Characterization of the deposition of silicone copolymers on keratin fibers by streaming potential measurements

Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

Characterization of the deposition of silicone copolymers on keratin fibers by streaming potential measurements Anne D. Dussaud ∗ , Peter C. Breen, Kalman Koczo Momentive Performance Materials, 769 Old Saw Mill River Road, Tarrytown, NY 10591, USA

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

• Cationic silicone copolymer deposition on keratin monitored by streaming potential. • Amino silicone polyether block copolymer screened charges more efficiently than conventional aminosilicone. • Long ethylene oxide block facilitated the silicone block copolymer unfolding. • At low pH, silicone block copolymers overcompensated substrate surface charges.

a r t i c l e

i n f o

Article history: Received 10 February 2012 Received in revised form 29 April 2013 Accepted 30 April 2013 Available online 14 May 2013 Keywords: Cationic silicone Aminosilicone Deposition Streaming potential Keratin

a b s t r a c t This paper compares the deposition of two aminofunctional silicones on hair that are both widely used as textile finishes and hair conditioners. The two polymers had different structures: a linear amino-polyether-silicone block copolymer (ABn ) with hydrophilic character and a conventional pendant aminodimethicone (AMD) with the same amine content. The deposition (irreversible adsorption) on keratin fibers as a function of pH, concentration, and treatment time was studied using streaming potential measurements. The polymer deposition was also measured independently by analysis of the silicon element content of the treated hair samples. With ABn , the deposition reached a plateau after the reversal of the hair charge (overcompensation). In contrast, with AMD, the reversal of charge was not observed and the deposition was much higher, suggesting a more prominent role of hydrophobic interactions. A recent model of streaming potential was used to interpret the data. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The deposition of silicones from aqueous dispersions on solid substrates is of practical importance in textile and cosmetic industries. Numerous studies have addressed the deposition of uncharged silicone, typically polydimethylsiloxane, from aqueous dispersion containing anionic surfactants and organic cationic polymers where the deposition is controlled by the formation of insoluble coacervates [1,2]. Much less attention has been devoted

∗ Corresponding author. Tel.: +1 914 784 4832; fax: +1 914 784 4803. E-mail address: [email protected] (A.D. Dussaud). 0927-7757/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfa.2013.04.071

to the deposition of cationic silicone dispersions in water in the absence of organic cationic polymers although this chemistry is increasingly used in textile finishes and in hair conditioners [3,4]. Most cationic silicones are polydimethylsiloxanes containing amine groups which are either quaternized (permanent charge) or ionized as a function of pH. In conventional aminosilicone polymers, the amine functionality is present as side chains of the linear siloxane backbone (Fig. 1a). The number of siloxane units between amino groups is usually kept high (m > 20), because longer siloxane chains provide the enhanced spreading properties and a low coefficient of friction, which are beneficial for the finish performance. These copolymers remain very hydrophobic, requiring the addition of surfactants to disperse the aminosilicone in water.

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

CH3

CH3

Si

H3C

O

m

CH3

Si

103

CH3 O

CH2

n

Si

CH3

CH3

CH2 CH2 NH CH2 CH2 NH2

a. Conventional aminosilicone (AMD)

CH3

Si CH3

CH3 O

Si

x CH3

C3H6

O

H

OH

H

R

H

H

C

C

C

N

C

C

H

H

H

CH3

H

(PO)(EO) a b

n

b. Amino silicone copolymer (ABn) Fig. 1. Aminosilicone chemical structures. (a) Conventional aminosilicone (AMD). The parameters m and n denote the number of siloxane unit and amino functionalized siloxane unit respectively (b) ABn copolymer. The parameters x, a, b denote the number of siloxane unit, the number of propylene oxide unit, the number of ethylene oxide unit respectively.

In hair conditioners, amino functional silicones are usually deposited on the hair during washing with shampoos or rinseoff conditioners [5]. Jachowicz and Berthiaume [6] have studied in detail by streaming potential measurements the deposition on hair fibers of amino functional silicones emulsified by surfactants. The authors demonstrated that due to the strong negative charge of the hair surface, the electrostatic interactions controlled the deposition of the cationic silicone drops in the initial stage of the deposition process. However, it was observed that even after the reversal of the hair surface charge (the surface charge became positive), the deposition of aminosilicones continued despite the electrostatic repulsion barrier. Due to this behavior, conventional amino functional silicones have a tendency to accumulate on the hair after multiple washes, bringing a greasy feel to the hair. This problem can be solved by using more hydrophilic linear aminosilicone [AB]n block copolymers which are comprised of alternating silicone blocks and hydrophilic polyether blocks (Fig. 1b) [7]. These copolymers do not accumulate on the hair (small deposition) and an additional advantage is that they can be dispersed in water without the use of emulsifiers. In this study, we used streaming potential measurements to understand the deposition of a hydrophilic linear aminosilicone [AB]n block copolymer on hair fibers. This study aimed also at identifying the differences between the two aminosilicone structures, the conventional aminosilicone and the ABn block copolymer.

2. Materials and methods 2.1. Copolymer emulsions Two silicone polymers were studied: a conventional aminosilicone (amodimethicone, AMD) (Fig. 1a) and an amino [AB]n block copolymer (ABn ) (Fig. 1b). Their respective properties are given in Table 1. The samples were obtained from Momentive Performance Materials (Columbus, OH, USA). AMD was synthesized using the well-known equilibration method [8]. An oil-in-water emulsion containing 20% silicone was

Table 1 Relevant properties of silicone particles and keratin fibers. Block copolymer ABn Average molecular weight Amine content (meq/g) Siloxane block MW Polyether block MW Wt% of ethylene oxide (EO) unit in polyether block Dispersion particle radius (nm) Pendant aminosilicone AMD Average molecular weight Amine content (meq/g) Polyether content Dispersion particle radius (nm) Keratin fiber Estimated surface area (cm2 /g)

50,000 0.2 7600 2000 83% 68 38,000 0.2 None 150 500

104

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

12

ABn

% intensity

10

AMD

8 6 4 2 0 0.1

1

10

100

Size r (nm)

1000

10000

Fig. 2. Drop size distributions of the AMD emulsion and the ABn copolymer dispersions obtained by dynamic light scattering.

produced using 10% of a nonionic surfactant (Trideceth-6) according to the method described in [9]. The particle size distribution of AMD diluted emulsion is shown in Fig. 2. The average droplet radius was 120 nm. The ABn was synthesized according to the method described in [10] by reacting an epoxy end-blocked polysiloxane with a polyetheramine precursor. The epoxy end-blocked polysiloxane contained an average of hundred siloxane units. The polyether diamine precursor (Jeffamine ED-2003 from Huntsman, The Woodlands, TX, USA) was a linear chain containing 83% of ethylene oxide and had a pKa of 9.5. The ABn copolymer was blended in dipropylene glycol to form a 30% (w/w) clear dispersion which could then be further diluted easily in water to form droplets of 68 nm average radius (Fig. 2). The particle size (average radius) of the emulsion droplets was measured by dynamic light scattering (DLS) at an angle of 173◦ with a Malvern Nanosizer Nano ZS (Worcestershire, UK).

2.2. Buffer solutions Reagents were purchased from Fisher and used without further purification. In order to obtain buffers with low ionic strength, the following solutions were prepared. An electrolyte solution at pH 6 was prepared by dissolving potassium chloride in deionized water. A pH 4 buffer stock solution was made to contain 0.06% (w/w) acetic acid and 0.0175% NaOH (w/w, to adjust pH) in deionized water. The pH 4 buffer was prepared by diluting 100 mL of the pH 4 buffer stock solution with 60 mL 0.33% (w/w) KCl solution and deionized water to a total volume of 1 l. The pH 8.7 buffer was prepared by dissolving 1.2 g solid Tris(hydroxymethyl)aminomethane and 0.068 g solid KCl in 900 mL water. The solution was titrated to pH 8.7 with 0.2 M NaOH, and deionized water was added to give a total volume of 1 l.

2.4. Microelectrophoresis The electrophoretic mobilities of the silicone droplets were measured in a capillary cell in a Malvern Nanosizer Nano ZS (Worcestershire, UK). A 10−3 M KCl buffer was used and the pH was adjusted with dilute NaOH or HCl solution. The concentration of silicone was 1 mg/g. This concentration was determined to be optimal for measurement by the instrument, as concentrations too low or high caused difficulty in fitting data to the experimental model. The electrophoretic mobility was averaged over 100 runs for each data point. Electrophoretic mobilities were converted to  potential using the Smoluchowski equation,  = 4/ε where  is the viscosity of the dispersion,  is the measured electrophoretic mobility, and ε is the dielectric constant. 2.5. Deposition treatment protocol Concentrated silicone dispersions were diluted in buffer at constant ionic strength. The buffered silicone dispersion of dilute concentration was poured into a 100 mL jar containing 3 g dry hair and immediately shaken on a table shaker at 200 min−1 for a time ranging from 5 min to 2 h (“deposition time”). The hair was then separated from the treatment solution by filtration using a ceramic funnel. The treated hair sample was subsequently rinsed with 500 mL fresh electrolyte and then placed into 100 mL fresh electrolyte. After sitting overnight, the hair was rinsed again with 500 mL fresh electrolyte and packed into the streaming potential cell. This protocol insured that the deposition of silicone was irreversible and produced stable signals of streaming potential. 2.6. Streaming potential measurement The streaming potential was measured using a Zetacad apparatus from CAD Instrumentation (Les Essarts le Roi, France). A glass cylinder of 10 cm length and 1.5 cm inner diameter was used. A porous plug was formed by packing wet hair fiber snippets (3 g dry) into the cylindrical cell. The flow of the electrolyte, back and forth through the plug was controlled by the pressure applied into the electrolyte reservoirs by nitrogen gas. The voltage across two Ag/AgCl electrodes mounted at the inlet and outlet of the cell was measured and recorded as a function of the applied pressure. The electrical conductivity and the temperature of the electrolyte were also measured. Prior to any measurement, the hair plug was conditioned by flowing fresh electrolyte through the plug at low pressure for 10 min. Then, the measurement was started, and the applied pressure was increased incrementally from 5 to 300 mbar. At steady state, for a constant pressure applied, a constant potential (“streaming potential”) was observed. Provided that the potential (V) varied linearly with the applied pressure (P), the streaming potential could be calculated from the slope of V versus P according to the Helmholtz–Smoluchowski equation: =

2.3. Hair keratin fiber samples The deposition studies with keratin fibers were performed using yak hair purchased from International Hair Importers & Products (Glendale, New York, USA). Undamaged, clean hair was cut into 2–4 mm segments. 3 g of dry hair was used for each measurement. A plug of hair was formed by random packing of cylindrical fiber segments each having an average diameter of 70 ␮m. Using gravimetric methods, the fiber geometrical parameters, and assuming the hair surface as smooth, the estimated hair surface area was calculated to be 500 cm2 /g.

V  P

..

 εεo

(I)

where  is the viscosity of the electrolyte solution,  is the conductivity of the solution, and ε and εo are the dielectric constant of water and the vacuum dielectric permittivity respectively [11]. The plug pore size was much larger than the double layer thickness (∼10 nm), justifying the use of the above equation [11]. 2.7. Total silicon content analysis The hair sample used for the streaming potential measurement was collected after the measurement, dried in an oven, and analyzed for total silicon element content. The silicone polymer

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

80

Zeta potential (mV)

70 60 50 40

ABn

30

AMD

20 10

105

For AMD, two transitions were observed: at pH < 4, the zeta potential was highly positive with a value around +70 mV and decreased sharply around pH 5.5. Between pH 6 and pH 9, the zeta potential decreased monotonically with pH. Electroneutrality was obtained at pH around 9.5. The two transition regions at pH 5 and pH 9 probably corresponded to the pKa -s of the primary and secondary amines of the aminosilicone side chains, respectively (Fig. 1a). The lower pKa value corresponded to the weaker base, which was the primary amine.

0 -10 2

4

6

-20

8

10

12

pH

Fig. 3. Zeta potential of silicone drops determined by microelectrophoresis as a function of pH in a 0.001 M KCl electrolyte, at 1 mg/g concentration.

was depolymerized using boron trifluorodiethyl etherate to liberate volatile methyl fluorosilane, which was then quantified by GC according to the method described in [12]. Since the silicone copolymers were too large to penetrate the hair fiber [13], the total silicon element content reflected the amount of silicone deposited on the surface of the fiber. The deposition will be expressed in mg silicone polymer per g hair. 3. Results and discussion 3.1. Zeta potential of silicone drops The zeta potential of the silicone drops ( p ) in 10−3 M KCl electrolyte measured by microelectrophoresis is shown in Fig. 3. For the ABn drops, the zeta potential slowly decreased from +45 mV to +25 mV between pH 4 and pH 8, and then decreased sharply between pH 9 and pH 10. Electroneutrality was achieved around pH 10. This pH dependence of zeta potential was likely related to the ionization of the amine groups present in the backbone of the copolymer. At low pH (pH < 6.5), most of the amine groups were protonated (maximum positive charge). At higher pH, the amine groups were only partially protonated and the protonation decreased as the pH increased.

3.2. Streaming potential measurements 3.2.1. Voltage vs. applied pressure Typical plots of the voltage vs. pressure drop for an untreated hair sample and two, silicone treated hair samples are shown in Fig. 4. A linear behavior was observed, demonstrating that the deposited polymer did not desorb under the shear flow conditions of the streaming potential measurement. Curve a represents the untreated hair sample exhibiting a highly negative surface charge ( ∼ −50 mV). The negative surface charge was probably carried by acidic amino acid residues of the hair fiber keratin proteins [13]. Curve b represents a silicone-treated sample with a negative surface charge and a reduced streaming potential absolute value; the negative surface charge of the substrate was partially screened by the cationic groups of the deposited silicone polymer. Curve c represents a silicone-treated hair sample with a positive surface charge ( > 0). The deposited polymer screened the surface charge of the hair and brought an excess of positive charge (overcompensation). In the following sections, only the streaming potential values calculated from the slope of voltage-pressure plot are reported. 3.2.2. Kinetic experiments with the ABn silicone copolymer The polymer deposition and streaming potential are plotted as a function of the deposition time in Figs. 5 and 6, respectively, for different pH values and for a fixed concentration of silicone polymer (0.5 mg/g) in the treatment bath. At time t < 50 min, the silicone deposition increased sharply with the deposition time (Fig. 5), then, it leveled off. The kinetics of polymer deposition did not change significantly as a function of pH. This behavior seems consistent with the particle transfer mechanism by convection described by

Fig. 4. Typical plots of streaming potential voltage as function of pressure for untreated fibers and silicone treated fibers, respectively. Diamond symbols: untreated keratin fibers ( o  0), square symbols: silicone treated hair with  o <  < 0, triangle symbols: silicone treated hair when  > 0.

106

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

30 20

Deposion (mg/g)

0.8 0.7 0.6 0.5

pH 4

0.4

pH 6

0.3 pH 8.7

0.2 0.1

Streaming potenal (mV)

1 0.9

pH 4

10

pH 6

0 -10 -20 -30

pH 8.7

-40 -50

0 0

20

40

60

80

100

120

140

-60 0

Deposion me (min)

0.2

0.4

0.6

0.8

1

1.2

1.4

Deposion (mg/g) Fig. 5. Polymer deposition vs. deposition time for hair fibers treated with amino [AB]n for C = 0.5 mg/g, I ∼ 0.001 M, at different pH values (diamond symbols: pH 4, square symbols: pH 6, triangle symbols: pH 8.7).

Fig. 7. Streaming potential as a function of deposition for hair fibers treated with the ABn copolymer at different pH values and I ∼ 0.001 M (triangle symbols: pH 4, square symbols: pH 6, diamond symbols: pH 8.7, dashed lines: calculated curves from Eq. (V)).

30

Streaming potenal (mV)

20 10 0 -10 -20

pH 4

-30

pH 6

-40

pH 8.7

-50 -60 0

50

100

150

200

250

300

Deposion me (min) Fig. 6. Streaming potential vs. deposition time for hair fibers treated with the ABn copolymer for C = 0.5 mg/g, I ∼ 0.001 M, at different pH values (triangle symbols: pH 4, square symbols: pH 6, diamond symbols: pH 8.7).

Adamczyk et al. [14]. In this convective-diffusion model, the typical deposition kinetic was linear in the early stage of deposition and could be described by the relation N = kc nb t, where N is the deposition, kc is the mass transfer rate, nb is the particle concentration in the bulk and t is the deposition time. The mass transfer rate depended on the flow condition and the diffusion coefficient of the drops. In our case, the slope of the plot gave a mass transfer rate of 2.5 × 10−6 cm s−1 . According to this model, the initial deposition rate should not depend on the particle charge, as observed in the present study. However, the kinetics of the streaming potential showed a clear variation with pH (Fig. 6). At time zero, the streaming potential was highly negative ( o ∼ −45 mV). Shortly thereafter (t = 5 min), the absolute values of the streaming potential dropped due to the deposition of the positively charged polymer. The streaming potential change was much more pronounced at pH 4 and pH 6 than at pH 8.7. At pH 4 and pH 6, the kinetic curve leveled off and the streaming potential reached a positive value. At pH 8.7, a maximum was observed at around t = 10 min (overshoot).

The change of streaming potential with pH at fixed adsorbed amount could be explained by the dependence of the polymer charge on pH. At high pH, the amount of ionized amine groups per polymer was reduced, and therefore, for a fixed deposited amount, the charge screening by the polymer was less. In contrast, at pH 4, most of the amine groups were ionized, leading to not only a higher polymer charge and streaming potential change, but also more charge deposition than necessary to neutralize the substrate charge. This overcompensation has been often observed with the deposition of solid particles or soluble polyelectrolytes on oppositely charged substrates [15,16]. This will be discussed further in the next section. The overshoot effect at pH 8.7 could be probably explained by the poor polymer anchoring due to the small portion of charged segments. We observed that for hair treated at short contact time and left overnight in quiescent electrolyte, the deposited polymer did not desorb, whereas at long contact time (t > 60 min), the polymer that was loosely attached to the hair surface, desorbed under the action of the vigorous mixing. 3.2.3. Streaming potential as a function of the ABn silicone copolymer deposition In Fig. 7, each data point represents a different hair sample for which both the streaming potential and the ABn deposition were independently measured at various times. The streaming potential (Y-axis) is plotted as function of deposition (X-axis) at varying pH values. The streaming potential increased monotonically with the deposition and leveled off for all pH values studied (pH 4, 6, and 8.7). The plateau values of streaming potential at different pHs are shown in Table 2 (column #4). At pH 6, the streaming potential plateau corresponded to deposition in the range of about 0.6–1.2 mg/g. The deposition of spherical drops (particles) on a solid surface can be estimated by assuming that the drops deposit randomly until they form a single layer of close packed spheres. For hard spheres, the maximum surface coverage is  j = 0.547 (jamming coverage)

Table 2 Comparison of the plateau values of streaming potentials obtained on keratin fibers treated with silicone dispersion with the zeta potential of silicone drops measured in bulk and the fitting parameter values of model (Eq. (V)). pH

Untreated hair zeta potential  o (mV)

Plateau value of streaming potential  s (mV)

 s / o

Silicone particle zeta potential  p (mV)

 p / o

Fitting parameter a

Fitting parameter b

Regression coefficient R2

ABn

4 6 8.7

−44 −45 −48

22 9 −27

−0.50 −0.20 0.56

44 37 22

−1.00 −0.82 −0.46

46.8 13.6 23.1

7.3 13.6 8.7

0.959 0.947 0.925

AMD

4

−44

−8

0.18

68

−1.55

–

–

–

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

(II)

where Ah is the hair surface area (per gram of hair) and r is the radius of the spherical droplet. The ABn copolymer deposition N in gram per gram of hair can be estimated by N ≈ n · v. ≈

4  · Ah .r · 3

(III)

where n is the number of ABn drops per gram of hair, v is the ABn drop volume and is the density of the copolymer ( = 0.96 g/cm3 ). Then, from (II), the deposition at saturation is: Ns ≈ ns · v. ≈

4 max · Ah .r · 3

(IV)

With an estimated hair surface area of 500 cm2 /g, the deposition should be in the order of 1.5 mg/g. This estimated value is reasonably close to the observed deposition at saturation (∼1.2 mg/g). (One always has to remember that such polymers are never uniform and always contain a distribution of species with a range of a, b and x values – see Fig. 1b, and therefore a close agreement cannot be expected.) In contrast, if we assume that the ABn copolymer deposits as individual random coils, using the polymer density and molecular weight to estimate the polymer random coil radius (rp ∼ 2.7 nm), a single layer of individual polymer coils would lead to a deposition of about 0.1 mg/g, which is one order of magnitude smaller than the measured value. These calculations indicate that the silicone polymer probably deposited as whole droplets on the solid substrate, as also suggested by Jachowicz and Berthiaume [6], rather than as individual molecules. Therefore, it seems reasonable to compare our data with the streaming potential model proposed by Adamczyk et al. [14], who postulated a three-dimensional charge distribution near the solid surface, induced by deposition of isolated spherical particles. This model predicting the streaming potential as a function of the surface coverage could be approximated by the following exponential equation [14]: ς = ςo

ς  s

ςo

+ e−a −

ς  s

ςo

e−b

(V)

where  o is the streaming potential of the untreated hair,  s is the plateau value of the streaming potential obtained at saturation,  is the surface coverage, the constants a and b reflect the electrostatic contributions from the interface and from the adsorbed drops, respectively. The surface coverage  can be calculated from the deposition data using the relation (III). The experimental data of streaming potential versus deposition (Fig. 7) and the calculated curves are plotted using the reduced streaming potential /␨o as a function of surface coverage in Fig. 8. The pH dependence of the curves resembles the experimental data obtained with the deposition of amidine polystyrene latex on mica [16]. To fit the model (Eq. (V)),  s was assumed to be equal to the experimental value of the streaming potential plateau. The parameters a and b in (V) were determined by a regression algorithm (Sigmaplot). The fitted value of the parameters a, b and the regression coefficient of the fits are shown in Table 2. The values of a and b have a similar order of magnitude as the theoretical values calculated in [14]. Fig. 8 shows that Adamczyk’s model describes reasonably well the silicone copolymer particle deposition data. However, some

Reduced streaming potenal

max Ah ns ≈  · r2

1.2 1 0.8 0.6 0.4 pH 4

0.2

pH 6

0

pH 8.7

-0.2 -0.4 -0.6 0

0.05

0.1 0.15 0.2 Surface coverage

0.25

0.3

Fig. 8. Reduced streaming potential, / o as a function of deposition for hair fibers treated with [AB]n copolymer at different pH (I ∼ 0.001 M) and comparison with the model described by Eq. (V) (triangle symbols: pH 4, square symbols: pH 6, diamond symbols: pH 8.7).

aspects of this model are in disagreement with our results. In Adamczyk’s model [14], at very high coverage, the streaming potential () of the solid substrate converged toward a constant value which was equal to 0.7 p , where  p is the zeta potential of the particle, measured by microelectrophoresis. In our case, however, at saturation the streaming potential converged toward a much lower value than  p . This could indicate that the amine ionization of the copolymer deposited on the hair fiber surface was significantly reduced compared to the ionization of the copolymer inside the droplets dispersed in the bulk. A similar decrease of amine ionization induced by adsorption on a solid surface has also been observed for the hydrophilic polyelectrolyte polydimethylaminoethyl methacrylate polymer adsorbed onto a TiO2 wafer [17]. In addition, at pH 8.7 the streaming potential leveled off and stayed negative whereas in the theoretical model, for uncharged drops, the streaming potential should converge toward 0. This may indicate that when a particle had a very low charge, it coalesced with another deposited droplet instead of depositing on unoccupied surface. 3.3. ABn copolymer deposition as a function of ionic strength Fig. 9 shows that the deposition decreased as the ionic strength increased for a fixed ABn copolymer concentration of 0.5 mg/g, at pH 6. These data are consistent with the data reported by Jachowicz and Berthiaume for the deposition of emulsified aminosilicones on hair [6]. The reason for this effect is that small cations screened the charges on the hair substrate whereas small anions screened the charge of the silicone particle, and this way decreased the deposition. This screening-reduced deposition has also been observed 1 y = 0.0869x -0.457 R² = 0.9318

Deposion (mg/g)

when the electrostatic repulsion between the spheres is negligible. In the presence of lateral electrostatic interactions, the maximum coverage is smaller and depends on the Debye layer thickness. Here, with an ionic strength of 0.001 M, the double layer length is equal to 9.6 nm and the maximum coverage,  max predicted by the random sequential model is around  max = 0.35 [14]. Assuming that the ABn dispersion is monodisperse (r = 68 nm) and the spreading of the droplets on the keratin surface is negligible, the number of ABn copolymer drops in a saturated layer (per gram of hair) ns is:

107

0.1 0.001

0.01

0.1

1

Ionic strength (M) Fig. 9. Polymer deposition as a function of ionic strength for the ABn copolymer at pH 6 and C = 0.5 mg/g.

108

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109

Streaming Potenla (mV)

30 20 10

deposion 0.8 mg/g

0 -10

ABn -20 AMD -30

deposion 9 mg/g

-40 -50 0

0.2 0.4 0.6 0.8 1 Treatment bath concentraon (mg/g)

1.2

Fig. 10. Streaming potential as a function of the treatment bath concentration at pH 4 and ionic strength 0.001 M for the ABn copolymer (diamond) and AMD (square), respectively. Depositions at 0.5 mg/g treatment bath concentration for ABn and AMD are shown below their respective plots.

experimentally with the adsorption of low charge density polymers [18]. This result was also predicted by Dobrynin and Rubinstein in their model of polyelectrolyte adsorption with high charge solid substrate and low charge density polymers [19,20]. However, their predicted power law of the dependence of ionic strength had a power of −0.75, whereas our experimental data indicated a power of −0.45. 4. Comparison of streaming potential and deposition of ABn and AMD polymers In Fig. 10, the streaming potential as a function of copolymer concentration at pH 4 is shown after 1 h deposition time for ABn and AMD polymers. The deposited amount was measured for a bath treatment polymer concentration of 0.5 mg/g. AMD amine content was identical to the one of ABn (0.2 meq/g). At concentrations below 0.1 mg/g, the streaming potential increased sharply with the polymer concentration of the treatment bath, and then leveled off for both polymers. However, the two streaming potential curves are very different. For the ABn , the streaming potential changed sign and reached a quasi-steady plateau value of +20 mV, whereas with the AMD, the streaming potential kept increasing with the polymer concentration, but was lower than with the ABn and stayed negative (  0 mV). Remarkably, for a fixed bath treatment polymer concentration of 0.5 mg/g, with the AMD, the deposition was 9 mg/g, which was more than ten times higher than the ABn deposition (0.8 mg/g). Since the particle size of AMD was almost twice the size of the ABn particle, according to (III), the deposition would be expected rather to be twice as much, if we assume a deposition of a single layer of drops (with r = 120 nm). From the particle size we can estimate the deposition of AMD as 4 mg/g, which is significantly lower than the measured value. In addition, for AMD, despite the very high deposition, there was a strong discrepancy between the zeta potential (+68 mV) of the AMD drops and the plateau value of the streaming potential (∼−8 mV) (Table 2). This is in disagreement with Adamczyk’s model, where the plateau value of streaming potential should be close to the particle zeta potential. This may indicate that the deposition of AMD was very heterogeneous, leading to many uncompensated charges on the hair surface. 5. Concluding remarks This study showed clear differences between the deposition of the emulsified conventional aminosilicone AMD and the surfactant-free dispersed aminosilicone block copolymer ABn . Similar to Jachowicz and Berthiaume [6], we found a large amount of

aminosilicone (AMD) deposited on the hair surface, although our AMD did not cause charge reversal. This may be due to the low amine content of our AMD (0.2 meq/g). In other experiments, not shown here, we also observed charge reversal with an aminosilicone with a much greater amine content (0.8 meq/g amino). Treating with the ABn copolymer, we not only observed charge reversal (at low polymer concentrations), but also an order of magnitude lower deposition than with the AMD polymer. The streaming potential data, as a function of deposition at pH 4 and pH 6, was consistent with near monolayer adsorption of ABn copolymer drops. At saturation, the hair surface charge was overcompensated by the charged amine groups and became positive. After a positive streaming potential plateau value was reached, further deposition was significantly reduced, due to the formation of an electrostatic repulsive barrier. In spite of its lower adsorption, the ABn copolymer works well as a hair conditioner because the long siloxane sections are sufficient to cause a silky feel and reduce combing force [7]. In contrast to the results with ABn , AMD did not reverse the surface charge of the hair fiber, despite the AMD droplets having a high positive charge (+68 mV) and depositing much higher amount than ABn . This behavior is not consistent with the Adamczyk’ model [14], which postulates that higher particle charge should lead to greater overcompensation of the substrate surface charge. A major difference between the two silicone polymers is that the ABn copolymer contains significant amounts of polar (polyethyleneoxide – see Fig. 1b) groups, and therefore it self-dispersed in water. The AMD polymer, however, contains no hydrophilic groups, apart from the amines, which are themselves low (Fig. 1a). Consequently, a significant amount of non-ionic surfactant (Trideceth-6) was necessary to form an emulsion. During the deposition of the AMD polymer emulsion, the silicone polymer drops and the surfactant molecules/micelles are competing to bind to the hydrophobic hair surface. Therefore it is likely that the aminosilicone droplets can only cover part of the hair surface and can only compensate a subset of the negative charges. The high deposition of AMD polymers can be explained only if multiple drops deposit on the same spot. The conformation model of hydrophobic polyelectrolytes recently elucidated in the literature [20] may offer additional explanation for the differences between the surface charge of AMD and ABn droplets. At low charge density (fraction of charge f ∼ 0.01), in the absence of surfactant, a hydrophobic polyelectrolyte polymer, such as AMD, may adopt a globular conformation burying a fraction of the amine groups in the collapsed siloxane backbone, and thereby reducing ionization. In contrast, for ABn , the different solubilities of the silicone and the polyethylene glycol blocks probably led to block segregation, allowing the ABn copolymer to adopt a more unfolded conformation than AMD. The ABn , for example, could adopt a necklace conformation, such as proposed by Halderin [21] for multiblock copolymers, allowing the polymer amine groups greater access to the negative charges of the solid substrate. This study confirms the advantage of a more hydrophilic ABn copolymer structure vs. the AMD polymer as a mean to avoid excessive accumulation of silicone conditioners on hair. The formation of a repulsive electrostatic barrier can explain why the ABn silicone copolymer did not have the tendency to build up on hair when used in conditioners whereas AMD polymer interaction with hair was driven more by hydrophobic interactions and consequently AMD tended to accumulate on hair. Further studies would be necessary to reveal the structure of the adsorbed aminosilicone drops. References [1] E. Desmond Goddard, J.V. Gruber, Principles of Polymer Science and Technology in Cosmetics and Personal Care, Marcel Dekker, Inc., New York, 1999.

A.D. Dussaud et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 434 (2013) 102–109 [2] J.V. Gruber, B.R. Lamoureux, N. Joshi, L. Moral, Influence of cationic polysaccharides on polydimethylsiloxane deposition onto keratin surfaces from a surfactant emulsified system, Colloids Surf. B: Biointerfaces 19 (2000) 127–135. [3] P. Purohit, P. Somasundaran, R. Kulkarni, Study of properties of modified silicones at solid–liquid interface: fabric–silicone interactions, J. Colloid Interface Sci. 298 (2006) 987–990. [4] M.C. Burrell, M.D. Butts, D. Derr, S. Genovese, R.J. Perry, Angle-dependent XPS study of functional group orientation for aminosilicone polymers adsorbed onto cellulose surfaces, Appl. Surf. Sci. 227 (1-6) (2004). [5] R.A. Lodge, B. Bushan, Effect of physical wear and triboelectric interaction on surface charge as measured by Kelvin probe microscopy, J. Colloid Interface Sci. 310 (2007) 321–330. [6] J. Jachowicz, M.D. Berthiaume, Heterocoagulation of silicon emulsions on keratin fibers, J. Colloid Interface Sci. 133 (1989) 118–133. [7] A. Dussaud, L. Fieschi-Corso, S. Gonzalez, Influence of functionalized silicones on hair fiber–fiber interactions and on the relationship with the macroscopic behavior of hair assembly, SOFW J. 3 (2009) 18–26. [8] A. Michael, Brook in Silicon in Organic, Organometallic and Polymer Chemistry, John Wiley & Sons, Inc., New York, 2000. [9] R. Gee, Oil in water microemulsions from association structures of surfactant, water and aminosilicone polymer oil, Colloids Surf. 137 (1998) 91–101. [10] A.M. Czech, US5807956, September 15, 1998. [11] R.W. O’Brien, Electroosmosis in porous materials, J. Colloid Interface Sci. 110 (1986) 477–487.

109

[12] G.W. Heylmun, J.E. Pikula, Analysis of methyl fluorosilanes from methypolysiloxanes by gas chromatography, J. Gas Chromatogr. 26 (1965) 6–268. [13] C.R. Robbins, Chemical and Physical Behavior of Human Hair, 4th ed., SpringerVerlag, New York, 2002. [14] Z. Adamczyk, K. Sadlej, E. Wajnryb, M. Nattich, M.L. Ekiel-Jezewska, J. Blawzdziewicz, Streaming potential studies of colloid, polyelectrolyte and protein deposition, Adv. Colloid Interface Sci. 153 (1–29) (2010). [15] Y. Shin, J.E. Roberts, M.M. Santore, The relationship between polymer/substrate charge density and charge overcompensation by adsorbed polyelectrolyte layers, J. Colloid Interface Sci. 247 (2002) 220–230. [16] Adamczyk, M. Nattich, M. Wasilewska, M. Zaucha, Colloid particle and protein deposition-electrokinetic studies, Adv. Colloid Interface Sci. 168 (2011) 3–28. [17] N.G. Hoogeveen, M.A. Cohen-Stuart, G.J. Fleer, Polyelectrolyte adsorption on oxides I. Kinetics and adsorbed amounts, J. Colloid Interface Sci. 182 (1996) 133–145. [18] G. Durand, F. Lafuma, R. Audebert, Adsorption of cationic polyelectrolytes at clay-colloid interface in dilute aqueous suspensions-effect of the ionic strength of the medium, Prog. Colloid Polym. Sci. 266 (1988) 278–282. [19] A.V. Dobrynin, M. Rubinstein, Theory of polyelectrolytes in solutions and at surfaces, Prog. Polym. Sci. 30 (2005) 1049–1118. [20] A.V. Dobrynin, M. Rubinstein, Adsorption of hydrophobic polyelectrolytes at oppositely charged surfaces, Macromolecules 35 (2002) 2754–2768. [21] A. Halderin, On the collapse of multiblock copolymers, Macromolecules 24 (1991) 1418–1419.