Specific Ion Effects at Two Single-Crystal Planes of Sapphire

May 24, 2013 - sapphire aqueous electrolyte interface is a nonpolar surface with negative charge. The nonpolarity finds repercussions in the weak...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/Langmuir

Specific Ion Effects at Two Single-Crystal Planes of Sapphire J. Lützenkirchen* Institut für Nukleare Entsorgung − INE, Karlsruher Institut für Technologie (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany S Supporting Information *

ABSTRACT: Experimental results on specific ion effects at the c- and r- singlecrystal planes of sapphire obtained by zeta-potential measurements at pH 5.8 are reported. Both crystal planes have negative electrokinetic charge at pH 5.8 and their intrinsic isoelectric points are found close to pH 4. The water structure “making” surface (i.e., r-plane, based on surface diffraction and surface complexation modeling) causes cation specificity in the order Li+ > Na+ > K+ > Rb+ > Cs+ in chloride systems while no anion sensitivity occurs in sodium systems (Cl−, NO3−, and BrO3−) as expected. The cation series concurs with the simple idea of structure making ions being adsorbed more strongly on structure making surfaces and also concurs with the sequence found for particulate alumina for the cation series in nitrate systems. On the structure breaking basal plane (i.e., c-plane, again based on surface diffraction and surface complexation modeling), no cation specific effects are observed in chloride systems, but the structure breaking properties are retrieved in the cation series in nitrate systems. Surprisingly, anion specificity is observed on sapphire-c. Furthermore, the chloride ion shows unexpected behavior that suggests chloride adsorption onto the negatively charged surface. Based on these experimental observations in conjunction with generic results from published MD simulations, the c-plane sapphire aqueous electrolyte interface is a nonpolar surface with negative charge. The nonpolarity finds repercussions in the weak water ordering and the observed ion specific effects. The low isoelectric points of the cuts cannot be explained by the respective surface chemistries of the ideal surfaces. Relation to “inert” surfaces and concomitant dominance of hydroxide ion adsorption is a possible explanation for the low isoelectric points of both cuts. The reported ion specific effects occur at concentrations below 10 mM. Overall, the results support the idea that ion specific effects are largely governed by surface hydration.



INTRODUCTION There is a fundamental interest in understanding the behavior of different crystal planes of a given oxide (the following statements all pertain to hydroxylated surfaces). The MUltiSIte-Complexation (MUSIC) model1−4 predicts that there are substantial differences between various faces of a given mineral in terms of pristine points of zero charge and variable charge/ potential patterns. Figure 1a illustrates this for three cuts of sapphire, that is, the c-cut (001 face), the r-cut (012 face), and the a-cut (110 face). Figure 1b displays the structures of the three cuts for (hydroxylated) oxygen terminations. While the perfect c-plane is entirely dominated by doubly coordinated hydroxyl groups, which are fairly unreactive in the normal pHrange according to the MUSIC model, the presence of singly coordinated groups on the two other cuts causes strongly pH dependent charging behavior (Figure 1a). The c-plane shows more or less zero electrostatic surface potential over a wide pH range (between pH 4 and 8) and little charging beyond and within the commonly studied pH range. The two other cuts exhibit the typical oxide behavior with near-Nernstian slopes. A sharp point of zero surface potential occurs for the r- and a-cuts at about pH 8. On the basis of Figure 1, it is expected that the surface chemistry of the different cuts has strong repercussions on many observable properties. In the context of this work, © XXXX American Chemical Society

particular interest is in the water structure at these surfaces. Water structure is expected to be affected by the hydrogenbonding network of water at the interface (encompassing hydrogen bonding between water and the surface hydroxyls, among the surface hydroxyls and among the interfacial water molecules) and by electrostatics (the surface potential orienting the water dipoles). A balance between those factors will give a resulting water structure that can be studied experimentally. On the basis of Figure 1, it is expected that in the water structuring processes surface potential plays a minor role on the c-face compared to the two other cuts. In a crystal truncation rod (CTR) comparison study,5 water structuring on sapphire-c faces was indeed found to be much weaker than that on the respective a- and r-cuts (see the Supporting Information). In the following, we will use the term structure making for the rcut because the CTR study has provided evidence of a water layer adjacent to the surface that exhibits much stonger ordering than sapphire-c, which will be considered as structure breaking. The body of investigations on sapphire-c in aqueous electrolyte solutions available in 2010 has been recently Received: September 20, 2012 Revised: May 24, 2013

A

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Figure 1. (a) Surface potential at three faces of sapphire as a function of pH for 1 mM 1:1 electrolyte solution (001 is the c-cut, 012 the r-cut, and 110 the a-cut). The diffuse layer model was used for electrostatics. (b) Surface structure of three faces of sapphire (001 is the c-cut, 012 the r-cut, and 110 the a-cut). The three different kinds of blue oxygens pertain to surface positions and their respective coordination. Sapphire-c includes only doubly coordinated groups. Sapphire-a includes singly and triply coordinated groups, whereas sapphire-r includes all three kinds of groups.

reviewed,6 and a dual charging mechanism was suggested to reconcile many experimental data sets from many different methods. The dual mechanism started from the assumption that directly adjacent to the surface there is an ordered water film on sapphire-c7,8 referred to as 2D-bilayer in the following. This 2D-bilayer would impede interactions with subsequent water. As a consequence, the water adjacent to the 2D-bilayer would resemble that of an “inert” surface and also cause the electrokinetic charging properties of such an interface. The implication of the dual-charging model is that two patterns are observed as a function of pH: the first is that of the intrinsic oxide surface, which can be explained by the MUSIC model (i.e., Figure 1) and suggests a more or less uncharged oxide surface between pH 4 and 8 and relatively low charge due to protonation and deprotonation of the aluminols beyond that

pH range. This behavior was observed in contact angle measurements6 and sum frequency generation (SFG) spectra9 on sapphire, and recently reported for surface potential measurements on hematite.10 The second pattern corresponds to that of an inert surface. Such surfaces (a prominent example is the air−water interface) are not expected to show pHdependent charging whatsoever, but all have isoelectric points (IEPs) between 2 and 4.11 While the origin of the pH dependent electrokinetic charge is not settled, we assume in the absence of better explanations that it is due to the preferential (physical) accumulation of hydroxide ions at such interfaces.11−13 The most recent streaming potential study on sapphire-c corroborates our IEPs of about pH 4 for sapphirec and shows that heating at high temperature leads to even lower IEPs.14 Independent results supporting the pH-dependB

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

In a more recent SFG study,19 the decrease of the dangling bond with increasing pH is used to define the zero surface charge and the authors conclude that the point of zero charge of that sample is between pH 5 to 6. The dangling OH is effectively vanishing at pH 7 and the interpretation is that the aluminol-groups that cause the dangling bond lose a proton (therefore there is no OH-stretch left) and contributes to negative charge at pH > 7.19 A concomitant change in the water band amplitude agrees with this. A decline in the amplitude of the dangling bond may have been observed in an earlier study9 though at much higher pH, but it was not explicitly discussed.9 The recent paper19 deals with a sample that was treated in a different way compared to the studies that were finally considered in the design of the dual charging mechanism model. The strong decline of the dangling bond has not been previously reported in the available studies, and rather the presence of the dangling bond has been subject of discussion6 and can also be traced to sample pretreatment.20 With respect to sample preparation, Zhang et al.21 have very recently criticized previous pretreatments. However, they do not obtain the same cleaning efficiency in terms of the measured contact angle and carbon contamination when applying our procedure, while others22 reproduced our cleaning results. In summary, some controversary will remain concerning sample preparation, but also with respect to sample origin and even the significance of the experimental data as obtained with different methods. A consistent multimethod study on one sample would be extremely helpful to exclude some of the prevailing uncertainties in the results that are currently available. The r-cut has been studied less frequently compared to the ccut. Kershner et al.23 reported an IEP of the r-cut at pH 4.5, insensitive to ionic strength. Fitts et al.15 obtained an average value of 5.2 from second harmonic generation experiments. Franks and Meagher39 reported IEPs of 4.9 and 5.5 from streaming potential and 6.5 from force distance measurements. SFG studies on this cut resulted in a point of zero charge of about 6.7.24 All these experimental points of zero charge are below the expected value based on the MUSIC model (see Figure 1a). One theoretical study25 advances the presence of singly coordinated surface hydroxyls on tetrahedral aluminum, which could explain points of zero charge between 5.5 and 7. However, this cannot explain measured IEPS below pH 5. As a general summary of the available data for the two cuts that were studied in the present paper, it can be concluded that the reported IEPs are consistently lower than what is predicted based on the expected surface chemistry of the ideal surfaces. Direct experimental investigations on interfacial water structure and comparisons between different cuts are available from SFG and CTR studies. These are related to each other in the Supporting Information. Briefly, there is surprisingly little difference in the pH-dependent patterns on sapphire-c9 relative to sapphire-r and -a24,26 (see Supporting Information Table SI1). Since the interpretation of the vibrational spectra is under debate, the CTR results are used to relate our zeta-potential measurements to independent information on interfacial water structure The observed differences between the c- and r-cuts in the CTR study5 should profoundly affect the interaction of monovalent ions with these two surfaces. Specific ion effects at oxide water interfaces have been studied for oxide particles using various methods,27−30 one being agglomeration studies, which are extremely sensitive to electrolyte composition and concentration.31−33 Single crystal surfaces have also been

ent streaming potential and the sharp IEP around pH 4 come from second harmonic generation measurements.15 The assumption that hydroxide ions are physically adsorbed is an unconventional perception of an oxidic surface. Recently, negative charge due to the presence of acidic groups was put forward to explain low IEPs by Yang et al.14 However, this should generate the typical oxidic behavior with strong water structuring at pH 6. MUSIC-type observations (like contact angle titrations) cannot be explained by this. Since the proposal of the dual charging mechanism involving the 2D-bilayer for sapphire-c, several papers have appeared, which appear to either confirm or refute it. MD simulations16 on sapphire-c suggest that a limited amount of water prefers to form a wetting layer on which a drop remains, very much in agreement with the previous suggestion. However, the surface diffraction study mentioned above5 shows that sapphire-c exhibits much weaker water binding directly adjacent to the surface hydroxyl groups than what is observed on two other cuts of the same material. This study disagrees with the presence of a 2D-bilayer, though the exclusive use of specular data in the surface diffraction study does not allow the resolution of the full 3D structure of such water layering. The previous interpretation was based on the information available at that time. The contact angle measurements with maximum of about 30° is on the one hand probably not so different from the one obtained in MD simulations,16 but on the other hand it is not sufficiently high, if the first water layer were to make the surface truly hydrophobic for subsequent water. This would imply that there has to be some interaction between the first wetting layer and the water that forms the drop on top of it. Recent theoretical work17 studied in much detail the water structure on the c-plane surface in an effort to explain the published vibrational spectroscopy data,9 which we also used for our conceptual model. The calculations17 suggest that the first water layer consists of four coordinated waters in agreement with our far more simplistic phenomenological interpretation of the spectra. The authors17 assign parts of this first layer water to the “ice”-like contribution (H-bond donating waters) while the H-bond accepting waters in that layer would contribute to the liquid-like band. This is not in agreement with our simplistic model which postulates only ice-like water in the first layer. Owing to the potential complexity of interfacial water and the contributions of different water species to the liquidand ice-like bands that are for example found for the air−water interface,18 such controversy is not really surprising. The terms “ice-like” and “liquid-like” are to assign contributions in the 3250 and 3450 cm−1 wavenumber range, respectively. Another debatable issue in our previous interpretation is the zero SFG amplitude of the ice-like band in the pH range 5−7. To explain a net zero contribution of the waters and aluminols potentially contributing to this band (despite strong water ordering) would require cancellation. This might be possible if in the 2D-bilayer water molecules with dipoles pointing away and toward the surface with similar vibrational fingerprints would be arranged in such a way that the overall contribution would be zero. Since the proposed 2D-bilayer would be hydrophobic to subsequent water layers, the water adjacent to the 2D-bilayer would contribute to the measured vibrational spectra in a way similar to the spectra obtained for inert surfaces. So for this ideal case the situation would be more complex than sole dipole cancellation in the 2D-bilayer.The absence of the 2D-bilayer could agree with the CTR investigation. C

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

CO246,47 a constant flow of purified argon over the solution was assured. Prior to reaching the solution, the argon was sent through a hydroxide solution to remove trace carbon dioxide from the gas and subsequently through Milli-Q water to minimize water evaporation from the solutions during the experiments. The equilibration time between each addition of titrant solution was 10 min. This was found to be sufficient in test experiments involving longer equilibration times. Furthermore, in some experiments, several data points were collected twice for identical conditions (i.e., without addition of salt solution), thus doubling the equilibration time and no difference was observed. At least six points per measurement condition were collected, and an average and standard deviation were obtained. The upper salt concentration for the experiments was about 10−15 mM. For higher salt concentrations, standard deviations became too high to observe significant differences within a given series. Some attempts were made to study anion effects in more detail as well. However, and despite the contrary statement of the manufacturer of the streaming potential device, measurements including bromide salts turned out to be very difficult. The data were not well reproducible at bromide concentrations above ca. 2 mM. This was found to be due to interaction of bromide with the Ag/AgCl measuring electrodes. In agreement with the manufacturer’s statement it was possible to obtain reproducible results with small standard deviations at sufficiently low bromide concentrations in the CTABsapphire-c system.48 More recent studies on AgBr single crystals resulted in color changes on the measuring electrodes indicating AgBr formation at higher bromide concentrations. This problem is expected with other structure breaking anions such as thiocyanate and iodide. Thiocyanate was tested with a quartz surface and resulted in substantial measurement errors even at submillimolar salt concentrations (where errors in chloride solutions were 10 times smaller). Consequently, the structure breaking anions could not be studied in more detail with the device. The bromide data (in KBr) that were deemed reliable did support the chloride effects on sapphire-c that will be discussed below. Experiments as a function of pH were carried out to obtain the intrinsic IEP of the samples and to cross-check some of the specific ion effects. In these experiments, drops from freshly prepared NaOH or KOH solutions were added to the working solution to increase the pH and the solution was then titrated by HNO3 or HCl solutions to pH about 3. Two selected salt systems were also studied as a function of pH in the acidic pH range to obtain the IEP for 20 mM salt concentrations. These experiments were started at about pH 5, and by addition of HNO3 or HCl the systems were titrated to about pH 3. The observed shifts of the IEPs were significant for the salt systems studied. Procedures were the same as given above for the salt titrations. The calibration procedures for pH and conductivity electrodes were as previously described.6

studied, mainly by force measurements34−36 and more recently second harmonic37 and sum frequency38 generation. The latter studies were all done on silica. A systematic study comparing two different cuts of one oxide with respect to ion specific effects has not been published to the best of our knowledge, although there are some experimental data from other groups,23,39 which were not discussed in that sense. Specific ion effects have been associated with the Hofmeister series,40,41 treated in terms of the lyotrophic series or interpreted in the simple structure breaking structure making picture for colloidal particles (and thus for oxide surfaces).31−33 Note that the structure-making/breaking notion used here is not necessarily in agreement with the sequences established by Marcus in his recent review.42 The notion is rather based on previous papers dealing with oxides.31,32,43 A recent overview for oxides can be found in a broader review article.44 The goal of this experimental study was to investigate whether sapphire-c and sapphire-r would exhibit different ionspecific effects, based on our view of the sapphire-c electrolyte interface as being related to “inert” surfaces and that of sapphire-r being more oxide like in the sense that more reactive groups (i.e., singly coordinated oxygens) are present on this surface.



EXPERIMENTAL SECTION

The sapphire substrates were obtained from TBL Kelpin (Germany) as precisely cut 20 mm by 10 mm substrates. Properties and specifications of these samples have been detailed by Rabung et al.45 The samples were cleaned as previously described6,45 and then mounted to the gap cell of the SurPass apparatus (Anton Paar) for streaming potential measurements. Streaming potential measurements on single crystal substrates of sapphire-c and sapphire-r in a variety of electrolyte solutions were carried out at pH ≈ 5.75 (the pH of Milli-Q water to mimic the CTR study in the presence of deionized water). This is a convenient pH, since it avoids complications that would arise from the addition of acid or base to the electrolyte solutions to keep the pH constant. Such complications occurred in a study on hematite particles:43 the authors reported that their HBr had included a contamination which had falsified previous results.31 We used salts of maximum purity and heated them prior to use to remove organic contaminants and hydration water. The salt solutions were prepared by weighing a known amount of dried salt into a volumetric flask and adding Milli-Q water. The pH of the salt solutions was verified prior to use and found close to that of the Milli-Q water assuring minimal variation of the pH during a salt titration (see the Supporting Information). For a salt titration, known volumes of a salt solution of known concentration were added via the high precision syringes mounted to the SurPass to a known volume of Milli-Q water (500 mL) after an initial measurement in Milli-Q water. Thus, the salt concentration was continuously increased. The pH was rather stable for the salt-titration experiments, and maximum drift was 0.2 pH units in a single run. Separate runs with one salt showed that the zeta potential versus salt concentration curves were not significantly affected by the slight variation in pH (see the Supporting Information). Both the setup and the syringes delivering the salt solutions of known concentration (all approximately 1 M) were extensively rinsed with fresh Milli-Q water before switching to a new salt system. As noted in the earlier paper, there is always leakage of KCl from the pH electrode.6 Some control experiments were therefore done in nitrate systems without the pH electrode being present and the pH was only measured at the end of the experiment. Furthermore, the conductivity of the solutions was monitored to cross-check that the additions of the salt solutions actually occurred. The experiments were carried out at room temperature, and the temperature was constantly monitored. To minimize interference from



RESULTS AND DISCUSSION pH Dependence and Intrinsic Isoelectric Points. The charging behaviors of the two cuts as a function of pH are presented in Figure 2. In agreement with previous results, the IEP of the c-cut is close to pH 4. For the r-plane, a similar result is obtained, which is corroborated by independent measurements on samples of the same origin in another laboratory (R. Zimmermann et al., unpublished results), and literature data also do support the low IEPs of these surfaces.23 On the basis of the ion-specific effects to be discussed next, it is possible to assign the data in Figure 2 to non-ion-specific conditions and the IEPs can be considered intrinsic. The IEPs should not be confused with points of zero surface potential which are expected to be different from pH 4 (see Figure 1). However, a comprehensive model involving the point of zero surface potential at pH 8 and an IEP of pH 4 for the r-plane requires adsorption of negatively charged solute within the shear-plane. D

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

Figure 2. Zeta-potentials of sapphire-c and sapphire-r as a function of pH.

It is clear from Figure 2 that, at pH 5.8, where the ion specific experiments were carried out, both sapphire-c and -r exhibit an intrinsic, negative zeta-potential. Specific Ion Effects on Sapphire-r. Figure 3 shows the variation of the zeta-potential for sapphire-r as a function of salt concentration for the anion series in sodium form (Figure 3a) and the cation series in chloride form (Figure 3b). The anion series shows no effect as expected for a negatively charged surface. Figure 3a also illustrates that experimental errors remain very small. The latter changes when the zeta-potential approaches zero or if the salt concentration increases further. The cation series (Figure 3b) shows that the most structure making ion (Li+) most strongly decreases the absolute value of the zeta-potential. The least strongly hydrated cation (Cs+) affects the zeta-potential least. This is the expected result for a structure making surface and in that sense the results agree with the outcome of the surface diffraction study.5 Specific Ion Effects on Sapphire-c. Figure 4 summarizes the results for sapphire-c, which are are unexpected for the anion series (Figure 4a) in sodium form in the sense that (i) a significant anion effect on a negatively charged oxide surface occurs, (ii) the zeta-potential variation with increasing NaCl concentrations disagrees with normal shielding behavior, and (iii) theNaBrO3 and NaNO3 additions decrease the absolute values of the zeta-potential clearly more strongly than in the case of sapphire-r (Figure 3a). A nitrate effect (relative to that of chloride) had already been observed in an earlier study6 and was studied here more systematically. It is opposite to observations for goethite, where chloride allows stronger shielding than nitrate.49 The anion effect in going from chloride to nitrate on sapphire-c is confirmed in the potassium salts as shown in Figure 4b. Figure 4b includes three separate KCl experiments showing that the data were very well reproducible. Due to the particular behavior of the chloride systems, the anion data do not follow a particular series. Figure 4c shows that (i) there is virtually no effect of the variation in the cation identity on the zeta-potential in the chloride series, and (ii) for the other cations the unexpected chloride behavior was retrieved.

Figure 3. (a) Zeta-potential of sapphire-r at pH 5.8 for sodium salts as a function of salt concentration. For NaCl the standard deviations are shown. They were typically 1 mV or lower. (b) Zeta-potential of sapphire-r at pH 5.8 for chloride salts as a function of salt concentration.

Possible causes for anion specificity encompass issues such as ion size, ion polarizability, or mono- versus polyatomic character of the ions.42 Further discussion of the anion effects would be rather speculative, in particular since more structure breaking monovalent anions such as bromide, iodide, and thiocyanicde could not be studied as explained in the experimental part and shown in Supporting Information. The observed anion effects (i.e., the occurrence of anionspecificity and the nonmonotonous behavior of the chloride systems) were entirely unexpected, but results from recent MD simulations50 are quite comparable: a spectrum of generic surface properties was simulated with respect to the concomitant specific anion effects. In the MD study, the nonmonotonous behavior experimentally observed in the chloride systems (Figure 4) was obtained for nonpolar surfaces with slightly negative charge in the presence of structure breaking anions. In the simulations, the effect occurs at 10 times higher salt concentration. However, the simulations are typically run at high salt concentrations, because lower concentrations would require larger systems and thus unrealistic calculation times. Therefore, the comparison E

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

obtain an anion effect.50 Interestingly, in a sum frequency generation study on OTS covered silica, chloride adsorption on a hydrophobic surface was indirectly inferred.51 Unfortunately, direct evidence of chloride adsorption for the systems studied here (i.e., single crystals with low surface area exposed, relatively high concentration of above 5 mM that cause the adsorption) is very difficult to obtain if possible at all. The structure breaking/making anion series was extended in sodium systems to include perchlorate as more structure breaking and bromate as a very structure making anion. The results are included in Figure 4a and confirm the series that would be expected from the simulation study50 (i.e., BrO3 < NO3 < ClO4) as translated to the structure breaking/making picture. The effect of the bromate ion is very strong. It was recently debated whether the related iodate ion could be considered as a strongly hydrated cation.52 Since the sodium nitrate as opposed to the sodium chloride resulted in more conventional screening, the cation series was repeated with nitrate as counter-ion. The results in Figure 5 reveal the cation sequence expected for a structure breaking negative surface.

Figure 5. Zeta-potential of sapphire-c at pH 5.8 for nitrate salts as a function of salt concentration.

Thus, in the nitrate system, we confirm the structure breaking property of the sapphire-c−water interface in agreement with the surface diffraction study.5 A very recent simulation study retrieves the inversion of cation specificity from hydrophobic to hydrophilic SAMs.53 To verify to what extent the ion specific behavior could be retrieved in shifts of the IEP, two selected salt systems were studied as a function of pH in the lower pH range (Figure 6). It is obvious that the IEPs in these systems are shifted to lower pH values compared to Figure 2, which pertains to the ion-unspecific concentration range and the intrinsic IEP at pH 4. Comparison between the chloride and nitrate systems suggests that chloride has a specific effect in the sense that it is adsorbed on sapphire-c and shifts the IEP to lower pH values at sufficiently high salt concentrations. In the LiNO3 system, a shift of the IEP is also observed compared to the sodium system, which was previously shown to cause no shift of the IEP6 in agreement with a common intersection at pH 4 in second harmonic generation data15 for sodium nitrate systems.

Figure 4. (a) Zeta-potential of sapphire-c at pH 5.8 for (a) sodium salts, (b) potassium salts, and (c) chloride salts as a function of salt concentration.

obviously is only tentative. The simulations suggest that the initial decrease in the absolute value of the zeta-potential is due to the shielding while the change in tendency stems from the adsorption of the anion. For the same surface, the authors F

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

specifically adsorbed to a negatively charged oxide surface, which is not uncommon in general,57 but unexpected to the extent observed in the present study for an oxide surface. It should be added, in this respect, that in a spectroscopic study nitrate accumulation on a flat-plate sample silica is reported above its IEP.58 No conclusive explanation is available for these counterintuitive observations. Overall, it is noteworthy that the reported ion specific effects occur at rather low concentrations and are clearly visible below 10 mM. While this is not usually the case, the most recent study by Morag et al.59 does report ion specificity at silica surfaces in 10 mM chloride systems. Furthermore, the force distance measurements by these authors support the relevance of surface hydration in that they observe an inversion of the cation series in going from high pH, where silica is strongly deprotonated and hydrophilic, to lower pH, where deprotonation is less pronounced or even absent and silica is sometimes termed hydrophobic. These results do agree with independent second harmonic work.37



Figure 6. Zeta-potential of sapphire-c at about 20 mM electrolyte concentration as a function of pH for some selected chloride and nitrate salts.

CONCLUSIONS Electrokinetic experiments were carried out on two distinct sapphire surfaces to study ion specific effects at pH about 5.8. Both surfaces have a net negative electrokinetic charge in MilliQ water at this pH, and the IEPs are at pH of about 4. Published surface diffraction data for the two crystal planes suggest significant differences in water structuring, which concur with expectations based on the MUSIC model. Sapphire-r is expected to be water-structuring and shows behavior that is in agreement with the simple structure breaking/structure making picture; that is, structure making cations show the strongest interactions with this surface. Sapphire-c shows much more complex behavior in various ways. Anion effects are observed on this negatively charged oxide surface. Chloride was found to surpress the cation effect that is retrieved in nitrate solutions. Chloride appears to be adsorbed on intrinsically negatively charged sapphire-c above about 5 mM concentrations. It is noteworthy that ion-specific effects were observed here at concentrations below 10 mM. Overall, the work supports the idea that surface hydration largely governs ion-specific effects, which is clearly emerging in the recent literature.53,55,59,60

The shift of the IEP in LiNO3 to lower pH is in agreement with the trend shown in Figure 5. In the Supporting Information, a discussion of the structure breaking properties of sapphire-c is included. Briefly, the measured oxygen profile reported by Catalano5 compares very favorably to generic simulations for hydrophobic surfaces.54 Compared to our previous interpretation the inclusion of the surface diffraction data5 then requires a modification, in the sense that the “inert”-surface behavior would be associated with the surface itself. It would still explain the low IEP of sapphire-c by hydroxide adsorption but without the need to involve a 2Dbilayer water. The dominance of surface chemistry for sapphirer certainly disagrees with the measured IEP, which should occur close to the point of zero surface potential. Interestingly, the water profile from the surface diffraction study indicates secondary water ordering on sapphire-r, that is, a weakly ordered layer of water, beyond the strongly ordered water adjacent to the solid (see the Supporting Information). The secondary water is very reminiscent of the primary ordering on sapphire-c, that is, a nonpolar surface. If this secondary water favored hydroxide adsorption, it could explain the low IEP. Overall the results suggest that the ion specific effects are largely driven by solvation thermodynamics, that is, entropic effects. This conclusion was drawn in a comparison study dealing with monovalent ions on hydrophobic and hydrophilic colloids, where experimental and simulation work was presented and where the relation to water structure profiles was made.55 Here, similar conclusions are possible for two cuts of the same oxide. Variation of both cation and anion in a full series is not frequently reported in experimental studies on ion specificity. Because of the cooperative effect of nitrate/chloride on sapphire-c, a short discussion is included in the Supporting Information on data on the air−water surface56 (i.e., another inert surface). Briefly, surface potential data in chloride and nitrate by Chartier and Fotouhi56 reveal a cooperative effect, in the sense that the monovalent cations behave differently in nitrate and chloride systems. A possible interpretation of the chloride data on sapphire-c involves the counterintuitive assumption that chloride is



ASSOCIATED CONTENT

S Supporting Information *

Information dealing with the comparison of CTR data for three cuts of sapphire, discussing SFG data for three cuts of sapphire, plotting literature data for calculated water structure profiles on hydrophobic surfaces in comparison to measured CTR data for sapphire-c, a comparison of ion specific effects between the air−water and sapphire-c−water interfaces, and experimental data that show that pH changes during the salt titrations are small and their effect on zeta-potential vs salt concentration curves are not significant. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Telephone: +49 721 608-24023. Fax: +49 721 608-23927. Email: [email protected]. Notes

The authors declare no competing financial interest. G

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir



Article

(17) Gaigeot, M. P.; Sprik, M.; Sulpizi, M. Oxide/water interfaces: how the surface chemistry modifies interfacial water properties. J. Phys.: Condens. Matter 2012, 24 (12), 124106. (18) Pieniazek, P. A.; Tainter, C. J.; Skinner, J. L. Interpretation of the water surface vibrational sum-frequency spectrum. J. Chem. Phys. 2011, 135 (4), 044701. (19) Hsu, P. Y.; Dhinojwala, A. Contact of Oil with Solid Surfaces in Aqueous Media Probed Using Sum Frequency Generation Spectroscopy. Langmuir 2012, 28 (5), 2567−2573. (20) Braunschweig, B.; Eissner, S.; Daum, W. Molecular structure of a mineral/water interface: Effects of surface NanoRoughness of αAl2O3(0001). J. Phys. Chem. C 2008, 112 (6), 1751−1754. (21) Zhang, D.; Wang, Y.; Gan, Y. Characterization of critically cleaned sapphire single-crystal substrates by atomic force microscopy, XPS and contact angle measurements. Appl. Surf. Sci. 2013, 274, 405− 417. (22) Gentleman, M. M.; Ruud, J. A. Role of Hydroxyls in Oxide Wettability. Langmuir 2010, 26 (3), 1408−1411. (23) Kershner, R. J.; Bullard, J. W.; Cima, M. J. Zeta potential orientation dependence of sapphire substrates. Langmuir 2004, 20 (10), 4101−4108. (24) Sung, J. H.; Zhang, L. N.; Tian, C. S.; Shen, Y. R.; Waychunas, G. A. Effect of pH on the Water/α-Al2O3 (110̅ 2) Interface Structure Studied by Sum-Frequency Vibrational Spectroscopy. J. Phys. Chem. C 2011, 115 (28), 13887−13893. (25) Tougerti, A.; Methivier, C.; Cristol, S.; Tielens, F.; Che, M.; Carrier, X. Structure of clean and hydrated α-Al2O3 (11̅02) surfaces: implication on surface charge. Phys. Chem. Chem. Phys. 2011, 13 (14), 6531−6543. (26) Sung, J.; Shen, Y. R.; Waychunas, G. A. The interfacial structure of water/protonated α-Al2O3 (112̅0) as a function of pH. J. Phys.: Condens. Matter 2012, 24 (12), 124101. (27) Shimizu, K.; Shchukarev, A.; Kozin, P. A.; Boily, J. F. X-ray photoelectron spectroscopy of fast-frozen hematite colloids in aqueous solutions. 4. Coexistence of alkali metal (Na+, K+, Rb+, Cs+) and chloride ions. Surf. Sci. 2012, 606 (13−14), 1005−1009. (28) Rosenholm, J. B.; Kosmulski, M. Peculiar charging effects on titania in aqueous 1:1, 2:1, 1:2 and mixed electrolyte suspensions. Adv. Colloid Interface Sci. 2012, 179−182, 51−67. (29) Franks, G. V.; Johnson, S. B.; Scales, P. J.; Boger, D. V.; Healy, T. W. Ion-specific strength of attractive particle networks. Langmuir 1999, 15 (13), 4411−4420. (30) Johnson, S. B.; Franks, G. V.; Scales, P. J.; Healy, T. W. The binding of monovalent electrolyte ions on alpha-alumina. II. The shear yield stress of concentrated suspensions. Langmuir 1999, 15 (8), 2844−2853. (31) Dumont, F.; Vantan, D.; Watillon, A. Study of Ferric-Oxide Hydrosols from Electrophoresis, Coagulation, and Peptization Measurements. J. Colloid Interface Sci. 1976, 55 (3), 678−687. (32) Dumont, F.; Watillon, A. Stability of Ferric Oxide Hydrosols. Discuss. Faraday Soc. 1971, 52, 352−&. (33) Kallay, N.; Colic, M.; Fuerstenau, D. W.; Jang, H. M.; Matijevic, E. Lyotropic Effect in Surface-Charge, Electrokinetics, and Coagulation of a Rutile Dispersion. Colloid Polym. Sci. 1994, 272 (5), 554−561. (34) Chapel, J. P. Electrolyte Species-Dependent Hydration Forces between Silica Surfaces. Langmuir 1994, 10 (11), 4237−4243. (35) Dishon, M.; Zohar, O.; Sivan, U. Effect of Cation Size and Charge on the Interaction between Silica Surfaces in 1:1, 2:1, and 3:1 Aqueous Electrolytes. Langmuir 2011, 27 (21), 12977−12984. (36) Dishon, M.; Zohar, O.; Sivan, U. From Repulsion to Attraction and Back to Repulsion: The Effect of NaCl, KCl, and CsCl on the Force between Silica Surfaces in Aqueous Solution. Langmuir 2009, 25 (5), 2831−2836. (37) Azam, M. S.; Weeraman, C. N.; Gibbs-Davis, J. M. Specific Cation Effects on the Bimodal Acid-Base Behavior of the Silica/Water Interface. J. Phys. Chem. Lett. 2012, 3 (10), 1269−1274. (38) Flores, S. C.; Kherb, J.; Konelick, N.; Chen, X.; Cremer, P. S. The Effects of Hofmeister Cations at Negatively Charged Hydrophilic Surfaces. J. Phys. Chem. C 2012, 116 (9), 5730−5734.

ACKNOWLEDGMENTS The author is grateful for the review of a previous version of the manuscript by three anonymous referees. Their comments led to substantial improvements. In particular, pointing out reference 59 was very supportive. The author is indebted to Jeff Catalano for sharing the numerical data of his previously published CTR data (ref 5) and to Pavel Jungwirth for making available the numerical values of water profiles (ref 54).



REFERENCES

(1) Hiemstra, T.; Dewit, J. C. M.; Vanriemsdijk, W. H. Multisite Proton Adsorption Modeling at the Solid-Solution Interface of (Hydr)Oxides - a New Approach 0.2. Application to Various Important (Hydr)Oxides. J. Colloid Interface Sci. 1989, 133 (1), 105−117. (2) Hiemstra, T.; Vanriemsdijk, W. H.; Bolt, G. H. Multisite Proton Adsorption Modeling at the Solid-Solution Interface of (Hydr)Oxides - a New Approach 0.1. Model Description and Evaluation of Intrinsic Reaction Constants. J. Colloid Interface Sci. 1989, 133 (1), 91−104. (3) Hiemstra, T.; Venema, P.; VanRiemsdijk, W. H. Intrinsic proton affinity of reactive surface groups of metal (hydr)oxides: The bond valence principle. J. Colloid Interface Sci. 1996, 184 (2), 680−692. (4) Venema, P.; Hiemstra, T.; Weidler, P. G.; van Riemsdijk, W. H. Intrinsic proton affinity of reactive surface groups of metal (hydr)oxides: Application to iron (hydr)oxides. J. Colloid Interface Sci. 1998, 198 (2), 282−295. (5) Catalano, J. G. Weak interfacial water ordering on isostructural hematite and corundum (001) surfaces. Geochim. Cosmochim. Acta 2011, 75 (8), 2062−2071. (6) Lutzenkirchen, J.; Kupcik, T.; Fuss, M.; Walther, C.; Sarpola, A.; Sundman, O. Adsorption of Al-13-Keggin clusters to sapphire c-plane single crystals: Kinetic observations by streaming current measurements. Appl. Surf. Sci. 2010, 256 (17), 5406−5411. (7) Ranea, V. A.; Carmichael, I.; Schneider, W. F. DFT Investigation of Intermediate Steps in the Hydrolysis of α-Al2O3(0001). J. Phys. Chem. C 2009, 113 (6), 2149−2158. (8) Thissen, P.; Grundmeier, G.; Wippermann, S.; Schmidt, W. G. Water adsorption on the α-Al2O3(0001) surface. Phys. Rev. B 2009, 80, 24. (9) Zhang, L.; Tian, C.; Waychunas, G. A.; Shen, Y. R. Structures and charging of α-alumina (0001)/water interfaces studied by sumfrequency vibrational spectroscopy. J. Am. Chem. Soc. 2008, 130 (24), 7686−7694. (10) Chatman, S.; Zarzycki, P.; Preočanin, T.; Rosso, K. M. Effect of surface site interactions on potentiometric titration of hematite (αFe2O3) crystal faces. J. Colloid Interface Sci. 2013, 391, 125−134. (11) Healy, T. W.; Fuerstenau, D. W. The isoelectric point/point-of zero-charge of interfaces formed by aqueous solutions and nonpolar solids, liquids, and gases. J. Colloid Interface Sci. 2007, 309 (1), 183− 188. (12) Beattie, J. K.; Djerdjev, A. M. The pristine oil/water interface: Surfactant-free hydroxide-charged emulsions. Angew. Chem., Int. Ed. 2004, 43 (27), 3568−3571. (13) Creux, P.; Lachaise, J.; Graciaa, A.; Beattie, J. K.; Djerdjev, A. M. Strong Specific Hydroxide Ion Binding at the Pristine Oil/Water and Air/Water Interfaces. J. Phys. Chem. B 2009, 113 (43), 14146−14150. (14) Yang, D.; Krasowska, M.; Sedev, R.; Ralston, J. The unusual surface chemistry of α-Al2O3 (0001). Phys. Chem. Chem. Phys. 2010, 12 (41), 13724−13729. (15) Fitts, J. P.; Shang, X. M.; Flynn, G. W.; Heinz, T. F.; Eisenthal, K. B. Electrostatic surface charge at aqueous/α-Al2O3 single-crystal interfaces as probed by optical second-harmonic generation. J. Phys. Chem. B 2005, 109 (16), 7981−7986. (16) Argyris, D.; Ho, T. A.; Cole, D. R.; Striolo, A. Molecular Dynamics Studies of Interfacial Water at the Alumina Surface. J. Phys. Chem. C 2011, 115 (5), 2038−2046. H

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX

Langmuir

Article

(39) Franks, G. V.; Meagher, L. The isoelectric points of sapphire crystals and alpha-alumina powder. Colloids Surf., A 2003, 214 (1−3), 99−110. (40) Kunz, W. Specific ion effects in liquids, in biological systems, and at interfaces. Pure Appl. Chem. 2006, 78 (8), 1611−1617. (41) Kunz, W.; Lo Nostro, P.; Ninham, B. W. Hofmeister phenomena. Curr. Opin. Colloid Interface Sci. 2004, 9 (1−2), 9. (42) Marcus, Y. Effect of Ions on the Structure of Water: Structure Making and Breaking. Chem. Rev. 2009, 109 (3), 1346−1370. (43) Dumont, F.; Warlus, J.; Watillon, A. Influence of the Point of Zero Charge of Titanium-Dioxide Hydrosols on the Ionic Adsorption Sequences. J. Colloid Interface Sci. 1990, 138 (2), 543−554. (44) Lyklema, J. Simple Hofmeister series. Chem. Phys. Lett. 2009, 467 (4−6), 217−222. (45) Rabung, T.; Schild, D.; Geckeis, H.; Klenze, R.; Fanghanel, T. Cm(III) sorption onto sapphire (α-Al2O3) single crystals. J. Phys. Chem. B 2004, 108 (44), 17160−17165. (46) Lutzenkirchen, J. Influence of impurities on acid-base data for oxide minerals - Analysis of “Observable” surface charge and proton affinity distributions and model calculations for single crystal samples. Croat. Chem. Acta 2007, 80 (3−4), 333−343. (47) Lutzenkirchen, J. Influence of impurities on acid-base data for oxide minerals - Analysis of Observable surface charge and proton affinity distributions and model calculations for single crystal samples (vol 80, pg 333, 2007). Croat. Chem. Acta 2010, 83 (4), vii−vii. (48) Maczka, E.; Luetzenkirchen, J.; Kosmulski, M. The significance of the solid-to-liquid ratio in the electrokinetic studies of the effect of ionic surfactants on mineral oxides. J. Colloid Interface Sci. 2013, 393, 228−233. (49) Lutzenkirchen, J.; Boily, J. F.; Gunneriusson, L.; Lovgren, L.; Sjoberg, S. Protonation of different goethite surfaces - Unified models for NaNO3 and NaCl media. J. Colloid Interface Sci. 2008, 317 (1), 155−165. (50) Schwierz, N.; Horinek, D.; Netz, R. R. Reversed Anionic Hofmeister Series: The interplay of Surface Charge and Surface Polarity. Langmuir 2010, 26 (10), 7370−7379. (51) Tian, C. S.; Shen, Y. R. Structure and charging of hydrophobic material/water interfaces studied by phase-sensitive sum-frequency vibrational spectroscopy. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (36), 15148−15153. (52) Baer, M. D.; Pham, V. T.; Fulton, J. L.; Schenter, G. K.; Balasubramanian, M.; Mundy, C. J. Is Iodate a Strongly Hydrated Cation? J. Phys. Chem. Lett. 2011, 2 (20), 2650−2654. (53) Schwierz, N.; Horinek, D.; Netz, R. R. Anionic and Cationic Hofmeister Effects on Hydrophobic and Hydrophilic Surfaces. Langmuir 2013, 29 (8), 2602−2614. (54) Vacha, R.; Zangi, R.; Engberts, J. B. F. N.; Jungwirth, P. Water structuring and hydroxide ion binding at the interface between water and hydrophobic walls of varying rigidity and van der waals interactions. J. Phys. Chem. C 2008, 112 (20), 7689−7692. (55) Calero, C.; Faraudo, J.; Bastos-González, D. Interaction of Monovalent Ions with Hydrophobic and Hydrophilic Colloids: Charge Inversion and Ionic Specificity. J. Am. Chem. Soc. 2011, 133 (38), 15025−15035. (56) Chartier, P.; Fotouhi, B. Surface-Potential of Dilute AqueousSolutions - Case of Ferro-Cyanide and Ferri-Cyanide and Some Others Dissymetrical Electrolytes. J. Chim. Phys. Phys.-Chim. Biol. 1974, 71 (3), 335−345. (57) Barany, S. Complex electrosurface investigations of dispersed microphases. Adv. Colloid Interface Sci. 1998, 75 (1), 45−78. (58) Hayes, P. L.; Malin, J. N.; Konek, C. T.; Geiger, F. M. Interaction of nitrate, barium, strontium and cadmium ions with fused quartz/water interfaces studied by second harmonic generation. J. Phys. Chem. A 2008, 112 (4), 660−668. (59) Morag, J.; Dishon, M.; Sivan, U. The Governing Role of Surface Hydration in Ion Specific Adsorption to Silica: An AFM-Based Account of the Hofmeister Universality and Its Reversal. Langmuir 2013, 29 (21), 6317−6322.

(60) López-León, T.; Santander-Ortega, M. J.; Ortega-Vinuesa, J. L.; Bastos-González, D. Hofmeister Effects in Colloidal Systems: Influence of the Surface Nature. J. Phys. Chem. C 2008, 112 (41), 16060−16069.

I

dx.doi.org/10.1021/la401509y | Langmuir XXXX, XXX, XXX−XXX