ARTICLE pubs.acs.org/Langmuir
Modifying the Adsorption Properties of Anionic Surfactants onto Hydrophilic Silica Using the pH Dependence of the Polyelectrolytes PEI, Ethoxylated PEI, and Polyamines Xiaoli Zhang,† Diana Taylor,† Robert Thomas,† Jeffrey Penfold,†,‡ and Ian Tucker§ †
Physical and Theoretical Chemistry Laboratory, Oxford University, South Parks Road, Oxford ISIS Facility, STFC, Rutherford Appleton Laboratory, Chilton, Didcot, OXON § Unilever Research and Development Laboratory, Port Sunlight, Quarry Road East, Bebington, Wirral ‡
bS Supporting Information ABSTRACT: The manipulation of the adsorption of the anionic surfactant, sodium dodecyl sulfate, SDS, onto hydrophilic silica by the polyelectrolytes, polyethyleneimine, PEI, ethoxylated PEI, and the polyamine, pentaethylenehexamine, has been studied using neutron reflectometry. The adsorption of a thin PEI layer onto hydrophilic silica promotes a strong reversible adsorption of the SDS through surface charge reversal induced by the PEI at pH 7. At pH 2.4, a much thicker adsorbed PEI layer is partially swelled by the SDS, and the SDS adsorption is now no longer completely reversible. At pH 10, there is some penetration of SDS and solvent into a thin PEI layer, and the SDS adsorption is again not fully reversible. Ethoxylation of the PEI (PEI-EO1 and PEI-EO7) results in a much weaker and fragile PEI and SDS adsorption at both pH 3 and pH 10, and both polymer and surfactant desorb at higher surfactant concentrations (>critical micellar concentration, cmc). For the polyamine, pentaethylenehexamine, adsorption of a layer of intermediate thickness is observed at pH 10, but at pH 3, no polyamine adsorption is evident; and at both pH 3 and pH 10, no SDS adsorption is observed. The results presented here show that, for the amine-based polyelectrolytes, polymer architecture, molecular weight, and pH can be used to manipulate the surface affinity for anionic surfactant (SDS) adsorption onto polyelectrolyte-coated hydrophilic silica surfaces.
’ INTRODUCTION Surfactant and mixed surfactant adsorption at the solidsolution interface play an important role in many key processes, such as lubrication, detergency, surface conditioning, mineral floatation, dyeing, and oil recovery.1 The use of modern surface sensitive techniques, such as X-ray and neutron reflectivity,2 ellipsometry,3 and AFM,4 have considerably advanced our understanding of the nature of adsorption at the solid-solution interface and of the structure of the adsorbed layer.5,6 However, it remains an important area of research, and there is now an emphasis on the study of more complex systems, different types of surfaces, and the ability to modify or manipulate surface properties and surface adsorption. Specific interaction with the solid surface modifies the adsorption behavior compared to that observed at the simpler airwater interface. For example, adsorption onto a hydrophilic surface, such as silica, is cooperative, and surface self-assembly that resembles or is related to the corresponding bulk aggregated phase takes place. In contrast, adsorption at the hydrophobic r 2011 American Chemical Society
solid-solution interface is similar to that at the air-water interface, and is usually in the form of a monolayer.5,6 Furthermore, the development of tailored surfaces with specific functionalities can result in subtle and sensitive manipulation of surfactant and mixed surfactant adsorption and in the structure of the adsorbed layer.7 This provides the potential to develop surfaces with specifically designed functionalities, “smart surfaces”, and this is the impetus of much of the current research in this area. A recent and important manifestation of this, and of direct relevance to this paper, is the use of polymers or polyelectrolytes to manipulate surface properties and surface adsorption. This could take the form of in situ grown polymer brushes with different functionalities8 or the use of adsorbed polyelectrolytes.9 Polyelectrolytes adsorbed onto a hydrophilic surface of opposite charge will tend to charge reverse the surface,9-11 and this is the Received: November 24, 2010 Revised: January 10, 2011 Published: March 02, 2011 3569
dx.doi.org/10.1021/la1046723 | Langmuir 2011, 27, 3569–3577
Langmuir mechanism responsible for the formation of polyelectrolyte multilayer structures that are now being exploited for applications such as encapsulation and controlled release.11 We have previously demonstrated how the polyelectrolytes, poly(dimethyl diallyl ammonium chloride), polyDMDAAC, and poly(styrene sulfonate), PSS, can be used to manipulate cationic, anionic, and mixed surfactant adsorption at the hydrophilic silica surface,7,9 and there is other related work in this area.12-14 For example, the adsorption of polyDMDAAC onto hydrophilic silica results in a charge reversal of the surface and the subsequent strong adsorption of SDS onto the polyelectrolyte modified surface, where in the absence of the polyelectrolyte adsorption no SDS adsorption would take place.9 In this paper, we have extended that initial work to different types of polyelectrolyte, where the charge on the polyelectroltye is pH sensitive and where we have also modified the polyelectrolyte to manipulate the strength of the ionic surfactant/polyelectrolyte interaction. We focus here on the pH-sensitive polyelectrolyte polyethyleneimine, PEI, related polymers, and oligomers. Our recent studies on the adsorption of PEI/SDS mixtures at the air-water interface15 show a strong surface complexation and strong surface adsorption down to very low SDS concentrations, and over a wide pH range, from pH 3 to 10. At low pH (pH 3), the interaction is dominated by charge attraction between the two oppositely charged species, whereas at high pH (pH 10), a strong interaction is still observed, resulting often in trilayer and multilayer adsorption at the interface. Following recent studies on the adsorption of polyamine/SDS mixtures,16 we can now attribute the interaction that is dominant at high pH (where the PEI is essentially a neutral polymer) to a short-ranged dipole-ion attraction between the amine nitrogen and the SDS sulfate group, which is further enhanced by the hydrophobic interaction between the alkyl chains of the neighboring surfactant molecules. Ethoxylation of the PEI disrupts the strength of that interaction at both high and low pH, and especially at high pH for attached EO groups > EO3, when the ethoxylated PEI now behaves more like a neutral polymer.17 In this paper, we consider the effect of pH on the adsorption of PEI, ethoxylated PEI, and the polyamine, pentaethylenehexamine, at the hydrophilic silica-solution interface, and on its subsequent impact on the adsorption of the anionic surfactant SDS to that interface. We will demonstrate how pH, ethoxylation of the PEI, and through the polyamine, the polyelectrolyte molecular weight, can all have an important impact upon the surfactant adsorption. These results lead to the possibilities of using such effects to tailor the manipulation and modification of surfactant adsorption onto different surfaces.
’ MATERIALS AND EXPERIMENTAL DETAILS Materials. The protonated SDS (h-SDS, C12H25SO4-Naþ) was obtained from PolySciences. The deuterated SDS (d-SDS, C12D25SO4-Naþ) was synthesized as described previously and purified prior to use by recrystallization from ethanol.18 The chemical purity of the surfactants was checked using surface tension measurements and no minimum in the surface tension was observed at the cmc. PEI is most commonly synthesized in its branched form, and for consistency with other work, we have used that form here. The 25k molecular weight (MW) branched PEI was obtained from BASF, and used as supplied. The 2k MW branched PEI was obtained from BASF and dehydrated by a 72 h freeze-drying process under vacuum before use. The PEIs used were hence highly branched, with primary, secondary, and tertiary
ARTICLE
nitrogens in the ratio 1:2:1.19 The polymers have a broad MW distribution (as stated by the manufacturers), but were not measured separately. The modified PEIs (PEI-EO7 and PEI-EO1) were synthesized from 2k MW branched PEI. The ethylene oxide (h-EO, C2H4O) was supplied by Aldrich and was used as received. The alkoxylation process was conducted through the vapor phase, which avoids some of the dangers of normal pressurized synthesis, as described in detail elsewhere.20 The reaction used a synthetic route similar to that described Watson et al.,21 and the degree of ethoxylation was verified using the proceedure defined in ref 20. The ethoxylated chains are attached to the primary and secondary nitrogens replacing the hydrogen atoms, to give effective MWs of ∼16k for PEI-EO7 and ∼4k for PEI-EO1. The polyamine, pentaethylenehexamine, was obtained from Sigma-Aldrich, and used as supplied. Ultrapure water (UHQ) was used throughout to prepare solutions and to clean all glassware. Deuterium oxide (D2O) was purchased from Aldrich. Most of the measurements were made in D2O, and the solution pH was adjusted by the addition of HCl or NaOH. The liquid-solid sample cells used for neutron reflectivity measurements were soaked overnight in Decon 90 and then washed in UHQ. The surfaces of the silicon crystals used were made hydrophilic using the “mild piranha” treatment.22 The degree of ionization of the PEI varies with pH, and at pH 3, >70% of the nitrogens are positively charged, which reduces to ∼30% at pH 5-6 and to 3 10-4 M the reflectivity shows a pronounced interference fringe, consistent with increasing SDS adsorption with increasing SDS concentration (see Figure 1). At each SDS concentration measured, the SDS was removed by rinsing in D2O. Measurements were also made for a different isotopic combination, d-SDS/D2O, where the SDS is now effectively “index matched” to the D2O. At each SDS concentration, the measured reflectivity for d-SDS/D2O is similar to that for the bare silica surface, and indicates that the SDS layer does not swell or penetrate the PEI layer that is adsorbed onto the silica surface. The solid lines in Figure 1 represents the three-layer model used to fit the reflectivity for the h-SDS/D2O measurements, and the key model parameters are summarized in Table 1. In all the modeling of the data, the oxide layer, as described in Table 1 in the Supporting Information, is also included.
ARTICLE
Figure 1. Neutron reflectivity for h-SDS/D2O at pH 7 on PEI coated silica surface for (red) 6 10-5 M, (blue) 3 10-4 M, (green) 7 10-4 M, (yellow) 3 10-3 M, (pink) 7 10-3 M, and (cyan) 10-2 M SDS. The solid lines are model fits using the parameters summarized in Table 1.
The three-layer model of surfactant adsorption onto a solid hydrophilic surface has been extensively used and described in detail,30 and is the simplest model that describes what is effectively a “fragmented bilayer” or “flattened” micellar structure adsorbed at the interface. The key model parameters are the thicknesses of the headgroup regions adjacent to the solvent and solid phases, d1, d3, and the hydrophobic alkyl chain region, d2, the fractional coverage, fc, and the density of the surfactant regions, characterized by a area/molecule, A, (in this case taken as a constant at ∼40 Å2). The adsorbed amount is then given by Γ ¼ 2fc =Na A
ð2Þ
where Na is Avogadro’s number. The key model parameters are broadly similar to those previously reported for SDS adsorbed onto cationic hydrophilic surfaces.8 The adsorption isotherm for SDS onto the PEI coated surface at pH 7 is shown in Figure 2, and compared with that previously reported for SDS adsorption onto a polyDMDAAC surface.7 The two isotherms are broadly similar, and so both polymers are providing a similar role in charge reversing the silica surface, as in the absence of the polyelectrolyte no SDS adsorption would be observed. Furthermore, it is evident from the measurements in D2O after rinsing and from the measurements for d-SDS/D2O that the adsorption is reversible. Measurements at pH 10. Figure 3 shows the reflectivity data for SDS adsorbed onto a PEI coated surface at pH 10 for h-SDS/ D2O and in Figure 3 in the Supporting Information for d-SDS/ D2O. Although the data in Figure 3 are superficially similar to that in Figure 2 for h-SDS/D2O adsorbed onto PEI at pH 7, the details of the adsorption behavior are somewhat different. First, the simple model of SDS aggregates adsorbed reversibly onto a PEI coated surface, used to describe the data at pH 7, is not consistent with the data at pH 10. The data in Figure 3 in the Supporting Information provide a clear indication of the origin of this difference. At pH 10, the reflectivity for d-SDS/D2O is no longer identical to the surface prior to surfactant adsorption, which would be the case if the d-SDS was just adsorbed onto the PEI 3571
dx.doi.org/10.1021/la1046723 |Langmuir 2011, 27, 3569–3577
Langmuir
ARTICLE
Table 1. Key Model Parameters for SDS Adsorption onto PEI Coated Surface at pH 7 SDS concentration (M)
d1 ((2 Å)
d2 ((2 Å)
d3 ((2 Å)
fc ((0.02)
Γ ((0.2 10-10 mol cm-2)
3 10-4
4
37
8
0.38
3.2
7 10-4
6
36
5
0.39
3.2
3 10-3
9
31
3
0.62
5.2
7 10-3
9
31
6
0.62
5.2
10-2
9
28
6
0.64
5.3
Figure 2. SDS adsorption isotherm (•) onto PEI coated silica surface at pH 7 and (o) onto polyDMDAAC coated surface (from ref 7).
Figure 3. Neutron reflectivity for SDS adsorbed onto PEI coated silica surface at pH 10. (a) h-SDS/D2O, (line) bare silica surface, (red) 5 10-4 M, (blue) 10-3 M, (green) 5 10-3 M, (cyan) 10-2 M.
layer, as the d-SDS is closely index-matched to the D2O. This implies that the SDS and solvent partially penetrate the PEI layer and swell it. Furthermore, measurements in D2O, post adsorption and rinsing, show that the adsorption is now not completely reversible. A detailed analysis of the reflectivity data at pH 10 for h-SDS and d-SDS in the concentration range 10-3 to 10-2 M (see Table 2) illustrates clearly the more complex nature of the adsorption of the SDS and its impact upon the adsorbed layer of PEI.
The simplest model consistent with all the data and both of the “contrasts” measured (h-SDS/D2O and d-SDS/D2O) is a twolayer model. This comprises a relatively thin inner layer ∼15 to 20 Å and an outer layer ∼30 to 60 Å. The scattering length density of the layers indicate that there are SDS, solvent, and PEI in both layers. The partial irreversibility of the SDS adsorption and the limited “contrasts” measured makes it difficult to reliably estimate the amounts of the different components present in each layer, but the measurements made do indicate a rather different surface structure. The structure is consistent with the initial compact PEI layer being swelled and extended by the SDS and solvent to form a thicker surface layer. Measurements at pH 2.4. At pH 2.4, PEI deposited onto a hydrophilic silica surface from a 20 ppm PEI/D2O solution results in a pattern of adsorption quite different to that observed at either pH 7 or 10 (see Figure 4a). The adsorption of the PEI at pH 2.4 results in a substantially thicker PEI layer being formed, as indicated by the model parameters in Table 3. In Figure 4b, the impact of the adsorption of SDS at pH 2.4 (from h-SDS/D2O solutions) in the SDS concentration range 10-4 to 10-2 M is illustrated. The data for 10-4 M h-SDS and d-SDS are almost identical to that for the PEI coated surface prior to adsorption, and this is consistent with very little SDS adsorption and very little impact of the surfactant upon the PEI layer. In the concentration range 5 10-4 to 10-2 M, there is a progressive change in the reflectivity (although the data for 5 10-3 and 10-2 M are essentially the same) consistent with SDS adsorption and incorporation into the PEI layer, resulting ultimately in a swelling of the PEI film, by solvent and surfactant. The parameters in Table 3 summarize the simplest two- or threelayer models that describe the deposited PEI layer and the combined PEI/SDS layers following h-SDS adsorption at surfactant concentrations of 5 10-4, 10-3, and 5 10-3 M. The adsorbed PEI layer is relatively thick, ∼140 Å, with a gradient of composition, which has been approximated with a two-layer model and with substantial solvent penetration in to the layer. The scattering length densities of the two layers (see Table 3) are consistent with a polymer volume fraction, jpei, of ∼0.53 and ∼0.4 for the inner and outer layers, respectively. The reflectivity data for the “contrast” d-SDS/D2O are very similar to that for the bare interface (data not shown here); but this is not consistent with the PEI being removed from the surface, as SDS adsorption still takes place, as shown by the measurements for h-SDS/D2O. Hence, it is most likely that due to the incorporation of d-SDS and D2O into the PEI layer the scattering length density is sufficiently close to that of the solvent or substrate that it becomes essentially invisible. The parameters in Table 3 (and the relatively subtle changes in the reflectivity profiles in Figure 4) show that in the concentration range 10-4 to 10-3 M SDS very little happens to the surface layer, except some changes in the scattering length density of the adsorbed layer. These changes (and taking into account that the 3572
dx.doi.org/10.1021/la1046723 |Langmuir 2011, 27, 3569–3577
Langmuir
ARTICLE
Table 2. Model Parameters for SDS Adsorbed onto PEI Coated Surface at pH 10 SDS concentration (M) 10-3 5 10-3 10-2
surfactant “contrast”
d1 ((2 Å)
F1 ((0.2 10-10 Å-2)
d2 ((2 Å)
F2 ((0.2 10-10 Å-2)
h-SDS
20
3.3
50
4.1
d-SDS
15
4.5
55
5.5
h-SDS
20
4.7
33
3.5
d-SDS
20
3.3
57
5.2
h-SDS
20
4.4
30
3.0
d-SDS
20
3.8
60
5.9
mixtures. For the ethoxylated PEI/SDS mixture,s a more limited range of measurements were made, but sufficient to discern the general trends and any differences and similarities compared to PEI/SDS. Specifically, the measurements were made only at the extremes of pH, pH 3 and 10, and for a more limited range of SDS concentrations. SDS/PEI-EO7. The deposition of PEI-EO7 at pH 3 and at pH 10 from dilute solution (20 ppm in D2O) produces a reflectivity profile that is not perceptibly different from that of the bare interface. This does not necessarily mean that no polymer is adsorbed, as demonstrated earlier for PEI and previously for polyDMDAAC.7 However, at pH 10 subsequent measurements with the addition of SDS, at 5 10-4, 10-3, 5 10-3, and 10-2 M h-SDS/D2O, show no evidence for SDS adsorption. Measurements for d-SDS/D2O also indicate no effect upon any possible adsorbed polymer layer. This implies that either there is no polymer adsorption or the SDS does not adsorb onto it. It is known from previous neutron reflectivity measurements at the air-water interface that the SDS does interact with PEI-EO7 over a wide pH range, and hence, we can conclude that no PEIEO7 adsorption occurs at pH 10. From neutron reflectivity measurements for 5 10-4 M and -3 10 M h-SDS/D2O at pH 3, a slight adsorption of the SDS is observed (see Figure 4 in the Supporting Information). In Figure 4 in the Supporting Information, the neutron reflectivity for the bare silicon oxide surface in D2O, the surface after exposure to 20 ppm PEI-EO7 at pH 3 in D2O, and for 5 10-4 and 10-3 M h-SDS/D2O are shown. The adsorptions for 5 10-4 and 10-3 M DS are almost identical and can be modeled by a layer of thickness ∼50 Å and a scattering length density, F2, ∼5.5 10-6 Å-2. From eq 3, the adsorbed amount can be estimated. Figure 4. (a) Neutron reflectivity for (black) silica/D2O and (red) PEI deposited onto silica at pH 2.4, in D2O, (b) Neutron reflectivity for h-SDS/D2O onto PEI coated silica surface at pH 3 for (red) 10-4 M, (blue) 5 10-4 M, (green) 10-3 M, (pink) 5 10-3 M, and (cyan) 10-2 M SDS. The solid lines are model fits using the parameters summarized in Table 3.
h-SDS/D2O and d-SDS/D2O measurements were done sequentially) are consistent with some adsorption of the SDS into the PEI layer partially replacing solvent, which is not entirely reversible. At 5 10-3 and 10-2 M SDS, the most significant change in the reflectivity takes place. Here, the data are consistent with both a swelling of the PEI layer by the SDS and D2O, and the adsorption of an additional layer of SDS/solvent on the top of the swollen PEI film, with a total thickness of the PEI film ∼170 Å. Ethoxylated PEI/SDS Mixtures. Neutron reflectivity measurements were made for two ethoxylated PEI/SDS mixtures (PEI-EO7 and PEI-EO1) similar to those made for the PEI/SDS
Γ¼
dðFs - F2 Þ Na V ðFs - Fa Þ
ð3Þ
where d is the thickness of the adsorbed layer, F2 its scattering length density, Fs the scattering length density of the solvent phase, Fa the surfactant scattering length density, V the surfactant molecular volume, and Na Avogadro’s number. In this case, it is ∼2.9 10-10 mol cm-2, which is similar to that observed at 5 10-4 M SDS on PEI and on polyDMDAAC.7 Measurements for d-SDS show that the thin PEI-EO7 layer is unaltered, and measurements post rinsing in D2O show that the adsorption is also reversible. Furthermore, measurements at higher surfactant concentrations (10-2 M) show that no surfactant adsorption exists, and this implies that at these higher surfactant concentrations the surfactant and polymer desorb from the surface. This is similar to what is observed for some polymer/surfactant mixtures at the air-water interface.26 SDS/PEI-EO1. For PEI-EO1 at pH 3 the adsorption pattern is broadly similar to that for PEI-EO7. At an SDS concentration of 3573
dx.doi.org/10.1021/la1046723 |Langmuir 2011, 27, 3569–3577
Langmuir
ARTICLE
Table 3. Key Model Parameters for PEI/SDS Adsorption onto Hydrophilic Silica at pH 2.4 d1 ((5 Å)
F1 ((0.2 10-10 Å-2)
d2 ((5 Å)
F2 ((0.2 10-10 Å-2)
-
49
3.0
91
3.8
-
-
10-4
49
3.0
91
3.8
-
-
5 10-4
47
4.5
91
5.2
-
-
10-3
43
3.3
90
4.1
-
5 10-3
63
3.2
106
4.6
15
2.6
10-2
63
3.2
106
4.6
15
2.6
SDS concentration (M)
5 mM, the neutron reflectivity data for h-SDS/D2O shows a surfactant adsorption similar to that shown in Figure 4 in the Supporting Information for SDS onto the PEI-EO7 coated surface. Modeling the reflectivity as a single layer gives a d of ∼53 Å, F2 of ∼5.5 10-6 Å-2, and an adsorbed amount, Γ, of ∼3 10-10 mol cm-2. At pH 10, the pattern of adsorption of SDS onto the PEI-EO1 coated surface is different, as illustrated in Figure 5. For measurements with h-SDS/D2O at 5 10-4 M SDS, there is no adsorption, but for 10-3 M and 2.5 mM SDS, some adsorption, which increases slightly with increasing surfactant concentration, occurs. At 5 mM, a much stronger SDS adsorption, characterized by a prominent interference fringe in the reflectivity data, is observed. However, at 10 mM there is no adsorption, and again it is consistent with both the SDS and polymer having been removed from the surface. Measurements for d-SDS/D2O indicate that the polymer layer is unaffected by the addition of SDS, for SDS concentrations of
