(2) Am. Soc. Testing Materials, Philadelphia, Pa., Tentative Recommended Practice for Sampling Atmospheres for Analysis of Gases and Vapors, D-160558 T, 1958. (3) Henne, A. L., Chanan, H., Turk, A., J. Am. Chem. SOC.63.3474 (1941). (4) Henne, A. L., Turk, A:, Ibid., 64, 826 (1942).
( 5 ) Turk, A,, Chanan, H., Org. Syntheses
27, 7 (1947). (6) Turk, A,, Sleik, H., LIesser, P. J., Am. Ind. Hyg. Assn. Quart. 13, 23 (1952). (7) Turk, A,, Morrow, J., Levy, P. F., Weisaman. P.. I d . J . Air and Water Pollution, ’5, 14 (1961). ( 8 ) West, P. IT., Sen, B., Gibson, N. A.,
ANAL.CHEM.30, 1390 (1958).
RECEIVED for review October 11, 1961. Accepted February 13, 1962. Division of iTater and Waste Chemistry, 140th Meeting, ACS, Chicago, Ill., September 1961. Work was SuDDorted entirelv b r a research grant f r o g the iYation&l i n stitutes of Health, Public Health Service.
Analysis of 4,4’-Isopropylidenediphenol by Isotope Dilution L. H. GRIFFIN Houston Research and Developmenf laboratory, lndusfrial Chemicals Division, Shell Chemical
b An isotope dilution procedure is described for the direct assay of the The p,p’-isomer of bisphenol A. method is accurate to approximately f 0.5% p,p’-BPA and is applicable over a wide concentration range.
T
importance of 4,4‘-isopropylidenediphenol (bisphenol A) as a raw material in the manufacture of epoxide resins and other polymers has accented the need for a more direct and comprehensive method of analysis, free from the limitations of existing procedures. Paper chromatographic analysis, described by workers of this laboratory, involves determining individual known impurities and then obtaining the major p,p’-BPA isomer by difference (1). ASTM precision freezing and melting point determinations are not generally applicable to low purity samples, are tedious to apply, and require considerable amounts of sample material for analysis. Isotope dilution techniques afford quantitative analysis without quantitative separation, provided that qualitative isolation of the component of interest in high purity is possible. Accordingly, an isotope dilution method applicable over a n-ide concentration range has been developed for direct assay of p,p’-BPA. The method applies not only for technical and high purity grades of bisphenol A, but for crude inprocess materials as well, and requires only a fraction of a gram of sample for analysis. HE
recovered by recrystallization. Comparative specific activities of the recovered sample fraction and of the labeled reference material are determined by liquid scintillation counting for calculation of the final result. APPARATUS AND REAGENTS
Apparatus. A manually operated Tri-Carb liquid scintillation spectrometer (Packard Instrument Co., LaGrange, Ill.) is used for counting. Counting vials are of low potassium glass, Kith a 20-ml. capacity. Reagents. The liquid scintillation counting solution has the following composition: 500 ml. of reagent grade toluene; 500 ml. of 1,4-dioxane treated with activated alumina and activated charcoal and then filtered; 0.3 gram of 2,5-diphenyloxazole; 0.05 gram of 1,4 bis[2-(5-phenyloxazolyl) 1benzene (POPOP); and 50 grams of naphthalene recrystallized from ethyl alcohol. The carbon-14-labeled p , p ’ BPA was synthesized by the catalytic condensation of phenol with 0.5 me. of carbon-14-labeled acetone a t a phenolto-acetone molar ratio of 10 to 1. Unconverted reactants were distilled from the crude product, then high purity p,p’-BPA was recovered by three successive recrystallizations from chlorobenzene a t a 5 to 1 weight ratio of solvent to bisphenol A. The final recovered crystalline fraction was 99.9% mt. pure p,p’-BPA with a specific activity of approximately 2 pc. per gram. Purity of the labeled reference material was checked by paper chromatography and by ASTlI freezing point methods. PROCEDURE
OUTLINE
OF METHOD
A known amount of high purity, carbon-14-labeled p,p’-BPA of known specific activity is intimately blended with a known amount of sample by dissolution in hot chlorobenzene; then a high purity fraction of p,p’-BPA containing a representative mixture of labeled and unlabeled molecules is 564
e
ANALYTICAL CHEMISTRY
Tagging and Blending. Accurately weigh into a 25-ml. glass-stoppered Erlenmeyer flask approximately 0.2 gram of sample and approximately 0.05 gram of carbon-14-labeled reference p,p’-BPA. Add approximately 15 ml. of reagent grade chlorobenzene and dissolve t h e mixture on a hot plate set a t low heat. During the solution and subsequent crystallization
Co., Deer Park, Tex.
steps, stopper the flask lightly to avoid solvent evaporation. Remove the flask from the hot plate and stir the solution gently with a narrow spatula to ensure thorough mixing of the labeled and unlabeled p,p’-BPA molecules. Avoid getting liquid on the sides of the flask. Crystallization. Allow the solution to cool slowly to room temperature; then add a single p,p’-BPA crystal to initiate crystallization. Crystallization is normally rapid, requiring less than 45 minutes for recovery of approximately SOTo T T ~ . of the available p,p’-BPh. The resulting crystals should separate from the mother liquor into clear, long, needle-like structures. An amorphous, gel-like mass enveloping the entire volume of sample may form during the crystallization step, particularly during analysis of high purity bisphenol h samples. If this happens, redissolve the mixture and repeat the crystallization step until satisfactory crystals are obtained. Formation of the gel-like material is objectionable because of possible coprecipitation of impurities and because of difficulty in removing the solvent during the filtration and drying steps. Filtration a n d Drying. As soon as a recoverable amount of crystals is available for counting, collect the crystals on a medium pore frittedglass filter attached to a vacuum system. Rinse the crystals once with several milliliters of fresh chlorobenzene t o remove occluded mother liquor, then follow this initial rinse with several milliliters of benzene to remove chlorobenzene, thus expediting crystal drying. Dry the crystals by air f l o ~under vacuum for about a half hour before weighing and counting. Counting. Accurately weigh 0.1 gram or less of the dry recovered crystals into a 20-nil. glass counting vial. Add scintillation counting solution, keeping the level of liquid just below the shoulder of the vial. I n a similar manner, prepare a standard counting solution containing approuimatelj the same amount of accurately weighed carbon-14-labeled reference p,p’-BPA tracer dissolT ed in liquid srintillator. Prepare :ithird vial con-
taining scintillation solution only for the background determination. Store the solutions in the freezer unit of the Tri-Carb spectrometer for a t least 1 hour prior to counting. Count all solutions until a statistically satisfactory number of counts is obtained. Calculations. Calculate counts per minute for each sample, correct for background, and use in the formula below to obtain per cent p,p’-BPA in the sample.
where SO= counts/minute/gram of reference p,p’-BPA S = counts/minute/gram of recovered p,~’-BPAfraction W * = weight of reference p,p’-BPA added originally W Z= weight of sample RESULTS A N D DISCUSSION
The precision and acruracy of results have generally approached the predicted limitations governed by counting statistics. As currently applied, the method is accurate to approximately *O.57, p,p’-BPA in the sample, as determined by analyses of numerous synthetic samples containing known blends of the p,p’- and the o,p’-isomers of bisphenol A. Typical results are shown in Table I. Comparative results of analyses by several methods obtained for a single sample of crude bisphenol A are shown in Table 11. Isotope dilution analysis is definitely competitive with the other methods and probably more reliable for samples of low p,iq’-BPA purity, since the determination is independent of other components,
known or unknown, in the sample. The elapsed time required for a single analysis is on the order of 11/2 hours. Operator time per sample is less than 1 hour. As a control measure on the over-all procedure, a sample of known p,p’-BPA content is analyzed with each series of sample unknowns. The labeled reference material will accumulate moisture as the result of repeated exposure to the atmosphere. Accordingly, small portions of the reference compound are recrystallized and dried intermittently, as needed, before use in analysis. Earlier work in these laboratories had shown that p,p’-BPA of 99.9% purity could be obtained by a single recrystallization from a large excess of chlorobenzene, on the order of 50-to-1 parts by weight, solvent to bisphenol A. The effectiveness of this scheme was checked by isolating fractions of carbon14-labeled p,p’-BPA from synthetic mixtures of the tagged p,p’-isomer and untagged impurities and then comparing the specific activity of the recovered fractions with that of the high purity reference material. To determine the maximum tolerable sample weight for counting without fluorescence quenching, varying weights of a given sample fraction were counted, and the count rate us. weight of sample was then plotted. A linear relationship was found to exist up to a sample weight of 0.1 gram; beyond this point the relationship became nonlinear because of sample quenching. The sample size for counting was therefore limited to 0.1 gram. -4statistical counting error amounting to no more than 0.3Y0 standard
Table 1.
Analysis of Samples of Known p,p’-Bisphenol A Content
Amount Present
(per cent weight) Amount Found Difference
QQ 7.
99 9
99.7 99.7 90.5 86.8 90.7
99.0 99.4 90.8
Table II.
+0.2 -0.7
-0.3 +0.3 -0.6
86.2
91.1
$0.4
Comparative Analyses of Crude Bisphenol A PIP’-
Bisphenol Method of Analysis Isotope dilution Paper chromatography ASTM freezing point
A
(per cent weight) 91 .o 90.5 89.8
deviation was maintained for a counting time of 10 minutes. I n order that the counting rates from 0.1-gram fractions for counting be sufficiently high, the weight of original sample unknown taken for analysis was limited to approximately 0.2 gram. The total tracer activity used for each analysis was on the order of 0.1 qc. LITERATURE CITED
( I ) Anderson, W. M., Carter, G. B., Landua, A. J., ANAL. CHEM. 31, 1214 (1959).
RECEIVED for review October 23, 1961. Accepted January 15, 1962. Division of Analytical Chemistry, 139th Meeting, ACS, St. Louis, ILlo., March 1961.
Determination of Iodine in Mineral Premixes PERRY KING, EMILY S. BRETZ, and THEODORE J. KNEIP] Mallinckrodt Chemical Works, St. louis 7, Mo.
b A method was developed for the determination of iodine in mineral premixes for livestock feed. A sample containing more than 5 mg. of iodine i s digested in a KC103-HCI mixture. Dilute formic acid i s added to the digestion mixture liberating iodine which is distilled off and collected in a sodium hypochlorite solution. The excess hypochlorite i s removed, potassium iodide is added, and 1 3 - i s titrated with standard sodium thiosulfate. The method was developed for premixes containing calcium iodate as the iodine source, but functions equally well when ethylenediamine dihydroiodide i s the
source. A preliminary ignition of premixes containing thymol iodide rendered them suitable for this method, which appears to b e flexible enough for general application to the determination of the iodine content of mineral premixes.
I
is essential to the production of thyroglobulin TI hich controls the rate of metabolism in the animal body. For livestock, iodine is commonly combined xith essrntinl minerals, and the resulting premix is blended n-ith the feed. The AOAC has two official ODISE
methods (1) for the assay of iodine in feeds fortified with minerals, but does not have a method for iodine in the premix. EXPERIMENTAL
Samples. Synthetic mineral mistures were prepared in this laboratory to simulate the compositions of premixes. The method was applied to three rommercial premises containing
Present address, Mallinckrodt Chemical Korks, Uranium Division, P. 0. Box 172, Saint Charles, Mo. VOL. 34, NO. 4, APRIL 1962
565
