I r D USTRIAL AiYD ENGIXEERl'iVG CHEMISTRY
October 15, 1930
Effect of Time of Bleaching
Bleaching tests were made according to the official method but varying the time from 1 minute to 72 minutes. The official time was 5 minutes. The results are shown in Table V.
377
With efficient stirring the bleaching reached its height in the first minute. This means that the bleaching action is very fast. Bleaching for more than 1 hour gave no added effect. Literature Cited
T a b l e V-Effect
of Variation of T i m e of Bleaching Test
RED
I
Min.
TOTAL
ing
nosity
43 44 45 44 44 45 44 45 45 44 44
4.3 4.4 4.5 4 4 4.4 4 5 4 4 4 5 4 5 4 4 4 4
70
7c
85 84 85 84 85 85 85 84 85 84 85
60.35 59.64 60.35 59.64 60 35 60.35 60.35 69.64 60.35 59.64 60.35
L = 53.97 per cent
70
70 83.65 83 04 83 85 83.04 83.75 83 85 83 75 83 14 83 85 83 04 83 75
%' 64.52 63.19
ti:!: 64.74 64.95
i:,:: 64.95 :::;!
(1) American Oil Chemists' Society, Official Methods, p. 17 (1919' (2) Bailey, J . Ozl Faf I n d . , 2, No. 1, 8 (1925). (3) Bailey and Allen, Cotton Ozl Press, 7, Xo. 8, 36 (1923). (4) Benedict, J . Ozl Fat I n d . , 2, KO. 2, 62 (1925). ( 5 ) Davidsohn, Seifcnsieder-Zfg., 60, 648 (1923). (6) Hardy, J. Oil F a f I n d . , 6, Xo. 9, 31 (1929). ( 7 ) Keuffelt, Ibid., 2, KO. 1, 14 (1925). (8) Kress and McNaughton, J. IND. ENG.CHEM.,8, 711 (1916). (9) Maynard and Mallory, Chem. Met. Eng., 26, 1074 (1922). (10) Parsons, J . Am. Chem. Soc., 29, 598 (1907). (11) Parsons, Bur. Mines, Bull. 71, 23 (1913). (12) Porter, U. S. Geol. Survey, Bull. 316, 268 (1907). (13) Priest, J . Oil Fat I n d . , 6, KO.9, 27 (1929). (14) Smalley, Cofton Oil Press, 2, Xo. 3, 42 (1918). (15) Wesson, Ibid., 2, No. 3, 52 (1918). (16) Wesson, Trans. A m . Ins!. Chem. E w , 3, 327 (1910).
Early Stages of Oxidation in Rubber' A Quantitative Application of the Pyrrole Test J. W. Temple, Sidney M. Cadwell, and Morris W. Mead, Jr. UNITED STATES RCBBERCOMPANY, DEVELOPMENT DEPARTMEXI', PASSAIC, N. J.
The pine splint-hydrochloric acid test for oxidation holding a pine splint in the products of rubber has been developed in a quantitative vapors. They obtained posithe simple products of way. The technic and special apparatus are fully tive tests with both raw and its ultimate oxidation described. cured rubber which had been many stages intervene. For Data from application of the test to a wide range of oxidized. I t has been found the practical rubber techcommercial stocks are given. Its use in following early possible, through a number nologist chief interest in this stages of oxidation as they occur i n accelerated aging of refinements, to increase oxidative process is centered of vulcanized rubber is described, and attention is greatly the sensitivity of the in that portion which has called to an apparent difference between oxygen aging reaction, and to use the inoccurred up to the time when and heat aging. Evidence is cited that the substance tensity of color developed as the rubber has so c h a n g e d which gives the test, presumably levulinic aldehyde, a comparative measure of that it has lost its value as is formed independently of, and previous to, the actual the amount of this particurubber. physical deterioration of the rubber. lar oxidation product in a This paper describes a semisample. The sensitivity is quantitative application of the well-known pine splint-hydrochloric acid test for pyrrole such that it is usually possible to get positive tests before derivatives which the authors have found useful in following there is any evidence of physical deterioration as shown by the early stages of oxidation in rubber, particularly in vul- tensile measurements. canized rubber. It consists essentially of digesting the aged Testing Technic rubber with fused ammonium acetate, distilling with steam, and treating an ether extract of the distillate with an alcoFigure 1 shows an apparatus which was developed for use holic pinewood extract to which has been added hydrogen chloride. A red color develops, which is compared with a in the test. LL'ith several of these in use simultaneously three or four analyses can be finished in from 20 to 30 set of permanent standards. Several different observers have noted that water will minutes. The rubber sample is cut into thin slices and accurately extract from oxidized rubber a substance which can be condensed with ammonium acetate, and thereupon gives the weighed. One gram is suitable for a stock only slightly pyrrole color reaction with a pine splint. Bruni and Pe- aged, with smaller samples in proportion for those which lizzola (.e) credit Gorter with being the first to record this are further advanced in oxidation. The sample is placed fact, and cite several others who verified it. The active in the short-neck flask and from 2 to 3 grams of crystalline substance is generally assumed to be lewlinic aldehyde, ammonium acetate are added. After the delivery tube has since the y-ketoaldehydes give the reaction, and since been placed in position, the flask is gently heated with a small Harries (8)and Whitby (6) have identified it among products direct flame until the acetate is completely fused, using care from rubber ozonide and oxidized rubber, respectively. t o avoid charring the rubber. Gentle heating is continued Bruni and Pelizzola ( 2 ) simplified the test by heating for from 3 to 5 minutes in such a way that the level of conthe rubber directly with fused ammonium acetate and densing vapors on the sides of the flask rises barely to the ground-glass fitting. After the flask has cooled, 8 to 10 cc. Presented before the Division of Rubber 1 Receired April 15, 1930. of water are added, and this is then distilled over into the Chemistry at the 79th Meeting of the American Chemical Society, Atlanta, long-stemmed flask as a receiver. Most of the ammonium Ga., April 7 t o 11, 1930.
ETWEES rubber and
B
ANALYTICAL EDITION
378
acetate should remain in the distilling flask. The distillate in the receiver is diluted with water until it is about onehalf way up the neck (it is convenient to make a mark on the neck of the flask indicating some definite volume). Two cubic centimeters of ether are added from a pipet. The flask is well shaken and two layers are allowed to form. The shaking and settling must be repeated several times in order t o complete the extraction. One-half cubic centimeter each of a!coholic pine-wood extract, alcoholic hydrogen chloride, and the above ethereal layer are then mixed in the order named in a small test tube. After standing from 5 to 10 minutes the mixture is compared for color with the previously made standards. Values are calculated for 1 gram of sample.
1-01. 2. so. 4
member was a very faint pink. Standards thus prepared, and sealed in glass, have not changed appreciably in the course of a year's standing. Cork- or rubber-stoppered tubes are to be avoided, since the substances in the stopper tend to react with the dye. Application to Typical Rubber Stocks
In Table I are listed some results with typical rubber stocks of various types and histories. It will be seen that there is an excellent qualitative correspondence between the pyrrole test and the amount of deterioration. The effect of antioxidant is strongly brought out both by the condition of the stock and by the magnitude of the pyrrole test. In the used tire it is of interest that, even in the interior of the carcass as well as in the tread, there was definite evidence of oxidation. The acetone extracts of the carcass and tread were 8.2 and 13.8 per cent, respectively. of Deterioration of Typical Rubber Stocks with Pyrrole T e s t ANTIPYnnoLn TYPEOF STOCK AGINGHISTORY O X I D A N T CONDITION TEST Thread 5 years shelf 90 Weak 1 8 Yes Verv strone 0 2 White shoe upper 4 months sun 5 years shelf Xo Weak, s t i 5 2 0 Yes Strong, pliable 0 . 4 Red shoe upper 5 years shelf So Strong 0 7 Yes Very strong 0 0 Inner tube 2 years weather So Weak 2 0 Yes Very strong 1 2 Carcass compound 4 months neather 5 years shelf h'o Very weak 2 1 Yes Strong 0 2 service, Used tire carcass 31/2 years 18,000 miles NO 0.7' service, 3112 years Used tire tread 18,000 miles NO 0.8 On a rubber basis this figure 0 Tested without separation of fabric. should be at least twice as large. Table I-Comparison
+
+
I
// ''0~7J/&6%.447€&
Figure 1-Apparatus
for Pyrrole T e s t
The pine-wood extract is made by extracting clean white pine shavings in a Soxhlet apparatus for 16 hours. New shavings should be substituted about three times during the operation. The concentration of pine-wood extract seems to be unimportant provided an excess is present in the final determination. However, it has been found that water in the alcohol tends to decrease the sensitivity of the test and, although ordinary 95 per cent alcohol can be used successfully, better results are obtained if the alcohol is practically anhydrous and solutions are protected from moisture of the air. The alcoholic hydrogen chloride is prepared by running dry hydrogen chloride into alcohol of the same composition as is used for the pine-wood extract, until it fumes strongly a t the surface when exposed to the air. The two solutions must be kept separate until the test is made. The accuracy of this method is not sufficiently great to demand the use of a colorimeter. For comparing colors small test tubes of about 1 cni. inside diameter have been used. Care was first taken to eliminate those of too large or too small diameter. For standards of comparison some difficulties arose. It was originally intended to run through the test with a piece of badly aged rubber and use the resulting solution and suitable dilutions of it as primary standards. The instability of the color, however, soon showed the impossibility of this, and the ultimate scheme was the use of watersoluble dyes in dilute buffered solutions of pH 7.0, and glasssealed in the small test tubes mentioned above. Basic fuchsin, modified by traces of malachite green and auramine, gave a close approximation of the color actually obtained with aged rubber. A solution of this was given an arbitrary value of 10, and it was diluted down through a graded series to a lowest member with a value of 0.1. The color of this last
Effect of Kind of Aging on Pyrrole Test
As a routine method for measuring deterioration, however, the pyrrole test can have only a limited application, since it will presently be shown that its relationship to physical properties is largely dependent upon aging conditions, which in natural aging may vary widely. The main use of the test has been in trying to follow the actual process of oxidation, and the results have been most interesting and suggestive. Table I1 gives data obtained with a simple rubber-sulfur (100 to 7 by weight) compound, which was calendered to 0.030-inch gage, and cured in open steam with a 30-minute rise and 130 minutes a t 142" C. Three methods of aging were used-in oxygen a t 21.1 kg. per sq. cm. and 60' C., in air a t 70" C. (Geer oven), and also in air a t 100" C. Table 11-Comparison of Physical Properties of a Stock a n d Pyrrole T e s t a f t e r Different Periods a n d Kinds of Aglng TEXSILE PYRROLE .kCETONE EXTRACT TxnfE OF ACING STREKCTH TEST Per cenl Kg. Der sq. cm. Unaged stock 143.4 0 5.1 Aged in oxygen bomb (hours): 136.4 0 1 ... 24 0.5 5.3 132.2 4s ... 125.8 7.0 72 120 92.1 12.0 ... Aged in air at 70' C. (weeks): 3 days 165.9 0.1 3.8 0.1 3.0 135.0 1 2.8 93.5 0.3 2 3 23.2 0.5 3.6 Aged in air at 100°.C. (hours): ... 0.2 4 129.0 ... 0.3 8 S5.8 ... 0.4 16 57.6 0.7 24 8.2 ... 0.7 36 4.4 0.6 48 4.4 ~~
~
...
...
October 15, 1930
I-VDCSTRIAL A N D ENGINEERISG C H E X I S T R Y
From this table it can be seen that the pyrrole test indicates the beginning of the aging process a t a much earlier stage than acetone extract data, and also that it is positive before there is any appreciable drop in tensile strength. The most interesting point, however, is that there seems to be a difference in character between aging in the oxygen bomb and in air at higher temperatures. In oxygen-bomb aging the pyrrole test continues to rise steadily throughout the course of aging. I n heat aging the pyrrole test is positive after the shortest period employed, but rises only very slowly as aging progresses and seems to tend to a maximum which is very little larger than the first value obtained. Roughly, the figures in oxygen aging are from ten to fifty times as great as the corresponding stages of deterioration due to heat aging. Comparison between the data in Tables I and I1 places the natural aging, as far as the pyrrole test is concerned, intermediate between the two types of accelerated aging, There is no evident quantitative agreement between physical condition and pyrrole test. This emphasizes the important distinction that the test cannot be regarded as one for rubber oxidation products in general, but is for a specific product or class of products-and, it is believed, a class which occurs early in the oxidative process. There is consequently no reason to expect any exact correspondence between the pyrrole test and the physical properties of naturally aged rubber, since natural aging conditions can and do vary considerably, and some conditions may be expected to favor more than others the accumulation of a specific product. Theory of Variation in Pyrrole Test
Bruni ( I ) failed to get any pyrrole test with stocks aged in a current of dry air at 77" C. and ascribed it to volatility of the product. Jones (4) apparently verified this conclusion by passing a slow stream of dry oxygen over heated rubber and condensing a t the outlet a small amount of substance which gave the pyrrole test. The present writers do not believe, however, that volatility of the product is the entire reason for the large differences noted in Table 11. If such were the case it would seem to be entirely possible so to regulate the conditions as either to diminish or accentuate the volatility loss and to secure corresponding variations in the result. This was done, first by aging the sample in a "dead air pocket," which was simply a bottle entirely closed except for a section of capillary tubing piercing the stopper; and second by passing a continuous current, of air over the sample. In both cases the temperature was 100' C. The results are given in Table IT1 together with data for similar stocks aged in the oxygen bomb. Despite the difference in conditions the two heataged samples gave pyrrole tests of the same general magnitude and of a magnitude far below that of stocks aged under oxygen pressure. Furthermore, in the case of the sample aged in moving air no condensate was obtained which gave a pyrrole test, either in the conducting tubes or in a condenser immersed in ice. of Aging in "Dead" and Moving Air on Physical Properties and Pyrrole Test TENSILE PYRROLE AGING TIM% STRENGTH TEST Hours KP _. der . so. . cm. 129 0 0.2 Closed bottle at 100" C. 4 24 8 2 0.7 113 2 0 3 Moving air st 100' C. 4 22 16 5 0 4 21.1 kg. per sq. cm. oxygen pressure at 60' C. 72 126 8 7 0 120 92.1 12.0 Table 111-Effect
The writers believe that a better explanation of the difference in pyrrole test results lies in the hypothesis that levu-
379
linic aldehyde or similar material is formed comparatively early in the oxidation, without involving any great physical deterioration, and that in the subsequent course of actual physical deterioration it changes so that it no longer gives the pyrrole test. A tentative expression of such a hypothesis would be as follows: Oxidation Rubber > No physical deterioration Heat ---f (A) Physical deterioration
(A) Positive to test
(B) Negative to test
JT7e might therefore expect that in the oxygen bomb, where conditions greatly favor oxidation, the first stage would proceed relatively faster, with an accumulation of early oxidation products. In heat aging the second part of the process might be favored, and the levulinic aldehyde d e stroyed almost as fast as it is formed. The last data in this paper are centered around this hypothesis. If a fresh piece of the above-mentioned stock is aged at 100' C. in nitrogen from which oxygen has been very carefully exc!uded, the rate of deterioration is extremely slow and is presumably nothing more than a slow overcuring of the stock. If, however, before being put in the nitrogen it is placed in the oxygen bomb for a short period, the deterioration in nitrogen is much faster, even though it lost little or nothing in strength while in the oxygen bomb. In the actual experiment a piece was aged in the oxygen bomb for 72 hours without showing any drop in tensile strength, but it did show a decided pyrrole test. This piece was then placed in a tube filled with oxygen-free nitrogen and kept for periods of 24 and 72 hours, with the results noted in Table IV. Table IV-Effect
of Aging in Oxygen Bomb Previous to Nitrogen Aging TENSILE AFTER AGINGAT 100' C. BLANKSAMPLE In air In nitrogen Kg. p e r sq, cm. K g . per sq. cm. Unaged 143.4 143.4 22 hours 8.2 136.7 112.6 7 days T€sT SAMPLE
TBNSILB
Kg. per 72 hours in oxygen Same plus 24 hours in nitrogen at 100" C. Same plus 72 hours in nitrogen a t 100" C.
PYRROLE TEST
sq. cm.
148.3 73.1 32.3
1.4 1.0 0.6
It is seen that the test sample suffered a drop in tensile greater than might be expected solely from heating in nitrogen. That is, although the sample was as strong before the test as the blank itself, it nevertheless contained potentialities for deterioration which were not present in the blank. The writers think these potentialities are represented by the pyrrole test figures, and it may be noted that as the deterioration in nitrogen continued the pyrrole figures grew smaller. The experiment has since been repeated with essentially the same results. I n this later experiment, in order to exclude oxygen which might remain dissolved in the rubber as a result of its having been in the oxygen bomb, the stock was subjected to a vacuum of 2 111111. for about 30 minutes before being placed in the tube containing nitrogen. Conclusion
The foregoing will illustrate some of the applications which have been made with the test as developed by the authors. Since the above work was completed, van Rossem and Dekker (5) have published their very significant paper on extraction of rubber with benzene-alcoholic potash mix-
VOl. 2, No. 4
AiVALY TICAL EDITION
380
ture, Both methods apply to rubber a t the same stage of its life, and it should be interesting to compare results from the same sample. I n general it h a y seem that their method is of somewhat higher accuracy, while the pyrrole test requires less time and labor and is more informative aa t o the actual processes of oxidation.
Literature Cited (1) Bruni. I n d i a Rubber J., 63, 814 (1922). (2) Bruni and Pelizzoia, I b i d . , 63, 415 (1922).
:ii ~ n ~ ~ e s ; , ~ . c ~ , ~ f ~ ~ ~ X , ~ ~ ~ 2 ~ , l (1925), and Dekker, Kautschuk, 5 , (5) Rossem, (6) Whitby, India Rubber J., 63, 742 (1922).
(1929).
Determination of Nitrogen and Acetyl Content of Cellulose Nitroacetate' Wyly M. Billing and John S. Tinsley HERCULES POWDBR COMPANY, EXPERIXENTAL STATION, KENVIL, N. J.
HE increasing wide-
A method has been devised for determining both the slowly rises even Over a period spread use of cellulose acetate and nitrate content of cellulose nitroacetate. of several days. The present esters has made it seem This method is based on the usual nitrometer procedure miters found that where nide&able to investigate fully for nitrogen. The sulfuric acid residue from the nitrate was present the sapo& the mixed esters of Some of trometer after adjustment with secondary sodium phosfication value varied with the the more commonly used phate to pH 3.7 is steam-distilled in a special flask. time and temperature of the types, particularly cellulose Titration of the distillate gives the acetyl content. treatment w i t h t h e a l k a l i nitrate. I n s t u d y i n g the The maximum variation on known samples was solution. 2 per cent, while the mean average error was *0.5 Attempts to d e t ermi ne preparation and properties of per cent. the cellulose nitroacetates it the nitrogen by the nitromewas found that a more exact ter and the acetate by disanalytical procedure was necessary. tillation of the nitrometer residue gave results that were always high in acetate and due undoubtedly to cellulose Previous Work decomposition. A closer investigation into this general prinAn analysis of cellulose nitroacetate was first published ciple, together with necessary modifications, led to a very by Berl and Smith (2) and later Odd0 ( 7 ) used this method, satisfactory procedure. pointing out its drawbacks. Atsuki (1) and Kishida (6), Theory of Method Adopted in their work on nitroacetates, passed over the analvsis with just a statement that they had kmployed the Lunge-nitromeWhile the treatment of cellulose esters with concentrated ter for nitrogen and the Ost (8) method for acetate. sulfuric acid may be held to be quite drastic, it must be The Berl and Smith procedure, as more fully detailed borne in mind that, should the cellulose be decomposed by it, by Oddo, consists in a nitrogen determination by a nitrome- carbon dioxide would be evolved. If carbon dioxide is ter and a steam distillation of a separate sample after liberated from nitrocotton in any appreciable quantities, hydrolysis with sulfuric acid. The distillate gives, by then the whole principle of the Lunge nitrometer is wrong. titration, the total acidity, from which the acetate is calcu- The writers made no attempt to prove or disprove this point. lated by difference. On the other hand, data are lacking on the effect of the I n attempting to test this procedure by using known nitrometer reaction on cellulose acetate and nitroacetate. The experiments on cellulose acetate show that no gas mixtures of cellulose nitrate and cellulose acetate, total acidity values were obtained which were always much too is evolved from it on shaking in the nitrometer. With nitrolow. The precaution of an all-glass apparatus, as suggested acetates the gases evolved were bubbled slowly through by Berl and Smith, did not eliminate the error. From the limewater without forming a precipitate. It is therefore work of Carpenter and Babor (3) on the vapor pressure of concluded that the acetates are no more subject to demixed acids it is not surprising that the nitric acid could not composition than the nitrates. Obviously, then, the residue from the nitrometer should be completely distilled, because of its very low vapor pressure be free of decomposition products, and these must develop over the weak sulfuric acid content of the distilling flask. To increase the sulfuric acid concentration in the distilling in the distillation. The effect of the acidity and the temperaflask would lead to decomposition of the cellulose residue, ture can be minimized by dilution of the residue with water, with the accompanying formation of volatile acids. steam distillation to maintain a dilute solution, and further Determinat,ion of the composition of the distillate by double by conversion of the sulfuric acid of the residue to sodium indicator methods was unsatisfactory, owing to the high sulfate by the use of sodium phosphate. A satisfactory application of these factors haa been dedilution of the acids in the distilled portion. Similarly, a direct saponification of the original nitro- vised and put to use in the following method. icetate, or mixture of nitrate and acetate of known compoApparatus sition, by the Knoevenagel (4) procedure for acetates was The essential feature of the apparatus is a steam generator found unsuitable. Cellulose acetate is completely saponified in a very short time and the reaction proceeds no further. connected t o a special distilling flask of the type shown. With nitrocellulose a saponification value equivalent to its This flask is 1 liter capacity and is provided with a sealed-in nitrate content as determined by the nitrometer method thistle tube for admitting the sample and the steam supply. is reached only after 1 or 2 hours, and from that Doint it The vapor Dasses through a suitable sealed-on trap which effectively prevents entFainment. The condenser is fitted Received April 26, 1930. Presented before the Division of Cellulose to the t'rap by a ground-glass joint, thus eliminating mbber Chemistry a t the 80th Meeting of the American Chemical Society, Cincinconnections. nati, Ohio, September 8 t o 12, 1930.
T
