The distribution and activity of

J. Soc. Cosmetic Chemists 18 207-214 (1967} 1967 Society of Cosmetic Chemists of Great Britain The distribution and activity of benzoic systems acid in some emulsified R. A. ANDERSON and CLARA E. CHOW* Synopsis--The distribution of benzoic acid between arachis oil, isopropyl myristate or liquid paraffin and water or 20 per cent glycerin in water has been investigated. The solute was found to dimerize in the liquid paraffin phase. The fungistatic activity against Aspergillus niger of benzoic acid in emulsified systems depends on the concentration free in the aqueous phase. The benzoate ion seems to have some activity. Increased concentrations of benzoic acid are required in the presence of glycerin. Propylene glycol enhances the activity of benzoic acid and at a concentration of 15 per cent prevents the growth of A. niger in the absence of benzoic acid. INTRODUCTION Many pharmaceutical and cosmetic emulsions require the inclusion of a preservative to prevent microbial deterioration. The activity of a preservative will be affected by its interaction with an emulsifying agent (1,2) and by its distribution between the oily and aqueous phases of the emulsion (3,4). Bean et al (4) have calculated the partition coefficients (o K w -- C Coil water ) for several preservatives and suggest that for any given overall concentration of preservative the concentration in the aqueous phase may be calculated from the expression, Cw = c ( +1) (Kw ø + 1) where Cw = concentration in the aqueous phase (per cent w/v), C = overall concentration (per cent w/v), *Pharmacy Department, University of Sydney, N.S.W., Australia. 207

2O8 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Kø w oil ' water partition coefficient, and = oil: water ratio. This expression is not appropriate when the preservative dimerizes in the oil phase, and does not allow for preservative inactivated by ionization or interaction with emulsifier molecules. EXPERIMENTAL Liquid paraffin, arachis oil, glycerin and propylene glycol complied with the requirements of the British Pharmacopoeiand isopropyl myristate with those of the British Pharmaceutical Codex. Brij $5 was used as received. Benzoic acid was recrystallized from water (m.p. 122ø). All chemicals were A.R. grade. Water was distilled from a Scorah all-glas still. Estimation of benzoic acid Aqueous solutions were diluted with 0.01 N HC1 and the absorbance measured in a Unicam S.P. 700 spectrophotometer at 274.5 m. Beer's law was found to hold up to at least 1.0 x 10-3 moles/i; glycerin and surfactants in the concentrations present in the diluted solutions did not interfere with the estimations. The concentration of benzoic acid in the oil phase was usually calculated from the decrease in concentration of the aqueous phase. For liquid paraffin phases this procedure was shown to be valid by direct spectrophotometric determination after suitable dilution with isooctane. DETERMINATION OF PARTITION COEFFICIENT Suitable volumes of oil and aqueous phase containing an accurately known concentration of benzoic acid in 0.01 N HC1 were pipetted into glass-stoppered bottles and agitated by a wrist-action shaker for about 1 hr at room temperature. The bottles were then transferred to a water bath at 25 + 0.1 ø for at least 7 days, being shaken 3 or 4 times during this period, and then the concentration of benzoic acid remaining in each aqueous phase was determined. Concentrations were redetermined after shaking and standing for a further one to three days to establish that equilibrium had been attained. For arachis oil and isopropyl myristate equal volumes of oil and of aqueous phase were used. The original concentration of benzoic acid in the aqueous phase was usually 20 millimoles/l, but in some experiments was

THE DISTRIBUTION AND ACTIVITY OF BENZOIC ACID 209 reduced to 10-15 millimoles/1 to show that the partition coefficient was independent of concentration. For liquid paraffin systems the paraffin to aqueous phase volume ratio was increased to 6:1, and because the distribution was concentration dependent, the original benzoic acid concentration in the aqueous phase was varied from 5 to 25 millimoles/1. In some experiments, 20 per cent glycerin was included in the aqueous phase. Microbiological evaluation A strain of Aspergillus niger was maintained at 25 ø on slopes of potato agar containing 2 per cent malt extract, and was transferred every two months. The organism also grew well at 25 ø on a medium containing 2 per cent malt extract and 0.1 M citrate ph 4.2 (or 5.2) solidified with 2 per cent agar. Attempts to determine a minimum inhibitory concentration of benzoic acid in this medium were unsuccessful; when the experiment was replicated it was found that the proportion of inocula which grew to form visible mycelia and spores gradually fell from 1 to 0 over a range of benzoic acid concentrations. However, the concentration required to prevent growth in half of six or more replicates was reproducible and this criterion, referred to later as FC 50, was chosen as a satisfactory means of assessing the effects of added agents on the fungistatic activity of benzoic acid. RESULTS AND DISCUSSION Partition coefficients of benzoic acid at 25 ø I Phases 0.01 N HC1 20 per cent glycerin in 0.01 N HC1 isopropyl myristate 9.87 7.0 Arachis oil 6.14 4,38 Liquid paraffin 2.6 1.7 per cent - per cent -t The partition coefficients of benzoic acid in arachis oil-water and isopropyl myristate-water systems are essentially independent of benzoic acid concentration and are calculated as the ratio of the concentration in the oil to that in the aqueous phase ( I). The value of 6.1 for the distribution between arachis oil and water is higher than the coefficient of 5.33

210 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS reported by Garrett & Woods (5); the discrepancy is probably due to the variable nature of vegetable oils. When liquid paraffin is the non-aqueous phase, the ratio of the concentrations in the two phases is not independent of concentration. However, the ratio of the square root of the concentration in paraffin to the concentration in the aqueous phase is constant ( II), suggesting that benzoic acid exists predominantly as dimers in the paraffin phase. Glasstone (6) found similar behaviour for the distribution of benzoic acid between benzene and water. Partition coefficient of benzoic acid in liquid paraffin-water systems at 25 ø aqueous Initial concentration Equilibrium concentration Equilibrium. Co K = Cw Co concentration aqueous phase* in paraffin* Cw (M/l) -t (mm/1)-t(%)- as mm/1 (Cw) as mm/1 (Co) as mm/1,, 25 6.17 3.14 0.51 9.08 0.287 2.6 20 5.31 2.45 0.46 9.32 0.295 2.67 15 4.53 1.75 0.39 9.23 0.292 2.64 10 3.50 1.08 0.31 9.4 0.296 2.68 *Ratio of oil: water ---- 6: 1 Average 9.3 0.29 2.6 II When the solute dimerizes in one of the phases, the numerical value of the coefficient depends on the units in which the concentrations are expressed and the value of 9.3 M-tl ( II) is equivalent to 0.29 mm-tl or 2.6ø/o -t. The last method of expression, though unusual, is probably most convenient in pharmaceutical and cosmetic formulation. The presence of glycerin in the aqueous phase lowers the partition coefficients, and the values in the presence and absence of 20 per cent glycerin are listed in I. FUNGISTATIC ACTIVITY The effect on antimicrobial activity of benzoic acid concentration is shown in III which records the results of eight separate experiments in which each concentration was replicated six times. The estimated FC50 from each experiment is shown at the bottom of the table. Other FC50 values given in s IV and V are determined from six replicates with the exception of the value for 20 per cent glycerin in the absence of oil and surfactant which is based on 24 replicates.

THE DISTRIBUTION AND ACTIVITY OF BENZOIC ACID 211 The activity of benzoic acid has been shown to vary with the ph of the system. The FC50s at ph 4.2 and 5.2 are listed in IV. The equivalent concentrations of unionized acid (using pk a = 4.2 in the Henderson buffer equation) have been calculated and are included in the table. These data The determination of the fungistatic activity of benzoic acid against A. niger: at 25 ø and ph 4.2 Total Numbers of positive results from six Number of concentration replicates positive results of benzoic from total of acid added (%) Expt. Expt. Expt. Expt. Expt. Expt. E t. Expt. 48 inocula 1 2 3 4 5 6 P 8 0.040 6 6 6 6 6 6 6 6 48 0.045 5 6 6 6 6 6 6 6 47 0.050 4 6 6 6 6 5 6 6 45 0.055 2 4 3 2 4 2 3 3 23 0.060 I 1 1 0 0 0 1 0 1 4 0.065 0 0 0 0 0 0 0 0 0 Estimated FC50 0.055 0.055 0.055 (percent) 0.0530.0570.0550.0530.0570.052 I III Fungistatic activities (as FC 50) of benzoic acid in the presence of propylene glycol or glycerin against A. niger at 25 ø IV FC 50 (per cent) ph Added agent and concentration Concentration Concentration of benzoic unbound and acid added unionized $.2 None 0.190 0.019 4.2 None 0.055 0.027 Propylene glycol 5 % 0.050... 10,, 0.035,, 15,, 0.000 Glycerin 20,, 0.072 0.040 supporthe conclusion of Evans and Dunbar (7) that less of the unionized acid is required at higher ph values because the benzoate ion also has some activity; they do not agree with the findings of Winsley and Waiters (8) that the antimicrobial activity depends only on the unionized benzoic acid over the range of ph 2.2 5.1.

212 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS All other determinations reported in this paper have been carried out at ph 4.2. Propylene glycol or glycerin is often included in pharmaceutical or cosmetic emulsions and the increased solubility of preservative in the aqueous phase leads to a reduction in the partition coefficient ( I). It has been suggested (9) that this makes more preservative available in the aqueous phase with consequent increase in antimicrobial activity. IV shows that propylene glycol augments the fungistatic activity of benzoic acid, and that concentrations above about 15 per cent have considerable activity even in the absence of any other agent. On the other hand, glycerin reduces the activity. The equivalent concentration of unionized benzoic acid shown in IV has been calculated using pk a' value (in 20 per cent glycerin) of 4.3 determined by the method of Albert and Serjeant (10). This FC 50 of unionized acid is about 1 times the FC 50 in the absence of glycerin. (The concentration of benzoate ion is only slightly greater in the 20 per cent glycerin.) It is suggested that in the same way that the increased solubility of benzoic acid in 20 per cent glycerin reduces the oil-water partition coefficients, it also reduces the availability of the preservative to the microbial biophase, and consequently a higher concentration is required. Brij $5 also reduces the availability of benzoic acid to the oil phase and to the biophase. Since Evans and Dunbar (7) and Mitchell and Brown (11) have shown that benzoic acid is distributed between the surfactant micelles and the true aqueous phase in a constant ratio which is independent of the overall benzoic acid concentration, it is possible to calculate from the solubilization data of Anderson and Slade {12) that 0.1 per cent Brij $5 will bind about 3.6 per cent of the unionized acid present in the aqueous phase. When 20 per cent glycerin is present in the aqueous phase the proportion of unionized acid bound by 0.1 per cent Brij is reduced to about 3 per cent {unpublishedata). These values have been used, along with the partition coefficients listed in I, to calculate the concentrations of unbound, unionized acid shown in V. Although the total concentration of benzoic acid required to give equal fungistatic activity ranges from 0.062 to 0.23 per cent in the aqueous phase (0.032 to 0.115 per cent overall), the eqhivalent concentrations calculated to be unbound and unionized in the aqueous phase vary over the much smaller range from 0.027 to 0.03 per cent. Similarly the results in the presence of 20 per cent glycerin range from 0.037 to 0.042 per cent unionized, unbound benzoic acid and show reasonable agreement with the 0.04 per cent found for a simple buffered system.

-- THE DISTRIBUTION AND ACTIVITY OF BENZOIC ACID 213 Fungistatic activities (as FC 50) against A. niger of benzoic acid in emulsified systems: at 25 ø and ph 4.2, V FC 50 (per cent aqueous phase) Phase ratio Concentration Concentration Aqueous phase Oil phase w: o of benzoic unbound and acid added unionized in aqueous phase Buffer only None 0.055 0.027 9.1 Brij 35 None 0.056 0.027 Liquid paraffin 2: 1 0.062 0.030 Liquid paraffin 1: 1 0.065 0.029 Arachis oil 2: 1 0.136 0.027 Arachis oil 1: 1 0.230 0.028 Isopropyl myristate 3:1 0.147 0.027 20 glycerin None 0.072 0.040 20 o glycerin None 0.076 0.042.nd Liquid paraffin 1: 1 0.080 0.041.1% Brij 35 Arachis oil 2: 1 0.145 0.036 isopropyl myristate 3: 1 0.152 0.037 These results suggesthat the fungistatic activity of benzoic acid in an emulsified system depends on the amount free in the aqueous phase, and that the amount required to compensate for losses into oil phases can be calculated if the appropriate partition coefficient is known. Contrary to the results of Bean et al (4, 13, 14), the phase volume ratio does not seem to affect the activity, although it should be pointed out that a different organism, a different method and a different preservative have been used in the present study; in addition the inclusion of surfactant molecules which concentrate at the oil-water interface would be expected to affect adsorption of microorganisms at the interface. ACKNOWLEDGEMENT This work has been supported by the New South Wales Pharmacy Research Trust. (Received.' 3rd October 1966) REFERENCES (1) Bolle, A. and Mirimanoff, A. J. Pharm. Pharmacol., 2 685 (1950). (2) Wedderburn, D. L. in Advances in Pharmaceutical Sciences, 1 195 (1964)(Academic Press, London).

214 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (3) Arkins, F. Mfg. Chem., 2t 51 (1950). (4) Bean, H. S., Heman-Ackah, S. M. and Thomas, J. J. Soc. Cosmetic Chemists, t6 15 (1965). (5) Garrett, E. R. and Woods, O. R. J. Am. Pharm. Assoc. Sci. Ed., 42 736 (1953). (6) Glasstone, S. Textbook of Physical Chemistry, 738 (1946)(Van Nostrand, New York). (7) Evans, W. P. and Dunbar, S. F. in Surface Activity and the Microbial Cell, S.C.I. Monograph No. 19, 169 (1965) (Society of Chemical Industry, London). (8) Winsley, B. E. and Walters, V. J. Pharm. Pharmacol., t7 225 (1965). (9) Hibbott, It. W. and Monks, J. J. Soc. Cosmetic Chemists, t2 2 (1961). (10) Albert, A. and Serjeant, E. P. Ionization Constants of Acids and Bases 80 (1962)(Methuen, London). (11) Mitchell, A. G. and Brown, K. J. Pharm. Pharmacol., t8 115 (1966). (12) Anderson, R. A. and Slade, A. H. Australian J. Pharm. Sci. Supp., 46 S53 (1965). (13) Bean, H. S., Richards, J.P. and Thomas, J. J. Boll. Chim. Farm., t01 339 (1962). (14) Bean, H. S. and Heman-Ackah, S. M. J. Pharm. Pharmacol., 16 58T (1964).