`
`Determination of micelle/water partition coefficients
`of cosmetic preservatives
`Optimisation of the capillary electrophoretic method
`A. Bianco Prevot a;(cid:3)
`, E. Pramauro a, M. Gallarate b, M.E. Carlotti b, G. Orio a
`a Dipartimento di Chimica Analitica, V. Giuria 5, 10125 Torino, Italy
`b Dipartimento di Scienza e Tecnologia del Farmaco, V. Giuria 9, 10125 Torino, Italy
`
`Received 28 July 1999; received in revised form 3 January 2000; accepted 8 January 2000
`
`Abstract
`
`The hydrophobicity of a group of molecules used as preservatives for aqueous cosmetics, containing surfactants in the form
`of micelles, was evaluated by means of micellar electrokinetic capillary chromatography (MECC). Five different preservatives
`mixtures were examined, with a composition analogous to that employed in the preservation of cosmetic formulations. Sodium
`dodecyl sulphate and sodium dodecyl(polyoxyethylene)1–4 sulphate were used as model surfactant and surfactant present in
`real cosmetic samples, respectively. The appropriate choice of the micellar velocity marker turned out to be a crucial point
`for the application of the MECC approach. © 2000 Elsevier Science B.V. All rights reserved.
`
`Keywords: Parabens; MECC; Surfactants; Partition coefficients
`
`1. Introduction
`
`The aim of preservation of aqueous based cosmetics
`is to ensure that they are microbiologically safe and
`stable [1]. Adequately preserved products help to pre-
`vent microbial contamination during manufacturing
`and use, which minimises the potential health hazards
`to consumers. Up to now the choice of optimal preser-
`vative molecules or preservatives mixture is based on
`empirical testing and is often viewed as an art rather
`than a science. As a consequence, no correlation has
`ever been established between structure and properties
`of the molecules normally employed as preservatives.
`
`Corresponding author. Tel.: C39-116-707634;
`(cid:3)
`fax: C39-116-707615.
`E-mail address: bianco@ch.unito.it (A. Bianco Prevot)
`
`In this research the attention has been focused
`on the preservation of water-based products such as
`shampoos, which mainly consist of micellar surfac-
`tant solutions. The partition between aqueous and
`micellar phase of a group of molecules that exhibit
`a preservative action against a wide spectrum of
`moulds, yeast and bacteria was examined. It can-
`not be excluded that their activity could be related
`to the extent of their presence in water, almost the
`only phase where the micro-organisms can develop.
`Five different preservatives mixtures were charac-
`terised; all of them contained four parabens (esters of
`4-hydroxybenzoic acid) and different combinations of
`other four antibacterial agents. This research can be
`considered as an analytical preliminary study concern-
`ing the assessment of a possible relationship between
`antimicrobial activity and micelle/water partitioning
`of aromatic preservatives. The micelle/water partition
`
`0003-2670/00/$ – see front matter © 2000 Elsevier Science B.V. All rights reserved.
`PII: S 0 0 0 3 - 2 6 7 0 ( 0 0 ) 0 0 7 5 5 - 8
`
`Page 1 of 8
`
`SENJU EXHIBIT 2094
`LUPIN v. SENJU
`IPR2015-01097
`
`
`
`142
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`coefficients of the preservatives were determined by
`means of micellar electrokinetic capillary chromatog-
`raphy (MECC) according to the model proposed by
`Terabe [2,3]. Such a method was applied to determine
`the micelle–solute interaction for a wide number of
`compounds [4], with an extension to cationic and
`anionic solutes [5,6]. It was also employed to study
`the effect of physico-chemical properties and molec-
`ular structure of the micelle–water partitioning [7].
`Moreover, MECC presents an interesting alternative
`to micellar liquid chromatography (MLC) to estimate
`n-octanol/water partition coefficients [8,9]. Compared
`to MLC, MECC has the advantages related to the
`absence of a stationary phase: faster pre-conditioning
`and cleaning of the system, higher reproducibility and
`decreased error caused by intercolumn variability [8].
`MECC is also a useful method for the separation
`and analysis of the parabens in real samples, in addi-
`tion to MLC [10,11]; it was also employed to deter-
`mine the imidazolidinyl urea, one of the mostly used
`preservatives, in cosmetic preparations containing the
`parabens [12,13].
`In the present study sodium dodecyl sulphate (SDS)
`was chosen as model surfactant system because it has
`been widely used in the detergency field and it has been
`extensively studied and characterised; moreover the
`analytical method here proposed has been developed
`using SDS micelles.
`Successively two selected mixtures were analysed
`also in a micellar solution of sodium dodecyl(polyoxy-
`ethylene)1–4 sulphate, which is actually used in deter-
`gency products as a valid alternative to SDS because
`it is less irritating.
`
`2. Experimental
`
`2.1. Reagents
`
`Louis, MO), sodium dodecyl(polyoxyethylene)1–4
`sulphate, Zetesol 1070® (SDES, aqueous solution of
`70% w/w SDES) by Zschimmer & Schwarz S.p.A.
`(Italy). Sodium dihydrogen phosphate and sodium
`tetraborate were purchased from Fluka (St. Louis,
`MO); methanol and Sudan III
`from Sigma (St.
`Louis, MO); 1–(4-n-butyl-2-pyridylazo)-2-naphthol
`(PAN-C4) and 1-(4-n-octyl-2-pyridylazo)-2-naphthol
`(PAN-C8) were synthesised in our lab.
`Bidistilled water from a Milli-Q, water purifica-
`tion system (Millipore Corp., Bedford, MA) was used
`throughout the work.
`
`2.2. Apparatus
`
`The electrophoretic analysis was carried out on
`an electrophoresis apparatus Eureka 2100 model
`(Kontron) equipped with Diode Array (190–500 nm)
`detector. A fused silica capillary with 62.5 cm of
`length (50 cm to the detection window) and 0.050 mm
`of internal diameter was used (Supelco Celect).
`A Metrohm 654 pH-meter, equipped with a
`combined glass–calomel electrode was employed for
`pH measurements. All buffers and samples were fil-
`tered (0.22 mm cellulose acetate filters, Millipore) and
`degassed by an ultrasonic bath (BRANSON 2200).
`Critical micellar concentration (c.m.c.) measurements
`were performed with a digital-tensiometer
`(K10,
`KRÜSS).
`
`2.3. Electrolyte solutions
`
`The proper amount of surfactant was dissolved in
`a buffer solution of sodium dihydrogen phosphate
`
`−2 M and sodium tetraborate 8.3(cid:2)10−3 M
`3.30(cid:2)10
`(pHD6.90). The employed surfactant concentrations
`ranged from 20 to 100 mM for both SDS and SDES.
`
`reagents were of analytical-reagent grade.
`All
`Methylparaben, ethylparaben, propylparaben, butyl-
`paraben, benzyl alcohol, 2-phenylethanol, 2-phenoxy-
`ethanol,
`4-chloro-3-methylphenol,
`4-chloro-3,5-
`dimethylphenol, o-phenylphenol were purchased from
`0
`0
`-trichloro-2
`-
`Sigma Aldrich (St. Louis, MO); 2,4,4
`hydroxy-diphenylether (Triclosan®) was supplied by
`Res Pharma (Italy). Their molecular structures are
`reported in Fig. 1. SDS was supplied by Sigma (St.
`
`2.4. Sample preparation
`A stock solution 1.0(cid:2)10
`−3 M of each analyte in
`buffered SDS or SDES 20.0 mM was prepared; the
`same surfactant solutions were used for the dilution
`of the sample. The mixture compositions were cho-
`sen on the basis of preservative products employed
`in the cosmetic field. An effort was made to supply
`the widest antimicrobial activity spectrum. The four
`
`Page 2 of 8
`
`
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`143
`
`Fig. 1. Examined cosmetic preservatives: molecular structure and formula.
`
`parabens present in all the mixtures were chosen be-
`cause of their activity against moulds, yeast and Gram
`(C) bacteria. The other preservatives were then added
`to broaden the spectrum of action also against Gram
`(−) bacteria. The preservatives concentration in the
`analysed mixtures was chosen taking into account the
`ratio preservative/surfactant generally encountered
`in detergency cosmetic products, keeping as refer-
`ence the parabens concentration (2.0(cid:2)10
`−5 M each).
`About 1% V/V of methanol and 1(cid:2)10
`−5 M of Sudan
`III (or PAN-Cn) were added to each sample solution
`as electroosmotic (veo) and micellar (vMC) veloc-
`
`ity marker, respectively. The overall concentrations
`of both preservatives and surfactant were consider-
`ably lower than in real samples (i.e. preservatives
`0.1–0.8% w/w, surfactant 10–20% w/w) to ensure the
`appropriate experimental conditions for MECC.
`Tables 1 and 2 report the five examined mixtures
`and the analytes concentration.
`
`2.5. Electrophoretic procedure
`
`When a new capillary was used, it was washed
`with water for 10 min, activated with 0.1 M NaOH
`
`Page 3 of 8
`
`
`
`144
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`Table 1
`Molecules present in the examined mixtures
`
`Mixture Composition
`ParabensC2-phenoxyethanolCTriclosan®
`ParabensC2-phenylethanolCTriclosan®
`ParabensCbenzyl alcoholC2-phenoxyethanol
`ParabensCbenzyl alcoholC4-chloro-3-methylphenol
`ParabensCo-phenylphenolC4-chloro-3,5-dimethylphenol
`
`1
`2
`3
`4
`5
`
`Table 2
`Analytes concentration in the examined mixtures
`
`Preservative
`
`R-paraben
`Benzyl alcohol
`2-Phenoxyethanol
`2-Phenylethanol
`o-Phenylphenol
`4-Chloro-3-methylphenol
`4-Chloro-3,5-dimethylphenol
`Triclosan®
`
`Concentration (M)
`2.0(cid:2)10
`−5
`3.2(cid:2)10
`−5
`2.5(cid:2)10
`−5
`2.8(cid:2)10
`−5
`4.1(cid:2)10
`−6
`4.9(cid:2)10
`−6
`1.1(cid:2)10
`−5
`3.6(cid:2)10
`−6
`
`for 10 min and rinsed with water for 10 min. Then it
`was equilibrated with the running buffer for 15 min at
`20 kV. The same treatment was applied daily before
`starting the analysis. Each time the buffer composition
`was changed, the capillary was washed with water for
`3 min, washed with methanol for 3 min, rinsed with
`water for 3 min, activated with 0.1 M NaOH for 3 min,
`rinsed with water for 3 min. Then it was equilibrated
`flushing the running buffer for 5 min and applying a
`constant potential of 30 kV for 20 min. The capillary
`was rinsed for 3 min with the buffer between runs.
`The sample was introduced by applying a negative
`pressure at the cathodic capillary end for 4 s. The sep-
`arations were run at 20, 22, 25, 27 and 30 kV keeping
`the capillary temperature constant at 35.0(cid:6)0.1
`(cid:14)
`C.
`The diode-array detection was performed in the
`200–350 nm wavelength range.
`
`3. Results and discussion
`
`3.1. Preliminary analytes separation
`
`the
`Before starting with the determination of
`partition coefficient, an electrophoretic analysis of
`a mixture containing all the examined compounds
`
`together with methanol and Sudan III was run. This
`was in order to establish their migration order and
`their resolution. All the compounds were identified
`by comparing the spectrum and the migration time of
`each peak with those recorded for the solutions of the
`pure standards. A 60 mM SDS solution buffered at
`pH 6.90 was chosen as the model surfactant system;
`the operating voltage was 30 kV; methanol and Sudan
`III were used as electroosmotic velocity marker and
`micellar velocity marker, respectively. Since at this
`pH all the examined molecules are in their neutral
`form their migration times differ only because they
`have different micelle/water partition coefficients. All
`the analytes were separated (see Fig. 2), falling within
`the elution window with the exception of Triclosan®,
`which migrates slower than Sudan III and partially
`coeluted with it.
`
`Fig. 2. Pherogram of the examined preservatives. SDS 60 mM;
`pHD6.90; applied voltageD30 kV; UV detection at 220 nm. Migra-
`tion order: (1) methanol, (2) benzyl alcohol, (3) 2-fenoxyethanol,
`(4) 2-fenylethanol,
`(5) methylparaben,
`(6) ethylparaben,
`(7)
`4-chloro-3-methylphenol,
`(8) propylparaben,
`(9) 4-chloro-3,
`5-dimethylphenol, (10) o-phenylphenol, (11) butylparaben, (12)
`Sudan III and (13) Triclosan®.
`
`Page 4 of 8
`
`
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`145
`
`3.2. Examination of the proposed analytical method
`
`The applicability of the electrophoretic model was
`checked by running preliminary measurements on a
`mixture containing only the four parabens (all present
`in every mixture further considered), dissolved in SDS
`together with methanol and Sudan III. Prior to the de-
`0
`) at differ-
`termination of the solute capacity factors (k
`ent surfactant concentrations, the existence of possible
`temperature variations was verified by examining the
`capacity factor behaviour as a function of both the
`applied voltage and the electroosmotic velocity.
`Fig. 3 reports the results obtained in the analysis
`run with 0.050 M of SDS at five different applied volt-
`0
`values and elec-
`ages. A linear correlation between k
`troosmotic velocity can be observed, with the slope
`increasing with the hydrophobicity of the molecule;
`0
`an analogous behaviour was observed when plotting k
`versus applied voltage. In order to eliminate the effect
`0
`, each line of the plot was extrapolated to
`of veo on k
`veoD0, yielding the corresponding k0
`0 values [3]. The
`same procedure was followed for the other examined
`SDS concentrations.
`Plots of k0
`0 as a function of the surfactant concen-
`tration are linear according to the following equation:
`D KMW Nv.Csf − c:m:c:/
`k0
`
`(1)
`
`0
`
`where KMW is the micelle/water partition coefficient,
`Nv is the surfactant partial molar volume and Csf is the
`total surfactant concentration. Fig. 4 shows that a good
`linear correlation exists although no common intersec-
`tion is observed. Moreover the values of the intersec-
`tion on the x-axis are negative resulting in meaningless
`negative values of the c.m.c.
`Considering the Terabe’s equation [3] adopted to
`calculate k0
`0, it clearly demonstrates the importance of
`an accurate evaluation of both electroosmotic and mi-
`cellar velocity; in particular, a slight variation of tMC
`exerts a strong effect on the k0
`0 evaluation. Acetone
`was used instead of methanol as veo marker, but no
`changes were observed. On the other hand, PAN-C4
`was chosen as vMC marker; this molecule possesses an
`amphiphilic structure, which allows a suitable strong
`interaction with the micelles. In absence of added sur-
`factant, PAN-C4 and Sudan III migrated with the same
`velocity as acetone and methanol, thus evidencing that
`their migration was only due to the electroosmotic ef-
`fect. In the presence of surfactant micelles PAN-C4
`migrates slightly slower than Sudan III. Due to its
`very low concentration (about 1.0(cid:2)10
`−5 M) it cannot
`significantly modify the aggregate. On the contrary,
`it was observed that more concentrated organic mod-
`ifiers, when added to the buffer to extend the elution
`window [14], can alter the micellar structure. In this
`
`Fig. 3. Capacity factor behaviour in function of the electroosmotic velocity for the four parabens. SDS 50 mM; pHD6.90; applied voltages:
`20, 22, 25, 27, 30 kV.
`
`Page 5 of 8
`
`
`
`146
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`Fig. 4. Partitioning coefficient for the four parabens in SDS micelles; micellar velocity marker: Sudan III.
`
`case the PAN-C4 slower migration can only be a con-
`sequence of its quantitative binding to the micelle, thus
`leading to a better evaluation of vMC.
`
`3.3. Determination of binding constants (KB) in SDS
`
`Fig. 5 reports the results obtained for the examined
`parabens in the presence of PAN-C4 as vMC marker.
`Lower k0
`0 values have been obtained when PAN-C4
`
`was used, with a shift of the k0
`0 versus veo plots lead-
`ing to their intersection on the positive side of the
`x-axis. This effect increases as the hydrophobicity
`of the investigated substrate increases. Similar plots
`were also obtained for the five examined preservative
`mixtures. From the slopes of each straight line the
`product KMW Nv can be obtained for each preserva-
`tive. These values are a good approximation of the
`solute/micelle binding constant, KB, defined by the
`Berezin’s equation [15]:
`
`Fig. 5. Partitioning coefficient for the four parabens in SDS micelles; micellar velocity marker: PAN-C4.
`
`Page 6 of 8
`
`
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`147
`
`Table 3
`Binding constants of the molecules in SDS in all the examined systems
`
`Compound
`
`Mixture 1
`
`Mixture 2
`
`Mixture 3
`
`Mixture 4
`
`Mixture 5
`
`42
`106
`300
`795
`
`22
`
`Methylparaben
`Ethylparaben
`Propylparaben
`Butylparaben
`Benzylic alcohol
`2-Phenylethanol
`2-Phenoxyethanol
`4-Chloro-3-methylphenol
`4-Chloro-3,5-dimethylphenol
`o-Phenylphenol
`
`KB D .KMW − 1/Nv
`
`38
`98
`255
`613
`
`22
`
`(2)
`
`In this way quantitative information on the
`solute–micelle interactions can be obtained also for
`surfactants whose Nv value is unknown, such as SDES
`examined later. Table 3 reports the KB values de-
`termined as described for all the components in the
`different mixtures. In the case of molecules present
`in more than one mixture their partitioning seems
`not significantly influenced by the global composi-
`tion: under the examined experimental conditions no
`‘salting out’ of the more hydrophilic molecule by the
`action of the more hydrophobic solutes was observed.
`Due to its high hydrophobicity (it elutes very close to
`the vMC marker) it was not possible to evaluate the
`partition coefficient for the Triclosan®.
`
`3.4. Determination of binding constants in SDES
`
`After the optimisation of the method with SDS, the
`attention was directed to the examination of SDES,
`which is widely used as primary surfactant in most
`commercially available shampoos and bath products.
`Two mixtures (no. 3 and 4 in Table 1) were exam-
`ined, where the more hydrophobic butylparaben was
`not included. The preservative action is in fact mainly
`exerted in aqueous media, where micro-organisms
`can generally find a suitable growing medium; com-
`pounds present almost only in the micellar phase
`have thus only a small practical interest as preserva-
`tive agents. Moreover, the mixture of parabens and
`phenoxyethanol (mixture no. 3) is present in many
`preservative kits commercially available. For these
`
`43
`102
`283
`790
`13
`
`23
`
`44
`113
`304
`878
`13
`
`170
`
`44
`110
`384
`970
`
`358
`460
`
`reasons, the evaluation of the antimicrobial activity of
`mixtures no. 3 and 4 is actually under investigation.
`Also in this case the choice of the vMC marker was
`essential in order to achieve results fitting the cho-
`sen model. Using PAN-C4 as vMC marker, the same
`problems already observed for SDS in presence of Su-
`dan III, appeared. The introduction of ethoxy moieties
`makes SDES more hydrophilic than SDS and in this
`system PAN-C4 binding could be not quantitative. For
`this reason the more hydrophobic PAN-C8 was tested
`and it gave rise to a more satisfactory vMC measure-
`ment. Plots analogous to those reported in Fig. 5 were
`obtained and the resulting binding constants are sum-
`marised in Table 4. Only a slight reduction in the par-
`titioning is observed for all the analytes with respect
`to their interaction with SDS micelles, without any
`change in their relative hydrophobicity.
`
`3.5. Determination of c.m.c. values of the examined
`micellar buffered systems
`
`Even if this method has not been widely applied to
`determine the c.m.c., these data were estimated from
`
`Table 4
`Binding constants for mixtures 3 and 4 in SDES
`
`Compound
`
`Mixture 3
`
`Mixture 4
`
`Methylparaben
`Ethylparaben
`Propylparaben
`Benzyl alcohol
`2-Phenoxyethanol
`4-Chloro-3-methylphenol
`
`40
`106
`270
`12
`20
`
`36
`90
`239
`11
`
`180
`
`Page 7 of 8
`
`
`
`148
`
`A.B. Prevot et al. / Analytica Chimica Acta 412 (2000) 141–148
`
`the plots k0
`0 versus Csf and compared with those ob-
`tained by independent experiments. Surface tension
`measurements were run at 35.0(cid:6)0.1
`(cid:14)
`C in buffered
`surfactant solutions. The contribution of preservatives
`was assumed to be negligible, due to their relatively
`low concentration (about 1(cid:2)10
`−4 M) with respect to
`the solution ionic strength.
`
`−3 and 6.1(cid:2)10−3 M
`C.m.c. values of 2.3(cid:2)10
`were determined for SDS and SDES, respectively,
`whereas MECC experiments gave values in the range
`
`−3 M for SDS and 4.5–7.5(cid:2)10−3 M for
`1.5–4.0(cid:2)10
`SDES. These results are quite satisfactory consider-
`ing that differences even of one order of magnitude
`can be noted between the c.m.c. values determined,
`for the same surfactant, using different techniques
`[16].
`
`4. Conclusions
`
`The above-discussed results clearly evidenced that
`MECC is a suitable analytical technique for the sepa-
`ration of a complex mixture of preservatives. In order
`to apply MECC to the evaluation of molecular hy-
`drophobicity, the proper choice of the vMC marker has
`been demonstrated to be of fundamental importance.
`An amphiphilic molecule with a tunable alkyl chain
`such as PAN-Cn seems to be, in this case, the better
`solution.
`Moreover, the obtained results could probably allow
`to establish a correlation between the micelle/water
`partitioning and the anti microbial efficacy of different
`preservatives mixtures. For this purpose, studies on
`
`microbial survival are actually in progress on systems
`analogous to those examined in the present research.
`
`Acknowledgements
`
`from Ministero Università e
`Financial support
`Ricerca Scientifica Tecnologica (MURST).
`
`References
`
`[1] D.K. Brannan, J. Soc. Cosmet. Chem. 46 (1995) 199.
`[2] S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchija, T. Ando,
`Anal. Chem. 56 (1984) 111.
`[3] S. Terabe, K. Otsuka, T. Ando, Anal. Chem. 57 (1985) 834.
`[4] M.A. Garcia, M.L. Marina, J.C. Diez-Masa, J. Chromatogr.
`A 732 (1996) 345.
`[5] M.G. Khaledi, S.C. Smith, J.K. Strasters, Anal. Chem. 63
`(1991) 1820.
`[6] J.K. Strasters, M.G. Khaledi, Anal. Chem. 63 (1991) 2503.
`[7] N. Chen, Y. Zhang, S. Terabe, T. Nakagawa, J. Chromatogr.
`A 678 (1994) 327.
`[8] B.J. Herbert, J.G. Dorsey, Anal. Chem. 67 (1995) 744.
`[9] M.A. Garcia, J.C. Diez-Masa, M.L. Marina, J Chromatogr.
`A 742 (1996) 251.
`[10] R.J. Geise, N.I. Machnicki, J. Cap. Elec. 2 (1995) 69.
`[11] J.F.
`Noguera-Orti,
`R.M.
`Villanueva-Camañas,
`Ramis-Ramos, J. Chrom. Sci. 37 (1999) 83.
`[12] R.J. Geise, N.I. Machnicki, R.M. Ianniello, J. Soc. Cosmet.
`Chem. 45 (1994) 173.
`[13] G.M. Michalakis, E.F. Barry, J. Soc. Cosmet. Chem. 45 (1994)
`193.
`[14] P.G.H.M. Muijselaar, H.A. Claessens, C.A. Cramers, J. Chro-
`matogr. A 696 (1995) 273 and references 10 and 11 here in.
`[15] I. Berezin, Russ. Chem. Rev. 42 (1973) 778.
`[16] P. Mukerjee, K.J. Mysels, Critical Micelle Concentrations
`of Aqueous Surfactant Systems, NSRDS-NBS 36, US
`Government Printing Office, Washington DC.
`
`G.
`
`Page 8 of 8