A STUDY OF THE PRODUCTS OB- TAINED BY THE SOLVENT AC- TION OF DIPHENYL ETHER ON A UTAH COAL BY JAMES REMINGTON SMITH THESIS FOR THE D EGREE OE BACHELOR OF SCIENCE CHEMISTRY COLLEGE OF LIBERAL ARTS AND SCIENCES UNIVERSITY OF ILLINOIS 1922 /92 2 . -S m 6 UNIVERSITY OF ILLINOIS I92_ J«t_ THIS IS TO CERTIFY THAT THE THESIS PREPARED UNDER MY SUPERVISION BY ENTITLED A_ _ L_l_ k*i _ jk il _ i ik _ _ j VIS. 4i~Q_A§ __________ Qv.4 kiiiU; Ar>+irin rk-P ni nhonnl tt •*->.«». nrnn -j TT-t-aVi _>£.■«. i. i ^ _ _-*.*»•_> IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE degree OF __?-che l_c_-__cX_r_aiiUi^ .sJZ— Instructor in Charge - V .C 7 > . Approved ACTING HEAD OF DEPARTMENT OF _5HMI_STRY. Digitized by the Internet Archive in 2015 https://archive.org/details/studyofproductsoOOsmit TABLE OB CONTENTS I. INTRODUCTION. Page 1. Purpose 1 3. Historical 1 II. EXPERIMENTAL 1. Description of Coal 10 2. Method of extraction. 10 3. Diagram of apparatus 13 4. Examination of Products: A. Ultimate analysis 13 B. Fractional Carbonizat ion 13 C. Coking tests 13 D. Iodine numbers 13 E. Special tests on the extract 14 III. DISCUSSION OF RESULTS 1. Diphenyl ether as a solvent 23 3. Coking tests 35 IV. CONCLUSIONS 37 V. BIBLIOGRAPHY 28 A STUDY OF THE PRODUCTS OBTAINED BY THE SOLVENT ACTION OF DIPHENYL ETHER UPON A UTAH COAL. I. INTRODUCTION 1. Purpose . The purpose of thi.s investigation is a study of the extractable material and the insoluble residue obtained by the solvent action of diphenyl ether upon a Utah coal of the fat bituminous type. 2. Historical . Considerable investigation has been carried on in the past along the line of determining the constitution of coal. These studies have been approached from three view points; from micro- scopic examination, from examination of the decomposition products obtained by fractional distillation, and from examination of the material obtained by the action of certain solvents and reagents upon coal. Each of these methods have contributed a share to our knowledge of coal. The use of solvents dates back to 1851 at which time Dr, Smythe 1 , in examining a brown coal from Brflhl, near Cologne, used the following solvents: benzene, chloroform, alcohol, ether and petroleum ether. He obtained a 3 per cent extraction with benzene, 1.8 per cent with chloroform and 2.4 per cent with alcohol. On the basis that these extracts were pure substances, he calculated from percentage composition, the formula of the chloroform extract a 3 C 15 H 24 O 2 and of the benzene extract C^H^gOg. - 2 - o Anderson and Roberts in 1898 subis cted Scotch coals to the saponifying action of caustic potash and found that with a semi -coking coal the re sine id bodies were saponified and all cok- ing tendency in the remaining residue destroyed while a strongly coking coal only partially lost it 3 coking property, or in other words, only part of the resinic material in a strongly coking coal is saponifiable with caustic potash. 3 Baker in 1901 extracted various coals with pyridine. He found that anthracite coal furnished very little material soluble in pyridine and that a bituminous coal was soluble to the extent of 20.4 per cent. An ultimate analysis did not show a concordant change in the proportions cf elements present in coal, extract and residue. 4 Anderson and Henderson in 1903 extracted coals from Bengal, Scotland and Japan with pyridine and found that the extract from all of these coal 3 were similar in character and possessed, after the removal of the solvent, a black lustrous appearance similar to bitumen. The percentage in extract of carbon, hydrogen and nitrogen was about the same as that of the original coal. They found that the coking property was entirely removed from a semi- coking coal but only partially removed from a strongly coking coal. 5 Professor Bedson in 1903 obtained with pyridine, an extraction of 33-33 per cent of a gas coal, and from 7-39 per cent of cannel coal. He shows that there is no relation between the volatile matter of the coal and the percentage of extractable material 3 S Professor Lewfcs in 1914 in his book on coal, reviews some of the work done up to the time on extraction of coal with pyridine. He states that the coking property of a serai-coking coal could be entirely removed aid that of a strongly coking coal only partially removed by the solvent action of pyridine. He believes that this constituent in a strongly coking coal which resists the solvent action of pyridine and, a3 Anderson and Roberts 0 found, resists the saponifying action of caustic potash, to be a "hydro- caroon derived from the resin bodies, and that in a strongly coking coal they condition the coking after the unaltered resins have been saponified ... ", Professor Lew&s criticises the U3e of pyridine as a solvent on the basis that in many cases the resulting insoluble residue contains a higher volatile matter percentage than the ori- ginal coal, and also an increase in nitrogen content, both of these facta leading to the inference that the pyridine forms an addition product with some constituent of the coal. Regardless, however, of this criticism, he still regards pyridine and aniline as being the two most successful solvents, 7 Wahl in 1913 examined the solvent action of pyridine upon a number of continental coals and confirmed Bedson's observa- tions that there is no relation between the weights of extractable material and of volatile matter of the coal. Those coals which yielded a large amount of extract did not lose all their coking property, 'but the coke from the remaining residue was much more compact than that obtained from the original coal. The extract - 4 - yielded an extremely ’’puffy” coke, and upon mixing with the residue, the original coking properties of the coal were reproduced. Vignon 3 in 1914 extracted various coals with aniline and quinoline. With aniline he obtained a 36,8 per cent extrac- tion of rich gas coals, 7.3 per cent extraction of medium coal and 1.8 per cent of a lean coal. The same coal which yielded 3S.8 per cent extraction with aniline, yielded only 17.3 per cent extraction with pyridine; however, with quinoline the coal gave a 47,3 per cent extraction. 9 Fraser and Hoffman" in 1913 tried the effect of a number of reagents and organic solvents on coal and found that pyridine, aniline, and phenol removed the largest amount of soluble material. Using phenol as the solvent they extracted 10.87 per cent of a Franklin County Illinois coal and then by further treat- ment of the extract with sodium hydroxide and organic solvents, they were able to isolate certain substances which they believed were approximately pure compounds but which could not be proven so on account of the small amount of material at their disposal. Parr and Hadley^ in 1914 also studied the solvent action of phenol on Illinois coals. They were able to obtain from the high volatile coals a 35 to 40 per cent extraction, from medium volatile coals a 30 to 35 per cent extraction and from the rela- tively low volatile coals a 30 to 30 per cent extraction. They found that the residue left after extraction did not coke and that the extract yielded a very fluffy coke. By mixing residue and extract they were able to reproduce the original coking properties of the coal. - O - 11 F. Fischer and W, Gludd in 1917 extracted coal with benzene under a pressure of 55 atmospheres and a temperature of 375°C and obtained a 6.7 per cent extraction of the raw coal. By pouring the extract dissolved in a little benzene into petroleum ether, they were able to precipitate out about 50 per cent of the extract. They found that this precipitate, when dried, was a brown o o powder which softens at 140 C, and melts and rune at 180 C. The portion soluble in the petroleum ether, upon evaporation, yielded an oily product 50 per cent of which was volatile in steam. 12 Pictet in 1918 extracted large amounts of a fat Sarre coal with benzene under atmospheric pressure, but was only able to obtain a. 25 per cent extraction. By anelvsis of the con- stituents of the extract, he 3hows that the extract is identical with vacuum tar and concludes, therefore, that the hydrocarbons in the vacuum tar must be present in coal as such. 13 Bone and Sarjant in 1919, working with pyridine under atmospheric pressure, extracted 30 to 33 per cent of coals of 33.3 per cent volatiles. By carrying on the extraction for a prolonged period in sealed tubes at a temperature of 13D to 150°C, they were able to increase the amount of extractable material to 57-63 per cent. They concluded from this that the solvent action of pyridine upon a coal substance involves in addition to a rapid dissolving of the resinous material, a simultaneous and much slower unbuilding or depolymer i zing of the whole coal structure, which is thus gradually brought into a more soluble condition. In summarizing the work on solvents we see that by this maan3 coal has been separated into two portions, the one - 3 - soluble in certain solvents and known as the resinic portion, and the other insoluble and known as the cellulosic or humus portion. In a semi-coking coal, the coking power is entirely lost with re- moval of resinic material, while in a strongly coking coal, the g coking property i9 only partially lost. This lead Professor Lewes' to infer that a portion of the resinic material resisted the solvent action of pyridine and saponification action of caustic potash, and that this constituent was a hydrocarbon. On the assumption that the bonding material i3 confined solely to the tar derived from the coal resins and hydrocarbons, Lewes states that the reason for the non-coking property of lignites and younger coals is due to an excess of humus bodies which on dis- tillation yt«ld no bonding material in the residue and the resinic cannot supply enough bonding material to give more than a friable mas3. As the coals undergo altered conditions of temperature and pressure, the humu3 bodies are further decomposed, resulting in a concentration of the resinic and hydrocarbon portion. When the pro- portion of resin and hydrocarbon bodies have reached the right ratio as compared with the humus and residuum, a strongly coking coal results. Professor Lewes believes that the resinic material is derived from the resins, gums and oils present in the original plant life. During the transition period from a semi-coking coal to a coking coal, a portion of the resinic material undergoes de- composition to yield a hydrocarbon which is responsible for the coking of the insoluble pyridine residues of coking coals. Dr. Hager 14 does not agree with Professor Lewes on the - 7 - question of the hydrocarbons being responsible for the difference between semi and strongly coking coals. He states that there is no evidence of such hydrocarbons beyond mere traces in coal, and says that "the same phenomena would be observed if the so-called resin bodies on 3low heating etc. are decomposed and yield among other products, a smaller quantity of a resin-like body which melts at a higher temperature. ... True coking coals have sufficient o rosin-like bodies to yield, after heating to 300 C., still suffic- ient resin-like bodies, or rather bodies that would melt to form Q coke when heated to 330 C and above, whilst the feebly coking coals on heating give the same kind and proportion of melted bodies from their resin but not sufficient in quantity to 'melt* or render liquid the whole material, or to give sufficient luting material to form a good coke." , 12 Arne Pictet claims that petroleums — at least some of them — are the verible tar from what is essentially a low tem- perature, dry distillation of coal. He bases his claim upon the physical resemblance between vacuum tar and petroleum, the identity of six of the saturated' hydrocarbons from coal with certain frac- tions of American petroleums, and the finding of melene in Galla- cian petroleum as well as coal. By identifying the benzene extract with vacuum tar, he concludes that the hydrocarbons which compose 9S per cent of the vacuum tar were present in the coal as such. His work is thus contradictory to Dr. Hager* s views. 15 R. Thiessen w in his report on the structure in Paleozoic Bituminous Coal^ which summarizes his extensive study of the structure of coals by microscopic examination, proves very con- - 8 - clusively that the bitumen of ooal is a degradation product of cellulose and his views are, in this country at least, supplanting those of Professor Lewes. n r* David White'*’” in 1909 put forth a Il/o ratio theory for the index of the coking power of coal. White shows that as the H/O ratio increases the coking power proportionally increases, that only a few coals with ratio between 53 and 55 will coke, that those between 55 and SO will yield only a fair coke, and as the ratio in- 17 creases to above 60 the quality of the coke increases. White advocates an algal theory of coal, i.e. that micro-algae form the bitumen part of coal, and that those coals that have a high vola- tile and high H/0 ratio, and correspondingly high coking power, have a high content of these micro-algae. R. Thiessen finds no evidence of algae and states that any such theory could not be demonstrated. From White 1 s work we see that coals of high oxygen content are non-coking or feebly coking coals, depending on the amount of oxygen. The effect of weathering or oxidation upon the coking property of coal has been known for a long time. Professor r% Lewes 0 mentions the work of Dr. Percy *who, more than fifty years ago pointed out the fatal effect of weathering upon certain coals and slacks, and who showed that if a fairly good coking coal were kept at a temperature of 300°C for a few hours, and wa3 later heated to redness, it did not swell and coke". Anderson and Roberts 2 found that the coking property of a 3 emi-coking coal was entirely destroyed by subjecting the coal to atmospheric oxidation, while the coking power of a strongly cok- ing coal was only partially destroyed by this oxidation. '■n Professor Lewes 0 believes that "when coal absorbs oxygen the compressed gas becomes chemically very active and soon commences to combine with the hydrogen and carbon of the resinic por- tion, converting them into carbon dioxide and water vapor". IS Boudouard' s work, however, shows that oxygen attacks the cellulosic constituents rather than the resinic and he reports the formation of humic acid as being coincident with the disappear- ance of coking power. 11 14 The work of both Hadley and Cherry substantiates Boudouard' s views. Cherry, as well as Hadley, extracted coal with phenol and examined the products of extraction. He found that upon oxidation cf the coal the first constituent to be affected was the cellulosic portion .and that in the first stages of oxidation, at least, the oxygen absorbed combines additive Iv with the unsaturated compounds of the cellulosic portion. His results showed that "oxi- dation of the cellulosic constituents of coal alone i3 sufficient tc modify the coking property of the coal, although the coking principle is contained in the phenol extract". He explains the above facts by assuming "that there is a reaction at fusion temperatures between the oxidized cellulosic constituents and the resinic bodies, whereby the latter are so altered that they cannot furnish luting material to bind the particles together". - 10 III. EXPERIMENTAL. 1, Description of Coa l . 0 The coal used was a high volatile bituminous coal from Castle Gate, Utah. Outcrops of resins are visible throughout the cc mass and since of apparently high resinic content, this coal is of particular interest as it retains the non-coking property character- istic of the western or younger coals. 2. Met hod o f Extraction The coal ground to pa33 through a 100 mesh sieve was dried in 100 grams samples for 3 hours at 110°C, Nitrogen was passed through the even to prevent oxidation of the coal. By dry- ing the coal before extracting, it was found that the extraction could be accomplished without the violent bumping that would other- wise attend its boiling. After drying, the coal was introduced into a 750 cc. Pyrex Erlenmeyer flask along with 300 cc. of warm diphenyl ether. The flask and contents were placed in an electric resistance furn- ace, the flask connected with condenser and air seal, and the air displaced from the apparatus with ethyl ether. All connections were coated with litharge and glycerol mixture to make tight joints. The temperature of the furnace was then raised to 350-330° C and kept between these temperatures for 48 hears. After extraction the contents were filtered on to a quantitative filter paper in a four inch Buchner funnel. A slight amount of coal will come through at first but as soon as a coal layer forms on the filter paper the solution come 3 through free of - 11 coal. The first filtrate i3 re-filtered to insure complete removal of the coal. The coal was then washed with pure hot diphenyl ether and replaced as rapidly as possible in the extraction flask, 500 cc. of warm diphenyl ether mixed with it and the coal again extracted. It was found that seven extractions of 48 hours each were necessary tc completely remove all the extractable material. The diphenyl ether solution of the extract was con- centrated to about 100 cc. by distilling off the excess solvent. The seven extractions concentrated in this manner were mixed to- gether and reconcentrated to about 100 cc. and then poured into a 300 cc. straight walled beaker. The beaker was stoppered and placed Q in the resistance furnace whose temperature was maintained at 350 C. Preheated nitrogen was passed into the beaker until all the diphenyl ether was volatilized off and no odor of it could be detected. The extract was then removed from beaker, weighed and preserved in nitrogen. The insoluble residue was washed first in the Buchner funnel with ethyl ether to remove the most of the diphenyl ether, and then washed in a beaker with several portions of ethyl ether. The ethyl ether was then filtered off and the residue dried in an o atmosphere of nitrogen for 4 hours at 110 C. The residue was then weighed and preserved under nitrogen. . - ■ - 12 3. Diagram of Apparatus . Resistance furnace male of 60 ft. of No. 16 Chrcmel wire (B) wound on a 5 x 10 inch sheet iron can (H) , the wires being im- bedded in alundum cement (C). The furnace is packed with crude fibrous »3oestos (D). The temperature of the furnace is controlled by the external resistance R. The extraction flask (A) is a 750 cc. Pyrex Erlenmeyer con- nected to a 3 ft. Pyrex condenser.(F.) Constant pressure and an atmosphere of nitrogen are main- tained by use of a i inch water seal (F) and a solution of alkaline pyrogallol (G). . - 13 - 4, Fx an; in at ion of Products . A. Ultimate Analysis. B.T.U. was obtained by a Parr Oxygen Bomb, total carbon by a Parr Total Carbon apparatus, nitrogen by Kjeldahl, sulfur, moisture and ash by the usual methods, and oxygen and hydro- gen calculated by Dulcng* s formula. B. Fractional Carbonization ana analysis of the gaseous products. Carbonization was carried on in a 150 cc. Pyrex dis- tilling flask provided with a nitrogen delivery tube and connected to first a tar well and then calibrated aspirator bottles containing saturated 3alt solution and in which the ga3e3 were collected and measured. The apparatus was swept cut before and after each determ- ination with a measured amount of nitrogen. The gases were analyzed in a modified Orsatt apparatus. Carbon dioxide was removed with caustic potash solution, oxygen with alkaline pyrogallol, unsaturated hydrocarbons with bromine water, aromatics with 15 per cent fuming sulfuric acid, hydrogen and carbon monoxide by means of combustion at 300°C with copper oxide, and the paraffins by slow combustion in an atmosphere of oxygen. C. Coking Tests. Coking tests were made in one gram portions in an Illium crucible and subjected to the temperature of the flame of a Meeker burner for seven minutes. D . I odi ne Uurabe r s . 19 The same method was used that Cherry 1 found to be well suited to coal. The iodine solution wa3 prepared by the Hanus - 14 method .and consisted of an iodine monobromide solution in glacial acetic acid. Twenty cc. of the solution was added to .5 gram of coal extract or residue in a 500 cc. Erlenmeyer flask. The flask was stoppered with a stopper moistened with 10 per cent potassium iodide solution and allowed to stand in a dark cool place for 1 hour. At the end of this time 10 cc. of 10 per cent potassium iodide solution were added, followed with 200 cc. of water, and the excess iodine titrated with standard l/lO F so diurn thiosulphate solution. Duplicate blanks were run with each determination. 5. Special tests on the extract. The melting point ms determined by supporting a small """ i lump of extract in a small wire loop and placing in the chamber of a resistance furnace. The temperature of the furnace was then gradually increased and the temperature noted when the extract be- came soft enough to run. Free Carbon was determined by extracting two grams of the extract in a Soxhlet extractor, using an equal mixture of benzene and toluene as the solvent. The residue left after complete removal of a soluble material was dried in an atmosphere of nitrogen for 3 hours at 120°C. Asphaltenes were determined on a 1 gram sample of extract by first letoing it stand for 48 hours with 40 cc. of petroleum ether, then filtering the solution and residue through \ two thicknesses of filter paper. The residue remaining on the fil- ! ter papers were then washed with hot benzene and the benzene wash- , ings collected in a taxed beaker. The benzene was volatilized off in an oven at 11C°C., the flask cooled in a desiccator and weighed. - 15 Paraffins were determined on a 1 gram sample of the extract as follows. The sample was placed in a 300 cc. ^rlenmeyer flask, 25 cc. of gasoline added and mixed thoroughly with the ex- tract and then 40 cc. of fuming sulfuric acid introduced. The mixture was allowed to complete sulphonation "by placing flask and contents in an oven at 110 c C for one hour and then contents were poured into Babcock milk dottles and enough concentrated fuming sulfuric acid added to fill the dottles. The dottles and contents were centrifuged and clear gasoline solution pipetted off the acid layer and placed in a tared flask. The gasoline was volatilized off in an oven at 15C U C, the flask cooled in a desiccator and weighed. - 16 TABLE I ULTIMATE ANALYSIS As rec' d basis Moisture and ash free basis Coal Coal Re si due Extract Moisture 3.35 Total Carbon 70.80 73.27 71.71 8 4.25 Sulphur .40 .42 .45 .38 A.sh 5.88 S . 08 S . 73 .10 Nitrogen 1.30 1.34 1.32 1.40 Oxygen 12.65 13.08 14.00 5.09 Hydrogen 5.62 5.81 5, 80 7.78 B.T.U. 13,835 13,259 13,965 16,633 Volatile Matter •48.82 44.31 40.50 67.75 Fixed Carbon 47.95 49.61 53.78 33.15 - 17 TABLE II ULTIMATE ANALYSIS (Moisture Free) CALCULATED TO COAL BASIS Residue Per cent in coal 93. SO Total carbon 67.33 Sulphur .43 A3h 6.31 Nitrogen cv • • J- • w • 13,174 Extract Sura Coal 6.10 100.00 100.00 5.14 73,47 73.37 .03 .44 .42 .00 S • 31 6.08 .09 1.33 1. 34 .37 13.53 13.08 .47 5.93 5.81 1,015 13, 189 13,259 - 18 - TABLE III GASEOUS PRODUCTS OF FRACTIONAL CARBONIZATION. For temperatures between 20° and 375°C. Coal Re 3 idue Extract Per oent Gas Per cent Gas Per cent Gas C0 o i 1 C3 • i to I to 1 54.4 34.6 °2 3.0 12.3 19.3 c n H 3n 3.9 6.5 1.9 C 6 E s .9 .0 3.8 h 2 4.2 4.3 11.6 CO 11.5 17.0 1.9 C n H 3n+-2 38.3 5.3 23. S Cc * 3 of gas evolved per 10 grams - - 135 70 75 - 19 - TA3LE IV GASEOUS PRODUCTS OF FRACTIONAL CARBONIZATION. For ■temperatures ■between 375° and 430° C. Coal Residue Extract Per cent Gas Per cent Gas Per cent Gas C0 2 9.9 31.1 P P C • £> °3 3.7 2.3 0 to C n H 3n 9.4 10.2 7,3 C S H S . 5 .4 3.1 H 3 7.3 7.0 1.7 CO 7.3 8.7 5.9 C n H 3n+3 31.9 50.4 70.1 Cc* 3 of gas evolved per 10 grams — 320 270 360 - 30 - TAELS V AMOUNT OF MATERIAL EXTRACTABLE WITH DIPHENYL ETHER Weight of coal before extracting IOC grams Moisture lost in drying 3.35 grams Weight of residue 91.02 n Weight of extract 5.30 " Burn ICO. 27 " Per cent extraction on as rec'i basi3 5.90 Per cent on moisture free basis 3.10 Per cent on moisture and ash free basis 6.48 TABLE VI IODINE NUMBERS Coal Residue Extract Gram 3 of iodine absorbed per 100 grams of sample 30.57 36.60 21.34 TABLE VII SPECIAL TESTS ON T 'HE EXTRACT Free Carbon 34.10 per cent Asphaltenes 31.68 ft n Paraffins 1.15 tt ft Melting Point .... 220° - 350° C. Description of Extract Color black, shiny Fracture concho id al Streak chocolate brown . r •: . . t i 31 TA 2LE VIII COKIITC- TESTS Source of residue > from Coun xylene extraction of Franklin ty Illinois coal. Source of extract from Utah diphenyl ether coal. extraction cf a Test Residue Extract Volatile Mtr. Character of re- No. Per cent Per cent Per cent maining residue. 1 100 0 45.33 Powder . 3 95 5 43. 30 Powder 3 SO 30 47,00 Fair coke, silvery 4 70 30 49.70 Resembles test 3 onl coke more voluminous 5 50 50 58.10 Resembles test 4 only coke more voluminous 7 0 100 S7.75 L i ght and f luf f y, shows no coke structure . - 23 - TABLE IX COKING TESTS Source of .residue Source of extract extraction of Utah coal Test Residue Extract Vo 1. Mat ter Character of remain- no. Per cent Per Qent Per c^nt ing residue. 1 95 5 43.00 Powder. 3 70 30 48 , 55 ?Feak cake, with no coke structure. 3* 70 30 — Same as test- 3. TABLE X COKING TESTS Source of residue diphenyl ether extraction of Utah coal. Source of extract benzene extraction of Franklin County Illinois coal. Test Residue Extract Character of remaining no. Per cent Per cent residue. 1 70 30 Weak cake, with no coke structure 2* 70 30 Same as test 1. ♦in test 3 Table IX and test 2 Table X, the unaltered Utah coal wa3 substituted for the residue. - _ .IL/ *“ III. DISUUSSION OF RESULTS 1. Diphenyl ether as a solvent . From Table II it will be noted that the sum of each constituent in the ultimate analysis of residue and extract approaches the percentage of that constituent in the original coal within experimental error. This proves that diphenyl ether is a neutral solvent and neither combines with or decomposes any of the products present in the coal and is, therefore, as a true solvent superior to pyridine, aniline and quinoline. Diphenyl ether is non-corrosive to the skin, is of pleasant odor, is easily removed from both extract and residue with- out decomposition, and in all respects is an easy solvent to handle. Table V shows that the maximum amount of solvent retained by ex- tract and residue combined is .27 per cent. The percentage of material extractable by its solvent action, though not as great as compared with the amount obtained with some of the more chemically active solvents, compares very favorably with the best of neutral solvents. Because of its high boiling point it permits a higher obtainable temperature under atmospheric pressure than the lower boiling solvents 3uch as benzene, toluene, etc., and because of this fact a larger percentage of extractable matter can be obtained with diphenyl ether than the lower coiling solvents, extracting under atmospheric pressure. The fact that the diphenyl ether no longer removes any portion of the remaining residue after seven extractions, shows that only a definite portion of the original coal is soluble in this solvent. Table I shows that the analysis of this soluble 24 constituent is of the general composition of resinic material. From Table VII it will be noted that the asphaltene content i3 comparatively high and the paraffin content low, and in this respect the resinic material resembles a true bitumen. From analysis of the gaseous products of carbonization of extract, residue and coal in Tables III and IV, two facts are in evidence. One is that diphenyl ether removes a constituent which decomposes to liberate paraffin hydrocarbons below 375°C and the other, that the remaining residue evolves large amounts of CO and C0 o below 375°C. Since the melting point of the extract, though not definite, is in the neighborhood of 320° to 250°C, and since para- ffins are evolved to 3ome extent below 375°C with the percentage in- creasing rapidly as the temperature increases above 375°, we know that the resinic material starts to decompose below 375°C and that its pasty period is, therefore, below this temperature. From the second fact mentioned above it is evident that during the pasty per- iod of the resinic material, the cellulosic portion is evolving large volumes of CO and COo. By the comparison of iodine numbers we note that the coal and extract are nearly e-qual in unsaturation while the residue is decidedly more unsaturated. Likewise in the analysis of the gaseous products of carbonization it can be noted that the residue evolves a larger percentage of unsaturated hydrocarbons than the 19 extract. This data collaborates the results obtained by Cherry " and proves that the cellulosic portion of the coal is more unsaturated than the resinic. 2« Cokina; Tests , Table trill shows that by mixing the extract from the Utah coal with residue from Illinois coal, a coke is obtained that is equal in strength and structure to the coke formed when an equal amount of extract from Illinois coal i 3 mixed with the same residue. This proves that the resinic portion of the Utah coal ha 3 a bond- forming value equal to the resinic portion of the Illinois coal. 3 According to the theory of Professor Eswes the non- coking cf the younger coals is due to the predominance of the humic or cellulosic portion over the resinic and that the latter is, therefore, not present in large enough quantity to furnish the necessary bonding material. If his theory is correct, a coking property should be obtained by merely increasing the resinic con- tent of the coal. Tables IX and X show that the result of increas- ing the resinic content to 30 per cent, or 10 per cent more than was necessary to coke an Illinois residue, did not increase the coking tendency in the least. In making these determinations extract from both Illinois coal and Utah coal were mixed with first Utah residue and then Utah coal to prove that it was neither the Utah extract or a possible oxidation cf the residue during extraction that was responsible for the non-coking. All the above fact 3 prove conclusively that the non- coking of the Utah coal lies in the fault of the cellulosic por- tion and not in the resinic content. As mentioned previously, we note a large volume of CO 3 and 00 liberated by the residue under 3?5 c C or during the range in which the pasty period of the extract is in evidence, and it - 26 seams possible that the gases liberated by the cellulosic portion during the forming of the bonding material may be responsible for the destruction of the coking property. Now since 00, C0_ and Oo are all liberated by the resinic portion, as well a3 the cellulosic portion, it does not 3sem probable that these same gase3 liberated by the cellulosic portion should have any chemical effect upon the bond-forming constituent. It doe 3, however, seem plausible that a liberation of a large volume of gas around the cellulosic portion during the pasty stage of the resinic portion, would have a pro- tective effect on the cellulosic portion and thus prevent the bonding material from reaching the material to be bonded. As the temperature raises, the bonding material decomposes and loses its bonding power. Since oxygen and its oxides of carbon constitute 84 o per cent of the gases evolved under 375 , it can be inferred that the high oxygen content of the Utah coal is responsible for its non-coking, due to the liberation of the oxygen, combined or free, in the form of gases during the critical time when the resinic por- tion is in the bond -forming stage. This theory will explain the fact that all young coals, being high in oxygen, and likewise oxi- dized coals, will not coke because of the liberation of the gaseous products of oxygen in large volumes, causing a physical repulsion of the bonding material away from the material to be bonded. 37 IV. CONCLUSIONS 1. Diphenyl ether is a true solvent and neither com- bines with or decomposes any constituent of the coal. It is easy to handle and can he freed from both extract and residue without decomposition. 2. The resinic portion of the Utah coal possesses a bonding power equal to that in the resinic portion of a Franklin County Illinois coal and is, therefore, not responsible for the non-ccling of the Utah coal. 3. The non-coking of Utah coal lie 3 in the character of the cellulosic portion. 4. The hi gh oxygen content of the coal i.3 responsible for the liberation of a large volume of gas during the pasty period which by physical repulsion of the bending material away from cellu- losic portion, prevents the formation of coke structure. I . s . 1 28 V. BIBLIOGRAPHY I. J. Soc. Chem. Ind. 27, 14S (1908). 3. " " " " 17, 1013(1898). 3. " " » " 30, 789 (1901). 4. " " " " 31, 343 (1902) . 5. " " " " 27 ^ 140 (1908). 3. The Carbonization of Coal. (1914). 7. Co rip. rend. 154 . 1094 (1913). 8. »' " 158, 1421 (1914). 9. Tech. Paper 5, U.S. Bureau of Mines, (1913) . 10. Uni. of 111. Rxp. Star. Bull, 73 (1914). II. Chain, abstracts, 11, 1739 (1917). 13. Ann. Chim. 10, 349 (1918). 13. Proc. Roy. Soc. London, 98 A, 119 (1919). 14. Gas World SO, 18 5 (1914). 15. Bull. 117, U. 8. Bureau of Mines (1920). 13.. Bull. 383, U.S. Geo. Survey, (1909). 17. Bull. _38, U.S. Bureau of Mines, (1913). 18. Bull, fe la Soc. Cherr.. 96 , 365 (1909). IS. Thesis, Uni. of 111. (1919).