Composition of Shale WILLIAM E. CADYl
AND
HERMAN S. SEELIG
Research Department, Standard Oil Co. (Indiana), Whiting, lnd.
T
HE oil shale deposits of Shale oil differs from crude petroleum in that it contains and for conjugated diolefins the United States are a relatively large quantities of nitrogen, sulfur, and oxygen. by maleic anhydride addiKnowledge of its composition is essential for the producpotential source of tretion. Essentially all samples tion of usable petroleum products. NTU shale oil has mendous quantities of petrowere analyzed for nitrogen, leum substitutes. The most been examined by a combination of the following techsulfur, and oxygen. Modified niques: distillation, extraction with a metal chloride in micro-Kjeldahl methods (6, important deposit is located concentrated hydrochloric acid, adsorption, extractive in Colorado, U t a h , and 1 7 ) were used for nitrogen crystallization with urea, and infrared spectrometry. Wyoming and is called the analysis, the conventional This shale oil consists of 397' hydrocarbons and 61% Green River formation ( 2 7 ) . bomb method for sulfur, and The Mahogany Ledge of this compounds of nitrogen, sulfur, and oxygen. The hydroa modifiedUnterzaucher techcarbons are 6 % normal paraffins, 6 % normal olefins, formation covers 1000 square nique (8) for oxygen. M o 570 isoparaffins plus naphthenes, 12% iso-olefins plus miles in western Colorado, is lecular weights of many samcyclo-olefins, 4 % monocyclic aromatics, and 6qo polyabout 100 feet thick, assays ples were determined by boilcyclic aromatics. The nonhydrocarbons are 36y0nitrogen ing point elevation in either a 25 or more gallons of oil per compounds, 6 % sulfur compounds, and 19% oxygen ton, and contains 126 billion Menzies-Wright apparatus ( d o ) or a semimicro appacompounds. The removal or conversion of the nonhydrobarrels of oil ( 2 ) . It is this Mahogany Ledge t h a t is becarbons in shale oil to produce high-quality petroleum ratus described by Matteson products will require extensive processing that is not ing worked by the Bureau of ( 1 9 ) . Determinations of rerequired in conventional refining of petroleum. Mines a t Rifle, Colo., inits profractiveindex,specific gravity, and specific dispersion were gram of mining and retorting made on most of the samples. oil shale and refining shale oil. Distillation. A partial separation by molecular weight was Because shale oiyis unlike crude petroleum, specific knowledge effected by distillation. After removal of about 0.5y0 water by of its composition is essential for the production of usable prodheating to 205" C., 6 kg. of shale oil were distilled at 0.1 to 0.3 ucts. The composition of shale oil naphtha has been reported by Ball and associates ( 1 , 6); they have also reported progress mm. into five successive 10% fractions and a 50% residue in the analysis of shale oil fractions in the 200' to 325" C. range There was no evidence of decomposition; the recovery was 99.470. The specific gravities, molecular weights, and elemental (7'). Nottes and Mapstone ( 2 1 ) have analyzed the gas-oil fraction, boiling from 240" t o 320" C., of Australian shale oil. Fraganalyses of the fractions are presented in Table 11. Nitrogen content increased with molecular weight; sulfur 1% as almost mentary data on individual organic compounds and hydrocarbon uniformly distributed; maximum oxygen content occurred in isomers or types in various shale oils exist (3,4>10-16, 18, 23, 24). the 20 t o 30% fraction. No comprehensive information on the composition of shale oil Extraction with Metal Chloride Solution. The distillation frachas yet been published. tions mere treated with a metal chloride solution to separate the The shale oil used in this study was obtained by retorting hydrocarbons from the compounds containing nitrogen, sulfur, or Mahogany Ledge oil shale in an NTU retort (9) a t the Bureau of Mines Oil-Shale Demonstration Plant a t Rifle (27). Inspecoxygen (82). These nonhvdrocarbons are referred to as NSO tion data are tabulated in Table I. This oil is characteristic of compounds. Raffiates from metal chloride extraction are called hydrocarbon concentrates. the product that has been produced in relatively large quantities The extraction reagent was a concentrated hydrochloric acid by the U. S. Bureau of Mines. In small scale experiments, solution containing either 30% zinc chloride or 33% ferric chloride. shale oils have been produced that differ markedly from the ( A number of other metal chlorides were found t o be less effective.) NTU product ( 2 8 ) . Batch extractions were performed on 25 grams of the 20 t o 30% EXPERIMENTAL The analysis of shale oil involved six steps: Table I. 1. Distillation of moisture-free shale oil a t high vacuum into five successive 10% fract,ions and a 50% residue 2. Extraction of the fractions and residue with ferric or zinc chloride dissolved in concentrated hydrochloric acid 3. Silica-gel percolation of the raffinates from step 2 4. Urea extraction of the nonaromatic hydrocarbon eluates from step 3 5. Infrared analysis of some of the extracts and raffinates from step 4 and of selected monocyclic-aromatic concentrates from step 3 6. Microfractionation under vacuum of the urea extracts from step 4
On the appropriate samples obtained during the six steps, tests were run for unsaturation by bromine number or iodine number 1 Present address, California Research Corp., 576 Standard AT'^., Richmond 1, Calif.
2636
Inspection of Colorado Shale Oil
Specific gravity 15.6/15.6' C . Distillation, C . Init. 10 % 20 % 30 70
40 70 50 % 60 % 70 %
80 %
Elemental analysis, wt. % Carbon Hydrogen Nitrogen Subfur Oxygen Total €I/C atomic ratio
INDUSTRIAL AND ENGINEERING CHEMISTRY
0.934 182 262 304 345 380 413 448
470 483 84.25 11.41 2.02 0.72 1.61 100.01 1.61
Vol. 44, No. 11
PETROLEUM-SHALE OIL fraction and on 200 t o 500 gramn of the other fractions. For ease in handling, the 0 t o 10% and 10 t o 20% fractions were diluted with 3 volumes of technical grade pentane; the other fractions were diluted with 7 volumes. In each case, 2 volumes of t h e resulting solution were treated with 1 volume of the metal chloride reagent. ~~~~
~
Table 11. Fraction, Wt. % 0- 10 10- 20 20- 30 30- 40 40- 50 50-100
Characterization of Distillation Fractions Specific Gravity
15.6/15.6'
0.861 0,881 0.902 0.919 0.922 0.984
C.
Ilqolecular Weight 192 221 250 280 330 586
Elemental Composition, Wt. % N S 0 1.14 1.33 1.84 1.99 1.94 2.42
1.00 0.80 0.80 0.77 0.68 0.75
1.40 1.72 1.88 1.50 1,12 1.22
Yields, molecular weights, elemental analyses, and compositions of the hydrocarbon concentrates are listed in Table 111. The extraction proved t o be most specific for the nitrogen compounds. The compositions of the hydrocarbon concentrates from the five 10% fractions were relatively constant; the 50 t o 100% concentrate contained less than half as much hydrocarbon as the five 10% concentrates. Hydrocarbon content of the distillation fractions declined as molecular weight increased. No attempt was made t o identify the compounds that formed adducts with the metal ohlorides. Silica-Gel Percolation. The silica-gel percolations of the hydrocarbon concentrates accomplished three separations: ( 1 ) hydrocarbons from the N 8 0 compounds not removed by the metal chloride extraction; (2) nonaromatic hydrocarbons from aromatics; and (3) monocyclic aromatics from polycyclic aromatics. Samples of 100 t o 300 grams of each of the five hydrocarbon concentrates were dissolved in 0.5 to 1 liter of technical grade pentane. The solutions were passed first through a tapered column containing 1000 grams of 28- t o 200-mesh silica gel and then through a smaller tapered column containing 400 grams of silica gel. The samples were eluted successively by 7 to 11 liters of pentane, 4 to 5 liters of benzene, and 1.5 to 4.5 liters of methanol. Eluates, in multiples of 0.4 liter, were collected. The eluents were removed by distillation and the refractive indices, specific gravities, and specific dispersions of the residues were determined. Hydrocarbons were composited by type as indicated by refractive index and specific dispersion. In most casw, a sharp separation of high-purity nonaromatic hydrocarbons was achieved. The aromatic composites were essentially free of nitrogen compounds, but they contained some oxygen compounds and higher concentrations of sulfur compounds than were present in the hydrocarbon concentrates. Good separation of monocyclic aromatics from polycyclic aromatics was obtained. A typical adsorptogram is shown in Figure 1 in which refractive index and specific dispersion are plotted as functions of the weight of product eluted. The eluents employed are also indicated. The breaks in Lhe curves when about 60% of the sample has been eluted are due t o the appearance of monocyclic aromatics; the transitions in the vicinity of 70% indicate a shift from monocyclic t o polycyclic aromatics.
Table 111. Metal Yield, ~ ~ Chloride Wt.% Weight 200 0- 10 ZnCln 84 236 79 10- 20 ZnCln 280 63 ZnCln 20- 30 325 30- 40 62 FeCla 375 59 FeCla 40- 50 540 50-100 34 FeCla From bromine numbers. b Assuming one N, S,0 atom per molecule.
Fraction
November 1952
The 20 t o 30% and 50 t o 100% fraction8 were handled differently than the other four fractions. I n t h e case of the 20 t o 30% fraction, a time-consuming silica-gel percolation on a large scale had been completed before the metal chloride studies were begun. The benzene eluate, consirting of aromatic hydrocarbons, had been further percolated through Florisil, a synthetic magneaium silicate (W6), to complete the separation of the aromatics from the nitrogen and oxygen compounds and t o separate the monocyclic aromatics from polycyclic aromatics. The 68%
I
I
0
I
20
.
40
,
60
80
100
WEIGHT % ELUTED
Figure 1.
Silica-Gel Percolation of the 10 to
20% Hydrocarbon Concentrate
of the fraction eluted from silica gel by pentane and benzene waa considered equivalent t o a hydrocarbon concentrate as obtained from metal chloride extraction. The 50 t o 1 0 0 ~ ohydrocarbon concentrate wa$ subjected t o an additional percolation through 500 grams of Florisil in aeries with the silica gel in order t o aepnrate a dark material-presumably nitrogen and oxygen compounds-from the nonaromatic hydrocarbons. The distribution of the adsorption composites by hydrocarbon types are preeented in Table IV. Most of the adsorption coniposites contained NSO compound^. Urea Extraction. Straight-chain hydrocarbons were separated from branched-chain and cyclic hydrocarbons by urea extraction (89). Samplw of 40 to 170 grams of the nonaromatic hydrocarbons from the percolation experiments were diluted with 0.5 t o 2 liters of pentane and reacted for 2.5 hours with 100 t o 900 grams of urea in the presence of 50 t o 150 ml. of methanol activator. The urea adducts were removed by filtration and decomposed by hot water. The diluent and activator were removed by distillation.
Characterization of Hydrocarbon Concentrates Elemental l Analysis, ~ Wt. ~ % N S 0 0.13 0.10 0.15 0.11 0.13 0.18
0.88 0.90 0.75 0.94 0.78 0.94
1.13 1.14
0.66 0.99 1.12 4 . 02
~
unsaturation", l Wt. % 81 86 92 100 100 101
conjugated Calculated Compositionb, Wt. % Diolefins, ~ ~ N S 0 HydroWt. % cmpds. cmpde. cmpds. carbons 0.3 0.0 1.0 0.0 0.0 0.0
INDUSTRIAL AND ENGINEERING CHEMISTRY
4 5 4 6 6 18
11 12 8 14 19 38
84 82 86 78 72 36
2637
Table IV.
Distribution of Adsorption Composites by Hydrocarbon Types
KonFraction aromatics 0- 10 62.3(100)b 10- 20 5 9 . 2 (100) 20- 30 6 0 . 8 100) 30- 40 54.6 Loo:, 40- 50 5 6 . 8 (99) 50-100 4 7 . 8 (97)
Hydrocarbon TypesQ, Wt. % Monocyclic Mixed aromatics Mixed 5 5 98) 3 : 2 [92)
8.0(77) 7 . 4 (78)
b:i( b i j
i:i ( i 4 j 7 . 5 (72) 1 0 . 4 (56)
0.6 (88)
... . . ,
...
...
4 . 8 (77) 2 1 . 4 (81) 1 . 3 (81)
., .. ., .. ,. ..
sso
Polycycllc aromatics
Compounds
15.8(66) 1 5 . 8 (60) 9 . 0 (69) 2 4 . 3 (67) 2 5 . 2 (50) 2 2 . 9 (21)
8.4(0) 9 . 6 (0) 8 . 8 (0) 6 . 4 (0) 4 . 9 (0) 18.9 (0)
a Based on refractive index and specific dispersion data. b Parenthetical figures indicate the actual concentration of hydrocarbons n the composite.
Table V.
Fraction 0- 10 10- 20 20- 30 30- 40 40- 50 50-100
Urea Extraction of Nonaromatic Hydrocarbons
Charge UnsatuMolecular ration, weight nt. % 220 73 257 71 315 71 359 69 415 60 42 490
Extract (Straight-Chain Hydrocarbons) UnsatuYield, ration, wt. % wt. % 54 34 55 36 59 45 58 57 46 5a 47 42
Raffinate (Branched plus Cyclic Hydrocarbons) UnsatuYield, ration, wt. % wt. % 66 81 64 80 81 55 ai 43 ao 42 40 58
The results of the urea extractions are summarized in Table 1'. I n the five successive 10% fractions, t h e straight chain-hydrocarbon content almost doubled, the unsaturation of the branched and cyclic hydrocarbons remained constant, and t h e unsaturation of the sttaight-chain hydrocarbons was essentially constant in the first four fractions and lower in the 40 to 50% fraction. The 50 to 100% fraction differed unexpectedly in t h a t it contained less straight-chain hydrocarbon than the three preceding fractions; furthermore, the unsaturation of the branched and cyclic hydrocarbons was only half that of the 0 t o 50% fractions, Hydrocarbons containing 20 or more carbon atoms and having a methyl branch near the end of the chain also form adducts with urea (935). Inasmuch as the hydrocarbons in the 20 t o 100% portions of shale oil contain more than 20 carbon atoms, urea extracts of these portions might be expected t o contain methylbranched hydrocarbons. Infrared analysis did not detect tertiary olefins; this suggested that methyl-branched hydrocarbons were not present in the adducts in significant amounts. Infrared Analysis. Selected straight-chain hydrocarbons and the corresponding branched plus cyclic hydrocarbons from the urea extractions were examined to determine the olefin types present. Certain of the monocyclic-aromatic concentrates from the silica-gel percolations also were examined to determine the types of ring substitution. Except for the straight-chain terminal and trans-internal olefins, infrared analysis yielded only qualitative data because of the complexity of the samples. Qualitative results of the examination of the straight-chain hydrocarbons of the 0 to lo%, 20 to 30%, 40 to 50%, and 50 to 100% fractions aresummarizedin Table VI. Terminal and trans-internal olefins were present. Tertiary terminal and tertiary-internal olefins were not detected and would have been present only if appreciable quantities of branched hydrocarbons had been extracted by urea. Conjugated diolefins were not detected; this result agrees with the values previously reported for the metal chloride raffinates (Table 111). Of the three olefin types that cannot be identified by infrared, nonconjugated diolefins may have been present, quaternary olefins would not be expected in urea extracts, and cis-internal olefins probably were present. Thermodynamic equilibrium data ( 1 6 ) indicate that the concentrations of cisand trans-internal olefins should be about equal. Quantitative data on the distributions of terminal and transinternal olefins in the straight-chain hydrocarbons are shown in
2638
Table VII. With a doubling of molecular weight, the trans-internal olefin content increased and the terminal olefin content decreased, so that the ratio of trans-internal t o terminal olefins increased threefold. The total olefin contents from infrared analysis agreed very well with those calculated from bromine numbers; both methods showed a maximum in the 20 t o 30% fraction. This agreement indicates that cis-internal olefins were present in relatively small amounts t h a t were far from thermodynamic equilibrium. Examination of the branched plus cyclic hydrocarbons yielded the following qualitative observations: I n comparison with the spectra of the corresponding straight-chain hydrocarbons, the intensity of the absorption a t 13.8p, which is characteristic of long straight chains, was weaker; a t 7 . 3 p, characteristic of methyl groups, it was stronger; and a t 10.0 and 11.0 p , which are characteristic of terminal olefins, it was weaker. The observed absorption in the region of 8.5 p was due t o branched structures and that in the region of 12 p could be due t o tertiary-internal olefins. Monocyclic-aromatic concentrates from the silica-gel percolations of the 0 t o lo%, 20 to 30%, and 50 t o 100% fractions were also examined in the regions of 11 t o 15 p for benzene-ring substitution. Of the substituted ring types, the para-substituted were predominant and lesser amounts of the meta-substituted were present; the complexity of the sample spectra precluded the possibility of estimating the ortho-substituted rings, Monosubstituted rings were present t o about the same extent as metasubstituted. -4s expected, the aromatic-ring content decreased with increasing molecular n-eight; the corresponding increase in side-chain carbons is accounted for primarily by an increase in the length of the side chains rather than by a n increase in the number of side chains. No substitution on the alpha carbon of t h e side chain on the mono-substituted aromatic ring was found in any fraction; this indicates t h a t aromatics are not alkylated during the retorting process. A11 the monocyclic-aromatic concentrates analyzed contained trans-internal olefin bonds in about the same volume concentration; an increase in the molec-
Table
VI.
Infrared
Examination Hydrocarbons
of
Straight-Chain
Identifying Absorption Olefin Type
Formula
Terminal
R-CH=CHz
Trans-internal Tertiary-terminal
imns-R-CH=CH-Ri R-C=CHz
Tertiary-internal
R-C=C-Ri
Band, p
Result
10.0, 11.0 10.33 11.25
Present
11.8to 12.5
S o t detected
6.2
Not detected
Present S o t detected
~
R1
!
I
H Rz R--CH=CH-CH=CH-Ri
Conjugated diolefin Konconjugated diolefin Quaternary
R-CH=CHCH=CH-R1 R-C=C-Ri
I
(CHz)n-
1
Rz R,E cis-R-CH=CH--Ri
Cis-internal
Table VII.
Sone
...
h-one
...
Kone
...
Olefin Contents of Straight-Chain Hydrocarbons Olefin Contenta, Mole % Infraredb
Fraction 0- 10 20- 30 40- 50 50-100 5
b
Mol. wt. 220 315 415 490
Transinternal 16 26 24 25
Terminal 37 37 27 21
Total 53 63 51 46
Bromine NO. 54 50
46 47
Calculated as mono-olefins. Estimated accuracy of & l o % of the value reported.
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 44, No. 11
PETROLEUM-SHALE OIL 20
t-
0 - IO % FRACTION
10
-
10
-
10
-
10
-
IO - 20 %
2 W 0
a 0 W
w
f-I
0
>
eo
10
30
40
CARBON NUMBER
Figure 2.
Distribution of Straight-Chain Hydrocarbons by Carbon Number
with reasonable assurance, and the distributions of aromatics according t o ring content with less assurance. Although the total NSO-compound contents of shale oil and its fractions are known with considerable accuracy, the distributions of the nitrogen, sulfur, and oxygen types are uncertain. T o calculate the composition of shale oil it was necessary t o make numerous assumptions concerning the separation processes, methods of calculation, molecular weight relationships of hydrocarbons and NSO compounds, and the concentration of polyNSO compounds. Those assumptions related t o separation processes or to the methods of calculation are tabulated and discussed in Table VIII. Initially it was arbitrarily assumed t h a t t h e hydrocarbons and NSO compounds in the same fractions had t h e same molecular weight, but this yielded inconsistent results. The molecular weights of the shale oil distillation fractions (Table 111, of t h e hidrocarbon concentrates (Table 111), and of t h e nonaromatic hydrocarbons (Table V) were determined experimentally. Molecular weights of the NSO compounds were then calculated from the results of these experiments. It was assumed that the aromatic and nonaromatic hydrocarbons in a given fraction had the same molecular weight. The molecular weights of the hydrocarbons, of the NSO compounds, and of the total fractions are shown in Figure 3. NSO compounds in the 0 t o 50% fractions had much lower molecular weights than t h e corresponding hydrocarbons, but for the 50 t o 100% fraction the NSO compounds had a much higher molecular weight than the hydrocarbons. I n the preliminary calculations of the NSO-compound contents, it was assumed that the NSO compounds contained only one atom of nitrogen, sulfur, or oxygen per molecule. However, when the molecular weights of the NSO compounds and the ele-
ular concentration of aromatics having trans-internal olefinic side chains seems t o occur with increasing molecular weight. Microfractionation. The boiling ranges of the shale oil fractions gave only an approximation of the carbon-number distributions. The molecular weights of the hydrocarbon concentrates were of limited assistance in further defining the average molecular weight, because of the NSO compounds that were also present. However, straight-chain hydrocarbons obtained by urea extraction were suitable materials for determination of the carbonnumber range of shale oil hydrocarbons. The straight-chain hydrocarbons from the five lO%fractions .were fractionated in miniature Table VIII. Assumptions Required for Calculating Shale Oil Composition Hyper-Cal, Podbielniak conSeparation Process Assumption Comment Distillation centric-tube, or Piros-Glover Proportional distribution of losses Wt. % recovery of 99.43 spinning-band columns. SamOf total shale oil Vol. 70recoveries of 92.0 to 97.7 Of urea adducts ples of about 35 ml. were No recovery of extracts charged to columns operating Extraction with metal chloride solution 100% recoveries No extraction of 4ydrocarbons Demonstrated experimentally Ne ligible chemical reaction withLow chloride contents of raffinates at 1 t o 20 mm., the pressure t i e hydrocarbons depending on the fraction being Silica-gel adsorption Proportjonal distribution of losses Wt. 70recoveries of 96.5 t o 100.9 distilled. At total reflux and o r gains a t m o s p h e r i c pressure, these Urea extraction Proportional distributions of losses Wt. 70recoveries of 88.8 to 99.9 columns have an efficiency Method of Calculation Experimental NSO-compound con- NSO-compound content equals sum equivalent, t o about 60 theotents of distillation fractions of: material removed by metal chloride, compounds separated by retical plates. From the distilpercolation, and compounds found lation data and the molecular in mixed percolation fractions weights of related fractions, Distribution of W S and 0 com- N, S, and 0 compounds are present pounds in nongydkooarbon porin atomic ratio of N, S, and 0 the carbon-number distributions of distillation fractions tions of the urea extracts were NSO-compound contents of fractions All NSO compounds have the same ... from adsorption experiments molecular weight calculated for estimated, with the results prethe total NSO compounds i n the sented in Figure 2; the yields particular distillation fraction ... Oxygen contents of nonaromatic Infrared spectra show no C=O and average carbon numbers of fractions equal zero 0-H o r -CHI-O-CH~Nitrogen contents below 0.05’% Represints lower limit of sensithe distillation residues are also equal zero tivity of method shown. As their molecular Hydrocarbon content of any fraction 100% - % NSO compounds Always gives a minimum value weight increased, the fractions Straight-chain hydrocarbon contents Urea extracts consist only of straight- Extraction of some methylcovered a wider carbon-number chain hydrocarbons branched hydrocarbons is possible range. Adjacent fractions overOlefin contents of urea charges, ad- Only mono-olefins present High values of diolefins are present lapped extensively. ducts, and raffinates Molecular weights of extracts and ... raffinates same as those of charges
METHODS OF CALCULATION
From the data obtained, the hydrocarbon contents of shale oil and its fractions can be calculated with considerable certainty, the type compositions of the hydrocarbons
November 1952
Nonaromatic and monocyclic-aromatic contents of mixed adsorption fractions Iso-olefin plus cyclo-olefincontents of certain mixed adsorption fractions Monocyclic- and polycyclic-aromatic contents of mixed aromatic adsorption fractions
Based upon average refractive in... dices and specific dispersions, some literature and some experimental Nonaromatics are iso-olefins plus , Unsaturation of the fractions cyclo-olefins range from 92 to 133% Monocyclic-polycyclic splits based ... on literature values of refractive index and specific dispersion of mono- and dicyclic aromatics: effect of NSO compounds on physical properties of aromatic fractions eliminated by calculation
INDUSTRIAL AND ENGINEERING CHEMISTRY
2639
mental analyses of the samples were used to calculate the percentage of these compounds present, the calculated values were always higher than the experimentally observed values. The calculated values could be high only if poly-NSO compounds were present. Therefore, it was necessary t o express the nonhydrocarbon contents in terms of mono-NSO and poly-NSO compounds.
I
o
Figure 3.
IO
m
I
*
I
I
30 40 so SHALE-OIL DISTILLATION FRACTION, WT.%
LOO
Molecular Weights of Shale Oil Fractions and Components
T o arrive at the poly-NSO content, two assumptions were made. The poly-NSO compounds could contain two, three, or more atoms of nitrogen, sulfur, or oxygen, but in the absence of specific knowledge concerning these materials, it was assumed that they contained only two of these atoms per molecule. Such di-NSO compounds could containnitrogen, sulfur, or oxygen atoms in like or unlike pairs. It was also assumed that the nitrogen, sulfur, and oxygen atoms doubled up t o an equal extent. From these assumptions, t h e percentage of di-NSO compounds in any sample was calculated by subtracting the experimentally observed percentage of NSO compounds from the calculated percentage since the following relationships hold:
% NSO compounds (calcd.)
=
% mono-NSO
+ 2 ( % di-NSO)
because the calculated percentage of KSO compounds counts all di-NSO molecules twice, and
% ;“\‘SOcompounds (exptl.)
=
% mono-NSO f yo di-NSO.
SHALE OIL COMPOSITION
Shale Oil. On the basis of the present studies, the composition of shale oil is as given in Table IX. With increasing molecular weight, the nonhydrocarbon content of shale oil increases and the hydrocarbon content decreases. The increase in the nonhydrocarbon content is due primarily t o the nit,rogen compounds. I n the five 10% fractions the increase in nitrogen compounds is almost exactly matched by the large decrease in iso-olefins plus cyclo-olefins and the small decrease in isoparaffins plus naphthenes; %-paraffins, n-olefins, and total aromatics are essentially constant. Shale oil contains 61% nonhydrocarbons, of which about 60y0 are nitrogen compounds, 10% sulfur compounds, and 3oY0oxygen compounds. Only 39% of shale oil is hydrocarbons. Isoolefins plus cyclo-olefins are predominant and constitute one third of the total hydrocarbons. The remaining two thirds is divided almost equally among the other types, Nonhydrocarbons. Little is known of the nature of the nonhydrocarbons. Ball (1) has shown t h a t nitrogen may be present as pyrroles, pyridines, and quinoline homologs; sulfur as thiophene homologs; and oxygen as phenols and fatty acids. Distribution of nonhydrocarbons according t o elemental and molecular types is given in Table X. The lower boiling half of 2640
shale oil contains 34% of the NSO compounds and, according t o t h e calculations, all of the mono-NSO compounds and 7% of the di-NSO compounds. There appears t o be a maximum in the concentration of di-NSO compounds in the 20 t o 30% fraction. Because the higher boiling half apparently contains only di-NSO compounds, there must be a distinct increase in the percentage of these materials as the molecular weight increases. It is less likely that the higher boiling half actually contains exclusively di-NSO compounds than that mono-, tri- , and tetra-NSO compounds are also present. In total shale oil, the nonhydrocarbons appear to be 30% mono-NSO compounds and 70% di-NSO compounds.
n -PARAFFINS I
0
I
1
I
I
&_
30 40 90 SHALE-OIL OlSTlLLATlOM FRACTION, WT. %
Figure 4.
20
10
LOO
Hydrocarbon Composition of Shale Oil Fractions
Hydrocarbons. The distribution by weight of the hydrocarbon classes in the shale oil fractions, exclusive of NSO compounds, is shown in Figure 4. I n the five 10% fractions, the normal paraffins double, the normal olefins and polycyclic aromatics increase slightly, the iso-olefins plus cyclo-olefins decrease by 40%, and the isoparafhs plus naphthenes decrease slightly. The 50 t o 100% fraction reveals an unexpected increase in isoparaffins plus naphthenes and decreases in the iso-olefins plus cyclo-olefins and polycyclic aromatics. I n total shale oil, only the monocyclic aromatics are constant.
Table IX.
Composition of Shale Oil Fraction, %
0-10
Nonhydrocarbons, wt. % S compounds S compounds 0 compounds
12 5 12
10-20 20-30 30-40 40-50
14 4 17
21 4 18
26 5
17
33 5 16
0-50 50-1000-100
21 5
51 7 22
36 6 19
-
-
-
-
-
-
__
-
Total 29 Hydrocarbons, wt. % n-Paraffins 8 Isoparaffins plus naphthenes 7 n-Olefins 10 Iso-olefins plus cyolo-olefins 31 Monocvclic aromatiis 6 Polycyclic aromatics 9
35
43
48
54
42
80
61
8
8
8
10
8
3
6
6 9
4
3
11
11
3 9
5 10
5 3
5 6
25
19
14
12
20
5
12
8
6
6
5
6
2
4
-
9
-
9
-
10
-
7
-
9
-
2
-
6
-
71
65
57
52
46
58
20
39
Total
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
16
Vol. 44, No. 11
PETROLEUM-SHALE OIL
_ I
CONCLUSIONS
The observation that the iso-olefin plus cyclo-olefin and the isoparaffin plus cycloparaffin contents of the five 10% fractions of shale oil decrease as the nitrogen contents increase leads to the conclusion t h a t in the degradation of kerogen during the retorting of oil shale, the yield of is0 plus cyclic compounds depends primarily upon the denitrogenation of nitrogen compounds, and t h a t this is less complete for compounds of hi her molecular weight. However, in the higher boiling half of sfale oil, an unexpected increase in t h e isoparaffins plus naphthenes and the decreases in iso-olefins plus cyclo-olefins and in polycyclic aromatics suggest t h a t the sequence of steps in t h e degradation of the nitrogen compounds originally formed from kerogen during retorting is a8 follows: nitrogen compounds t o isoparaffins $us naphthenes, t o iso-olefins plus cyclo-olefins, t o aromatics. hese results also suggest t h a t the polycyclic naphthenes of higher molecular weight do not dehydrogenate as readily as those of lower molecular weight.
Nonhydrocarbons in Shale Oil
Table X.
Elemental Type Distribution, Wt. % S 0
Fraction
N
0- 10 10- 20 20- 30
41 40
30- 40
40- 50 0- 50 50-100 0-100
49
54 61
50 64 59
17 11 9 10 9 12 9 10
42 49
42 36
30 38 27 31
Molecular Type Distribution, Wt. ’% Mono-NSO Di-NS0 3 97 94 6 79 21 19 81 87 13 13 87 0 100 30 70
The constancy of the hydrocarbon composition of the lower boiling half of shale oil leads to the conclusion t h a t kero en consists predominantly of homologs of a complex structure tfat upon degradation yield hydrocarbons having the same distribution of structural types. T h e trends i n molecular weight distribution of the hydrocarbons, the NSO compounds, and the total shale oil fractions suggest t h a t there are no hydrocarbons in the higher boiling portion of the 50 t o lOOyo fraction of shale oil. If the hydrocarbon molecular weight curve in Figure 3 is extrapolated to 490, which is the avera e molecular weight of the hydrocarbons found in the 50 t o lOO# fraction, this molecular weight corresponds t o a distillation point of 60% and all these hydrocarbons would be contained in a 50 to 70% fraction. Therefore, approximately
the highest boiling 30% of shale oil must contain essentially no hydrocarbons. ACKNOWLEDGMENT
The sample of shale oil studied was obtained through a cooperative agreement between Standard Oil Co. (Indiana) and the U. S. Bureau of Mines. The authors are grateful for discussions with A. P. Lien and for the infrared analyses and interpretation by P. J. Launer. LITERATURE CITED
Ball, Dineen, Smith, Bailey, and Van Meter, IND. ENG.CHEW, 41, 581 (1949).
Belser, U. S. Bur. Mines, Rept. Invest. 4769 (1951). Botkin, Chem. & Met. Eng., 24, 876 (1921). Ibid., 26, 398 (1922).
Clark, “Semimicro Quantitative Organic Analysis,” p. 37, New York, Academic Press, Inc., 1943. Dineen, Bailey, Smith, and Ball, Anal. Chem., 19, 992 (1947). Dineen, Thompson, Smith, and Ball, Ibid., 22, 871 (1950). Dinerstein and Klipp, Ibid., 21, 545 (1949). Dundas and Howe, U. S. Patent 1,469,028 (1923). Garrett and Smythe, J. Chsm. SOC.,81, 449 (1902). Ibid., 83, 763 (1903). Gavin, U. S. Bur. Mines, Bull. 210 (1922). Gray, J . SOC.Chem. Ind. (London), 21, 845 (1902). Guthrie, U. S. Bur. Mines, Bull. 415 (1938). Horne and Bauer, U. S. Bur. Mines, Rept. Invest. 2832 (1927). Kilpatrick, Prosen, Pitser, and Rossini, J . Research Natl. Bur. Standards, 36, 559 (1946). Lepper, “Official Methods of Analysis of the Association of Official Agricultural Chemists,” p. 742, Washington, D. C., Assoc. of Official Agricultural Chemists, 1950. McKee, “Shale Oil,” New York, The Chemical Catalog Co., Ino., 1925. Matteson, Anal. Chem., 22, 172 (1950). Mensies and Wright, J . Am. Chsm. SOC., 43, 2314 (1921). Nottes and Mapstone, J . Inst. Petroleum, 37, 259 (1951). Robinson, J . Chem. Soc., 127, 768 (1925). Robinson, Trans. Roy. SOC.Edinburgh, 28, 561 (1879). Ibid.. 29. 265 (18801. Schlenk,’Ann.,’ 565; 204 (1949). Smith, Smith, and Dineen, Anal. Chem., 22, 867 (1950). U. S. Bur. Mines, Rept. Invest. 4457 (1949). Ibid., 4471 (1951). Zimmerschied, Dinerstein, Weitkamp, and Marschner, IND. ENQ.CHEM.,42, 1300 (1950). REC~IVE for D review January 18, 1952.
High Temperature Shale
A C C E P T ~September D 10, 1952.
Oil
PRODUCTION AND UTILIZATlON F. E. BRANTLEY, R. J. COX, H. W. SOHNS, W. I. BARNET, AND W. I. R. MURPHY Petroleum and Oil-Shale Experiment Station, Bureau of Mines, Laramie, Wyo.
M
ANY retorts have been developed for producing a shale oil from which gasoline and other liquid fuels can be derived. These are operated in such a way t h a t the oil is not subjected t o a high temperature after it has been formed and released from the shale, thereby avoiding excessive cracking with consequent low liquid yields. I n the best of these retorts quantities of oil and gas almost equivalent t o the yields obtained from the same shale by the Fischer assay method of analysis ( 7 , 8) are produced. For example, when Green River shale is processed, about 66 weight % of the organic matter is converted. to oil, 12% to gas, and 22% t o a carbonaceous residue which re-
November 1952
mains on t h e spent shale. Quality of oil varies with retorting conditions, b u t generally high specific gravity, high pour point, and high sulfur and nitrogen contest are characteristic. T h e naphtha fraction may contain 50% or more olefinic and approximately 20% aromatic hydrocarbons, although little if any benzene or toluene, the most valuable aromatics in motor fuel and chemical manufacture, are found. Aromatic content, including nitrogen, oxygen, and sulfur compounds, increases with boiling point of the fraction, reaching 65% in the heavy gas oil, whereas percentages of paraffinic, cycloparaffinic, and olefinic hydrocarbons decrease. The major constituents of the gas produced
INDUSTRIAL AND ENGINEERING
CHEMISTRY
2641
