer aluminum process; the particular sample 0 5 7 Sn f 0 1 % C1 (as NHaClj 577 86.8 .. . 0.44 0 : 5 4 Sn + 0:5% C1 (as "&I) 579 used contained 16% ferric oxide and 5% titanium dioxide. This 8 6 . 0 . . . 0.26 XHdC1) 0 . 5 9 Sn + l . O $ C1 (as 580 0 . 5 g Sn + 1.5% C1 (as NHaC1) 83.7 12.7 0.21 material is roughly equivalent to the German "Bayermasse," 611 0.1% Sn + 0.1% C1 (as WHaCl) 70.2 ... 0.57 which was successfully used in German coal hydrogenation plants. 612 0.17 Sn + 0.6% C1 (as NHnC1) 79.9 13.8 0.40 Filings of type 303 stainless steel, which had previously been re615 0 . l d Sn f 1.?% C1 (as NHaCIj 79.0 12.3 0.33 616 0.1% 6 n + 1.0% C1 (as NHIC1) 77.6 13.5 0.29 ported to be useful in increasing the throughput of a coal hydrogenation plant (9),were found to have no catalytic effect. The next group of experiments in Table I1 demonstrates that promoting effect of ammonium chloride, Ammonium molybdate under the proper conditions, a t least 90% of the tin in a tinammonium chloride catalyst can be replaced by zinc without imgives results with Bruceton coal even poorer than those obtained pairing the catalytic activity. When zinc constitutes the bulk with molybdenum trioxide, of the catalyst, it seems to be desirable to increase the amI n the third group of experiments, a number of metals and monium chloride concentration above the usual amount in ordcr metal alloys were tested in combination with ammonium chloride. to obtain equivalent results. This finding may have practical Of the metals tested, arsenic and zinc show some catalytic acimportance if tin is considered as a catalyst for commercial plants, tivity, but antimony, cadmium, selenium, and bismuth are withsince supplies of zinc are much more available to this country out value for Bruceton coal. The excellent results with zinc-tin than are tin supplies. alloy (run 303) are undoubtedly largely due to the tin content. I n the next group of tests a study was made of different forms The next group of experiments shows that the presence of a of zinc. The use of the oxalate or of any alloy with zinc providcs high surface-area material is not sufficient to guarantee catalytic no better results than those obtained with powdered, metallic, activity, Of the group of materials tested, only thermofor crackzinc. ing catalyst shows any catalysis of the liquefaction, and, even here, a high asphalt production is observed. It is possible that The final set of experiments listed in Table I1 constitutes an investigation of various ways of adding tin. At low concentration levels, such as 0.1 % tin, there seems to be a marked dependence OF CHLORINE-CONTAIBING PROMOTERS of liquefaction on mode of addition of stannous chloride or tin TABLE IV. COMPARISON plus ammonium chloride. The most efficient way of using stan(lOOO-lb./sq. inch initial He pressure: 1 hr. a t 450' C.: 1% Sn) nous chloride is to impregnate the coal with it. Sddition of Chlorine-Containing Bspka!t powdered stannous chloride is less effective, and the use of Compounds __ yo per L nit Run % Lique% Liquepowdered tin plus an equivalent amount of ammonium chloride of 01 faction Gas faction No. Type is quite inefficient. Presumably, this is a question of achieving 0.21 ... 1.O 86.8 h"aC1 509 proper distribution of the small amount of tin among the coal 14.1 0.18 88.3 1.5 NHiCl 526 0.51 13.9 75.6 0.1 H C1 527 particles which have to react. A t a level of 0.50j, tin, impregna0.34 ... 85.9 0.5 HC1 528 0.25 13.3 86.5 1.0 HC1 tion with stannous chloride is still superior to the use of powdered 531 0.21 ... 81.4 1.5 HC1 530 stannous chloride, but a t 1% tin there no longer seems to be any 0.55 15.1 65.0 NaCl 0.5 533 0.49 14.0 1.0 71.3 NaCl 511 difference between the two modes of addition. Stannous chloride 0.47 13.7 1.5 67.2 NaCl 537 0.58 ... 70.9 CCljCOOH 0.1 supported on fuller's earth is essentially equivalent to unsup540 0.47 86.2 0.5 CClsCOOH 541 ported, powdered stannous chloride. Apparently tin can be 0.15 CClsCOOH 88.8 1.0 548 0.16 ... 86.3 1.5 CCIsCOOH 551 used in almost any finely divided form; tin tetraphenyl and Bruceton Anthraxylon stannic oxide (ignited metastannic acid) are as effective as 13.3 0.08 0.0 29.5 None powdered tin. 48Za 0.57 14.5 85.3 0.0 None 480 Hydriodic acid as a promoter for tin is as good as, or better 0.08 11.7 22.6 0.4 NH4C1 783a 0.44 14.6 92.9 0.4 hTHaCl 484 than ammonium chloride. Sulfuric acid, on the other hand, 0.33 14.8 93.5 "4Cl 0.7 486 16.3 0.19 91.0 1.0 has almost no promoting effect. I t is not possible a t the present 487 NHiCl 0.20 18.6 91.0 1.3 NHiCl 488 time to state with any certainty the reason for the peculiar suit0.39 93.3 13.6 cc14 0.5 506 0.27 92.2 ... CCla 1.0 507 ability of halogen acids. 0.33 91.6 ... 491 HC1 0.4 0.25 Runs 912 and 913 show that the physical distribution of tin is 15.1 92.9 0.7 489 HC1 0.17 ... 90.6 492 1.4 HC1 of great importance, a t least in autoclave experiments with dry 20.4 0.12 502 0.5 74.0 c12 20.9 0.16 1.0 77.0 503 C12 coal. The use of a single piece of stick tin or of tin shot, instead 0.30 14.0 CHClsCOOH 0.5 94.4 1129 0.52 of powdered tin plus ammonium chloride, gives results essentially 13.1 CClsCOOH 0.5 92.2 1141 13.3 0.44 NaCl 0.5 83.7 490 identical with those in the absence of any catalyst. This behav0.47 15.2 NaCl 499 1.0 85.0 ior does not agree with that observed by Booth ( 5 ) , but the dif0 No 6 n was used in runs 482 or 783. ference may be due to the absence of vehicle in these experiments;
E!:(:
:::;
INDUSTRIAL AND ENGINEERING CHEMISTRY
February 1950
use of an acidic vehicle might result in the solution of appreciable amounts of tin from a single piece. The last experiment of Table I1 shows that resulbs almost as good as those with tin plus ammonium chloride can be obtained with germanium dioxide plus ammonium chloride. Data of the Fuel Research Board (4) lead to a similar conclusion. Germanium seems t o be the only element of the periodic table which, used in the manner described, is as satisfactory as tin for coal hydrogenation. It is, of course, even less available than tin. s
TABLEV.
1
Run No. 43 1 1113 43 2 453 430 428 463 406 1121 1122
CATALYSTS FOR ROCKSPRINGS COAL
(lOOO-lb./sq. inoh initial Hz pressure: 1 hr. a t 450' C.) Aspha1,t % per'unit Lique% Liquefaction Gas faotion Catalyst None 30.9 13.3 0.01 13.6 0.17 44.2 1% i n 0.01 33.5 0 5 7 NH&l 1'5.3 0.37 8n 0.5% NH4Cl 88.6 0.04 16.9 36.4 1% Moos 0.10 16.0 38 8 1% MoOa 0.5% NHrCl 0.44 36.1 ... 17' M o o t o n F.E." 0.02 ... 34.7 MoOa on F.E." 0.5% "acl 0.03 11.9 33.7 Ammonium molybdate (1% Mol Ammonium mol bdate (1% Mo) 0.10 13.9 38.9 0.5%
+
lk
+
14
+
+ NHJI
5% Moos on fuller's earth.
STUDY OF CATALYST CONCENTRATION
.
In Table I11 are summarized data showing the effect of varying the concentration of zinc, tin, and ammonium chloride in the hydrogenation of Bruceton coal. Both with zinc and with tin, increase of ammonium chloride concentration a t fixed metal concentration results in an increase of liquefaction to a maximum value and then a slow decrease; the decrease is almost within experimental error. (With 1% tin, however, the position of any maximum must occur a t a chlorine-as ammonium chlorideconcentration greater than 1.5%.) In all of the experiments, the importance of ammonium chloride in promoting the reduction of asphalt is obvious; the asphalt production per unit of liquefaction decreases continuously with increase of ammonium chloride concentration. Increase of the zinc concentration from 1 to 3%, with the ammonium chloride concentration maintained a t 2.5%, results in a relatively small increase of liquefaction from 76 to 83%. The extent of asphalt production is only slightly dependent on the zinc concentration in this range. At a given concentration level of ammonium chloride, 0.5% tin is almost as effective as 1% tin, but the use of 0.1% tin results in significantly poorer liquefactions. The data of the upper part of Table IV show the effect of various concentrations of several chlorine-containing compounds. These experiments were all made with Bruceton coal in the presence of 1% tin. With the exception of run 530, which is a little out of line, it is clear that ammonium chloride, hydrochloric acid, and trichloroacetic acid are essentially equivalent a t the same concentration of total chlorine. Sodium chloride is inert as a promoter for tin. It is probable that all of the active chlorine-containing compounds form hydrochloric acid under reaction conditions and that hydrochloric acid is the real promoter in these experiments. The experiments of the lower part of Table IV constitute a similar study of chlorine-containing materials carried out on Bruceton anthraxylon rather than on whole Bruceton coal. Here again ammonium chloride, hydrochloric acid, carbon tetrachloride, dichloroacetic acid, and trichloroacetic acid show approximately the same promoting action when compared a t the same chlorine concentration. Sodium chloride is inert, as expected, but elemental chlorine exerts a definitely deleterious action. In view of the I'apid reaction rate of chlorine with coal, it
333
seems likely that chlorination of the coal occurred well before the reaction temperature of 450' C.was reached (6). Run 484, Table IV, is quite surprising. Although anthraxylon is generally considered the petrographic constituent of coal most easily hydrogenated, it is clear that this is true only in the presence of an appropriate catalyst. In the absence of any added catalyst, whole Bruceton coal shows a 55% liquefaction under the given conditions (run 311, Table 11); in the absence of catalyst Bruceton anthraxylon is only 30y0 liquefied. Addition of ammonium chloride alone (run 783) makes the liquefaction even poorer (compare run 518, Table 11), while use of tin alone (run 483) produces a tremendous increase in liquefaction to 85%. CATALYSTS FOR ROCK SPRINGS COAL
The data of Table V show a comparison of several catalysts for the hydrogenation of Rock Springs coal, which is of lower rank than Bruceton coal. The experimentti of Table V were all carried out under an initial hydrogen pressure of loOC, pounds per square inch. Rock Springs coal resembles Bruceton anthraxylon in the respect that in the absence of catalyst, hydrogenation a t 1000 pounds per square inch initial pressure results in only about 30% liquefaction, The use of 0.5% ammonium chloride alone has very little effect on the liquefaction in this case; 1% tin used alone exerts only a moderate catalytic effect. The synergistic effect of the tin-ammonium chloride combination is very striking with Rock Springs coal; the liquefaction jumps t o almost 90% in the presence of 1% tin plus 0,5y0ammonium chloride. As is the case with Bruceton coal, molybdenum trioxide is found to be an inefficient catalyst and not susceptible to promotion by ammonium chloride. A series of catalysts was also tested with Rock Springs coal for 1 hour a t 450" C. a t an initial hydrogen pressure of 3700 pounds per square inch. At reaction temperature, the hydrogen partial pressure in these experiments roughly corresponds to that in a commercial, 700-atmosphere coal-hydrogenation plant, and it was hoped that the catalyst comparisons might be useful in the operation of the Bureau of Mines Demonstration Plant a t Louisiana, Mo. Results of the high-pressure tests are summarized in Table
VI. TABLE VI.
Run
No.
859 903 906 855 878 941 937 887 884 890 893 849 857 866 872 875 845
CATALYSTS FOR RQCK SPRINGS COAL
(3700-lb./sq. inch initial H1 pressure; 1 hr. a t 450° C.) Asphalt % ' per Unit Lique'70 Liquefaction Gas faction Catalyst None 1% Sn 0.5% NH4C1
+ + 0.05% NH4Cl SnClz (0.1% Sn) 0.1% Sn 1% Zn + 1% NH&l 0 . 5 7 Zn + 0.5% NH&I o 1.3 Zn + 0.2% N H ~ C I 0.09% Zn + 0.01% Sn + 0.2% NH4Cl 0.1% Sn
0
17 MoOa
F e ~ 8 4 0.1% Fe) FeS04 { l %Fe) 1%,Fe Pyrite (1% Fe) 1 % dried red mud
Use of the higher hydrogen pressure markedly increases the liquefaction in the absence of catalyst (compare run 859 with run 431, Table V). As usual, tin plus ammonium chloride is outstanding as a catalyst combination, almost complete liquefactions being achievable. The use of 1% zinc plus 1% ammonium chloride gives results which are essentially as good as those with 1% tin plus o.5Y0 ammonium chloride, b u t this relationship does not seem to be true a t lower concentration levels of zinc.
INDUSTRIAL AND ENGINEERING CHEMISTRY
334
.kt a metal concentration of O . l % , however, it is again found possible to replace 90% of the tin by zinc without losing catalytic activity (runs 906 and 893); the presence of ammonium chloride is, of course, necessary. At concentrations of O.l%, neither molybdenum trioxide nor ferrous sulfate shows any cataIytic effect. Of the iron catalysts tested a t 1% concentration, ferrous sulfate, pyrite, and dried red mud (Bayermasse) are all about equally effective. Pyrite and red mud should be available in quantity a t low cost and should be considered for use in largescale plants. Other iron ores may prove to be equally suitable. Any comparison of the catalyst concentrations used in these autoclave experiments with those used in a continuous plant should take into consideration the fact that in the c~onventional
Vol. 42, No. 2
LITERATURE CITED
(l) Abe,
R.,Huzikawa, S., Kakutani, T., arid Okaniura, T., J
SOC.Chem. I n d . Japan,
41,
supplenientary binding,
417-18
(1938).
( 2 ) Berthelot, C., A n n . chim. et phys., 20, 526 (1870).
(3) Booth, PI‘. J., J . Sac. Chem. I n d . London, 63,i ( 1 9 4 4 ) . (4) Dept. Sci. Ind. Research (Brit.), Fuel Research, “ILept. for Period Ended March 31, 1932,” p. 54; “Rept. for Year Ihidrd March 31. 1933.” D. 100. ( 5 ) Eccles, A , and McCulloch, A., J . SOC.Chem. Ind., 49, 377T, 383T (1930). (o) Fisher, C. H., Sprunk, G. C., Eisner, A , O’Donnell, EI. ,J,, Clarke, L., and Storch, H. IT., U . 5’. Bur. iMines. T e c h . Paper 642 (1942). (7) Hlavica. B.. Brennstoff-Chem.. 9. 229 (1098), ( 8 ) Kurokawa, M., Hirotla, W., Fuiiwnra: K., and Aqaoka, N , .I. Fuel Sac. J a p e n , 18, 31 (1939). Storch, €1. H., ISD. ENG.CHR:M., 37, 340 (1945). .Varieti, T. E., Bowles, K. W., and Gillnore, R. E,. 1x1).I;vc,. I
_
( Coal Hydrogenation Catalysts)
MECHANISM OF COAL HYDROGENATION SOL WELLER, E. L. CLARK, ANI) AI. G. PELIPETZ Bureau of M i r z e s , P i t t s b u r g h , P a .
Consideration of the influence of catalysts on the hydrogenation of bituminous coal and asphalt has led to the formulation of a descriptive theory of coal hydrogenation. I t is postulated that coal (or asphalt) is thermally split to form reactive fragments, the splitting being catalyzed by halogen acids. The fragments either polymerize to form benzene insoluble products or are stabilized by the addition of hydrogen to form soluble products. The hydrogenation stabilization is catalyzed by tin.
phoric acids were added as G .Y :iqueous ~olutions. h rernovablo Pyrex KO.774 glass liner mas employed in all cases. Three inaterials were studied in t,hcse investigations: whole Bruceton coal (from the Bureau’s oxporimeiital mine a t Bruceton, Pa.), hand-picked Hruoeton anthraxylon (from thc si~nic mine), and crude asphalt isolated froiii the products of a coal hydrogenation pilot plant run 011 Bruceton coal. (Asphalt is dcfined here as material soluble in beiiecne but insoluble in n-hesane.) The ultimate analyses o f these substances nois is ti ire-frcc basis) are prescrited in Table I.
T
HE suggestion has been made ( 3 )that the over-all hydrogen-
ation of coal to distillable oil proceeds via the formation of asphalt as an intermediate. The coal-to-asphalt conversion appears to be a relatively rapid reaction which is accompanied by the elimination of the bulk of the oxygen in coal, primarily as water. The asphalt-to-oil step seems to be relatively slow, requiring higher temperatures and longer rraction times than does the primary liquefaction. It is also known, a t least for bituminous and subbituminous coal, that while tin used alone is a fair hydrogenation catalyst and halogen acid alone is a negative catalyst (with the exception of hydriodic acid), the combination tin-halogen acid constitutes perhaps the most effective known coal-hydiogenation catalyst (4). Consideration of these and other results described below has 1 ~ d us t o the formulation of a descriptive theory of coal hydrogenation. hlany of the separate dements of the theory are not unique, nor is it supposed that all the details of coal hydrogenation will be covered by the crude picture presented. It is hoped, however, that the hypothescs will permit a correlation of the major steps in the over-all hydrogenation and of the role played by the separate catalyst constituents, and that they will provide a framework for the design of future experiments. PROCEDURE AND MATERIALS
A detailed description of the hydrogenation equipment and the analytical procedures used has been published (1, 4). In all of the experiments described herein, an initial hydrogen pressure of 1000 pounds per square inch and a reaction time of 1 hour at the temperature designated were employed. No oil vehicle was added t o the powdered coal or asphalt; hydrochloric and phoa-
Laboratory analysis of the crude asphalt used showed it bo rontain 0.4% water, 1.4% benzene insolubles, 93.8% asphalt, and 4.4% oil (material soluble in n-hexane). CATALYSTS FOR ASPHALT IIYDROGENATIQN
A study was made of the irifluencc of n number of catalysts o n the hydrogenation of crude asphalt. In all of these experiments, ,50 grams of asphalt were hydrogenated for 1. hour a t 450” C. ut an initial hydrogen or helium prcssure of 1000 pounds per square inch gage. Result,s of the tests arc summarized in Tatilc I1 and in Figure 1. It is clear that of the metal catitlysts and combinations teated, tin plus ammonium chloride is as superior for the hydrogenat,ion of asphalt as it is for the hydrogenation of coal (4). It follows that the excellence of tin plus ammonium chloride for coal hydrogenation is not associated with any peculiar structure of coal as a solid, as, for example, the possible existence of graphitic layers between which the catalyst might be intercalated.
