1692
ANALYTICAL CHEMISTRY
Table I. Determinations of Testosterone Propionate
Soliltion
D
n
E E F F
G G
n
c
Labeled Strength, Rlg./hII.
25 25 25 25 25 25
5
5 5
5
Vehicle
benzyl alc Sesame oil Sesame oil Sesame oil Sesame oil Sesanie oil Sesame oil Sesame oil Sesame oil Sesame oil Sesame oil
Found. Mg.,'Ml.
24 23 24 24 24 23 4 4 5 4
Found.
Melting Range of Derivative,
96.1 95.1 99.5 99.5 99.7 95.1 90.9 98.5 101 2 Y-l 4
209-210 209-210 21 1-212 211-211 2 15-2 1ti 2 15-2 16 212-213 208-209 108-209 207-208
02 78 88 88 93 78 545 925 060 720
%
c.
TESTOSTERONE PROPIOSVATE SEMICARBAZONE
Five hundred milligranis of testosterone propionate U.S.P. were dissolved in 15 ml. of methanol and added to 15 ml. of 0.225 111 semicarbazide acetate solution. The mixture was refluxed for 2 hours, cooled t o room temperature, and precipitated with 400 nil. of chilled water. The precipitate was filtered off, washed neutral with water, and dried to constant weight a t 105" C. The yield was 578.4 mg., or 99.3%oof theoretical: melting point (capillary inserted in bath a t 200 , bath temperature raised 3" per minute) 215-216" corrected. Part of this material was recrystallized from benzene-petroleum ether (about 1 to l ) , giving off-white crystals melt,ing a t 215216" corrected by the above method, and a t 212-213" correcte; with the method modified by reducing the rate of heating to 1 per minute; [oljYo = +190 (lye dioxane). These and all other melting point determinations of this substance were characterized by decomposition and by a bright red color in the melt. A sample of the recrystallized seniicarbazone gave the following analysis: calculated: for C23H350313: carbon, 68.82; hydrogen, 8.78; nitrogen, 10.47. Found : carbon, 68.67; hydrogen, 8.75; nitrogen, 10.27. The ultraviolet absorption spectrum was.determined in a 0.001% solution in methanol, in which maximum absorption occurred at 268 to 269 mg (specific absorption, E = i25 i 5%) and neg1igil)le a1)sorption from about, 320 m p upward. Testosterone propionate seniicarbazone was found to be soluble in acetone, alcohols, ether, benzene, chloroform, pyridine, and acetic acid. It is insoluble in water and iso-octane, and very slightly soluble in petroleum ether.
;7m,
ASSAY PROCEDURE
If the solution to be tested contains less than 10 mg. of testosterone propionate per ml., a sample is measured which contains
50 mg. of testosterone ropionate. This is mixed with 40 ml. has been saturated with 90% ethanol. The petroleum ether solut,ion is extracted with eight 15-ml. portions of 90% ethanol which has been saturated wit,h petroleum ether. [This extraction method is based on t,he U.S.P. XI\method for progesterone injection, and includes the modification propowd by Umberger ( S ) . ] The combined alcohol extracts are then evaporated to dryness; the residue is transferred, with the aid of a little methanol, to a 25- to 50-ml. round-bottomed flask, and treated R-ith semicarbazide acetate solution and worked up in the manner described below. If the oil t o be t,ested contains 10 mg. or more of testosterone propionate per nil., a sample containing 50 mg. is accurately measured. This is placed in a 25- to 50-ml. round-bottomed flask x i t h 3 ml. of the 0.225 A I semicarbazide acetate solution; the mixture is refluxed for 2 hours and cooled to room temperature. Ten milliliters of iso-octane are added and the resulting solution is poured into 100 nil. of ice cold water. A little methanol is used to rinse the flask. The mixture is thoroughly stirred and allowed to stand, cold, for 2 t o 3 hours. The hydrocarbon layer is not separated; the entire mixture is filtered through a tared, medium porosity, sintered-glass filter, and the precipitate is washed with iso-octane, sucked dry, and washed with water. The precipitate is again sucked dry and then dried to constant weight in an oven a t 105". The assay is calculated as follows:
of petroleum ether whic!
Llg. of semicarbazone recovered 344.48 x 401.3.3 - y = ml. of samplc mg. of testosterone propionate per nil. ~~
~~
The melting point of the recovered seniicarbazone should be between 207' and 217" corrected, when determined in a bath heated at 3" per minute, with the capillary inserted :Lt 200" C. Table I shows results obtained with various commercial samples of festosterone propionat,e injections. ACKNOWLEDGMENT
The authors x-ish to thank Wilbur S. Felker, quality control manager, and Felix J. Pheiffer, analytical laboratory manager, for their encouragement and assistance in the execution of t,his ivork, Bradley Whitnian of the Chemical Research Divipion for advice and suggestions, and Edwin Comer of the Chemical Research Division for the microanalytical tlet.ermiriatioii. .LITERATURE CITED
(1) Klein, D., Weher, S . ,and Gordon, S.AT., A ~ I .CHEJI., . 20,1746 (1948). (2) Pheasant, R., and Zenno. E. E., unpublished research, (3) Umberger, E. J., J . A m , Pharm. .4ssoc., Sci. E d . , 36,700 (1950). RECI:I\-F.D .\pril 3 . 1H:l
Plasticizer Quality Test CARL .J. \IAL\l,
LEO B. GENUNG,
AND
MhURICE L. TOWNSEND
Eustman Kodak Co., Rochester, N.
L-UTICIZERS used in films, lacquers, plastics, and hot Pnielt coating compositions should not contribute to the breakdown or discoloration of the product under the conditions encountered in formation, application, and use. It is desirable, therefore, t o test plasticizers for their suitability for the purpose intended. In some cases, the test is made by mixing the plasticizer with :L cellulose ester, resin, etc., heating, and observing the change in color or some critical physical property. Salts, acids, or other impurities introduced during manufacture of the polymer may diminish or accentuate the results of the test. These impurities may vary from batch to batch of the polymer and, therefore, prevent a reliable test of the quality of the plasticizer. I n other cases, plasticizers may be tested by determining some undesirable component, such as sulfur when sulfuric acid has been
Y.
used as catalyst in the manufacture of the pljsticizer. The revult may or may not give an accurate indication of the quality of the plasticizer. An independent test is desired which can be applied to the plasticizer, and which is a simple, sensitive, and direct measure of objectionable impurities. The following test was found to fulfill the requirements acceptably. THE TEST
Procedure. Place 5 ml. of the plasticizer in an 18 X 150 nini. borosilicate glass test tube and insert a 0.5 X 2 inch (1.25 X 5 em.) strip of ashless filter paper (Whatman No. 42 was used for the experiments described). The strip is thus about half immersed. Heat the tube and contents unstoppered for 1 hour in an oil bath a t 180" C., and examine the filter paper for discoloration, which indicates the presence of undesirable impurities.
1693
V O L U M E 23, NO. 11, N O V E M B E R 1951 Table 1. Effect of Acidic Contaminants as IIIeasured by the Test Contaminant Added
I’alJer Color
Plasticizer, Diethyl Phthalate Sone Sone 0,OOOl‘ib diethyl sulfate None 0.0017c diethyl srilfate Slight 0 . 0 1 % diethyl sulfate Brown 0 . 1 % diethyl siilfate Black 1 .OYc diethyl sulfate Black 1 . O % phthalic acid Sone Plasticizcr, Triplicnyl Pliosilhate 12one Lt. yellow-brown o.o1g’, 1’OCla Black 0 . 0 2 % POCla Black 0 , 2 % FOCI1 Black 2 , oyc POCla Lt. yellow-hrown 0 , 0 1 7 HCI Black 0.7%’HCI Black 0 . 0 1 % IIsPOa Black 0.1Sb IIiPO4 Sone
Limits. Thc limits set should be based on use requirenients and on the purity of materials available. The limits should not be stricter than needcd for the purpose, in order to keep t’he cost of the plasticizer down and to avoid a burden on the manufacturcr. For rpccification purposes, color standards or a t least “go or no-go” standards are needed. Easily prepared standards can be madc hy immersing a strip of ashless filter paper for 10 seconds in a standard iodine solution, as prepared for a volumetric rea ent. IVithdraw the paper and compare with the test sample witkn 20 seconds. Acccptahle samples should show no more coloration of the paper than tlic chosen standard. The strength of the iodine solution used depends on the limit dcsired. Paper treated Kith 0.005 N iodine is light yellowbrown, while 0.01 N solution gives a light brown color. If the characteristic starch-iodine color appears, it is an indication that t,he paper is unsatisfactory for this purpose, and a starch-free paper should lie suhstit,uted.
test are sho\vn in Table I. The test readily detects 0.017, or less of the undesirable impurities. Phthalic acid did not color the paper. The corresponding effects on the intrinsic viscosities of cellulose ester compositions are shown in the next section. Many salts have no visible effect on the paper color. Two per cciit additions of magnesium acetate, magnesium carhonate, sodium chloride, sodium carbonate, OF zinc acetate to diethyl phthalate do not discolor the paper, whilesodiumbicarbonategives a very slight coloration, and zinc chloride blackens the paper. However, salts niay neutralize acidity in the plasticizer. Diethyl phthalate containing 0.01 yo diethyl sulfate, which ordinarily Fvould hro\vn the paper, gave no coloration whcn magnesium carbonate, sodium carbonate, sodium bicarbonate, or zinc acetate \viis present. Similar additions of sodiuni chloride, sodiuni tiicarbonate, sodium carbonate, and sodium hydroxide, or varying ;rinounts of phenol to triphenyl phosphate produced 1 1 0 xppreciable color on the filter paper. CORREL4TION O F PLASTICIZER QUALITY WITH VISCOSI’I’Y BREAKDOWN OF CELLULOSE ACETATE CO.MPOSITIONS
Correlation of plasticizer quality with the stability of conipositions n-as investigated by measuring viscosity breakdown under various coiiditions of heating. Samples of cellulose acetate sheeting were prepared from acetone solutions of cellulose acetate with plasticizers containirig varying amounts of added acidic impurities. This cellulose acetate sample \vas prepared using zinc chloride catalyst, aud \vas free from traces of sulfuric acid and sul)staiitially free from sitlts that might interfere.
‘Table 111.
Effect of 3Zagnesium Carbonate on Stability ---____Heat Treatment 1 hour 1 hoiir 72 h o l m looo C. in 180’ C. In 205’ C. 01 en metal block ~ i i e t a lblock .Icid ~..-. - Intrinsic Visrosity in Acetic ~. ._ ____. In
Contanunant
Tune
~~~
_______
Table 11. Effect of Plasticizer Quality on Viscosity Breakdown of Cellulose Acetate Compositions H e a t Treatment 72 hours 1 hour 1 hour in looo C. in 180° C. in 205O C. C‘ontaliiinant Kone oven metal block metal block Intrinsic Viscosity in Acetic Acid Celliilose Acetate ( 5 Parts), Diethyl Phthalate (1 P a r t ) Sollr 1,59 1.34 1 46 1 09 1.32 1.20 1.57 1.54 l.Oyc phthalic! acid 1.26 1.43 0.99 0 , 0 1 7 diethylsulfate 1 . 5 4 1.59 0.27 0.49 0.16 O.l%b%iethylsulfate 1 . O % diethyl sulfate 1.51 0.04 0.04 0.08 Celliilose Acetate (6 Parts), Triphenyl Phosphate (1 P a r t ) 1,59 1.51 1.51 1.02 None 0 . 0 1 % POCls 1.62 1.54 1.46 0 95 1.46 1.62 1.34 0.89 0.1% P o c k 1.69 0.79 0.19 0.66 1 .OTc POCla
Variations. T h r tmt conditions were chosen somewhat arbitrarily. Othcr tinies, temperatures, and color standards can he selected to fit specific requirements. Heating can be done in a metal block instead of an oil bath, but a longer time or a lower h i t would be selected. Unstablc plasticizers or plasticizers containing acidic impurities uruallr discolor or char the immersed portion of the paper. A sample o f tri-2-chloroethyl phosphate, however, blackened the papei abovc the surface of the liquid. Generally, any appreciable darkening or charring of the paper may be taken as a warning. CAUSES OF INST4BILITY
StronK witiic arid acid-producing impurities arc particularly objectionable in plasticizers and compositions. The effects of varying aniounts of such impurities on the color produced in the
Salt-Free Celliilose Acetate ( 3 Parts), Dicthyl Phthalate ( I Part) 1.32 1. 3 2 1 23 Sone 0.86 1 29 1.34 1.23 O.Ol(ibdiethyIsiilfate 0.89 1.29 0.12 0 08 0 , 1 % diethyl sulfate 0 08 Cellulose Acetate ( 5 P a r t < ) Containing 0 08% AIgCOa. Diethyl (1 P a r t ) 1.34 1 37 1.23 Sone 1 3’2 1.34 1.23 O . O l ~ odiet,hrlsulfate 1 34 1 32 0 , 1y0diethyl sulfate 1.26
Phthalate 1.02 1.05
0.99
Portioiis of each o f the plasticized films were given different heat treatnicnts representative of conditions that might be encountered in variouP practical uses. Int,rinsic (or inherent) viscosity measurements were then made on solutions containing 0.25 gram of sample in 100 nil. of glacial acetic acid. \lscosity 2.3 log relative viscosity results were calculated using (?)o.rs = 0.25 -’ and are presented in Table 11. The viscosity loss is considerable if the plasticizer impurity exceeds a critical value. Fortuuately, the filter-paper color test is more sensitive, as shown bj- coniparison with Table I. rlcidity in plasticizers, shoivn by tlic quality test, may be neutralized by salts. Salts present in cellulose acetate may have a similar stabilizing eff ect on acidic plasticizers. A coniniercial cellulose acetate was washed with dilute hydrochloric acid and distilled water to remove salts already present, and a portion of this material was then treated t o contain 0.08% magnesium carbonate. Films were coated from acetone solution containing 5 parts of treated cellulose acetate plus one part of plasticizer with varying amounts of diethyl sulfate. Table I11 presents intrinsic viscosities in acetic acid of these samples after varying heat treatments. The viscosity data show that 0.01% diethyl su1fat.e in the plasticizer does not produce appreciable loss in viscosity under
1694
A N A L Y T I C A L CHEMISTRY
Table IV.
Viscosity Stability of Film Containing Tri-2-chloroethyl Phosphate Intrinsic Viscosity in Acetic Acid No heat 1 hour 1 hour treatment at 180° C. a t 205' C.
Cellulose Acetate Film Plasticized with: Tri-2-chloroethyl phosphate (unstable) Diethyl phthalate (stable)
1 .23 1.26
0.49
B .OS
1.08
0.89
Table V. Effect of Stabilizer on Plasticizer Quality Test Results Diethyl Sulfate .4dded, % None
0.01 0.01 0.1 0.1
Table VI.
Effect of Stabilizer on Viscosity Stahility Heat Treatment 72 hours 1 hour 1 hour in 100° C. in 180' C. in 205' C . None oven metal block metal block Intrinsic Viscosity in Acetic -4cid
-_
[Film composition. Salt-free cellulose acetate (5 parts). diethyl phthalate with 0.1% diethyl sulfate (1 part)] S o stabilizer 1.59 0.27 0.49 0.16 1.O% stahilizer 1.57 1.54 1.54 1.05
(Plasticizer, diethyl phthalate) Stabilizer Added, % None None 0.25 0.25 1 .o
Paper Color None Brown None Black None
conditions of the experiment , and 0.08% magnesium carbonate present in cellulose acetate counteracted the effect of 0.1% diethyl sulfate in the plasticizer. The data in Tables I1 and I11 provide a basis for setting test h u t s and also illustrate possible difficulties of measuring plasticizer quality by tests on mixtures with other materials. One plasticizer sample, tri-2-chloroethyl phosphate, when tested by the stability test was found to darken the filter paper above the surface but not the submerged portion. Its effect on viscosity stability was tested by preparing a film csonsisting of 5
parts of a commercial cellulose acetate and 1 part of this plasticizer. A similar film containing good diethyl phthalate was prepared for compariuon. Intrinsic viscosities of acetic acid solutions of the two samples after varying heat treatments are shown in Table IF'. Again, the plasticizer showing darkening of the filter paper also shows poor viscosity stability. This plasticizer-quality test may be uaeful for evaluating the stabilizing values of coniniercial materials. A commercial material (an epoxy compound which is an acid acceptor) was tested for stabilizing action by adding it to diethyl phthalate containing diethyl sulfate, and running stability tests. The results, given in Table F', show that the effect of the diethyl sulfate was counteracted by the use of a sufficient amount of the stabilizer. Table VI shows that the presence of this stabilizer also maintained the viscosity stability of the cellulose ester film. RECEIVED M a y 4,1951. Presented before the Division of Paint, Varnish, and Plastics Cheniiatry a t the 119th 5Ieeting of the ~4IEBICAx C H E H I a a L SOCIETY,Boston, Mass.
Benzotrifluoride as a Cryoscopic Solvent for Fluorinated Compounds L. J. EI4LS AND 11. G. BRYCE Minnesota Mining & Manufacturing Co., S t . Paul, J I i n n .
HE molecular weights of many fluorinated compounds can T be determined by vapor density measurements ( 3 ) . This method can be extended to compounds of relatively high molecular weight, owing t o the high degree of thermal stability (6). Vapor density methods cannot be used on partially fluorinated compounds when they are not of sufficient thermal stability. Cryoscopic methods have been described by Bernstein and Miller ( 1 ) and Lit ant ( 2 ) in which unsymmetrical difluorotetrachloroethane, CFZCICC13, was used as a solvent. This material, although pwsessing a convenient freezing point and a high molecular freezing point constant as well as being a good solvent for fluorinated compounds, is not readily available and is difficult to free of its isomen. After a number of possible solvents had been examined on the basis of their ability to dissolve fluorinated compounds, cyclohexane and benzotrifluoride were chosen for detailed examination. Benzotrifluoride is a good solvent for fluorinated materials. It is relatively inexpensive, is readily available, and is easily purified in a laboratory column. Cyclohexane, while having good aolvent properties and a convenient freezing point, was found to be very difficult to purify. I n consequence consistent freezing point data could not be obtained. Benzene is not a satisfactor), solvent for compounds of this type.
weie purified by fractional distillation. thalene Feie both of reagent grade.
The camphor and naph-
APP4RATLS
The conventional Bechnann-type apparatus waa fouud to be satisfactory. This consisted of a 25-mm. glass test tube, surrounded by a 37-mm. test tube, and equipped with a suitable automatic stirrer. Because of the low freezing point of benzotrifluoride, it was necessary to construct a low temperature thermostat. This consisted of a 2-liter transparent Dewar flask filled with acetone. The acetone was cooled by circulating dry ice-cooled acetone through a coil of 3 / 8 inch copper tubing immersed in the bath. By means of a thermostatically controlled solenoid valve, the flow of cold acetone could be controlled so as to maintain the bath temperature a t -33" 5 0.5' C. The temperature of the cryoscopic solution was measured by mean3 of a No. 14A Thermistor manufactured by The Western Electric Co. The resistance of the Thermistor was measured by means of a Shallcross No. 630 bridge in conjunction with an external galvanometer which had a sensitivity of 0.003 ma. per millinieter of deflection. The curve shown in Figure 1 U R S taken from the calibration curve supplied by the manufacturer. The calibration was checked at two points, the freezing points of triply distilled mercury ( -38.89" C.) and pule benzotrifluoride (-28.16" C ) PROCEDURE
REAGENTS
Commercial grade benzotrifluoride was obtained from the Hooker Electrochemical Co. Material with a constant freezing point of -28.16' C. wa8 obtained by fractionation in a 50-plate column. The fluorinated solutes were direct or derived products of the electrochemical fluorination process ( 4 ) . All these compounds
The solutions weie prep:lred by adding to a weighed quantity of pure benzotri5uoride a weighed amount of the solute. In cases where the solutes had appreciable vapor pressures, the material was weighed out in a thin-walled glass ampoule, which was then broken under the cooled solvent. A known amount of the resulting solutio~iwas introduced into the apparatus and then placed in the thermostatically controlled bath.