UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Solvent extraction experiments on Hat Creek coal Ekman, Frank 1946

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1946_A7 E4 S6.pdf [ 2.77MB ]
Metadata
JSON: 831-1.0062166.json
JSON-LD: 831-1.0062166-ld.json
RDF/XML (Pretty): 831-1.0062166-rdf.xml
RDF/JSON: 831-1.0062166-rdf.json
Turtle: 831-1.0062166-turtle.txt
N-Triples: 831-1.0062166-rdf-ntriples.txt
Original Record: 831-1.0062166-source.json
Full Text
831-1.0062166-fulltext.txt
Citation
831-1.0062166.ris

Full Text

SOLVENT EXTRACTION EXPERIMENTS ON HAT CREEK COAL Submitted in partial fulfilment of the requirements of the course in Chemical Engineering leading to the degree of Master of Applied Science. Graduate Tear, Faculty of Applied Science, The University of British Columbia. September 1, 19J+6 Frank Ekman ACHBOWLEDGMENTS I am pleased to acknowledge the help and guidance of Dr. W. F. Seyer during the past year. I am also indebted to Mr. Sun Yip, II.A,Sc., for advice based upon preliminary work he and others did for Dr. Seyer. TABLE OF CONTENTS Page 1. Introduction 1 2. Theoretical Considerations 2 a. The Origin of Coal 2 (1) The Lignin Theory 2 (2) The Cellulose Theory 6 b. Bie Chemistry of Goal 7 3. Apparatus and Plan of Work 12 4. Experimental Results 14 a. Preliminary Analyses of the Coal 14 b. Atmospheric Extraction 15 c. Pressure Extraction 22 d. Other Solvents 30 e. Separation of Extract and Solvent 32 f. Time-Extraction Relationships 34 g. Ultimate and Calorific Analyses 35 5. Summary 39 6. Conclusion 40 7. Bibliography 41 LIST OF ILLUSTRATIONS Plate 1 Plate 2 - Electric Furnace - Pressure Bomb. l i INTRODUCTION At Hat Creek, near Ashcroft, there is a large seam of lignite coal, This coal seam has a great economic asset; i t can be mined by open-pit methods. There are, however, disadvantages. The coal* as mined, has too high an ash and water content and too low a calorific value to be commercially attractive. In addition, l ike a l l lignites, i t slakes and crumbles on prolonged storage. Dr. Seyer proposed a research program with the object of removing these disadvantages and producing a product suitable to the market. A feasible plan appeared to be the solvent extraction of the coal. By this method, a l l or most of the organic substances could be separated from the inorganic material. This extracted organic matter could then be utilized as fuel or as road-paving material. It is not improbable that part, or a l l of the extracted material could be used as a source of chemicals, 5to« s resaarch has been devoted entirely to the solvent extraction of coal. Although the problem has not been explored very fully, owing partly to inadequate equipment, enough results have been obtained to Justify further investigation. 1 2 2. THEORETICAL C0HSIDERATI0M3 a) The Origin of Coal (1) The Lignin Theory Although papers on the various chemical aspects of coal date back at least 100 years; i t was not until 1929 that a paper dealing with the chemistry of the origin of coal was publishes. In that year Hans Tropsch read a paper at the^University of Prague* expounding the theory that coal is derived from lignin alone, and from none of the other plant substances. He had amassed so many facts to support his contentions that his views, "The Lignin Theory," are generally considered the post plausible existing to-day. Tropsch divided coal into three sub-groups: 1) humus coal, 2) eapropel coal, and 3) mother of coal. Humus coal is that part of the coal derived from woody substances; sapropel coal is that part derived from fats and albumins; and mother of coal i s mineral charcoal. Very l i t t l e i s known of mother of coal, or of mineral charcoal; Tropsch considered only the humus coals, which are, however, the most abundant. There are three characteristic stages in the formation of humus substance: a) In Peat -The humus substances here exist as humic acids, 1Problems in the chemistry of coal, Chemical Reviews, vol . 6, 1929. Feats are acidic and dissolve in cold alkali , b) In Brown Coals or Lignitic Coals The huraic acids of peats have now been transformed into humins, which are anhydrides and lactones of the humic acids. These are not soluble in cold alkali but can be dissolved by boiling in lye, a process which converts them to the sodium salts of humic acids. The addition of hydrochloric acid or of other mineral acids precipitates out the humic acids. This process can be reversed; humic acids heated to 250°G. form humins. c) In Hard Coals The humus portion of hard coals is not soluble in boiling lye. However, Fischer and Schraeder converted these humus substances to compounds resembling humic acids by auto-oxidation at higher temperatures. The conversion can also be effected by pressure oxidation.,in the presence of a lkal i , or by moderate oxidation with H2O2.. The ease of interconvertibility suggests that there is no fundamental alteration in chemical structure. When a l l these substances are further degraded, benzene carboxylic acids and lower aliphatic acids are obtained. One-third of the identifi-able acid mixture consists of benzol derivatives. These can be converted to aromatic carboxylic acids by heat and pressure. As these same condi-tions of heat and pressure are known to degrade aromatic compounds, i t is concluded that coal is largely aromatic in. structure. Under pressure oxidation, lignin is the only one of the vegetable substances to behave in a like manner. Under the pressure-oxidation conditions already referred to, cellulose behaves quite differently to lignin. Kb humi.c acids are formed. 4 but large quantities of aliphatic acids such as acetic, oxalic, fumaric, and succinic are. In addition, there are formed unidentified non-volatile acids which yield furan upon further oxidation. Under similar conditions no furan can be obtained from coal or from lignin. As i t has been demonstrated that the furan ring is stable under these conditions, Tropsch. declared this to be conclusive evidence that no furan exists in coal. Humic acids can be formed art i f ic ia l ly from various substances other than phenols, for example, carbohydrates; but these reactions proceed most smoothly from the phenols. It is known that phenols can be formed directly from cellulose and carbohydrates, and i t Is therefore assumed that the formation of phenols is a necessary intermediate stop. Tropsch did not think this significant; i t i s , however, the weak point of his theory and the point upon which his opponents have centered their attack. The work of biologists and biochemists has helped to establish Tropsch*s theory. They have shown: (1) Cellulose is consumed by bacteria during plant deca y and disappears as GGj, CH^, H 2 O , and water soluble acids. (2) Lignin forms huimic acids. (3) laxes and resins form bitumens. In 1917, Hose and Lisse examined fresh, half-decayed, and fully decayed wood. They showed that the methoxyl content doubles during the process of decay, a good indication of an increasing percentage of lignin, Cellulose, on the other hand, disappeared. Bray and Andrews showed that the lignin content of wood remained practically constant and that the loss of weight during decay was the same as the weight of cellulose 5 originally present. Biey also showed that methoxyl splits off from lignin in the process of decay, thus explaining the. absence of methoxyl groups in coal. These methoxyl groups are more resistant than cellulose, and, consequently, do not begin to split off until the cellulose is largely decomposed. Time alone cannot bring about the transformation of peat to hard coals; heat is also necessary. In Russia, there are brown coals which, from their geological strata, should have been converted to more mature coals. That they have not been converted is due to the fact that they had not been exposed to heat. Erdmann had demonstrated this fact in the laboratory by heating lignite with oner-half its weight of water in an autoclave for 100 hours at 280°G. The brown lignite containing 64 per cent carbon was converted to a black coal containing 91 per cent carbon. He concluded from this that a temperature of about 300°C. is necessary to convert low-rank coals to those of higher rank. Goal bitumen i s generally assumed to be derived from the waxes and resins of the coal-forming plants. Tropsch has isolated C25» Qg], and acids from bitumen; higher acids are believed to exist. It is rather remarkable that odd-numbered acids from Cgj to are also found in beeswax. These odd-numbered acids are not found in animal and vegetable fats, in which the glycerides of lower even-numbered acids predominate. It has been shown by cracking paraffins that the odd-numbered compounds are much more stable than the even ones. This leads to the conclusion that a l l even-numbered acids originally present in the vegetable matter disappeared in the course of plant decay. 6 2) The-Cellulose Theory This theory is generally thought less reasonable than the l ignin theory. Although there is sojfte evidence to support i t , top-ranking coal chemists do not believe i t to be sound, 2 E . Berl and W. Koerber attacked the lignin theory because Fischer and his co-wprkers had considered only cellulose coals. They produced cellulose coals by heating cotton linters with 0.05 H MaOH in o an autoclave at 32 5 0. From this coal they then obtained aromatic compounds by the usual pressure-oxidation methods. They believed these experiments provided the.lignin theory in error and that cellulose is •a the chief sources of at least some coals. However, Thiessenv has shown that, in the moderate conditions prevailing in Wisconsin, cellulose decays in the f i r s t 10 feet of a peat swamp. And as, furthermore, 30 feet of peat are required to produce one foot of hard coal, i t would seem that even i f the cellulose theory were accepted, cellulose coal would only exist on the top four inches of a coal bed. b) The Chemistry of Coal The ultimate composition of any coal can be determined by analysis, but these data do not give much clue to the chemical structures and functional groups present. Some knowledge of structure and ^Partly aromatic constitution of ar t i f i c ia l carbohydrate coals, 3Thlessen, R., and Johnson, R.C. , Ind. Eng. Chem., Anal. Bd., 1, 216-28, 7 functional groups can be obtained from examination of the degradation ( products of coal, but i t is doubtful i f coals ean ever be exactly classified chemically. Some lines of endeavour, however, have yielded important additions to our knowledge. One of the most fruitful of these 4 ie examination of the primary degradation products, humic acids. These oxidation products of bituminous coal, because of their solubility in aqueous alkali and other solvents, represent the least degraded oxidation product of coal to which i t is possible to apply classical methods of characterization. They may be prepared from bitu-minous coals by alternate oxidation and alkaline extraction processes, as for example (1) by alternate exposure to air and extraction with a basic solvent, (2) by intermittent treatment with an acidic oxidizing agent such as nitric acid or sulphuric acid and an alkaline extracting medium, or (3) by continuous treatment with an oxidizing agent in the presence of a basic solvent medium, as treatment with (a) air in the presence of aqueous alkali , or (b) alkali permanganate. Since the regenerated humic acids consist of a series of molecular sizes analogous to polymeric homologous series, molecular weight determinations on any given sample can represent only an average value and the particular type of average wi l l depend upon the method used. The average value obtained by colligative methods, such as freezing point, boiling point, and osmotic pressure which yield number averages, is particularly sensitive to the presence of small weight percentages of low molecular species. Because of the difficulty of obtaining suitable solvents, few data are available on the measurements of the molecular weight of humic acids from any source. Odeh determined 4 Chapter 9, Chemistry of coal utilization, pp. 346-76. 6 equivalent weights of humic acids extracted from peat by eleetrometric titration, and estimated the molecular weight from the basicity, calculation of the basicity being made from the empirical conductance rule of Ostwald, Oden concluded that i t is very probable that this humic acid preparation was t r i - or tetrabasic. As an equivalent weight of approximately 340 had been found, this would lead to an estimated value of molecular weight for these acids of 1000 to 1350* Other workers have found its molecular weight to be near this range, Relatively l i t t l e work has been done on the determination of the functional groups in humic acids, the least degraded substances with which i t is possible to work, Formation of salts and determination of the amount of nation combined have been employed for determining the functional groups present in humic adds. Francis and wheeler prepared barium, iron, silver, and copper salts and estimated the equivalent weight of their acids to be 170. Through reactions with acetyl chloride, benzoyl chloride, and Grighard reagents, i t was established" that 860 grams was the weight of the unit associated with one hydroxyl group and that four carboxyls were present in this aame unit* Other than hydroxyl and carboxyl groups, no functional groups have been shown to be present in humic acids prepared by the oxidation of bituminous coal, but the presence of carboyxl groups has been reported in the primary oxidation products of humic acids from brown coals, ,A study-* baa been made relating the yield of m e l l i t i C ; acid obtained on the oxidation of coals to the rank of coals. The procedure of oxidation selected was: prolonged fefluxing with nitric acid ^Chemistry of coal util ization, p. 369. (s.g. 1,5) followed by alkaline permanganate. The results of the applica-tion of this method to a series of carbonaceous materials are shown in Table 1. Table 1 Data on Oxidation by Nitric Acid Followed by Alkaline Permanganate Materials* Carbon per Residue after Total Acids Metallic 100 grams Nitric Acid after Acid grams Oxidation in Permanganate Recovered grams Oxidation grams grams Coals and Graphite Illinois No. 6 Seam Coal 69.50 68.0 19.3 4.3 Pittsburgh Seam Coal 78.27 70.1 28.1 5.5 High Splint Seam Coal 78.70 53,2 14J4 5.0 Pocahontas #3 Seam Coal 85.10 92.4 25.8 10.9 Anthracite Coal 82.47 77.2 32.6 17.7 Natural graphite 97.40 78.5 26.9 21.7 1,000°C Cokes Illinois No. 6 Seam 83.78 52.3 26.2 19.4 Pittsburgh Seam 86.74 50.9 29.4 22.5 High Splint Seam 89.96 48.1 28.3 20.8 Pocahontas Seam 88.68 49-5 30.5 22.1 Anthracite 87.73 48.6 26.1 19.8 Other Cokes and Charcoals Pittsburgh Seam, 500°C. 80.17 95.5 29.7 11.9 Pittsburgh Seam, 700°C. 79.80 79.4 39.7 24.1 Domestic Coke 84.23 46.6 27.5 19.8 Metallurgical Coke 80.28 48.3 27.2 17.9 Cellulose char, 1000° 91.59 39*5 30.0 24.8 Active carbon 83.IO 39.0 33.1 19.7 Carbon black 93-91 56.5 42.9 32.2 Petroleum coke 92.31 139*6 49.9 20.7 Coke from low-temperature pitch, 7GO°C.-W- 88.05 99.2 47.3 22.9 Pitches and Known Compounds Pitch from low-temp, tar § 83.52 64.5 22.0 4.1 Pitch from high " " w 91.52 135.5 23.0 7.9 Triphenylene 94.70 169.6 77.8 67.2 Hsxamethylbenzene 87.72 11.0 2.9 0.0 (Legend on following page) 10 Legend for Table 1: •5- 100 grams used in each case, + 7.6 grams of unoxidized graphite remained, 4 Prepared by carbonizing the pitch from the low-temperature tar at 700°C. ^ The residue remaining at 400°C from the disti l lation of tar from a low-temperature carbonization process, it The residue remaining at 400°C from the disti l lation of tar from a high-temperature carbonization process. In the series from Illinois coal to natural graphite, there is a general increase in the yield of mellitic acid with rank. Mel-l i t i c acid can be formed only by oxidation of condensed carbocycllc structures, of V7hich the simplest that has been shown to furnish mellitic acid on oxidation is triphenylene or of compounds having a benzene ring completely alkylated such as H C - \ J - C H 3 hexaEKsthylbenzene or of mixtures of these two types of compounds. As both ultimate com-position and behaviour on pyrolysis indicate a decrease rather than an increase in aliphatic structure in higher rank msterials, tha most reasonable interpretation of the increased yield of mellitic acid with rank is that i t reflects an increase in the amount of condensed carbocycllc structures in the higher-ranking materials-. It is obvious, however, that there will be a certain structure from whieh the 11 maximum yield of mellitic acid could be expected. From triphenylene, 12 out of 16 carbon atoms present in the hydrocarbon could appear in mellitic acid. From ooronene, the maximum would be 12 out of 24, and with larger condensed structures the fraction appearing as maLlitic acid would become smaller and smaller until such a size was reached that conceivably the oxidation would effect a scission and more than one molecule of mellitic acid could be formed per molecule of hydrocarbon, ?he yield of nusllitlc sai<3 from triphenylene, the simplest purely cyclic structure from which i t has been prepared, was 67.2 per cent of that theoretically possible. Hexamethylbenzene, by this oxidafcion procedure, formed carbon dioxide almost exclusively; and no mellitic acid was recovered. Summarizing, the outstanding chemical characteristics of coal are (1) high molecular weight, (2) aromatic rings, (3) carboxyl groups, and (4) hydroxyl groups. It was thought that an ideal solvent for coal should approach coal as closely as possible in chemical analysis and in function-al groups. For this reason i t was decided that mixtures of phenol and 2-ethyl n-butyric acid would be worth investigation. Both solvents can be obtained in the large quantities that would be required i f a commercial plant were developed. Phenol contributes a hydroxyl group and an aromatic ring to the mixture; 2-ethyl n-butyric acid contributes a carboxyl group and a moderately high molecular weight. Later wcrlc showed that mixtures of these two solvents proved more satisfactory under the conditions of these experiments .than the solvents usually used in commercial plants-. 12 3. , APPARATUS AND,, gLAN OF WORK, In the f irst series of experiments, the coal was extracted in a Pyrex Soxhlet apparatus. When i t was felt that further work with this type of apparatus would be unproductive, samples were sealed in thick-walled Pjrrex combustion tubing; &ad were heated in the small electric furnace shown in Plate 1. This furnace was constructed by winding 25 feet of No. 18 Chromel wire about an 18-inch piece of l § - i n c h iron pipe* It was then heavily insulated, and transits base-plates were attached at each end. A rheostat and ammeter were placed in series with the heating element to control the current. The ends of the furnace were plugged with small pieces of asbestos paper so that an explosion caused by the bursting of the glass tubing would not be too closely contained. A calibrated thermocouple was placed inside the furnace to measure temperatures^ This furnace proved very satisfactory. A current of 1.25 amps generated a o temperature of 250 C and the maximum current of 2.7 omps produced a temperature of about 500°C. Unfortunately, the sealed glass tubes did not withstand the pressures accompanying such temperatures. Although a few tubes withstood heating to 400°C, a large percentage burst when subjected to 300°Gi In o later work i t was found advisable to reduce the temperature to 250 C; even at this temperature about one-third of the tubes burst. A small bomb was put into service to circumvent the inadequacies of the sealed tubes. A plan is shown in Plate 2. A l l surfaces exposed to the solvent, except the interior of the blow-off valve, were heavily / ' A I/O AC A PLATE / PRESSURE BOMB PLATE 2 13 copper-plated. As the bomb had no drainage valve, i t was necessary to put a container in the bomb to hold the coal. Porous Norton alundum thimbles were used so that natural circulation would be hindered as l i t t l e as possible. Repeated weighings showed that these thimbles did not react with the chemicals present. It was found that this bomb began to leak where the brass gauge pipe threaded into the top cover when 300 psi were applied. As the sealed glass tubes would also withstand this pressure, the bomb was not used extensively. It was thought that glass tubes should be used as much as possible because the different conditions of extraction in the bomb and the tubes would not allow a direct comparison. The coal used was carefully sampled, ground, and screened, Only the portion passed by a 20-mesh sieve and retained by a 42-mesh sieve was used in these experiments. The screened coal was then dried exactly one hour at 105°C and stored in a dessicator. Weighing bottles were used for a l l weighings. A weighed sample of coal was introduced into a glass tube sealed at one end,. A measured quantity of solvent was added, and then the other end of the tube was sealed. The tube ivas then heated in the furnace for the required time. Bhen cool, i t was withdrawn, and the top was cut off. The contents were poured through a weighed alundum thimble and the solvent filtered off. The thimble and coal residue were placed In a Soxhlet and were washed 24 hours with 50:50 alcohol-benzene solvent. The thimble was then dried 48 hours at 100°C, cooled in a dessicator, and weighed In a weighing bottle. The loss of weight was assumed to be the amount of coal dissolved. When the bomb was used, a weighed thimble and contents were 14 put into the bomb and a measured amount of solvent added. Apart from this, standard procedures v/ere used. Solvent was separated from the extract by vacuum disti l lation. A quantitative separation could not be made because the vapor pressure of the most volatile constituents of the extract approached that of the 2-ethyl n-butyric acid. This wil l be discussed in more detail later. 4. EXPERIMENTAL RESULTS a) Preliminary Analyses of the Coal Eight samples of Hat Creek coal, selected by Dr. Seyer, were received at the University. A composite sample of Hat Creek coal had the following characteristics: Percentage Moisture, air-dried 11.4 Moisture, oven-dried 13.1 Volatiles 32.9 Fixed carbon 29.4 Ash 24.5 Sulfur 0.74 Nitrogen 1.35 B.T.U. / lb . 7560 Amount of extract with solvent B-" 13*0 $ Amount of extract with C^H^ less than 1.0% Amount of extract with G^HjpH 8-.0$» 15 Low Temperature Distillation Percentage Coke 73 Tar 3^76 11*34 Light Gi l 0.50 Gas 11.35. The eight samples were also analyzed separately. The data are compiled in Table 2. to be noted that the ash of a l l samples contains vanadiunu Table 3 summarizes the data obtained. extracted in a glass Soxhlet apparatus at atmospheric pressure. A l l experiments were continued until the refluxing solvent was colorless. The samples were then extracted in a Soxhlet for 24 hours, with 50:50 alcohol-benzene to remove the original solvent. They were finally dried for 48 hours in the oven, cooled in a dessicator, and weighed. Loss of weight was taken to be the amount extracted. As may be expected, a l l solvents are not equ&ly powerful. Results are tabulated below, in Table 4. The ashes of each sample were analyzed qualitatively. It is b) Atmospheric Extraction In the f irst series of experiments, the coal samples were Table 2 PER GMT RECEIVED BASIS PER CENT DRIED BASIS CALORIFIG VALUE pgg C E ] SAMPLE ' 1 ; — ! DRIED BASIS SUITOR Moisture Volatiles Ash Fixed C Volatiles Ash Fixed C BTU Per LB.  1 19.79 31.40 8.65 40.16 39.15 10.78 50.07 10,580 0.670 2 18.19 22.01 33.40 26.40 26.97 40.92 32.11 6,055 0.645 3 22.10 28.00 l 6 . l l 33.69 . 35.95 20.68 43*37 8,770 I.064 4 20.28 25.30 23.97 30.45 31.78 30.10 38.12 8,025 0.574 5 11.05 16.89 45.23 26.83 18.99 50.74 30.27 6,275 0.354 6 17.72 22.14 35.22 24.92 26.91 42.81 30.28 5,730 0.637 7 15.47 18.85 46.38 19.30 22.30 54.89 22.81 4,2i0 0.340 8 22.40 25.43 20.15 30.02 31.27 26.73 42.00 8,325 0.306 17 Table 3 V m Al Mn Mg Ca Ti Fe Si 1 T T? S+ M S S M S+ s+ s+ 2 M T? n T n ii M tt fl 3 1S-S T? n M t> n M « n II 4 T T? II T n M T ti n n 5 T T? - - M? T T n it « 6 M-S T S+ T S S M tt tt fi 7 M ? tt M S M » tt ft 8 T- T « M re T M w ti T » Trace M o Medium, 0,5 to 1^ S » Strong, greater than 1$. Table 4 Solvent Per Cent Extraction Benzene 0.5 2-ethyl n-butyric acid 6.6 Hexoic acid 7.8 Phenol 7.0. In addition,.samples of coal were alloived to stand for a week in acetone, behzsJLdehyde,. decalin, ^aniline, and hydrochloric acid. Nona of these solvents was more than slightly colored at the end of the week.j 16 Finally, various mixtures of phenol and 2-ethyl n-butyric acid were used as solvents. Table 5 and Figure 1 show that extraction reached a maximum when the solution contained 55 volume per cent phenol. This maximum extraction was several times as great as that obtained by using either solvent alone. The mixture containing 55 volume pe? cent phenol and 45 per cent 2-ethyl n-butyric acid is hereafter referred to as solvent B. Table 5 Volume Per Gent Phenol Per Cent Extraction 0 6.6 20 8,3 30 8,5 40 8.8 50 25,3 §5 36.2 60 21.5 70 12.7 80 11.7 100 8.0 Ash content of extract = 2. A l l samples were extracted in a Soxhlet until the refluxing solvent was colorless; extracted 24 hours with 50:50 alcohol-benzene, dried 48 hours at 100° G, cooled in a dessicator, and weighed. The phenomenon of maximum extraction by a mixture, shown in Figure 1, is not unique. It may be recalled that certain cellulose 19 nitrates are soluble in a mixture of ether and alcohol, although they are insoluble in either solvent alone. Reilly^ has advanced the theory that surface-tension effects are responsible for this phenomenon. He reasoned that neither solvent •,,tvets'n as much as the mixture. Kiebler^ advanced the theory that one of the two solvents may depolymerize large molecules, which then become soluble in the second solvent. Tables 6 and 7, taken from a paper published by Reilly^ on his studies of Irish peats, show that the phenomenon is not unusual. Table 6 Extraction of Peat with Pure Solvents Solvent Yield Per Cent Acetone 11.4 Carbon tetrachloride 8.6 Chloroform 9.7 Ethanol 14.0 Methanol 9.2 Methyl Acetate 13.6 ^Reilly et a l , J . Soc. Chem. Ind., 56, 231-8T (1937) 7 'Chemistry of coal utilization* p. 679. 20 Table 7 Solvent Mixture and Percentage Weight Composi-tion. Boiling Point Boiling Point of 1st Component of 2nd Com-o r ponent Boiling Point Yield of Azeotropic Per Mixture Cent Or. Methanol (10} Chloroform (90) Acetone (80) Chloroform (20) Ethanol (l6) Carbon tetra-chloride (84) Acetone (86) Methanol (14) Acetone (55) Methyl Acetate (45) Carbon tetra-chloride (79.4) iiethanol (20.6) 64.5 56.1 7 8 . 5 56.1 56.1 76.8 61.2 61.2 76.8 64,5 57.1 64.5 53.5 64.7 64,9 55.9 56.1 55.7 16.7 15.0 15.4 14.2 15.6 11.0 Four samples of coal were completely extracted at atmospheric pressure., Percentage extraction varied with the ash content of the eoals; the coal that had the highest ash content had the lowest extraction. Results are tabulated in Table 8.-Indices of refraction were taken of samples of fresh solvent and reclaimed solvent to determine the sharpness of separation of solvent and extract.. Fresh solvent gave a reading of 62° 10* on a Pulfrich refract tometer at 20° C. Three samples of reclaimed solvent gave readings of 62° 2 0 ' , 62° 4 8 ' , and 62° 09'. This indicates that (1) the solvent can Table 8 RECEIVED BASIS DRIED BASIS TIME PER CENT bAfflPLE , — • — EXTRACTION EXTRACTION Moisture Volatiles Ash Fixed C Volatiles Ash Fixed C Hours DRIED COAL 3 22.10 28.00 16.11 33*69, 35.95 20.68 43.37 30 24.00 5 11.05 16.89 45.23 26.83 18.99 50.74 30.27 30 9.06 6 17.72 22.14 35.22 24.92 26.91 42.81 30.28 2? 16.83 8 22.40 25.43 20.15 30.15 31.27 26.73 42.00 24 22.35 22 be successfully reclaimed and (2) that the solvent is not appreciably affected by this temperature. However,.the last 5 ml. reclaimed from / • O a run gave a reading of 66 42* i- As this portion of the solvent was colored, the larger reading must have been due to dissolved coal substances.' c) Pressure Extraction The vapor pressure of solvent B at elevated temperatures was estimated. As latent heats of evaporation are not available for ; phenol or for 2-ethyl n-butyric acid, approximations had to be made. Latent heats can be estimated by Trouton's Rule, AH/Tg - 22, where "TB" is the boiling point in degrees Kelvin, although this is known to 8 be inaccurate for polar compounds. Kistiakowsky's equation L/T =8.75 + 4.571T, where "T" is the boiling point in degrees absolute, was considered superior, although i t , likewise, is strictly applicable only to non-polar compounds. The,temperature coefficient of the heat of vaporization may be calculated by the formula d ( A H) / dT, - ACp" inhere ACp = heat capacity of 1 mole of vapor ~ = heat capacity of 1 mole of l iquid. Again, there were insufficient data to uti l ize this formula. : Accordingly, 9 the integrated Clausius-Clapeyron equation, 23 1ogP 2 s _ ^ H m P i R ' • T2T1 T^ and Tg are in degrees Kelvin a cal* per mole S » 1.99 eels. P-^  and p£ are pres-surea, had to be used with the assumption that was conEtanti 8wenner, BiR., Thermochemieal Calculations, New York, McGraw-Hill; 1941jP23* 9Millard, Physical chemistry for colleges, New York, 1941, p.108; 23 Finally, Raoult's Law of Partial Pressures was applied to each component, of the mixture. The result, shown in Table 9 and Figure 2, i s , accor-dingly, only a f irst approximation. Table 9 VAPOR PRESSURES, LB./SQ. INCH Temperature, 1 11 ?C Phenol 2-Ethyl Solvent B n-butyric Acid 182 14.7 197 14.7 200 21.6 15.7 19.3 250 55.7 42.7 50.7 300 120 97.7 H I Great difficulty was experienced in obtaining reproducible results. The coal samples had to be weighed in weighing bottles to prevent absorption of moisture, as may be expected. In addition, i t ivas found that dessicated alundum thimbles absorbed water rapidly; they too had to be weighad in weighing bottles. Furthermore, the extracted coal residue ;a3sorbed solvent very tenaciously. This was shown in Run No. 9, A sample of coal was extracted for eight hours at 350° G. It was then washed 41 hours with commercial lighter fluid and dried 48 hours. The weight of coal and thimble was then 5.996 grams. The sample was then placed in a 24 steam-jacketed container and subjected to a 0.0001 mm vacuum at 100 C for 16 hours. After cooling in a dessicator, the sample s t i l l weighed 5.996 grams. The sample was then extracted with water for 48 hours, dried, cooled, and weighed. This treatment caused the weight to drop to 5.973 grams, a loss of 21.9 per cent of the coal residue originally-present. The tenacious solvent adsorption illustrated here is not 10 unusual; i t has been experienced by other experimenters. Kiebler's procedure of washing the sample 24"hours with 50*50 ethanol-benzene was finally adopted as the most satisfactory one available. Results given below will show that i t was not entirely adequate; some of the samples s t i l l gained weight after extraction, although a l l procedures were exactly duplicated. It ivas shown that alcohol-benzene extraction for 24 hours gave consistent results. Several samples were extracted a further 24 hours with alcohol-benzene solvent, dried, and reweighed; in a l l instances the samples neither gained nor lost weight. It was felt that the presence of an inert atmosphere during pressure-extraction may have led to more consistent results, but i t was not easy to displace the air in either the sealed tubes or in the bomb. Another phenomenon that was not fully explained was the 4 precipitation of coal substance during extraction. In the glass tubes a black precipitate was noticed at the liquid-glass interface. German-held patents^ state that this is caused by hydrogenation of. unsaturated extracts to form insoluble hydrocarbons. The precipitation may be due to retrograde condensation, or to supersaturation on a cooler surface, I°Kiebler , Extraction of a bituminous coal, Ind, Eng. Chern,, 32, 1389, 1940. n U . S. Patent 2,l6?,250 British Patent 480,644 British Patent 481, 108. 25 It was found that a high ratio of solvent to extract (circa 500/1} prevented formation of this precipitate. These proportions of solvent to extract were used thereafter to eliminate a possible variable. A steady increase of pressure with time was observed for a l l o solvents except anthracene, at, or above, temperatures of 250 C. This, of course, caused the bursting of a large number of glass tubes; and, in the bomb, raised the pressure beyond Its capacity. Possible causes are oxidation and thermal cracking. Summaries of the more successful experiments are given below. Times were varied so that the effect could be noted. Temperatures were o lowered in 50 C intervals in an unsuccessful attempt to find a pressure range which the available equipment could withstand. Run No. 3 Spontaneous precipitation of extract was f irs t noted. A precipitate collected along the glass-liquid interface and on the bottom of the tube. Paper thimbles were used, but were found unsatisfactory. The addition of excess commercial lighter fluid to the solvent and dissolved extract completely precipitated the asphaltic constituents of the extract. The coal was extracted 14 hr 40 min. at 300° C. The apparent extraction was 32.8 per cent. Unfortunately, disintegration of the paper thimbles makes this result inaccurate. The residue was subjected to three further extractions of nine hours each. Paper thimbles rendered the results inaccurate. However, each freeh portion of solvent became uniformly darkened, showing that the maximum amount of coal substance had not been removed in 41 hr 40 min. 26 Run No. 5 Sacks made of glass cloth were used to contain the coal sample, These sacks resisted solvent action but small particles of coal, broken down by solvent action, escaped through the fabric./ Run No. 17 The wet, extracted coal particles were seen to be swollen. They lost shape under the slightest pressure; they could be compared to small particles of putty. When dry, the particles lost their swollen appearance, but disintegrated to powder when handled. Run No, 16 Twenty-five mi l l i l i tres of solvent B and 0,445 grams of coal were heated 18 hours at 350° C. The residue was filtered into an alundum thimble and handled in the usual way. Extraction •» 10.5$. Run No. 18 In this run 500 times as much solvent as coal was used. Goal N 0.116 gm Solvent B 50 ml Time 45 hr Per cent Extraction 38 Ash content of coal 38.5 Per cent extraction of organic matter 6 l . 6 , A similar sample of coal was completely extracted at atmospheric pressure. The percentage extraction, on the basis of total weight, was 9.06 per cent. It is important to note that the pressure extraction was 27 not carried to completion, but the atmospheric extraction was. The residue was then extracted with coal-tar. The sample then gained weight. Bun Ho. 19 This sample was > digested in the bomb. Coal Solvent B Time Temperature Initial pressure Final pressure Per cent extraction 5.887 g 300 ml 45 hours 220° C 175 pai guage 250 psi guage 19.3 Run Ho, 20 (Bomb runty Coal Solvent B • Time Temperature Initial pressure Final pressure Per cent extraction 5.034 g 300 ml 39 hr 220° C 135 psi 235 psi 5.0 Run No. 80 Coal Solvent B Time Temperature Negative extraction. O.0651 20 ml (500/1 ratio) 24 hr 300° C 28 Run Ho. 81 Coal O.063 g Solvent B ~" 20 ml (500/1 ratio) Time 39 hr Temperature 300°C Per cent extraction 16.9. Run No. 82 Coal 0.0506 g Solvent B 20 ml Time 10 hr Temperature 300°C Per cent extraction 20.4. The results are tabulated in Table 10 for quick reference. These results, although not as complete as desired, demonstrate that high temperature and longer times increase extraction. One must conclude that apparatus capable of tvithstanding higher pressures would be most advantageous. Instances of "negative extraction," or of increase of sample weight after treatment, occur at the lower temperatures. Samples extrac-ted at 350° C or at 400° G (reported below) a l l lost weight; on the other hand, those extracted at 300° C or 250° C (reported below) frequently-gained weight. It is thought that higher temperatures lessen the possibility of negative extraction in two ways: (1) Higher temperatures remove more of the coal substances and so leave less to adsorb or combine with solvent; and (2) Higher temperatures tend to cause cracking and so Table 10 Run No. SOLVENT M l . WEIGHT GOAL WEIGHT RESIDUE TIME Hours TEMP 0 g PER CENT EXTRACTION SE1ARKS 3 50 2.266 1.523 14 2/3 300 32.8 Paper thimbles. Results inaccurate. 4 50 0.413 0.117 6 350 71.8 M ft tt ft 9 50 0.263 0.106 8 350 59-5 Glass cloth sack. Results inaccurate. 16 25 0.445 0.398 18 350 10.5 Results accurate. 18 50 0.116 0.072 45 350 38.0 Per cent extraction, dry, ash-free basis 61.6. 19 300 5.887 5.188 45 220 19.3 Bomb run. 20 300 5.0336 4.7852 39 220 4.9 80 20 0.0651 O.2557 24 300 Neg.J 81 20 0.0635 0.0528 39 300 16.9 ) Note inconsistencies. 82 . 20 0.0506 0.0403 10 300 20,4 ) 30 inhibit polymerization reactions. d) Other Solvents Various other solvents were used to compare their solvent power with that of the phenol 2-ethyl n-butja*ic acid mixture. Not even tetralin, widely used by the Germans, has proved as efficient as solvent B under the conditions of these experiments. Mixtures of commercial cresols and 2-ethyl n-hsxoic acid were tried, to see i f the larger molecular weight and Increased hydroxyl groups would dissolve a greater proportion of the coal, The results are summarized below: Hun No. 5 Temperature 400° 0 Time 8 hr Solvent - 70 volume $ cresol, 30$ hexoic acid Apparent extraction 28.2$, Hun No. 6 ' Temperature 400° C Tim© 8 hr Solvent 50 volume $ cresol-, 50$ hexoic acid. Apparent extraction 31.9$. Run No. 8 Temperature 400° C Time 8 hr Solvent - 30 volume $ cresol, 70$ hexoic acid. Apparent extraction 32.2$, 31 These experiments were carried out with paper thimbles and-were, therefore, inaccurate. However, a run with solvent B for eight hours at 350° gave an apparent extraction of 71*8 par cent. Although the results are admittedly inaccurate, the color of the solvent and the amount of degradation of the coal made i t evident that the figures were at least relatively correct. It was not felt necessary to repeat these runs when more accurate techniques were developed. Anthracene was used in the hope that a high molecular weight compound might increase the amount of extraction. A mixture of 75 par cent anthracene and 25 per cent phenol was tested at 280° C, the temperature which developed the maximum allowable pressure in the bomb (250 psi) with most solvents. In 16 hours the pressure ranged from an in i t ia l pressure of 25 psi gauge to a maximum of 50 psi gauge. Consequent-ly, i t would appear that the pressure increase noted when using solvent B was due to thermal decomposition of the 2-ethyl n-butyric acid component and not to either the phenol or the coal. Results were inconclusive because no solvent was found that would remove the anthracene from the coal. A mixture of 75 volume per cent decalin and 25 per cent phenol was tested. This mixture also yielded a negative extraction. Polymeriza-tion of the solvent or chemical union with the coal appeared responsible. A mixture of 50 volume per cent deealin and 50 per cent phenol caused the coal to gain weight slightly. A mixture of 225 ml. decalin and 75 ml. solvent B extracted 14,1 per cent in 23.5 hours at 220°C. A mixture of 225 ml. tetralin and 75 al« phenol (a mixture used in Germany)^ only extracted 2.4 per cent of the coal in 21.5 hours •^Pott and Broche, Solution of coal ...... Gluekauf, 69, 903-12, 1933. 32 at 22G G. After cooling thebomb there was a residual, pressure of 75 psi gauge, e ) Separation of Extract and Solvent Quantitative separation of extract and solvent is very difficult . Qualitative or commercial separation should be relatively easy. Extract and solvent were miscible in a l l proportions, so that no separation could be made by decahtation. Addition of excess commercial lighter fluid, or of other light hydrocarbons, causes a large proportion of the extract to precipitate as a floceulent dark-brown mass. The supernatant liquid i s s t i l l colored, but is colored only slightly. The precipitate is not crystalline, but i t filtered without difficulty under vacuum. This procedure was, of course, of l i t t l e use in quantitative work; but i t promises to be of great value commercially. The hydrocarbon-solvent-extract liquid could be easily rectified. Heads would be the hydrocarbons and tails the solvent and extract. The small amount of extract remaining in the solvent could be recycled until i t became concentrated enough to warrant disti l lation, Furthermore, as coal-tars and -oils are used in 13 Germany as solvents, the enriched solvent may prove superior to the pure material. Hydrochloric aoid solutions precipitate the extract as a black colloidal mass. Precipitation is apparently complete, because the super-natant liquid is water-white. Some of the liquid can be separated by ^See, for example, British Patent 493.447 (I. G, Farbenindustriej. 33 decantation. The precipitate can ba redissolved by the addition of NaQH. However, the colloidal precipitate cannot be filtered even under a high vacuum; i t must be centrifuged. This procedure is deemed to have l i t t l e value in a commercial process. It is doubtful i f centrifuging would be economical in a plant dealing with a low-cost product. Furthermore, the added HC1 and possibly HaOH would undoubtedly be troublesome in the utilization of the end-product. The method is not thought to have value in quantitative separa-tion, because, although the extract can be centrifuged and dried, there is every reason to believe that chemieals would be tenaciously adsorbed by the colloidal precipitate. Vacuum disti l lation was found to be the most suitable process for separating solvent and extract, when a vacuum of 0.0001 mm. was o applied, clear-colored solvent began to d i s t i l l over at about 55 C (depending upon the concentration of extract). As the solution became more and more concentrated, the temperature of the vapor rose gradually o to 92 to 95 C. When this temperature was reached most of the solvent had been disti l led over. The temperature then began to drop, owing to radiation losses from the mercury bulb, although small quantities of vapor s t i l l dist i l led over. At this vapor temperature the reclaimed solvent became darkened, showing that extract was beginning to d i s t i l l over in appreciable quantities. Repeated weighings of the residue •showed that there was no definite point at which the extract remaining equalled in weight the loss of the original coal. It was thought that controlling the oil-bath temperature might effect a sharp separation of solvent and extract.. Accordingly, a 34 o a thermo-regulator, controlling the oil-bath temperature to 100 C, was installed. This method of disti l lation proved more satisfactory, although a quantitative separation was not achieved. It was found by comparing weights of extract with the loss from the original coal that 90 to 95 per cent of the solvent could be dist i l led off without discoloration. In dist i l l ing the remaining solvent, however, some extract was driven over. Mien the weight of extract became the same or less than the loss of weight of coal, a change of odor was noted. The odor changed suddenly from the distinct smell of solvent B to an entirely different odor. This sharp odor was henceforth taken as the criterion of complete separa-tion. Undoubtedly a small percentage of extract had dist i l led over by this time, but the extract remaining was deemed free of solvent. The extracts from a l l runs were accumulated in one container. This product was a dark-black, lustrous, exceedingly viscous liquid at room temperatures. It closely resembled asphalt in a l l properties except its distinctive odor. It bad a pour-point near 100° C. The extract was found to eontain 15.65 per cent ash. The ash was found by qualitative analysis to eonsiet almost entirely of iron. As the solvent in the bomb came in contact with the unplated valve fittings, and as extract obtained by means of glass apparatus had only 2,88 per cent ash, i t was decided that the high percentage of ash was due to corrosion of the bomb, f) Time-Extraction Relationships Much time ivas spent in an attempt to obtain a relationship between time and per cent extraction. The difficulties that were, seemingly, impossible to overcome prevented the f u l l development of such 35 a relationship. In general terms, hoover, i t was established that longer times and higher temperatures increased percentage extraction. 14 Gregaznov developed the folloiving equation for extraction of coals: E 1 = at b where = per cent organic material extracted t = time "a" and "b" are constants to be established for each coal. For extraction with benzene, the following values are given: a = 9.5 b = 0.48. It i s possible to predict the extraction at infinite time i f a good curve is obtained for finite values of time. Reciprocal time is plotted against per cent extraction so that infinite time is read at the zero value of the abscissa instead of at infinite values. Our figures were not deemed complete enough to obtain significant data by these methods. A special series of runs was made in an effort to obtain a time-per cent extraction curve. The complete results are shown in Table 11, g) Ultimate and Calorific Analyses Carbon and hydrogen were determined by a combustion analysis, following the U. S. Bureau of Mines methodP The nitrogen content was determined by the Kjeldahl method for c o a l . ^ Sulfur was determined by the Eschka procedure}''' utilizing the ash left after oxygen bomb ignition. 1f*|[him Tverdogo Topllva, 7, 22, 2-9, 1936. •^snell, F. D., and Biff en, P. M., Commercial methods of analysis, Hew York, McGraw-Hill, 1944, p. 598. l 6 I b i 4 , p. 601. ^Ibid., p. 508. Table 1 1 Run No. TEMP. ° C . TIME Hr, SOLVENT M l . WEIGHT Coal, Gm. WEIGHT Residue, Gai. PER CENT EXTRACTION REMARKS 100 2 5 0 ~ 2 0 0 , 0 5 9 3 -« Exploded. 101 n 2 5 n O .0566 0 . 0 6 0 9 Neg, . 1 0 2 it « 0 . 0 5 4 0 Exploded. 103 ft 19 tt 0 . 0 5 3 1 O .0546 Neg, 104 tt 2 3 tt 0 . 0 4 2 1 0 . 0 3 8 1 9 . 5 105 ti *r tt 0 . 0 5 5 8 — Exploded 1 0 6 f» 2 2 « 0 . 0 5 4 3 0 . 0 4 5 1 16 .9 1 0 ? « Hh tt 0 . 0 4 1 3 0 . 0 3 7 9 8 . 2 .108 n 50§ n 0 . 0 4 6 4 0 . 0 3 3 3 3 1 . 7 Gxygoa vjaS'4e*KDa3i»Gd by difference. Calor i f ic values of coal and extract war© determined i n m i s iineroon oxygen-bOEib calorimator. • The organic material of the coal had the folloivlng u l t iaa t© analysis; Carbon . . . . . . . 65.8 Hydrogen * . . , . . , 5.25 Hitrogon . . . . . . 6.5 Sulfur . . . . . . 0,25 Oxygen . . . . . . . . 28.2 100.00 Ash content . . . . 49.15^ B.T.U. / l b Bone dry coal . . . 6,067 B.T.u, / l b Organic material . . . 11,920. The extract had tho following analysis; Carbon 67.7 Hydrogen . . . . . . . 9 .20 Hitrogen • . . . • , , 0.4 Sulfur . . . . . . . . . 0 .U Oxygen . . . . . . . . , 22.6 100.00 Ash content . . . . . 15*65$ B.5.U. / lb of extract 11,949 B.$.S. / l b of organic material . . 14,170. faese figures Show that solvent extraction produced a material having a ca lo r i f i c value of 11,920 / l b , almost twice that of the coal from which I t was derived. Bi ia value i s encouraging enough, laaaain, 1. 0., Quantitative Analysis, New York, McGraw-Hill, 1932, p.289. 38 in i tself to /warrant further research. It i s even more encouraging to realize that the use of corrosion-free equipment would produce a material practically free of ash (2.88 per cent). This material would have a calorific value in the neighbourhood of 14,170 (100.0 - 2.88) » 13.950 B.T.U. / lb , a value superior to that of many commercial coals. Solvent B was originally selected on the theory that "like dissolves l ike." Ultimate analyses show that the process of solution is not that simple. As the extract has a lower percentage of oxygen than the original coal, the process of solution must be more a process of thermal degradation than of true solution. The hydrogen content of the extract is greater than that of the coal from which i t was derived. It is possible that the extract was hydrogenated by the solvent, in spite of the fact that the solvent Is substantially unchanged during atmospheric extraction. Tetralin 1^ Is known to hydrogenate the extract i t dissolves; there is no evidence that solvent B does likewise. Selective extraction or differential thermal decomposition may well be the explanation of the apparent hydrogenation of extract. 19 British Patent 462,478 British Patent 480,644 British Patent 515, 335. 39 5« SUMMARY It has been shown: (1) That 55 volume per cent of phenol in ph.enol-2-ethyl n-butyric acid mixture is more effective than any other proportion, and is several times as effective as either of the two components. (2) That solvent B is more effective under the conditions of these experiments than the solvents more commonly used for the extraction of coal. (3) That at least 60 per cent of the organic material can be extracted by pressure extraction. It is only possible to extract about 18 per cent of this sample at atmospheric pressures. (4) That the calorific value of the extracted material is almost twice that of the original coal. It has been indicated that: (1) Equipment capable of withstanding higher temperatures and pressures would make i t possible to extract practically a l l of the organ-ic matter in this coal. (2) The process has commercial possibilities. However, data have not been assembled yet to show clearly: (1) Time-temperature-percentage extraction relationships. (2) The cause of erratic results. (3) The cause of instances of negative extraction. (4) An estimate of cost of production, (5) The chemical constituents of the extract. 40 6. CONCLUSION o An autoclave capable of withstanding at least 400 C and 1000 psi is a prerequisite to further research on this problem. It would be even more desirable to have one built for 450° C and 1000 atmospheres. Such an autoclave would meet the severest conditions that would be imposed on i t in solvent extraction. The next step should be, the writer believes, a final determination of the possibilities of the process in the laboratory. I f this process proves feasible here, i t should then be taken to the 20 pilot plant stage. British, German, and American patents in Dr. Seyer's private files indicate clearly how this may be done. 2 0See especially: British Patent 494,834 U. S. Patent 2,308,247. 41-. Bibliography 1. Berl, E . , and Koerber, W., Partly aromatic constitution of art i f ic ia l carbohydrate coals, Industrial and Engineering Chemistry, vol. 32, pp 676-7, 1940. 2. Hibbert, H.,, Status of the lignin problem, (unpublished.) y. Moore, E . S., Goal, New York, John Wiley and Sons, 2nd ed., 1940: 4. The National Research Council, Lowry, H. H . , Editor, The chemistry of coal utilization, New York, John Wiley and Sons, 1945, 2 volumes. 5. Tropsch, Hans, Problems in the chemistry of coal, Chemical Reviews, vol. 6, p. 90, 1929. 000000000000 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0062166/manifest

Comment

Related Items