From Repulsion to Attraction and Back to ... - ACS Publications

Feb 6, 2009 - Matan Dishon, Ohad Zohar, and Uri Sivan*. Faculty of Physics and the Russell Berrie Nanotechnology Institute, TechnionsIsrael Institute ...
0 downloads 0 Views 825KB Size
Langmuir 2009, 25, 2831-2836

2831

From Repulsion to Attraction and Back to Repulsion: The Effect of NaCl, KCl, and CsCl on the Force between Silica Surfaces in Aqueous Solution Matan Dishon, Ohad Zohar, and Uri Sivan* Faculty of Physics and the Russell Berrie Nanotechnology Institute, TechnionsIsrael Institute of Technology, Haifa 32000, Israel ReceiVed September 15, 2008. ReVised Manuscript ReceiVed December 23, 2008 The force between silica surfaces in NaCl, KCl and CsCl aqueous solutions is studied at pH 5.5 using an atomic force microscope (AFM). As ion concentration is increased, more cations adsorb to the negatively charged silica, gradually neutralizing the surface charge, hence, suppressing the electrostatic double layer repulsion and revealing van der Waals attraction. At even higher salt concentrations, repulsion reemerges due to surface charge reversal by excess adsorbed cations. Adsorption grows monotonically with cation radius. At pH 5.5 the smallest ion, Na+, neutralizes the surface at 0.5-1 M, K+ at 0.2-0.5 M, and Cs+ at ∼0.1 M. Titration with HCl to pH 4.0 shifts surface neutralization and charge reversal to lower salt concentrations compared with pH 5.5. When attraction dominates, the force curves are practically identical for the three salts, independent of their concentration.

Introduction The interaction between charged bodies immersed in aqueous electrolyte solution is relevant to fields ranging from molecular biology and bioengineering to colloid-based technologies found in water purification, cosmetics, inks, paints, and nutrients. At large separations between the charged surfaces, and millimolar concentrations of monovalent salts, the force is adequately described by the Poisson-Boltzmann (PB) equation. However, at separations smaller than the screening length or higher salt concentrations the interaction deviates significantly from the PB prediction due to multiple additional effects. One of the more intriguing aspects of the interaction in the latter regimes is broadly classified as ion-specific effects, namely, the marked differences in interaction between the same charged objects when immersed in different electrolyte solutions. Following pioneering experiments by Franz Hofmeister1 in the late 19th century, it became clear that aggregation and sedimentation of proteins in saline solutions vary markedly with ion type, even of the same valence and chemical nature. As time elapsed, ionspecific effects emerged as the rule rather than the exception in ample other cases including polymer and biopolymer sedimentation and adsorption to surfaces,2 molecule and colloid aggregation, gelation,3 ζ-potential and yield stress4 of colloidal suspensions, clay swelling,5 and, most relevant to our case, the force between two bodies in electrolyte solution as measured by surface force apparatus (SFA)6-8 and atomic force microscope (AFM).9-12 * To whom correspondence should be addressed. E-mail: phsivan@ technion.ac.il. (1) Kunz, W.; Henle, J.; Ninham, B. W. Curr. Opin. Colloid Interface Sci. 2004, 9, 19–37. (2) Cacace, M. G.; Landau, E. M.; Ramsden, J. J. Q. ReV. Biophys. 1997, 30, 241-277 and references therein. (3) Trompette, J. L.; Meireles, M. J. Colloid Interface Sci. 2003, 263, 522– 527. (4) Franks, G. V. J. Colloid Interface Sci. 2002, 249, 44–51. (5) Zhang, F.; Low, P. F.; Roth, C. B. J. Colloid Interface Sci. 1995, 173, 34–41. (6) Israelachvili, J. N.; Adams, G. E. J. Chem. Soc., Faraday Trans. 1 1978, 74, 975–1001. (7) Pashley, R. M. J. Colloid Interface Sci. 1981, 83, 531–546. (8) Chapel, J.-P. Langmuir 1994, 10, 4237–4243. (9) Hillier, A. C.; Kim, S.; Bard, A. J. J. Phys. Chem. 1996, 100, 18808– 18817.

Tadros and Lyklema13 were among the first to measure the surface charge of silica colloids in the presence of different alkali cations and variable pH. In the range of 6 g pH g 3 the negative surface charge was found to follow the series |σCsCl| < |σKCl| < |σNaCl|, indicating better adsorption of the larger cations to the charged silica surface. At pH g 7 the order of adsorption was reversed. The former order of affinity was consistent with earlier assays of cation affinity to silica gel.14 The same order was later observed in ζ-potential and yield stress measurements on silica colloids.4 The opposite ordering at pH 8 was also found by Dove et al.15 Better adsorption of larger cations to silica and mica at slightly acidic pH values was seen in published force measurements8,10–12 that studied short-range (e1 nm) hydration forces in the presence of different cations. Chapel8 has utilized SFA for measuring the short-range interaction between two pyrogenic silica surfaces in the presence of LiCl, NaCl, KCl, and CsCl solutions in the 10-4 to 10-1 M concentration range. Vakarelski and co-workers10,11 utilized AFM to study the adhesion and short-range interaction between a silica colloid and a mica surface in the presence of pure water and several alkali halide salts. Special emphasis was put in those publications on dynamic aspects of the interaction which the authors attributed to rearrangement of the structured cation/water layer at the liquid/solid interface. A strong adhesive force was found with highly hydrated ions (Li+, Na+), whereas poorly hydrated ions (K+, Cs+) led to weak adhesion. Donose et al.12 studied the effect of various alkali ions on the friction between a silica colloid and a silica wafer at different loading forces and scan rates. Smaller, tightly hydrated cations were found to provide better lubrication compared with layers of larger, weakly hydrated cations. Interestingly, in contrast to the case of silica colloid on mica surface,10,11 no adhesion was found in ref (10) Vakarelski, I. U.; Ishimura, K.; Higashitani, K. J. Colloid Interface Sci. 2000, 227, 111–118. (11) Vakarelski, I. U.; Higashitani, K. J. Colloid Interface Sci. 2001, 242, 110–120. (12) Donose, B. C.; Vakarelski, I. U.; Higashitani, K. Langmuir 2005, 21, 1834–1839. (13) Tadros, T. F.; Lyklema, J. J. Electroanal. Chem. 1968, 17, 267–275. (14) Tien, H. T. J. Phys. Chem. 1965, 69, 350–352. (15) Dove, P. M.; Craven, C. M. Geochim. Cosmochim. Acta 2005, 69, 4963– 4970.

10.1021/la803022b CCC: $40.75  2009 American Chemical Society Published on Web 02/06/2009

2832 Langmuir, Vol. 25, No. 5, 2009

12, suggesting that the interaction at vanishing separations could be different for silica on mica compared with two silica surfaces. No systematic study of the normal force as a function of salt concentration and type was reported in refs 10–12. Fair amounts of theoretical efforts16 have been invested in understanding the interaction of different monovalent ions with charged surfaces, but due to the scarceness of dedicated experiments the explanation of these dramatic effects remained undecided and invariably neglected in practical calculations of, e.g., protein-protein or protein-membrane interactions. In fact, some calculations17 predict better adsorption of small ions. Motivated by the lack of systematic experimental studies of such a fundamental phenomenon we have launched a series of experiments focused on characterizing differences in the interaction between two silica surfaces, as measured by AFM, in the presence of three monovalent salt solutions: NaCl, KCl, and CsCl. Using an extremely quiet measuring system and covering a broad range of parameters over many samples we were able to map ion-specific effects in great detail. The qualitative behavior was found to be similar for all salts. At concentrations below 2 mM no cation condensation was observed, and the two surfaces repelled each other due to overlap of their respective double layers. With the addition of salt, cations started to condense on the silica surfaces, eventually neutralizing them at a certain ion-specific concentration. When neutrality was approached, the diminishing repulsion was gradually displaced by van der Waals (vdW) attraction except, perhaps, at small separations. Addition of salt beyond neutralization led to excess cation condensation and charge reversal. As a result, electrical double layer repulsion reappeared, this time between the two positively overcharged surfaces. In accord with previous experiments we found that the tendency for condensation grew monotonically with the bare ion radius, Cs+ > K+ > Na+. Interestingly, the attraction at the point of neutrality was independent of ion type. The transition from negative to neutral and then to positive surface charge by specific cation adsorption resembles charge reversal as a function of pH, although the microscopic mechanism might be different. Our interpretation of the reemerging repulsion is different than previous identification of the same phenomenon with hydration repulsion.18,19 In particular, it should not be confused with possible hydration forces observed at subnanometer separations.7,8,20,21

Experimental Details The experiment was set as described by Zohar et al.22 Briefly, a 5 µm diameter silica bead (Bangs Laboratories) was glued to the AFM tip using glass bond (Loctite). After UV curing of the glue, the tip and a virgin silicon(100) substrate cleaned in ethanol and DI water were placed in oxygen plasma (Axic Multimode HF-8200, 50 mTorr, 100 W) for 50 min and immediately introduced into the AFM (MultimodesVeeco) fluid cell. Long plasma treatment turned out to be crucial for sample to sample reproducibility and stability over extended periods of time. The AFM was placed in an acoustic hood, and its piezoelectric crystal was driven by a low-noise HP3325B (16) Bostrom, M.; Deniz, V.; Frank, G. V.; Ninham, B. W. AdV. Colloid Interface Sci. 2006, 123, 5-15 and references therein. (17) Sverjensky, D. A. Geochim. Cosmochim. Acta 2005, 69, 225–257. (18) Fielden, M. L.; Hayes, R. A.; Ralston, J. Phys. Chem. Chem. Phys. 2000, 2, 2623–2628. (19) Valle-Delgado, J. J.; Molina-Bolivar, J. A.; Galisteo-Gonzalez, F.; GalvezRuiz, M. J.; Feiler, A.; Rutland, M. W. J. Chem. Phys. 2005, 123, 12. (20) Pashley, R. M.; Israelachvili, J. N. J. Colloid Interface Sci. 1984, 97, 446–455. (21) Grabbe, A.; Horn, R. G. J. Colloid Interface Sci. 1993, 157, 375–383. (22) Zohar, O.; Leizerson, I.; Sivan, U. Phys. ReV. Lett. 2006, 96, 177802– 177804.

Dishon et al.

Figure 1. Force divided by colloid radius vs separation for (red) NaCl, (green) KCl, (blue) CsCl, (solid black) DLVO with constant charge (CC) boundary conditions and the Hamaker constant, A ) 2.2 × 10-21 J, found in this study, (dashed black) DLVO with constant surface potential (CP) boundary conditions: (a) 1 mM pH 5.5; DLVO CC calculated with -0.015 e/nm2, DLVO CP calculated with -33.4 mV; (b) 10 mM pH 5.5; DLVO CC with -0.029 e/nm2, DLVO CP with -21.9 mV. Inset to panel a: magnified view of the small separation data in 1 mM.

synthesizer followed by a homemade low-pass filter with a 6 Hz roll-off frequency. Deflection signal and piezoelectric driving signal were recorded using an external DAQ board (PCI-6289, National Instruments) and later analyzed using Matlab code. This arrangement yielded superior data compared with the commercial AFM electronics, especially at small separations. Spring constants of the gold-coated silicon nitride cantilevers (Veeco model MLCTAUHW) varied between 0.1 and 0.5 Nm-1, as measured by analysis of their thermal fluctuations. Distances between the tip and the surface were calibrated by finding the relation between the bias applied to the piezoelectric crystal and tip deflection when resting on the substrate. Typical approach velocities were kept low (∼100 nm/s) to avoid hydrodynamic effects. All solutions were prepared from 18 MΩ DI water and analytic grade salts (NaCl and CsCl from Fluka, KCl from Merck). Each set of measurements began and ended by measuring the force in a calibration solution (NaCl 0.1 mM), confirming sample stability over the whole experiment. Rare cases where the force at the end of the measurements did not coincide with the initial curve were discarded. Statistics over nine different colloid-substrate pairs at 0.1 mM NaCl pH 5.5 yielded 6% variance in surface charge. We were therefore careful to compare data sets measured on the same sample or on two identical samples. Specifically, the NaCl and KCl curves in Figures 1-4 were measured with the same colloid-substrate pair, whereas the CsCl curve was measured on another pair. The data set presented in Figure 5 was measured with a third colloid-substrate pair.

Results and Discussion Figure 1a displays semilog plots of the force, F, divided by colloid radius, R, versus separation, h, for the three monovalent salts NaCl, KCl, and CsCl, all in 1 mM concentration (F/R ) 2π times the energy per unit area in Derjaguin approximation23). TheblacklinescorrespondtoDerjaguin-Landau-Verwey-Overbeek (DLVO)24 theory with constant charge (solid) or constant potential (dashed), both calculated with ε ) 78 for the water dielectric (23) Derjaguin, B. Colloid Polym. Sci. 1934, 69, 155–164.

Silica Surfaces in Alkali Halide Aqueous Solutions

Figure 2. Force divided by colloid radius vs separation for (red) NaCl, (green) KCl, (blue) CsCl at (a) 50 mM pH 5.5 and (b) 100 mM pH 5.5 salts. (black) vdW using A ) 2.2 × 10-21 J. Inset to panel b: same data as in the main figure (KCl and CsCl only) plotted in a double log scale. The black straight line depicts the theoretical expectation for vdW force, F/R ∼ h-2.

constant. At separations larger than a few Debye-Huckle (DH) screening lengths (λ ) 9.6 nm at 1 mM) all three curves decay exponentially with h/λ. The essentially identical force in all three cases indicates identical surface charge for the three salts. At smaller separations the three force curves split, indicating a different degree of surface charge regulation by the three cations. As seen in the inset to Figure 1a, charge regulation obeys the Hofmeister series, Cs+ > K+ > Na+. At very small separations, h e 0.3 nm, additional short-range repulsion appears in all curves. This repulsion is traditionally attributed to hydration forces.7,8,20,21 The effect of vdW interaction at short distances is pronounced in Cs+ and K+ and less so in Na+. Since the CsCl curve was measured on a different sample than the NaCl and KCl ones, the identical force curves at large separations testify to the reproducibility of our sample preparation procedure. The same type of data for 10 mM concentration is depicted in Figure 1b. While the characteristic decay lengths of the three curves remain equal and continue to agree with DH screening length the surface charge, as reflected in the magnitude of force, deviates between the three salts. We find the following order of surface charge, |σCsCl| < |σKCl| < |σNaCl|, and correspondingly an opposite ordering of cation adsorption: Cs+ > K+ > Na+. The same trend has also been observed in ζ-potential and yield stress measurements in suspensions4 of silica colloids and force measurements between pyrogenic silica sheets.8 Such ordering correlates with the level of hydration of the three cations suggesting that sodium, holding strongly to its hydration shell, is expelled from the Stern layer, whereas potassium, and even more so cesium, tend to lose their hydration shell and adsorb to the negatively charged silica. A dramatic change takes place when the concentration of salt is increased to 50 mM. As seen in Figure 2a, the interaction between the two negatively charged silica surfaces turns attractive in the presence of CsCl. The gap in data points below ∼1.2 nm reflects tip snapping to the surface, taking place when the differential attraction, dF/dh, exceeds the cantilever’s spring constant. We find then that at 50 mM CsCl, surface charge is (24) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press Inc.: San Diego, CA, 1991; p 246.

Langmuir, Vol. 25, No. 5, 2009 2833

suppressed by Cs+ adsorption to a point where vdW attraction dominates. Reduced repulsion at large separations and tip snapping to the surface due to short-range attraction are observed for KCl, indicating significant condensation of potassium ions on the silica surface. However, double layer repulsion due to the residual charge masks the vdW attraction at large separations. Tip snapping is restricted to small separations, where the attraction increases fast with shrinking distance. The strong repulsion in NaCl at all distances indicates little adsorption of Na+ to the silica surfaces. At 100 mM concentration (Figure 2b) the interaction in the presence of KCl turns attractive, whereas in NaCl repulsion persists for h e 3 nm. For h > 3 nm attraction is observed and the force converges to the KCl and CsCl curves. At those separations the double layer repulsion has already exponentially decayed (λ ≈ 1 nm), giving way to the algebraically decaying vdW attraction. Figure 2b proves that the same two objects may repel each other in a sodium-rich environment and attract in a similar potassium concentration. The variation of electrolyte composition between different environments in living organisms offers vast grounds for ion-specific effects. Many human cells, for instance, are known to function at high potassium (160 ( 10 mM) and low sodium (10 mM) concentrations,25 whereas extracellular fluids are typically characterized by an opposite composition (human blood plasma, for instance, contains 141 ( 6 mM Na+ and 4.1 ( 0.7 mM K+). Notwithstanding the marked differences between silica surfaces and biomolecules it is interesting to note that two silica surfaces may attract each other in potassium concentrations corresponding to intracellular environments and repel in sodium concentrations characterizing extracellular fluids. It remains to be checked whether such a transition from repulsion to attraction due to ion-specific effects occurs also between proteins, nucleic acids, membranes, etc., since it may affect our perception of the interaction between biomolecules in their natural environment. With the concentration of salt increased to 200 mM (Figure 3a) the interaction in the presence of NaCl turns attractive as well, except at small separations where short-range repulsion commences (λ ≈ 0.7 nm at 200 mM). At distances beyond 3 nm the interaction is dominated by vdW attraction. Note that for h e 3 nm the attraction in the presence of CsCl has grown smaller compared with KCl, indicating an onset of overcharging of the silica by excess Cs+ ions condensing on its surface. The same phenomenon is observed to greater extent in 500 mM solutions (Figure 3b). In analogy with measurements of force as a function of pH, we argue that cesium ions continue to accumulate on the, now, positively charged silica surfaces, generating pronounced short-range electrical double layer repulsion. At larger separations attraction dominates. The identity of the three force curves corresponding to the different salts supports the identification of the attraction with vdW interaction. Moreover, the fit of the attraction curves in different salt concentrations (Figures 2b and 3) to vdW force with the same Hamaker constant indicates independence upon salt concentration, as expected from this mechanism. Comparison between the KCl and the NaCl curves discloses a slight deviation of the former toward a less attractive interaction, indicating the onset of charge reversal by the condensed K+ ions. This slight deviation grows further at 1 M (not shown). Charge reversal by the same cations has been reported for silica colloids based on ζ-potential measurements.4 Since the silica colloid and the oxidized silicon wafer may display different surfaces it was important to rule out the possibility

2834 Langmuir, Vol. 25, No. 5, 2009

Dishon et al.

Figure 5. Force divided by colloid radius vs separation for two pH values and three 100 mM salts.

Figure 3. Force divided by colloid radius vs separation for (red) NaCl, (green) KCl, (blue) CsCl, (black) vdW using A ) 2.2 × 10-21 J at (a) 200 mM pH 5.5 and (b) 500 mM pH 5.5 salts. Inset to panel b: same data as in the main figure (NaCl and KCl only) plotted in a double log scale. The black straight line depicts the theoretical expectation for vdW force, F/R ∼ h-2.

Figure 4. Total surface charge vs salt concentration for the three salts studied in the experiment at pH 5.5. Surface charge was estimated by fitting the solution of the Poisson-Boltzmann equation to the exponentially decaying force at large separations.

that the observed phenomena originated from asymmetry between surfaces. To that end we have repeated some of the force measurements with the silicon substrate replaced by a silica colloid glued to the substrate. All the features displayed in Figures 1-3 repeated themselves in the force measured between the two identical colloids, proving the independence of our results upon small surface differences. To the best of our knowledge, the present manuscript is the first to use force measurements for studying overcharging of silica surfaces by monovalent cation condensation. It is also worth mentioning that charge reversal by trivalent cations has been studied experimentally,22,26 but in that case the driving force for condensation has been attributed to Coulomb correlation due to charge discreteness. Here, condensation and charge reversal are most likely driven by the entropy associated with the release of hydrating water molecules, especially in cases where they are weakly bound to the cation, and by the enthalpy of silanol-cation association. Generalization of site binding models27 to include positively charged silanol-cation centers is needed for quantify(25) Horne, R. A. Water and Aqueous Solutions: Structure, Thermodynamics and Transport Processes; John Wiley & Sons, Incorporated Publisher Record: Hoboken, NJ, 1972; Chapter 7. (26) Besteman, K.; Zevenbergen, M. A. G.; Lemay, S. G. Phys. ReV. E: Stat., Nonlinear, Soft Matter Phys. 2005, 72, 061501–061509.

ing the various free energy terms involved in the adsorption process. Measurements carried out on numerous substrate-colloid pairs revealed shifts by up to a factor of 2 in the salt concentration at which repulsion vanished, but the phenomenology emerging from Figures 1-3 repeated in all experiments. A rough measure of sample to sample variation can be gained by comparing the 100 mM curves in Figure 2b with the pH 5.5 curve of the same salt in Figure 5. The two experiments have been carried out under the same conditions on two different colloid-substrate pairs. The surface charge density, measured at 100 mM NaCl, was about 10% higher in the sample of Figure 5 compared with the sample of Figure 2b (-0.051 and -0.057 e/nm2, respectively). Taking advantage of surface neutralization and, hence, elimination of the electrostatic interaction between the two surfaces, we turn to verify the identification of attraction with vdW interaction and extract the Hamaker constant characterizing it. The inset to Figure 2b depicts a log-log plot of F/R versus h for 100 mM CsCl and KCl, together with best fit to the theoretical prediction for the vdW force between a sphere and an infinite plate in the Derjaguin approximation,28 F/R ) A/6h2; h , R. The remarkable agreement with the predicted inverse quadratic dependence upon h supports the proposed identification of attraction with vdW force, yielding A ) (2.2 ( 0.4) × 10-21 J compared with A ) 1.6 × 10-21 J as calculated from the known dielectric functions of silica and water.29 The same curves in 500 mM NaCl and KCl are plotted in the inset to Figure 3b, yielding again inverse quadratic dependence upon h and the same value for the Hamaker constant. The average value of the Hamaker constant and the standard deviation were obtained by averaging 10 measurements for each salt type at three or four concentrations each (about 100 measurements altogether). The weak dependence of the Hamaker constant upon salt type and concentration was again consistent with vdW attraction and inconsistent with other mechanisms of attraction that depend on ion type and concentration.4,16 Note that our results do not exclude the potential importance of the latter attraction mechanisms at distances below ∼2.5 nm. Our value for the Hamaker constant, as well as the one calculated in ref 29, is markedly smaller than A ) 1.2 × 10-20 J used in ref 10 to fit the force after subtracting extrapolated electrostatic contributions from the total force. The same is true for A ) 8.4 × 10-21 J assumed by several authors30-32 or A ) (27) Yates, D. E.; Levine, S.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1974, 70, 1807–1818. (28) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press Inc.: San Diego, CA, 1991; p 176. (29) Ackler, H. D.; French, R. H.; Chiang, Y.-M. J. Colloid Interface Sci. 1996, 179, 460–469. (30) Hunter, R. J. Foundations of Colloid Science; Oxford University Press Incorporated: New York, 1987; Vol. 1, p 222. (31) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831– 1836. (32) Hough, D. B.; White, L. R. AdV. Colloid Interface Sci. 1980, 14, 3–41.

Silica Surfaces in Alkali Halide Aqueous Solutions

Langmuir, Vol. 25, No. 5, 2009 2835

Table 1. Summary of Reported Force Measurements between Silica Surfaces in Alkali Halide Aqueous Solutions in Regard to Attraction (Observed or Not) at High Salt Concentrations

ref

method

surface

this work

AFM

silica sphere on oxidized silicon(100)

8

SFA

12

AFM

18

AFM

19

AFM

21

SFA

pyrogenic silica

silica sphere on oxidized silicon silica sphere on naturally oxidized silicon silica sphere on oxidized silicon (100 nm oxide)

pyrogenic silicasflamed pyrogenic silicassteamed

31

34 35 36

AFM

feedback balanced SFA AFM SFA

pyrogenic silica exposed to ammonia silica sphere on oxidized silicon (30 nm silica)

two glass filaments silica sphere on oxidized silicon pyrogenic silica

is attraction observed at high salt concentration?

electrolyte and surface potential NaCl

1 mM

pH 5.5

-33 mV

KCl CsCl NaCl KCl CsCl NaCl

1 mM 1 mM 10 mM 10 mM 10 mM 1.5 mM

pH pH pH pH pH pH

5.5 5.5 5.5 5.5 5.5 5.67

-33 -33 -22 -20 -16 -83

KCl CsCl NaCl KCl CsCl LiCl

1.6 mM 1.1 mM 8 mM 14 mM 11 mM 10 mM

pH pH pH pH pH pH

5.45 5.3 5.28 5.1 5.71 5.6

-69 mV -53 mV -63 mV -51.5 mV -54 mV -40 mV

mV mV mV mV mV mV

N/A

yes for NaCl > 100 mM, KCl > 100 mM, CsCl > 50 mM

no, up to 100 mM NaCl, 140 mM KCl, 100 mM CsCl

yes, at 1 M LiCl yes, at 1 M NaCl

NaCl

10 mM

pH 3

-10 mV

NaCl NaCl NaCl NaCl NaCl NaCl NaCl NaCl

10 mM 10 mM 10 mM 1.1 mM 11 mM 0.99 mM 10.1 mM 1.1 mM

pH pH pH pH pH pH pH pH

-17 mV -20 mV -20 mV -34 mV -25.2 mV -50.4 mV -45.9 mV -116 mV

NaCl

1 mM

pH ∼ 5.7

-53 mV

no, up to 100 mM NaCl

NaCl NaCl NaCl NaCl NaCl NaCl NaCl KCl

10 mM 1 mM 1 mM 1 mM 1 mM 1 mM 1 mM 1.1 mM

pH ∼ 5.7 pH 10 pH 7 pH 4 pH 3 pH 2.6 pH 2 N/A

-34 -67 -60 -48 -35 -35 -13 -45

yes, at 100 mM KCl

NaCl

1 mM

pH ∼ 6

-46 mV

no, up to 100 mM NaCl

NaCl NaCl NaCl

10 mM 1.1 mM 11 mM

pH ∼ 6 N/A N/A

-33 mV -32 mV -28 mV

no, up to 110 mM NaCl

1.02 × 10-20 J used by Grabbe and Horn21 in fitting their experimental data. Compilation of surface charge values at different concentrations of the three salts is depicted in Figure 4. At 1 and 2 mM, surface charge is independent of salt type (e.g., Figure 1a), indicating negligible cation adsorption to the charged silica surfaces. The more negative surface charge at 2 mM, compared with 1 mM, reflects enhanced screening by the solution, hence, reduced deprotonation energy cost of the silanol groups. At higher salt concentrations, ion-specific adsorption turns increasingly important following the series: Cs+ > K+ > Na+. In the case of CsCl, the absolute value of surface charge reaches a maximum of ∼0.021 e/nm2 at ∼10 mM and then grows smaller as the concentration increases. Since surface charge is estimated by fitting the exponential tail to PB solution at large separations, we were unable to quantify the charge beyond 50 mM CsCl. However, the small upturn of the corresponding curve in Figure 3a, compared with KCl, indicates observable buildup of positive charge at 200

5 7 9 4.8 4.8 5.5 5.5 7.6

mV mV mV mV mV mV mV mV

yes, for 1 M NaCl at all pH, 10 mM NaCl at pH 3

no, up to 110 mM NaCl

mM, suggesting an approximate neutralization of surfaces at ∼100 mM CsCl. Recalling the point of zero charge for protons on silica, pzc ≈ 1.5-2,33 we can now add it to our scale of charge neutralization by specific ion adsorption to silica, H+ g Cs+ > K+ > Na+. Similar phenomenology is observed in KCl at higher ion concentrations. Pronounced charge reversal is delayed in this case to ∼1 M concentration, but the onset of charge reversal already at 0.5 M suggests surface charge neutralization occuring between 0.2 M and 0.5 M. Condensation of Na+ cations on the silica surfaces requires even higher concentrations, but between 0.5 and 1 M repulsion is small (charge neutralization) and vdW attraction dominates. The suppression of electrical double layer repulsion and appearance of vdW attraction at these concentrations are consistent with published force curves in the presence of NaCl18,19 and KCl.34 The combined neutralization of surface charge by protons and cations is studied in Figure 5 for 100 mM solutions. Note that

2836 Langmuir, Vol. 25, No. 5, 2009

in this case titration with HCl from pH 5.5 to pH 4.0 affects the DH screening length by less than 1%. The suppression of double layer repulsion upon addition of protons is therefore attributed to neutralization of surface charge, much like addition of alkali cations. Fitting the two NaCl curves to the solution of PB equation one finds -0.057 and -0.040 e/nm2 surface charge density at pH 5.5 and pH 4, respectively. The number of publications dealing with force between silica surfaces at high salt concentrations is limited. Table 1 summarizes 10 such experiments, including the present one. Ref 8 is closest to our experiment in terms of salt types and concentrations, yet, with the exception of a comment concerning samples that were introduced directly into high salt concentrations, it invariably reports repulsion. Ref 12 reports attraction in 1 M LiCl and repulsion in 0.01 M of the same salt (no intermediate concentrations). Ref 18 reports attraction in 1 M NaCl (the only reported concentration in that reference), and ref 19 reports attraction at pH 3 already in 10 mM NaCl. Ref 34 reports attraction in 100 mM KCl (no mentioning of pH value). All other publications are limited to NaCl and show no attraction up to the maximal investigated concentration, ∼100 mM.8,21,31,35,36 An interesting question arises then: Is there a common feature dividing experiments that show attraction from those who do not? Examining the surface potential at 1 and 10 mM salt (33) Iler, R. K. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties and Biochemistry of Silica, 1st ed.; John Wiley & Sons: Hoboken, NJ, 1979. (34) Rabinovich, Y. I.; Derjaguin, B. V.; Churaev, N. V. AdV. Colloid Interface Sci. 1982, 16, 63–78. (35) Giesbers, M.; Kleijn, J. M.; Cohen Stuart, M. A. J. Colloid Interface Sci. 2002, 248, 88–95. (36) Horn, R. G.; Smith, D. T.; Haller, W. Chem. Phys. Lett. 1989, 162, 404– 408.

Dishon et al.

concentrations one arrives at a potentially important observation; all samples that show attraction are also characterized by relatively low surface potential compared with those who do not. Figure 5 shows that for the same concentration of KCl the interaction is repulsive at pH 5.5 and attractive at pH 4 when surface charge is lower. The effects of proton and potassium ion adsorption are, hence, found to be additive. In the case of cesium ions the surface charge at pH 5.5 is neutralized by 100 mM CsCl, whereas titration with HCl to pH 4 reverses the surface charge and generates small double layer repulsion. It is plausible that the sample of ref 8 would have shown attraction at higher salt concentrations as suggested by extrapolating Figure 9 in that reference. We conclude by commenting that the dehydration of potassium ions by localized negative groups, and the holding of sodium ions to their hydration shell, are not foreign to the principles governing ion transfer through potassium channels.37 There, the weakly hydrated potassium ions lose their hydration shell and pass the narrow negatively charged channel, whereas sodium ions, holding to their hydration shell, are too big for passing through. Acknowledgment. We thank Mr. Elad Brod and Mr. Elad Peer for fruitful discussions. Research was supported by the Israel Science Foundation (Grant Nos. 1370/07, 475/04), the German Federal Ministry of Education and Research (BMBF) within the framework of German-Israeli Project Cooperation (DIP), and the Israeli Ministry of Science. LA803022B (37) Doyle, D. A.; Cabral, J. M.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J. M.; Cohen, S. L.; Chait, B. T.; MacKinnon, R. Science 1998, 280, 69–77.