1600
INDUSTRIAL AND ENGINEERING CHEMISTRY ACKNOWLEDGMEKT
This investigation was carried out under the sponsorship of the Standard Oil Co. (Indiana), and the results are published with their permission. The authors are indebt,ed t o F. IT. Fink of the Battelle staff for his assistance in connection 73-ith the korrosion problem. LITERATURE CITED
(1) Bearse, A. E., and hlorin, R. D., U. 8.Patents 2,414,999 and 2,415,000 (Jan. 28, 1947). (2) Bouchardat and Lafont, Compt. r e n d . , 113, 551 (1891). (3) Dorris, T. B.. and Sowa, F. J . , J . Am. Chem. S o c . , 60, 358 (1938). (4) Dorris, T. B . , Sowa, F. tJ., and Kieurrland, J . d.,I b i d . , 56, 2689 (1934). (5) Dreyfus, H., Brit. Patent 485,108 (May 16, 1938). (6) Dreyfus, H., U. 5 . Patent 2,083,693 (June 15, 1937). (7) Ellis, C., “The Chemistry of Petroleum Derivatives,” Tal. I, p p ~348-58, 372, Nevi York, Chemical Catalog Co., 1934. (8) Ellis. C., U . S. Patents 1,365,050 and 1,365,052 (,Jan. 11, 1921). (9) E n n s , T. W,, Edlund, K. R., and Taylor, M. D., ISD. ESG. CHEM.,30, 55 (1938). (10) Frolich, P . K . , and Young, P . L., 1;. 6 . Patent 1,877,291 (Sept. 13, 1933).
Vol. 43, No. 7
(11) Grosse, A. V., and Linn, C. R.. J . Org. C h e m . , 3, 26 (1938). (12) Hilcken, V., U. S. Patent 1,902,364 (March 21, 1933).
(13) Hofmann, F., and Wulff, C., I h i d . , 1,898,627 (Feb. 21, 1933). (14)Kondakow. J . prakt. Chem., 48,479 (1893). (15) Kroeger, J. W., Sowa, F. J., and Nieuviland A c u d . Sci., 46,115 (1937). (16) Larson, A. T., U. S.Patent 2,093,695 (Sept. 21, 1937). (17) Lazier, W. A . , I b i d . , 2,174,895 (Oct. 3, 1940). (18) Lien, A. E’., Ihid., 2,399,126 (April 23. 1946). (19) Loder, D. J.. I b i d . , 2,253,525 (Aug. 26, 1940). (20) I b i d . , 2,265,946 (Dec. 9, 1941). (21) Meerwein, H., Be?., 66B,411-14 (1933). (22) Nieuwland, J. A., and Sow-a, F. .J.. U. 8. Patent 2,192,015 (Feb. 27, 1940). (23) Schneider, H . G., I b i d . , 2,065,540 (Dec. 29, 1936). (24) Spring, F. S.,Ann. R e p t s . o n Progress C k e m . , 39, 129 (1942). (25) Stanley, H. XI., and Youell. J. E., Rrit. Patent 641,056 (SOT-. 11, 1941). (26) Strange. E. H., and Kane, T., C . 8 . Patent 2,014,850 ( S e p t . 17, 1935). (27) Suida, H., E. S.Patent 1,536,135 (Doc. 15, 1932). (28) Woolcock, J. W., Brit. Patent 334,228 (April 30, 1929). RECEIVED October 31, 1Q.50. Presented hefore the Division of Potro!cuni Chemistry at the 117th Meeting of rhe . i \ r x a i c . ~ s C:HW\IICAI. S o c r m y , Houston, Tex.
GR-S Aging in Solution GIFFIN D. JOSES ~ N RALPH D E. FRIEDRICN Dow Chemical Co., Midland, Mich. T h e degradation and oxygen absorption of solutions of GR-S have been studied in an effort to explain the failure of antioxidants under ultraviolet light. It has been found that the photodegradation of GR-S in solution occurs only if oxygen is present. Two types of peroxide can be formed, a stable peroxide at elevated temperatures and a labile one, which may be polymeric, at room temperature. These peroxides can be differentiated by the use of aliphatic amines which cleave the latter an’d produce thereby a viscosity decrease. Aliphatic amines are shown to be good antioxidants for GR-S at 90” C. but poor antioxidants under ultraviolet light. The work has led to a better understanding of the viscosity changes which occur in GR-S solutions on standing and the mechanism of light stabilizer action.
I
T HAS long been recognized (1) that the viscosity changes in
rubber solutions under light depend in a complicated way on oxygen and on the effect of various promoters, retarders, and modifiers on the autoxidation and polymerization reactions involved. Stevens (10) has shown that a number of substances drawn from both pro- and antioxidant types have the effect of promoting photogelation when used in low concentration and retarding photogelation when used in high concentration. h partial answer to this seeming paradox appears to be in the discoveries connected with redox systems of polymerization. The use of aliphatic amines, for example, to initiate cold polymerization (9, 11) gives confirmation of the view (7) that the induced decomposition of a peroxide gives a free radical as a by-product. This free radical may be the cause of subsequent polymerization or autoxidation steps before it, too, interacts with the amine and presumably dehydrogenates the amine to form a nitrogen radical. This nitrogen radical attacks the peroxide and a new cycle of the chain reaction ensues. If an amine, such as n-butylamine is added portionwise to a solution of peroxidic GR-S, there is a nearly instantaneous decrease in viscosity. Equilibrium is reached within a few minuteq,
and a further decrease results aiter another portion OS aiiiiiie is added until an end point is reached; after this the further atidit,iou of amine produces little effect,. Figure 1 shdws amine titmtion curves for solutions (in ethylbenzene) of aged and unaged (;It-S. EXPERIMENTAL
9 107, solution of soluble GR-S (X274) in ethylbenzenc was prepared and precipitated in ethanol by means of a \\-&ring Blendor. The precipitated GR-S was then redissolved without drying, and the residual alcohol was distilled off under reducwl pressure. The operations were conducted under nitrogen containing less than O.OlYc oxygen and at room temperature or below. The final concentration of GR-S in ethylbenzene was -t.5cc. The intrinsic viscosity of thc GR-S in ethylbenzene was 2.1.
The GR-S solutions were nged by exposure to air a t various temperatures in diffuse daylight. The duration of exposure is indicated iri the key to Figure 5 and the peroxide content in Talde I10
, TITRATION OF 4 . 5 % G R - S SOLUTION
IN ETHYLBENZENE
50 0
WITH
ti-BUTYL
I
I
I
I
I
2
4
6
8
10
MILLIMOLES
n - B U T Y L AMINE
Figure 1
I. During the aging of the GR-S solutions some viscosity increase occurred (Figure 2). Viscosity Measurements. Two viscometers were used for the photodegradation measurements reported here (Figures 3 and 4). The data of Figure 6 were measured on the tilting viscdmeter shown in Figure 3, and those of Figure 7 were obtained with the 140
1601
INDUSTRIAL AND ENGINEERING CHEMISTRY
July 1951
I
I
Figure 6 in the tilting viscometer which was of a cell type and was closed. Some evaporation occurred during measurements in theopen viscometer. This amounted to 0.36% per hour and the GR-S concentration rose from 4.5 to 4.9% in a 24-hour period. No correction was made for the effect on viscosity of this change. The photogelling of natural ,rubber solutions in carbon tetrachloride has been studied by Buckingham and Planer ( 2 ) who found that the gel time varied considerably with rubber concentration over the narrow range studied (0.5 to 0.65%). I n the GR-S solutions studied by the authors there was no change in viscosity, in a closed cell containing oxygen, under diffuse daylight for 20 hours at 25" C., and the decrease in viscosity produced by the addition of n-butylamine was less than that of the unexposed sample. On sweeping the surface of the solution with air there was observed a viscosity increase. This was true also when the viscometer was illuminated by ultraviolet light. In this instance when the air stream was replaced by a nitrogen stream the viscosity continued to rise. This seemed to indicate that superficial evaporation favors cross linking.
40 0
8
16
24
HOURS
Figure 2
3%
40
48
TABLEI. PEROXIDE CONTENT OF AGED GR-S SOLUTIONS OF FIGURE 5 Peroxide Molarity" X 10-4
Sample
Unaged 1.3 apparatus shown in Figure 4. An AH-4 lamp was used in both A 1.9 B 3.8 cases. The cell of the tilting viscometer had glass windows l l / z C 4.6 mm. thick and 1.4 inches in diameter, with 1.6 inches length beD 9.4 tween windows; the windows were slightly convex. The visa Ferrous sulfate analysis. cometer was joined a t the l/2-inch neck opening and emptied through a 1.6-mm. capillary, dipping into the cell. An ultraviolet filter (Wratten No. 2) placed in the light beam with the Formation of Peroxide. Figure 6 shows the viscosity deapparatus of Figure 4 showed that degradation proceeded fairly crease obtained by the addition of amine to several GR-S solutions which had been aged in different ways. With the exceprapidly with visible light alone. In neither apparatus was agitation sufficient to give uniform absorption throughout the mass, tion of the solutions already degraded by severe aging, the viscosity decreased to approximately that of the unaged solution. but a t the time of measuring the viscosity the solution was mixed until consistent readings were obtained. (Three consecutive This may be a coincidence, or it may indicate that photogelling involves the formation of an intermolecular peroxide structure readings did not vary over 0.2 cp.) The light source was placed 2.2 inches from the window of the tilting viscometer, and the temthat is susceptible to induced scission. The viscosity of the GR-S aged a t 90" C. was relatively unafperature just inside the window was 32" C. With this cell there fected by the addition of amine. Inasmuch as this solution was was observed no initial viscosity rise. However, a little gel forhighly peroxidic (Table I), this fact suggests that two different mation on the window was observed. In the test of the effect of oxygen on photoscission the cell and light were immersed in types of peroxide are formed in GR-S. Previous workers have water to eliminate the possibility of thermally initiated autoxidanoted the apparent presence of both a labile and a stable peroxide tion. The tests reported in Figure 7 were made in the presence (4). of air. The first indication obtained in this work of the presence of two Viscosity measurements were made by loading the capillary types of peroxide in GR-S was in oxygen uptake rate measureand measuring the fall time. The loading was accomplished by ments made at 90" C. in the dark. The rate (Figure 8) was suction in the viscometer of Figure 4 and by tilting in that of highest at the beginning of exposure and it seemed that the samFigure 3. The visple, a GR-S solution cometers were caliaged at room temperature, contained a brated with a n s o l u t i o n of known higher peroxide conviscosity. centration than the VISCOSITYRISE steady state value ANOMALY. Comparitolerated a t 90' C . son of curve B of H o w e v e r , peroxide Figure 6 and curve A analysis showed that of Figure 7 shows a t the end of the run that there is an initial the peroxide content viscosity rise in the was higher than st OB AH4 one instance but not Cmling the start. The possiHS vapor in the other. This He16 st Lamp bility of a progressive discrepancy was development of a ret,raced to a n interesttarding impurity was ing phenomenon. d i s c a r d e d because Curve A of Figure 7 adding a portion of a was obtained in the d e g r a d e d sample to viscometer of Figure a fresh sample did 4, and curve B of n o t have a proFigure 3. Tilting Viscometer for Photodegradation Measurements ~~~~~~~
8
~~~~~
~~~
~~
1602
INDUSTRIAL AND ENGINEERING CHEMISTRY
nounced effect. Moreover, with a n unaged sample of GR-S solution the rate continuously rose until a steady rate was obtained. The absence of the initial high rate is no doubt due to the absence of the unstable peroxide,
.
-
120 E F F E C T OF O X Y G E N 3h F + O T G D E G R b D A T I O h
OF G R . 5
S A N P L E : 4 . 5 % G 9 - 5 I N E T H I I - B E h Z E Y E SOLUTIOlr COYDITI0hS:GE
A H 4 LAMP
.%T
3".O ' R E X
GEL\.
E
V SCOSIMETER
LAMP H O J S I V G A-
,/ \
L J
i
Vol. 43, No. 9
0
5
'\c
B
------
CELL
E
VISCOSHETER
1
I
'\\
-QLIBBER
SAMPLE SEALED INTO T O T l i L Y ENCLOSE3 C E L L AND V l S C C S I M E T E 9 UNDER VAC'JUM SA'4P-E S E A L E D IN'C U N D E R AN".
a
STOPPER
'..-=.:
CbPiLLARY H O J R S EXPOSURE
Figure 6
25'C
CONSTANT T E M P E RATURE
BATH
,/
hours. It was then titrated with 0.02 N titanous chloride. The titanous chloride was standardized with ferric chloride and potassium thiocyanate prior to each determination.
k5OC--/
GR-S IN ETHYLBEYZENE
Figure 4.
Alternate Viscometer
TABLE 11. PEROXIDE ANALYSIS OF OXYGEN UPTAKEREACTION OXYGEN UPTAKEMEASUREMENTS. To determine the rate of oxygen absorption, air w a s pumped a t a constant rate (95 cc. per hour) using a dual syringe pump (6) through 10 ml. (13 em.) of GR-S solutions a t 90' C. The oxygen concentration of the downstream gas, after passing through a condenser was determined by means of a Beckman (magnetic) oxygen analyzer (8); the difference in oxygen concentration of the exit gas and air permitted calculation of the rate of oxygen absorption by the rubber. A mole of GR-S was assumed to be 71 grams which is the weight required to provide a gram molecular weight of a polymerized butadiene unit. This figure is based on the assum tion that the soluble GR-S contains 76% butadiene. The stanfard deviation of the oxygen uptake mensuremerits was 0.13 X mole per mole of GR-S per hour.
ION
OF A G E D G i i - S
SAMPLES:4.5%
SOLUTIONS
GR-S I N ETHYLBENZENE
2 5 C C VOLUME
25'C
c K-x
BUBBLED WiTH AIR 2 5 DAYS AT
MIXTURE(FIQURE8 )
Peroxide Molarity X 10-4 Curve L4 Curve B 46 1.3
Start End
49
5111
The introduction into the unaged GR-S of a foreign peroxide shown to be subject to induced decomposition (tetralin hydroperoxide) did not give rise t o pronounced scission on treatment with amine (Figure 9). This shows that the mechanism of induced scission is more specific than a n incidental result of catalyzed autoxidation and t h a t probably peroxide links on or between polymer chains undergo induced scission. Acetic acid also gives rise to induced scission and it is probable that phenols, aromatic amines, and other types of compounds which promote the decomposition of peroxides will produce induced scission also. Light Stabilization. It seemed possible that light, like a chemical inducing agent, might seek out the labile peroxide and bring about scission by this mechanism. Figure 6 showed that exposure to oxygen was a necessary condition for the production of an alteration of viscosity under illumination, and Figure 7 showed that the higher the peroxide content of the sample, the earlier scission predominated over the tendency of viscosity to increase. If true, this effect might constitute further evidence that the scission of peroxide linkages is the explanation and not merely the 140
120
n
E5
IO0
t
w 80 0 I
QO 0
.Z
I .4
MILLIMOLES
I .6
I .E
i 1.0
n - BUTYLAMINE
z 2-
5
60
0 ul
Figure 5
PEROXIDE ANALYSIS. Peroxide analysis (Table 11) was made by a titanous chloride procedure. A solution of 0.1 N ferrous
ammonium sulfate (10 ml.), 1: 1 sulfuric acid ( 2 ml.), and 10% potassium thiocyanate ( 5 ml.) in methanol (100 ml.) was allowed to stand in the dark for 15 minutes, and was then decolorized with 0.02 N titanous chloride. The sample (25 ml.) was added and the solution shaken and allowed to stand in the dark for 2
g
40
20
n 0
I
I
I
I
I
I
I
I
I
8
16
24
32
40
48
56
64
72
HOURS E X P O S U R E
Figure 7
80
INDUSTRIAL A N D E N G I N E E R I N G CHEMISTRY
July 1951
T H E R M A L AUTO O X I D A T I O N
OF 0 R - S 90'C
SOLUTION TEMPERATURE,
"
A
1603
CCI 4 5 % G R - S IN ETHYLBENZENE BUBBLE0 WITH AIR 2 5 DPYS
AT 25'C B C--+-+
4 5 % GR-S I N ETHYLBENZENE
NOT EXPOSED TO A I R
w 0
0
gj
m IO
100
m
I
U
>. 8 0
8 0 "
t
/'
J'
LL
0
0 v)
4
In
60
6 ' 0
> 0
4
8
12
20
16
24
28
32
36
g In
0
0
40
EXPOSURE T I M E I N HOURS
40
4
20
2
Figure 8
trigger-mechanism of viscosity decrease. The general similarity of the curves of Figure 7 after the initial phase indicates that the effect of light is not very dependent on the type of peroxide initially present. It was not established whether in these samples the peroxide was the main absorber of the active wave length. (Even visible wave lengths produce a fairly rapid rate of degradation.) Most light stabilization appears to be achieved by the use of a substance which acts to absorb or reflect the actinic light in a layer near the surface and thus protect the remainder of the mass. It does seem possible, however, that, if it could be found, a compound which is a good antioxidant for GR-5 under illumination by ultraviolet light at moderate temperatures would also act as a stabilizer of the viscosity and perhaps of the physical properties in the absence of a solvent.
?!
TITRATION OF NON- POLYMERIC PEROXIDE SAMPLEi25C.C, OF 4 . 5 % G R - S IN ETHVLBENZENE PLUS 1.0 C.C. TETRALIN HYDROPEROXIDE
70
0
a
L
601
5 >
40
1 0
I .2
I
I
I
I
I
I
.4
.6
.8
1.0
1.2
1.4
MILLIMOLES n-BUTYLAMINE
Figure 9
A m i n e s as Stabilizers. The aliphatic amines, although not of themselves strong ultraviolet absorbers, produced a yellow product when added to the peroxidic GR-S solution. It is therefore possible t h a t they produce a photosensitizer effect. This may be the reason why they do not act as light stabilizers in GR-S, although they do in polystyrene (6). Table I11 indicates t h a t the rate of photoscission after the initial phase was not much affected by the presence of the amines. The GR-S solution used for these light stabilizer tests was unaged but the measurements were carried out in the presence of air. The polyamines produced a rise in viscosity; in fact with polyethylenimine some gel particles were observed. The polyethylenimine was dispersed in the ethylbenzene solution by adding the polyethylenimine in alcohol solution. It was expected that the amines would have some beneficial effect as light stabilizers because they did function as antioxidants at 90' C . in the dark. These tests (Table 111) were made with an aged 5% GR-S solution in ethylbenzene and 0.01 M peroxide concentration in the solution. Subsequently .it was observed
0
0
0
20
40
60
SO
I00
120
140
HOURS E X P O S U R E
F i g u r e 10. Autoxidation of GH-S under Ultraviolet L i g h t Sample: 30 ml. of 4.5% solution of GR-S (X-274) in ethylbenzene, Temperature,25' C.; ultraviolet light, AH4lamp at 3 inches, glass cell; air continuously recirculated through sample at 60 ml./min. A Viscosity of ample E Viscosity of sample containing n-butylamine (0.03 M) C Oxygen absorption of sample D Oxygen absorption of sample containing n-butylamine (0.03 M )
---
that the amines were not very effective in controlling oxygen uptake under photoexcitation at 25" (Figure 10). In these results the amount of oxygen absorbed at the time of maximum viscosity was of the general order of magnitude of one per chain. tert-BuTYLAMINE. tertButylamine was prepared by hydrolysin 8 grams of tert-butyl urea (8)in 75 ml. of diethylene glycol, 10 mf. of water, and an exceas of solid sodium hydroxide. This mixture was heated in a steam bath for 3 days; 3.7 grams of tertbutylamine were distilled off w it was formed (boiling point, 44" (3,). A 15ml. sam le wm prepared in this way, refluxed over caustic, and redistiied t o remove ammonia. AMINE TITRATION PROCEDURE. Amine titrations were performed in the viscometer of Figure 4. The amine was added in benzene solution (0.5 M ) to a 2.54. sample of GR-S solution.
It was possible to show in a general way t h a t the amine persists in the solution while photoscission proceeds. A solution of unaged GR-S containing 0.3 M n-butylamine was exposed in the viscometer for 49 hours. The visoosity dropped from 60 to 25 cp. It was then heated at 90" C., and the oxygen absorption rate was less than 10-8 moles per hour. An unstabilized sample had a rate of 2.104 moles per mole per hour. There was incidentally, a residual odor of amine after the light-aging period. There was also sufficient amine left to cause scission in an unexposed but aged G R S solution on adding some of the exposed solution. Several hypotheses may be formulated as to why the amines are less effective as antioxidants under photoexcitation than under thermal excitation. 1. The amine oxidation products act as photosensitizers. The temperature coefficient of the initial attack of RO. and R00.radicals on the antioxidant is considerably greater than that of attack on the GR-S; as a corollary, the steric factor is considerably less. 3. The peroxide formed under photoexcitation undergoes induced decomposition to yield free radicals faster than the amine (in its antioxidant role) can sweep them up. 4. The free radieals generated by photoexcitation are so energy-rich t h a t reaction requires no activation energy; hence i t is not selective for the antioxidant. 2.
INDUSTRIAL AND ENGINEERING CHEMISTRY
1604
TABLE I11 Amine (0.03 X ) None (blank) C n-Butylamine Dibutylamine Tributylamme tsrt-Butvlamine Tetraethylenepentamine Polyethylenimine
Initial Decrease in Viscosity, CP.
scission %/Hou;
....
20
Tetramethyldiaminisoqropanol
Ka,,Photo-
..
0.11 0.18 0.08
10 7
Steady State 0 2 Absorption h
oxidant which is not a photosensitizer should function as a light stabilizer. The cause of the failure of the antioxidants studied to function as light stabilizers of GR-S is shown t o be related to the occurrence of a very low rate of oxidation of the GR-S despite the presence of the antioxidant. The cause of the viscosity rise in GR-S solutions on autoxidation appears t o be the formation of an intermolecular peroxide bond which is subject to induced scission.
0.1 0 0
0 15 0.1
Ethylphenylethanolamine 0 First order rate constant. b Oxygen absorption expressed in 10-2 moles Odbase mole GR-S/hour after first 15 hours. C Original viscosity, 51 cp.
The second hypothesis ia subject to test, for if it were true the amines xould not function as antioxidants a t room temperature in the dark. The third hypothesis might be called paint-drier behavior for such is the probable mechanism of the action of oilsoluble copper, cobalt, and iron salts in promoting autoxidation of drying oils. It would have to be assumed in addition t h a t a t high temperatures either the induced decomposition proceeds less efficiently with respect to free radical generation or that the labile peroxide is not even transiently formed. CONCLUSIOK S
Photoscission of GR-S in ethylbenzene solution has been shown t o require the interaction of oxygen. I n principle then an anti-
Vol. 43, No. 7
LITERATURE CITED
(1) Blake, 3. T., and Bruce, P. L., ISD.ENG.CHEM., 33,1198 (1941). (2) Buclungham, R., and Planer 0. V., T r a n s . Inst. Rubber Ind., 21, 175 (1945). (3) Harvey, M. T., and Caplan, S , U. S. Patent 2,247,495 (Julz 1, 1941). (4) Lacau, J., and Magat, hl., Dzscussions Faraday SOC.,1947, No 2, p. 388. (5) Matheson, L. h.,and Boyer, R. F., U. S. Patent 2,287,188 (July 23, 1942). (6) Modern Metalcraft, Midland, Mich. (7) Sozaki, K., and Bartlett, P. D., J . Am. Chem. SOC.,6 8 , 1686 (1946). (8) Pauling, L., Wood, R. E., and Sturdivant, J. H., Science, 103, 388 (1946); J . Am. C h e n . Soc., 68, 795 (1946). (9) Spolsky, R., and Williams, H L., IXD.ENG.CHEM.,42, 1847 (1950). (10) Stevens, H. P., J . Soc. Chem. I d . , 64, 312 (1945). (11) Whitby, G. S.,Wellman, N., Flouts, V. W., and Stephens, H. I)., ISD. ENG.CHEM.,4 2 , 4 4 5 , 4 5 2 (1950). RECEIVED September 5 , 1950. Presented before the Division of High Polymer Chemistry, 118th Meeting of the AMERICANCHEMICAL SOCIETY Chicago, Ill.
Gum Formation in Shale-Oil Naphtha G. U. DINNEEN AND JV. D. BICICEL' U . S. Bureau of Mines, Laramie, Wyo. T w o distinguishing characteristics of untreated shaleoil distillates are their dark color and large gum content. The work reported in this paper was undertaken in an effort to establish the factors responsible for this apparent instability. It was found that the gums from shale-oil naphthas contain about 89'0 nitrogen. This is the principal difference in elemental composition between them and gums obtained from petroleum naphthas. The nitrogen appearing in the gum comes both from pyrrole and pyridine-type
compounds found in t-e naphtha. Oxidation is far more important than light or heat in the formation of gum i n shale-oil naphtha. The rate of gum formation is very rapid when the naphtha is first exposed to accelerated oxidizing conditions, but decreases sharply after about 1 hour. Information on the factors and compounds responsible for gum formation in shale-oil naphtha is necessary as a basis for the development of handling and refining methods for shale-oil products.
T
oxidation, probably peroxides (5, 7 , 10, 11, 15-15), catalyze the oxidation of normally less reactive hydrocarbons to increase greatly the rate of gum formation. Most of the oxidation products are soluble in the naphtha but decompose during evaporation to give gum composed largely of acidic material ( I S ) . Shale-oil naphthas contain larger quantities of nitrogen and sulfur compounds than most cracked petroleum naphthas. As certain types of these compounds are reactive, it appears probable that factors in addition t o those responsible for the gum in cracked petroleum distillates may be involved in the formation of gum in shale-oil naphtha. Mapstone (9) has shown t h a t addition of nonpurified tar bases to a refined naphtha increased the gu111 content. He attributed this increase to the presence of pyrroles in the tar bases. A brief investigation ( l a ) by this laboratory of gum formation in a thermally cracked shale-oil naphtha indicated
HE color of shale-oil naphtha when first distilled ranges from
very pale yellow to amber. However, if exposed to ordinarv room conditions, it will become dark purple and start to deposit gum on the sides of the container within a very short time. Large quantities of existent gum are produced after a fe\T days or even hours of such storage. Consequently, knowledge of the major factors involved in this gum formation is of primary importance in the development of handling and refining procedures. The formation of gum in cracked petroleum naphthas has been studied extensively and it seems to be generally agreed that oxidation of certain types of reactive, unsaturated hydrocarbons ( 6 , 6 , 8 ) is the primary cause with sulfur compounds possibly being a contributing factor in some instances. The products of the initial 1 Present address, American Smelting and Refining Co., Grand Junction, Cola.
