J. Soc. Cosmet. Chem., 32, 163-173 (May/June 1981) HPLC analysis of some bacteriostats in deodorant sticks and soaps RAJA G. ACHARI and DAVID CHIN, Research & Development Center, Bristol- Myers Products Division, 225 Long Avenue, Hillside, NJ 07207. Received February 5, 1981. Synopsis A stability-indicating HPLC method for the determination of some bacteriostats such as Triclosan (2,4,4'-trichloro-2'-hydroxydiphenylether) and TCC (3,4,4'-trichlorocarbanilide) is described. The liquid chromatographic separation is carried out using a •Bondapak Alkylphenyl column and the mobile phase consisting of 1:1 (V/V) acetonitrile:water. The method had been documented to be precise and accurate and has been successfully applied in assaying commercially available deodorant sticks and soap samples. INTRODUCTION Triclosan (I) (Irgasan DP-300 ©, 2,4,4'-trichloro-2'-hydroxydiphenylether) (Ciba-Geigy, Greensboro, NC) is a commonly used bacteriostat in deodorant sticks and soaps. TCC (II) (Monsanto, St. Louis, MO) (3,4,4'-trichlorocarbanilide) is mainly used in deodorant soaps. Various methods have been reported for the analysis of these bacteriostats in deodorants, but all of these methods suffer from certain drawbacks. Perfumes, other UV absorbing substances, or chemical breakdown products of the bacteriostats often interfere with ultraviolet and colorimetric methods (1,2). The gas chromatographic procedure of Demars and Yates (3) for the analysis of TCC is cumbersome and non-specific because the amines which are ultimately analyzed are also the probable chemical degradation products of TCC. One of the two reported HPLC procedures requires a gradient elution system (4) and the second requires a radial compression separation system (5). Both of the above systems are neither suitable nor available in all laboratories for routine analysis; moreover, neither of the above HPLC methods has been tested to determine whether it is "stability-indicating" due to the chemical degradation of the bacteriostats in the finished products. The present study is aimed at developing a stability-indicating simple isocratic HPLC method for assaying triclosan in deodorant sticks and TCC in deodorant soaps. Chromatographic parameters have been provided to assay both triclosan and TCC should these be present in combination. The chemical structures of triclosan and TCC are shown below. 163 164 jOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS O Cl C1 O • Cl C1 NHCNH C1 Cl HO TCC I II METHOD APPARATUS A modular liquid chromatographic unit consisting of a constant flow solvent delivery pump (Model 6000A, Waters Assoc., Milford, MA), a continuously variable UV-VIS spectroflow monitor with a capability of programming to read absorbance of an eluate at several wavelengths simultaneously, and of obtaining an absorbance spectrum between desired wavelengths (Model LC-75, Auto Control, Perkin-Elmer Corp., Norwalk, CT), a digital recorder (Model 56, Perkin-Elmer Corp.) and an integration device (Model SP-4000, Spectra-Physics, Santa Clara, CA) to calculate the area under the curve of the eluates, was used for chromatography. An alkylphenylsiloxane bonded column (•Bondapak Alkylphenyl, Waters Assoc.) was used for the separation of the bacteriostats with a mobile phase containing 1:1 (V/V) acetonitrile:water (HPLC grade, Baker Chemical Co., Phillipsburg, NJ). The bacterio- stats were monitored at 280 nm. STANDARD SOLUTIONS Appropriate amounts of triclosan and TCC were dissolved in methanol (HPLC grade, Baker Chemical Co.) and the appropriate aliquots of the solutions were further diluted using the mobile phase or 1:1 (V/V) methanol:water to yield a concentration range of 25 •tg/ml-75 •tg/ml for triclosan, and 6 •tg/ml-24 •tg/ml for TCC, respectively. SAMPLE PREPARATION A. Deodorant Sticks Approximately 1 g of the deodorant stick was accurately weighed into a 50-ml volumetric flask. 5 ml of methanol was added and the deodorant was dissolved by applying gentle heat on a hot plate. 25 ml of 1:1 (V/V) acetontrile:water was added. The solution gelled. It was warmed until dissolved and diluted to volume with additional 1:1 acetonitrile:water. The solution was cooled to room temperature and the volume was adjusted to mark if needed. The solution was mixed thoroughly and approximately 15 ml of the above solution was transferred into a stoppered centrifuge tube and placed in a methanol-ice bath for 30 min. The solution was centrifuged while cold to separate a clear supernatant from a small amount of gel which did not contain any bacteriostat. The supernatant was transferred to a sample vial and this solution was used for chromatography. HPLC ANALYSIS OF BACTERIOSTATS 165 B. Deodorant Soaps Approximately 1 g of soap was accurately weighed and transferred into a 50-ml volumetric flask. 10 ml of methanol was added and the soap was dissolved with gentle heating. The solution was diluted to volume with methanol. 5 ml of the methanolic solution was transferred to a 100-ml flask and brought to volume with 1:1 (V/V) methanol:water. The solution was thoroughly mixed and an appropriate amount of the solution was transferred to a stoppered centrifuge tube. The tube was cooled in a methanol-ice bath for 30 min and centrifuged while cold and the clear supernatant was transferred into a sample vial. This solution was used for chromatography. C. Chemical degradation of bacteriostats for "stability indicating" testing Several chemical degradation studies of the individual bacteriostats and the deodorant samples (containing bacteriostats) were carried out. The following tests were performed: 1) effect of acid; 2) effect of alkali; 3) effect of oxidation; 4) effect of heat; and 5) effect of UV irradiation. Triclosan and TCC and the deodorants were dissolved in methanol and to the methanolic solution acid (HC1), base (NaOH), oxidizing agent (Potassium monoper- sulphate, Oxone ©) were added, respectively. An aliquot of each of these solutions was transferred into an ampule and sealed. The ampules were heated at 40øC for 24-48 h. For UV irradiation the methanolic solution was applied on a clear glass plate and the solvent was evaporated in vacuum. The process was repeated several times to ensure that sufficient amounts of the sample had been added onto the glass plate. One half the glass plate was covered with aluminum foil to serve as a control and the other half was exposed to UV light. This was accomplished by placing a mercury lamp (Model //GE, G15T8) about 6 in above the glass plate. In addition, methanolic solutions of triclosan and TCC in sealed ampules were also exposed to UV light. QUANTITATION The calibration standards were injected in duplicate and a regression analysis of the drug concentration versus the area under the curve was carried out. The concentration of the bacteriostat was extrapolated from the regression curve and the percent of the bacteriostat was calculated in the following way: C x V x 100 = % Bacteriostat, SIV where C = extrapolated concentration of the bacteriostat in mg/ml, Sir/= sample weight in mg, V = sample volume. IDENTIFICATION OF THE BACTERIOSTATS Initial identification of the bacteriostats was carried out by comparing the retention of the bacteriostats with those of the authentic reference standards. Further confirmation 166 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS of the identity and purity of the chromatographic peak was effected by the absorbance ratio measurement technique which is accomplished in the following way. An authentic bacteriostat standard is chromatographed and the flow of the mobile phase is stopped approximately at the chromatographic peak maximum. The instrument is programmed to measure the absorbances of the eluate at pre-selected wavelengths and the ratios of the absorbances are calculated. Similarly, a sample is chromatographed and the absorbance ratios of the analytical peak are determined. If the absorbance ratios of the bacteriostat in the sample compare within _+ 10% with those of the reference standard, the identity and the purity of the bacteriostat is confirmed. RESULTS AND DISCUSSION Under the chromatographic conditions used, triclosan and TCC showed the chroma- tographic properties listed in Table I. Table I Chromatographic Properties of Triclosan and TCC • Rentention time (sec.): 483 468 k': 6.1 5.9 H.E.T.P. (mm): 0.111 0.123 Tailing factor (%): 70 70 •Chromatographic conditions: •Bondapak alkyl phenyl column, 1:1 (V/V)Acetonitrile:water, mobile phase; flow rate 2 ml/min., detection wavelength 280 nm. For calculation of tailing factor see Ref. (6). Both triclosan and TCC were tested for their linear detection range. The data were obtained using three calibration standards and each standard was injected in duplicate. The results are summarized in Table II. Table II Linearity Data of Triclosan and TCC • Amount Injected Slope Correlation Compound range 0tg) (counts, ng) Intercept Coefficient Triclosan 0.63- 1.89 467 +6600 0.99999 TCC 0.15 - 0.62 1,354 - 7800 0.99994 •Chromatographic conditions: As in Table I. For regression analysis, area under the curve was obtained in electronic counts. Chromatograms of a reference triclosan standard and that of a deodorant stick are shown in Figure 1. The chromatogram clearly demonstrates the separation of triclosan from perfumes and other ingredients. None of the deodorant sticks contained any TCC. Similarly, chromatograms showing the separation of TCC are shown in Figure 2. All of the commercial deodorant soaps except one analyzed in this study contained only TCC as shown in the product label; however, both TCC and triclosan can be separated using a modified solvent system as shown in Figure 3. The precision and accuracy of the procedure were determined. The within-run and between-run precisions of the chromatographic responses were determined by repli- HPLC ANALYSIS OF BACTERIOSTATS 167 Inject A Triclosan B Triclosan Figure 1. Chromatograms of (a) reference triclosan standard, and (b) that of a deodorant stick. cate injections of reference standard solutions of triclosan and TCC in a single day, and over a period of three to four days. Chromatographic response precision data are shown in Table III. Data indicate that precision of the procedure is very good. Accuracy of the procedure was tested by addition of known amounts of triclosan and TCC to deodorant sticks and deodorant soaps, respectively. For this study triclosan was added to a deodorant stick placebo sample, whereas TCC was added to a deodorant soap sample which was previously assayed. The data are presented in Table IV. 168 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Inject A •dTCC Iniect B Figure 2. Chromatograms of (a) reference TCC standard, and (b) that of a deodorant soap. The recovery data indicate that the accuracy of the •rocedure is excellent. The precision of the assay procedure was further determined by analyzing a commercial deodorant stick for triclosan and a deodorant soap for TCC. For five replicate assays the coefficient of variation for triclosan was + 3% and that for TCC was 0.38%. SPECIFICITY OF THE ASSAY PROCEDURE Although it was established that the procedure is capable of discriminating the bacteriostats from the inactive matrix, perfumes, etc., it was necessary to show that the procedure has the capability of discriminating the bacteriostats from the degradation products of the bacteriostats and from those of the perfumes or any other ingredient present in the deodorants. Table III Reproducibility Data of Chromatographic Responses • Within-Run (n = 6) Between-Run (n = 4) Retention time Area Retention Time Area Triclosan 483 _+ 8.1 572,870 _+ 10,983 496 _+ 18 548,166 _+ 24,654 % CV 0.84 0.96 1.8 4.5 TCC 468 _+ 2.4 at 340,877 _+ 1,811 473 _+ 9.8 365,393 + 29,748 % CV 0.26 0.27 1.04 8.1 •Retention time and area are expressed in seconds and electronic counts, respectively. The amounts of triclosan and TCC injected were 1.26 and 0.258/zg respectively. HPLC ANALYSIS OF BACTERIOSTATS 169 Inject TCC Figure 3. Chromatogram of (a) deodorant soap containing TCC and triclosan. Chromatographic conditions: mobile phase, 60:30:10 (V/V), methanol:water:ethylacetate; flow rate, 1 ml/min., other conditions as in Table I. For this purpose the bacteriostats and the deodorant samples containing bacteriostats were artificially degraded as described in the experimental section, and the degraded solutions were analyzed. In each case the identity and the purity of the bacteriostats were confirmed by the absorbance ratio technique. Based on the UV spectra of the 170 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS HPLC ANALYSIS OF BACTERIOSTATS 171 Table IV Recoveries of Triclosan and TCC from Spiked Samples • Triclosan 2 TCC 3 Amount Added Amount Added (/•g/injection) % Recovered (/•g/injection) % Recovered 1.007 100.6 0.155 100.2 1.26 100.3 0.206 99.9 1.51 100.4 0.258 99.9 •AII recovery data represent an average of three determinations. Range refers to recoveries from all determinations. 2Range: 99.8-101.1. 3Range: 99.4-100.6. bacteriostats in the HPLC mobile phase (Figure 4), the wavelengths selected for absorbance ratio determination were 266, 288, 300 nm for triclosan and 240, 270, 280 nm for TCC, respectively. The results of the chemical degradation of TCC and triclosan are described below: TCC No significant degradation was found upon heating of TCC. Some degradation products were observed upon hydrolysis (acid and base) probably from p-chloroanil- ine, 3, 4-dichloroaniline. Oxidation by potassium monopersulfate showed some degradation products. However, the identity of the degradation products has not been established here. Photolysis produced several degradation products. The degradation products were not identified in this study but previous studies on the photolysis of TCC in methanol have indicated that a loss of chlorine from TCC occurs, confirmed by the identification of chloride ion (7). This will probably result in yielding several dechlorinated carbanilide products. In all the degradation studies carried out here no interference was rendered to the analytical peak by any degradation products as determined by the absorbance ratios summarized in Table V. Data shown in Table V clearly demonstrate that the present procedure is stability- indicating for TCC. The analytical peak identified as TCC is pure and free from interferences in samples that have been subjected to vigorous degradation conditions of heat, acid and base hydrolysis, oxidation and photolysis. TCC degradation was observed in the latter cases. Triclosan Similar degradation studies were carried out for triclosan and the deodorant sticks containing triclosan. The absorbance ratios of the analytical peak were determined in all cases. The absorbance ratios of a reference triclosan standard between 288/266 nm and 288/300 nm were 3.17 and 2.78, respectively. No interference was observed in any of these studies and the triclosan peak was found to be pure in all cases. The absorbance ratios of the bacteriostat peak did not differ more than + 10% from those of the reference standard in any of the degraded samples. 172 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table V Absorbance Ratios of Reference TCC and Those of the Analytical Peak in the Degraded Solutions of TCC and Deodorant Soaps Absorbance Ratio Sample 270/240 (nm) 270/280 (nm) TCC Ref. standard 6.28 1.55 Degraded samples TCC (heat) 6.21 1.60 Soap 5.92 1.56 TCC (acid hydrolysis) 6.15 1.59 Soap 6.44 1.51 TCC (base hydrolysis) 6.13 1.58 Soap 6.18 1.56 TCC (oxidation) 6.03 1.55 Soap 6.31 1.55 TCC (photolysis) • 6.22 1.58 Soap 6.21 1.55 •Data from methanolic solution. The UV exposure to TCC on plates resulted in complete degradation of TCC and hence no absorbance ratios of the analytical peak could be provided. Three deodorant stick samples prepared in the laboratory (A, B & C) to contain 0.25% triclosan and two samples obtained commercially (D & E) were analyzed. The results are shown in Table VI. Table VI Assay of Triclosan in Deodorant Sticks (% W/W) Product Triclosan Deodrant Stick A 0.258 B 0.256 C 0.254 D 0.265 E 0.261 A number of deodorant soaps were analyzed for TCC. The results are summarized in Table VII. SUMMARY AND CONCLUSION The procedure described above is precise, accurate and specific for both qualitative and quantitative assay of triclosan and TCC. No interferences or any special difficulty have been encountered for the assay of the above bacteriostats in the commercial products we analyzed. However, it must be borne in mind that the deodorant samples (stick or soaps) contain many lipophillic components, and over a period of time these components tend to be deposited in the column. This could change both the nature and the efficiency of the chromatographic column. Thus, one should monitor the efficiency of the column periodically, and some form of clean up procedure should be adopted to regenerate the column. Also, wide varieties of perfumes are used in these products and these perfumes are composed of several components of different [...]... of a is not fullyknown,absorbance determination the analytical ratio of peakshould be doneto confirmboth identity andthe purityof the analytical peak;alternately, the analytical peakshouldbe collected and the 1stor 2nd derivative spectrum taken .In manyinstances procedure this may help in findinginterfering components the in analytical peak.These spectra shown Figure If anyinterference noticed are in. . .HPLC ANALYSIS OF Table BACTERIOSTATS 173 VII Assay TCCin Deodorant (%W/W)1 of Soaps Name TCC A 1.6 B 0.6 C 0.5 D 1.4 E 0.5 F G 1.03 0.47 H 1.04 I 0.39 J 0.16 tall soaps common names were brand obtained local from supermarkets "C"also Soap contained W/W 0.14% of triclosan Sample was laboratory "F" a prepared sample contain TCC to 1.0% hydrophobicity whichtendto interfere with the... anyinterference noticed are in 4 is a slightmodification the solvent of system alleviate problem mostcases may the in In certaincircumstancesshoulder a may be observed eitherin the leadingor tailingedge of theanalytical peak Quantitation peakheight by method prove will moreaccurate in thesecases thisreportwe haveprovided In dataon linearity, precision, etc.,by peak areaonly; however, have we tested... Determination germicides soaps detergents,J of in and Amer Chem Oil Soc., 513(1961) 38, (2) M B Graber,I I Domsky,and M E Ginn, TLC methodfor the identification germicides of in personal products,J care Amer Chem 46, 529(1969), Oil Soc., (andthereferences therein) (3) F X Demers and R L Yates, Antimicrobials: Identificationof 3,4,4'-trichlorocarbanilide and 4,4'-dichloro-3-(trifluoromethyl) carbanilide deodorant. .. 3,4,4'-trichlorocarbanilide and 4,4'-dichloro-3-(trifluoromethyl) carbanilide deodorant in bars,J Cosmet Soc Chem., 659(1977) 28, (4) T Wolf andD Semionow, Rapidliquidchromatography bacteriostats, Cosmet of J Soc Chem., 24, 363(1973) (5) E D George, HillierandS Krishnan, E.J Analysis trichlorocarbanilide triclosan soaps of and in by reverse phase pressure high liquid chromatography,J Oil Chem 57, 131 Amer... triclosan soaps of and in by reverse phase pressure high liquid chromatography,J Oil Chem 57, 131 Amer $oc., (1980) (6) T R Koziol, T JacobandR G Achari, J Ion-pair liquidchromatographic of decongestants assay andantihistamines,J Pharm.Sci., 1135 68, (1979) (7) Personal communication, MonsantoChemical Co., St Louis,Mo . taken. In many instances this procedure may help in finding interfering components in the analytical peak. These spectra are shown in Figure 4. If any interference. these perfumes are composed of several components of different HPLC ANALYSIS OF BACTERIOSTATS 173 Table VII Assay of TCC in Deodorant Soaps (% W/W) 1