ANALYTICAL EDITION
FEBRUARY 15, 1940
TABLEI. RELATION BETWEEN VOLUMES OF SOLUTION PYRETHRIN I CONTENT Pyrethrin I Determined cc. Mg 3 20 5 5.9' 10 7.96 15 10.3b 20 13.9b 25 28.2b 50 43.l a 75 100 57,O" 125 71.0b 150 85.8b 114.2' 200 0 Analysis b y J. J. T. Graham. b Analysis by D. A. Holaday. Aliquot
.
Pvrethrin I dalculated Using 100-Cc. Aliquot a s Standard
.4ND
Difference from Standard
'KO.
MQ.
2.9 5.7 8.6 11.4 14.3 28.5 42.8 57.0 71.3 85.5 114.0
10.3 +0.2 -0.7 -1.1 -0.4 -0.3 f0.3 0.0
-0 3 +0.3 +o 2
81
within the range investigated the reduction is linear. The linearity of results obtained by this modification shows that the erroneous values obtained using the original Wilcoxon procedure are not due to compounds other than chrysanthemum monocarboxylic acid reacting to give mercurous sulfate. The interferences are no doubt due to unsaturated organic compounds which absorb iodine during the titration with potassium iodate. These compounds which are precipitated along with the calomel are removed from the precipitate by washing with alcohol or acetone, followed by chloroform. The linearity of the results obtained indicates that the reduction of the Denigss reagent is a clear-cut reaction which is not affected by the quantity of chrysanthemum monocarboxylic acid within the limits of the experiment.
I
aliquot was acidified and analyzed by the Holaday ( 2 ) modification of Wilcoxon's method, except that alcohol was used instead of acetone, The three largest aliquots were extracted three times with petroleum ether, and 15 cc. of Denighs reagent were used in the reduction. The results are given in Table I. The determined pyrethrin content of each aliquot plotted against the volume of the aliquot (Figure 1) shows that
Literature Cited (1) Assoc. Official Agr. Chem., J . Assoc. OficiuZ A g r . Chem., 21, 78 (1938). (2) Holaday, D. A., ISD.ESG.CHEW,Anal. Ed., 10,5 (1938). (3) Martin, J. T., J . Agr. Sci., 28,Part 111, 456-71 (1938). (4) Wilcoxon, Frank, Contrib. Boyce Thompson Inst., 8, No. 3, 176-81 (1936).
Analvsis of Rubber for Certain Diarylamine J
d
and Ketone-Diarylarnine Antioxidants L. H. HOWLAXD
AND
E. J. HART, United States Rubber Company, Passaic,
The technique of semiquantitative anti'oxidant analysis which employs a turpentine oxidation method is described and the results of characteristic analyses are given. The antioxidant effect of BLE (a ketone-diarylamine reaction product) in purified turpentine as measured by the induction period is proportional to the BLE concentration whether added to the turpentine directly or as one of the components of an acetone extract of a rubber mixture. Certain qualitative tests are also given for the detection of antioxidants in a tire tread vulcanizate.
T
HE purpose of this paper is to present methods for the
detection and semiquantitative determination of certain types of antioxidants and anticracking agents in rubber articles. The specific compounds reported on are a ketonediarylamine reaction product (BLE), a ketone-diarylamine modification (BXA), N,N'-diphenyl-p-phenylene diamine (DPPDA) , and dimethylacridan (hereafter referred to as DMA) , but the semiquantitative method developed may be used for the determination of other antioxidants. During the past few years a number of methods have been published by Endoh (2-6) and Shimada (6) on the qualitative detection of antioxidants by color tests. These authors have described a number of color reactions characteristic of commercial antioxidants. These are very useful if the chemical can be isolated from the rubber and then tested. Craig (1) has described a semiquantitative method for the isolation and determination of diarylamine antioxidants by
N. J.
steam distillation from thinly sheeted rubber compounds. I n this method, 30 to 40 grams of' the unvulcanized or vulcanized rubber stock are treated wit,h steam a t 170" to 180' C. for 2.5 to 3 hours and the amine antioxidants isolated from the distillate. The recovery of amine was above 90 per cent for pure gum compounds, and as low as 16 per cent for tread compounds containing 0.5 part of N ,N'-diphenyl-p-phenylenediamine. The semiquantitative method given below for the analysis of BLE, BXA, DPPDA, and DRIA follows, in principle, the one described by Stern and PuFfett ('7) who measured the increase in weight and also the volume of oxygen absorbed by linseed oil containing rubber antioxidants as a function of time. I n these experiments Seoaone appeared to give induction periods proportional to its concentration in the linseed oil. When making semiquantitative antioxidant analyses by the procedure described in this paper, it is sometimes necessary to determine qualitatively what antioxidants are present in the rubber. A qualitative method that functions satisfactorily for the antioxidants under discussion is a color test based on an oxidation procedure. For the detection of the antioxidant present in a rubber stock, about 1 gram of the stock is cut into 2-mm. cubes and gently refluxed with 2 cc. of concentrated acetic acid for 2 minutes. After cooling, the acetic acid solution is decanted, 1 drop of concentrated sulfuric acid is added, and after mixing 2 drops of 30 per cent hydrogen peroxide are added. The colors developed are given in Table I. Experiments have shown that the oxidation of certain unsaturated compounds, such as rubber and turpentine, is autocatalytic in nature and very susceptible to the action of certain added inhibiting ingredients. I n general, antioxidants for rubber are also antioxidants for turpentine, and on this fact is based a method for their semiquantitative de-
INDUSTRIAL AND ENGINEERING CHEMISTRY
82
TO OXYGEN TANK
15MM.
11
hllb
II
DETAILS TURPENTINE C E L L
ROCKING MECHANISM 2 CYCLES/ SEC. SO'ARC
-
I
V ~ " S T * N C SSTEEL SHAFT
11 AMF MANOMETER
VOL. 12, NO. 2
From time to time considerable difficulty has been encountered in obtaining turpentine of sufficient purity for use in antioxidant analyses. I n these cases the foregoing treatment must be repeated once or twice to obtain suitable turpentine. As a test of turpentine adequacy, it should give an induction period of from 100 to 150 minutes per 10-6 gram of BLE per cc. of turpentine measured a t 80" C. This induction period has been taken as the distance on the time axis between the parallel parts of the curves obtained with turpentine alone in one case and turpentine plus the antioxidant in the other case. Purified DPPDA and DMA and standard commercial grades of BLE and BXA have been used for analyses in this paper. These chemicals are excellent inhibitors for thc oxidation of turpentine and readily adapt themselves for quantitative analyses by this turpentine method. Some rubber antioxidants and anticrncking agents may not be quantitatively determined by this method because of inadequate inhibitory action in turpentine.
Apparatus T h e apparatus used for the semiquantitative determination of antioxidants in rubber is shown in Figure 1. CO~sTAKT-TEafP E R a T v R E I v A T E R BATH. -411
determinations have been conducted a t 80.0 =t 0.1 O C. in an insulated None1 metal box 61.0 X 45.7 X 45.7 em. (24 X 18 X 18 inches). Approximately 2000 watts are required to maintain this temperature, 1000 of which operate on a relay actuated by a mercury thermoregulator. A cover is provided for all parts of the bath 5 0 CC. not in use. BURET ROCKINGMECHAKISM. Preliminary experiments demonstrated that measurements of oxygen absorption in unstirred or unshaken cells were dependent on the surface of turpentine exposed to the oxygen, u-hile if the absorption cell 0 was shaken, the rate of oxygen absorption was determined by the amount of turpentine present and not by the surface exposed to the oxygen. Lengthn-ise shaking a t a rate of 2 cycles per second of the cells through a 90" angle has been found sufficient for uniform and reproducible results. OXYGENABSORPTION EQUIPMEKT. The details FIGURE 1. APPARATUSFOR EVALUATION OF ANTIOXIDANTSIN RCBBER STOCKS of one unit are shown drawn to approximate scale in Figure 1. The measurements are made a t atmospheric pressure and the equalization of the pressures is- facilitated by leveling the mercury in the small manometer, E. The system is made of Pyrex except for the TABLE I. COLORTESTSFOR DETECTINQ ANTIOXIDANTS IN TIRE rubber connections to E and the absorption cell, A . The body TREADSTOCKS of the cell is made from 15-mm. tubing and is about 7.0 em. in Antioxidant 1 Drop HzSOd 1 Drop HzSO4 + 2 Drops 30% HzO? length. The cell is completed by sealing an Gshaped 8-mm. Present Immediately' Immediately 5 min. 30 min. tube into the center of the 15-mm. body. DPPDA Blue Blue (flash) Orange-red Tan I n using these cells, it is necessary to keep sulfur and rubber BLE Light green Green t o olive Light green Olive particles from the attached rubber tubes from falling into the BXA .... Green t o Brown Red-brown yellow-brown cells during the course of the experiment, since these materials DMA .... Green Green Green have considerable inhibiting povier against turpentine oxidation when present in sufficient quantity. The cells may be clamped directly t o the 1.27-em. (0.5-inch) rod and it has not been found necessary to protect the neck of the cell in any manner when termination. The existence of certain very simple relationreasonable care is taken. These cells and their respective maships between the amount of antioxidant added and the delay nometers and burets may be set up in batteries of eight units, in the turpentine oxidation reaction made the task of quanthus making possible the performance of seven quantitative antioxidant analyses simultaneously. ( A turpentine blank is titative determination simple if the antioxidants present in necessary for each series of analyses.) the rubber were known.
Materials Tur entine, the medium in which the antioxidant analyses are carrietout, must be essentially free from peroxides and naturally occurring antioxidants. Commercial turpentine is first oxidized by passing purified oxygen through it a t 50" C. in order to destroy any antioxidants that may be present. After 5 to 6 hours of oxidation 2 grams of finely divided sodium are added to ahout 1000 cc. of the oxidized turpentine in order to destroy peroxides. The turpentine treated in this manner is next fractionally distilled in a nitrogen atmosphere and collected in a nitrogen-filled flask.
Preparation of Sample For rubber compounds containing from a quarter to t\vo parts of BLE, BXA, DPPDA, or DnIA per 100 parts of rubber, two thinly sheeted samples of 1.00 gram each of the sample under investigation are placed in separate extractors. One is extracted with acetone and the other x i t h acetone to xx-hich has been added one half the amount of the antioxidant believed to be initially present in the 1-gram rubber sample. After 24 hours' extraction in Underwriters' apparatus, the samples are diluted with acetone to a concentration (based on the assumptions mentioned above) of 25 X 10-6 gram antioxidant per cc. of solution, and 0.20 cc. of this solution measured from a calibrated capillary pipet is then
FEBRUARY 15,1940
ANALYTICAL EDITION
83
characteristic curve for 0.7 X turpentine is also included.
0
TURPENTINE
n
0.7X I O - ~ G . BLEICC.
TURPENTINE
Effect of Concentration of BLE on Oxidation of Turpentine
~
ri 100\ 0
w m
a
0
cn
m
