132
Langmuir 1986, 2, 132-139
Characterization of the Silver-Aqueous Electrolyte Interface by Optical Second Harmonic Generation G. L. Richmond Department of Chemistry, Uniuersity o f Oregon, Eugene, Oregon 97403 Received J u n e 4. 1985. In Final Form: August 28, 1985 Optical second harmonic generation has been used to probe the electrical double layer structure of silver-aqueous interfaces at potentials within the double layer charging region. Reversible adsorption and desorption of simple ions and molecules at polished polycrystalline and single-crystal silver surfaces have been monitored. The potential dependence of the SH response exhibits a strong correlation with the excess charge density present at the metal surface. A comparison is made between the results from these smooth surfaces and those that have undergone electrochemical roughening. The results demonstrate the simplicity and potential of this technique in application to studies of the electrode-electrolyte junction. Introduction Characterization of metal-adsorbate interactions which occur a t the electrochemical interface has important relevance in both applied and basic research. In the past several years, optical techniques have been combined increasingly with standard electrochemical methods as a means of probing these interactions. Surface-enhanced Raman scattering (SERS) experiments measuring vibrational spectra of species a t or near electrochemically roughened surfaces comprise a large segment of these studies.’ An increasing body of information about the vibrational structure of adsorbed species is being obtained by infrared methods2 Recently, a growing interest has developed in understanding what kind of a role optical second harmonic generation (SHG) can play in providing information about adsorption, molecular orientation, and molecular structure a t the electrochemical interface. The attractive features of optical second harmonic generation are its experimental simplicity and its sensitivity to the interfacial region. Within the electric dipole approximation, these secondorder processes are restricted to regions where inversion symmetry is broken, such as is atomically present a t the interface.3 It is this feature that has been exploited in several recent studies where SHG has been used to monitor species electrochemically formed on ~ i l v e r , copper, ~ - ~ and golds electrode surfaces. Molecular adsorption a t the interface of the roughened surface has also been o b ~ e r v e d . ~ Shen and co-workers have discussed in detail the SH enhancement effects from roughened (1) For reviews, see: Van Duyne, R. P. In “Chemical and Biochemical Application of Lasers”; Moore, c. B., Ed.; Academic Press: New York, 1978; Vol. 4, Chapter 4. Furtak, T. E.; Reyes, J. Surf. Sci. 1980,93, 351. ( 2 ) See, for example: Foley, J. K.; Pons, S. Anal. Chem. 1985,57, 945A and references therein. (3) See, for example: Bloembergen, N. In “Nonlinear Optics”; Benzamin: New York, 1977; p 9. Bloembereen, N.; Pershan, P. S. Phys. Rev. 1968, 228, 606.
(4) Chen, C. K.; Heinz, T. F.; Ricard, D.; Shen, Y. R. Phys. Reu. B 1983,27, 1965.
(5) Richmond, G. L. Chem. Phys. Lett. 1984, 106, 26. (6) Richmond, G. L. Surf. Sci. 1984, 147, 115. (7) Chen, T. T.; von Raben, K. U.; Murphy, D. V.; Chang, R. K.; Laube, B. L. Surf. Sci. 1983, 124, 529. (8) Richmond, G. L. Abstr. Pap.-Am. Chem. SOC.1984, 188th. (9) Heinz, T. F.; Chen, C. K.; Ricard, D.; Shen, Y. R. Phys. Reu. Lett.
1982,48, 478. Heinz, T. F.; Chen. C. K. ; Ricard, D.; Shen, Y. R. Chem. Phvs. Lett. 1981. 83. 180.
:lo) Chen, C. K.;Heinz, T. F.; Ricard, D.; Shen, Y. R. Phys. Reu. Lett. (li)Boyd,~G.T.; Rasing, Th; Leite, J. R. R.; Shen, Y. R. Ph3.s. Reu.
1984. 46. 1010. ~~
B 1984, 30, 519. (12) Chen, C. K.; de Castro, A. R. B.; Shen, Y. R. Phys. Reu. Lett. 1981, 46, 145.
0743-7463/86/2402-Ol32$01.50/0
To exploit completely this optical method as a probe, several fundamental issues must be addressed concerning the relationship between the linear and nonlinear optical properties of both smooth and roughened electrode surfaces. It must also be understood how these properties are affected by the application of a bias potential to the electrode-liquid junction. Although this applied voltage is moderate (- 1 V), the actual electric field strength a t the interface can be as large as IO7V/cm since the field is dimensionally limited. It penetrates less than one atomic layer into the metal (Thomas-Fermi screening length) and decays over a few atomic radii on the solution side of the interface. Studies by Lee et were the first to note that second harmonic light reflected from the silver-aqueous electrochemical interface shows a strong potential dependence. These studies have more recently been extended in this l a b ~ r a t o r y land ~ , ~others,16J7 ~ demonstrating that the SH intensity appears to be related to the excess charge density which resides a t the metal surface portion of the interface. The work presented here is the first detailed investigation of the factors and effects that contribute to S H production at the silver-aqueous interface. A major portion of these studies pertain to the use of SHG to monitor reversible ionic and molecular adsorption at the electrode surface as it is polarized between the potential limits defined by solvent reduction and metal oxidation. A wide range of aqueous electrolytes have been studied which have different adsorptive behavior toward the silver electrode surface. The S H response has been correlated with these adsorptive effects and a charge density model described for quantitating the adsorption. Additional results pertaining to the effects of surface electrochemical cycling are also presented and discussed. Experimental Procedure The experimental procedure was similar to that described previously.6 The experiments were performed in a quartz electrochemical cell containing the silver electrode, a platinum counter electrode, and a saturated calomel reference electrode (SCE). All hias voltages reported here are relative to the SCE. The electrode potential was controlled with a potentiostat-galvanostat with the (13) Lee, C. H.; Chang, R. K.; Bloembergen, N. Phys. Reu. Lett. 1967, 18, 167. (14) Richmond, G. L. Chem. Phys. Lett. 1984, 110, 571. (15) Richmond, G. L. In “Proceedings of the International Conference on Lasers, ’84”; in press. (16) Corn, R. M.; Romagnoli, M.; Levenson, M. D.; Philpott, M. R. Chem. Phys. Lett. 1984, 106, 30. (17) Corn, R. M.; Romagnoli, M.; Levenson, M. D.; Philpott, M. R. J . Chem. Phys. 1984,81, 4125.
@ 1986 American Chemical Society
Langmuir, Vol. 2, No. 2, 1986 133
Silver-Aqueous Electrolyte Interface
Table I. Comparison of Silver PZC values wit.h the Onset of Anion Adsorption as Measured by SHG
c) Ag (1 10)
SO:-"
C10;
c1-
BrI-
Ag(poly) Ag(ll0) Ag(W SHG,Vb PZC,V SHG,V PZC,V SHG,V PZC,V -0.95 -0.94' -1.05 -l.Od -0.75 -0.72d -0.90 -0.95e -1.05 -0.99 -0.74 -0.735d -0.98 -0.938 -1.15 -1.05h -0.81 -0.83h -1.05 -1.15g -1.24 -0.95 -1.25 .
t
v)
5
t-
z
v) n
1 4 -1.2
-0.8
,v,
-0.4
(Volts)
Figure 2. SH intensity as a function of applied voltage for polycrystalline silver biased in (a) 80 and (b) 1.2 mM KBr, both adjusted to a constant ionic strength of 0.1 M with K2S04. h
,
VSCE(V0ltS)
Figure 3. SH response for Ag(ll0) and Ag(ll0) biased in a solution of 0.1 M K2S04 and M pyridine.
tration in solutions of either variable or constant ionic strength. An example of this behavior is shown in Figure 2, which compares 80 and 1.2 mM KBr. A constant ionic strength of 0.1 M was maintained by using K2S04. As shown, the onset in the rise in intensity and the potential a t which the saturation coverage occurs shifts to more negative potentials as the concentration of the bromide ion is increased. Additional studies were performed to investigate whether the SH response is sensitive to the specific adsorption of cations. As mentioned above, at voltages between -1.3 V and the PZC, a slightly voltage-dependent signal is observed for potassium-containing solutions. No detectable difference in the signal level can be observed in this region when the cation in solution is varied from the nonspecifically adsorbing Li+, K+, and Na+, to the less hydrated ions Rb+ and Cs+. Both chloride- and sulfatecontaining solutions were used for these studies. Molecular Adsorption on Smooth Surfaces. When pyridine is added in micromolar amounts to the electrochemical cell containing the silver electrode immersed in K,SO,, a difference in the SH profile is clearly detectable relative to the pyridine-free sample. Figure 3 compares the results for M pyridine in 0.1 M potassium sulfate solution for the various crystal faces. An increase in the second harmonic signal is observed for all surfaces as the pyridine is adsorbed at the negative potentials. With positive sweep, the SH signal decreases as the pyridine desorbs and then increases as the anion adsorption effects begin to dominate. The signal is reversible between the
Figure 4. SH intensity for polycrystalline silver cycled in 0.5 M KCIOl between -1.3 and +0.5 V.
two limits for these smooth surfaces. As shown in Figure 3, there is very little difference between the potential dependence of the adsorption on the different surface faces. There is also no observable difference when potassium chloride is used as the supporting electrolyte. Effect of Microscopic Changes in Surface Morphology. On the smooth surfaces described above, the signal at negative potentials is small relative to the intensity observed at more positive bias voltages. The SH signal throughout the voltage region is spectrally pure a t 532 nm and shows the expected quadratic power dependence. This behavior changes dramatically as the surface is altered by electrochemical cycling. Figure 4 demonstrates this effect for polycrystalline silver undergoing an oxidation-reduction cycle (ORC) in 0.1 M KC104. The scan is initiated at -1.3 V, reversed at +0.5 V, and subsequently returned to the initial potential. At potentials in the region between approximately +0.3 and -0?1 V on the forward scan, oxidation and reduction of silver occurs which corresponds to an average removal of approximately 80 atomic layers of flat silver surface. The amount of silver ions reduced to silver metal is less than the amount oxidized as a result of dissolution of the relatively soluble AgC104. As with previous there is a strong growth in signal in the region where the surface undergoes electrochemical modification. As the potential is scanned from -1.3 V to the onset of oxidation, the signal is identical with previous scans depicted in Figure 1. In contrast, after reduction is completed and the potential is returned to its initial negative limit, the SH signal is distinctly different relative to that prior to surface reformation. The overall SH signal level clearly increases between -0.5 and -1.3 V on the return sweep. In addition, a strong potential dependence in the SH signal between -0.9 and -1.3 V develops. This occurs for all of the electrolytes studied. The effect is even more dramatic for the halide electrolytes which cause a larger surface alteration when cycled between these potential limits. To understand this phenomenon in more detail, a set of experiments was performed in which a controlled amount of the silver surface was oxidized and reduced.25 Figure 5 depicts a sequence of scans between the limits of the double layer charging region after a controlled amount of oxidation and reduction of the silver surface. The measured oxidative current is expressed in silver monolayers for simplicity. After the slight oxidation of the surface, the potential was switched to -1.3 V and maintained for 5-10 min prior to the next scan. The re( 2 5 ) Richmond, G. L. Chem. Phys. Lett. 1985, 113, 359.
Langmuir, Vol. 2, No. 2, 1986 135
Silver-Aqueous Electrolyte Interface I
le\
.-
v)
CI
c
I
\I
;:I
t v)
6 Iz
5
z
0 -1.2
-0.8
-0.4
VSCE(VOltS)
Figure 5. Effect of slight surface roughening on the potential dependence of the SH intensity from polycrystallinesilver in 0.1 M KzS04 Curves (a-d) were recorded after oxidation-reduction cycles corresponding to 0, 1,5, and 12 monolayers of silver surface. M AgzSO, to Curve (e) was recorded after the addition of the solution of curve (d).
sults demonstrate that after only a few monolayers of surface have been altered, the surface begins to exhibit the characteristics observed upon more extensive electrochemical cycling. As with stronger oxidation, the overall S H intensity is found to increase throughout the double layer charging region. The SH signal at negative potentials grows progressively with additional surface reformation until the SH signal at negative potentials dwarfs the magnitude of the signal in the region of anion adsorption. As a consequence, the minimum in the SH profile shifts in the anodic direction relative to the original minimum on the unroughened surface. This signal shows a slight decrease when potentiostated at -1.3 V in the first minute after the ORC. However, after a few minutes the profile is relatively stable over many reversible scans between the potential limits of the double layer charging region. Not only is the region between -1.3 and -0.7 V sensitive to electrochemical modification, but it is also sensitive to surface preparation. Consistent behavior is observed when a rough surface polish or strong chemical etch is applied to the surface. At these oxidative levels, the enhanced signal is spectrally pure a t 532 nm with little or no contribution from metal luminescence. After several stronger ORCs, evidence of broadband background luminescence originating near 300 nm is observed. Further studies of these effects will be discussed in a later paper.26 Identical electrochemical roughening studies were performed on the single-crystal faces, giving results similar to the polycrystalline silver surface. For all three surfaces, the minimum in the profile for K2S04approaches -0.7 V. The S H profiles from the roughened single-crystal and polycrystalline substrates become indistinguishable with increased cycling. T o probe explicitly the effect of reduced silver ion on the S H signal, micromolar amounts of silver ions were added to the electrolyte solution. As the potential is applied to the roughened electrode surface, a measurable increase in SH intensity at the negative limit of the scan (26) Richmond, G. L.; Chu, P., unpublished results. (27) See, for example: Bockris, J. O’M.; Reddy, K. N. In “Modern Electrochemistry”; Plenum Press: New York, 1970; Vol. 2.
(3
I
cn
I
-1.2
,,v,
-0i8
I
1
-0.4
(volts)
Figure 6. SH intensity from polycrystalline silver mechanically roughened with a polishing alumina grit of 0.5-km average diameter. All electrolytes were 0.1 M in concentration.
is observed. This signal is 3-4 times higher in the first few seconds after the application of -1.3 V but stabilizes to the level shown in Figure 5e after several minutes. Addition of silver ion to a solution containing a silver sample that had not been electrochemically cycled results in a negligible amount of increase in SH signal at negative potentials even though the silver ion is reduced in an inhomogeneous manner on the relatively smooth surface. To investigate the adsorptive properties of the roughened surface in more detail, a similar SH profile was recorded for a variety of electrolytes. The results are depicted in Figure 6 and show that the mechanically roughened surface structure retains its sensitivity to the adsorbing characteristics of the anion in solution. Electrochemically roughened surfaces give similar results. The SH minimum in the profile for all the electrolytes is seen to shift approximately 350 mV positive relative to the reported PZC values for smooth surfaces. The most dramatic effect is observed for the iodide-containing solution. The new SH profile is also sensitive to concentration changes. As with the smooth surface, the minimum in the SH profile shifts increasingly negative as the concentration of halide ion increases. No shift in minimum is observed for sulfate- or perchlorate-containing solutions. Although the absolute SH signal at -0.1 V is not affected by concentration, the strong signal at negative potentials increases with increased ionic strength of the electrolyte. Variation of cation in the electrolyte does not affect the overall signal level. The S H signal generated as a result of pyridine adsorption is significantly increased by electrochemically cycling the surface. Figure 7 compares the SH profile from a slightly roughened surface and an unroughened silver substrate immersed in a K2S0, solution containing pyridine. For both, the profile is reproducible and independent of scan direction. Two observations can be made. As the surface roughness increases, the signal corresponding to pyridine adsorption between -1.3 and -0.7 V is enhanced. After oxidation and reformation of many layers of the surface, this intensity is no longer constant between -1.3
136 Langmuir, Vol. 2, No. 2, 1986
-1.5
-1.0
-0.5
Richmond
0.0
vSCE(volts) Figure 7. Comparison of the SH intensity from polycrystalline silver biased in 0.1 M K2S04and lo-* M pyridine: (a) polished surface; (b) after an ORC of several monolayers of silver surface; (c) after addition of lo-, M Ag,SO, to the sample in (b).
and -1.0 V but begins to mimic the stronger potential dependence of the S H signal when the pyridine is absent. When micromolar amounts of silver ions are added to the pyridinelroughened metal system, a very strong growth in SH intensity is observed at negative potentials as shown in curve c of Figure 7. This new enhanced signal is not stable but increases on several subsequent scans. Effect of Solvent Reduction. Care was always exercised t~ ensure that solvent reduction did not interfere with the measurements a t negative potentials. A sensitivity to this effect was first observed on the roughened surface where there is considerable signal at the negative limit. To investigate what effect hydrogen evolution has on the measurements, the solution measurements were performed in K2S04solutions a t a variety of different pH values. When all of the scans were initiated at -1.3 V, it was found that there is a substantial reduction in the signal throughout the range of the applied voltage. Steps were always taken to assure that this reduction in intensity was not a result of hydrogen bubbles a t the surface of the metal.
Discussion A. SHG at Ideally Polarizable Potential Regions. When a potential is applied to an electrode at voltages between the limits of solvent reduction and metal oxidation, the electric field resulting fro the opposing electrical charges on the liquid and metal interfacial layers is highly localized. The amount of static charge on both sides of this electrified interface is largely influenced by ionic species and molecules residing in the electrical double layer region of the solution. Differential capacitance's-24 and ellipsometric28studies of anions on silver surfaces indicate the trend for anion adsorption as I- > Br- > C1- > SO,*= C104-.21,23324This relative adsorptivity is manifested in a negative shift in the measured PZC with increased tendency toward adsorption. Such behavior is also observed in these SHG studies where the onset of growth in the harmonic signal arises a t more negative values with increased anion adsorption. Table I shows the similarity between the second harmonic rise and the reported PZC values measured by differential capacitance. Although the capacitance measurements were made with a variety of concentrations, the similarities between the S H behavior (28) Paik, W.; Genshaw, M. A,; Bockris, J. O M . J . Phys. Chem. 1970, 74 4266.
and these potentials is support that the second harmonic generation is sensitive to adsorption of anions a t submonolayer coverages. A distinction between the adsorptive character of the different surface preparations is clearly observed. For a given anion, the onset of the S H reponse from the Ag(l1O) surface is shifted approximately 350 mV negative relative to the signal from the Ag(111)face, in accordance with the known fact that the less densely packed Ag(ll0) surface is more strongly adsorbing. The polycrystalline surface which consists of a variety of microcrystalline structures is intermediate between the two with respect to its known adsorptivity and its differences in the SH response. Differential capacitance measurements show a similar ordering.1a-24Although mechanical polishing and the immersion of the crystal surface in the solution can alter the surface layers of the prepared sample, it is clear from the S H spectra that each surface maintains a considerable amount of it original single-crystal structure. This sensitivity to the double layer structure is also apparent in concentration studies. One indication of specific adsorption of charged species is a shift in PZC with a change in concentration of the specifically adsorbing anion." Within experimental error of these studies, it is not possible to detect a difference in the potential dependence of the SH response when the concentrations of the sulfate and perchlorate ions are varied from 1 mM to 0.1 M. Although weak specific adsorption has been reported for sulfate,ls the reported PZC shifts are within the experimental error of the S H measurements. In contrast, a negative shift in the onset of the S H signal is apparent for the halide ions over the concentration range of to 1 M. Although the relative shift for KC1 is small, the shift for KBr is easily measurable as shown in Figure 2. As the potential is stepped to values negative of the PZC, only a slightly increasing SH signal is observed. This lower signal is consistent with the lower adsorptivity of the cations for the surface. However, a distinction between cations such as potassium and sodium, which do not specifically adsorb, and cesium, which exhibits specific adsorption, is not observed. B. Can SHG Be Used as a Quantitative Measure of Anion Adsorption? These S H results consistently show that the production of harmonic light is sensitive to anion adsorption at the interface. T o quantitate this behavior it is necessary to understand the nonlinear optical properties of the interfacial region. At the interface between two centrosymmetric media, the SHG response can be-expressed as the square of the nonlinear polarizability, P"1s(2w)233where
x$jJ
The second-order nonlinear susceptibility tensor is a third-rank tensor with even parity under inversion. There are basically two contributions to in the interfacial zone, that contributed by the solution layer adjacent to the electrode and that contributed by the metal electrode surface. For ionic or molecular species that contact adsorb on the noninteracting substrate, the optical nonlinear polarizability is simply the sum of the intrinsic nonlinear optical properties of the species in the adsorbate layer. In the studies reported here, neither the probe nor second harmonic beam are resonant with optical transitions of the adsorbed ions or molecules. The contribution t o SHG from the underlying metal must also be considered. SHG from a bare silver surface is produced predominately by quadrupolar-type nonlinearities from both bound and free electrons involving interband and intraband transition^.'^,'^ For silver, this
x(2)
Silver-Aqueous Electrolyte Interface production of S H light is enhanced by the resonance of the 532-nm second harmonic light with the optical absorptivity of the silver metal.30 Application of a dc electric field to the electrode also appears to enhance the SH production. Such an effect has been previously observed in calcite crystal.31 This can be attributed to a third-order process with a nonlinear polarizability of 13,32333
Although third-order processes are dipole-allowed in centrosymmetric media this electric field effect occurs only at the surface of the metal where the applied field is significant. It is this latter effect, the SH response from the electrified metal surface, that appears to be dominant in these studies. The strongest evidence comes from the fact that appreciable SH signals result from solutions that either do not specifically adsorb, such as perchlorate,20 or that show only very weak specific adsorption, such as sulfate.ls In fact, the signal levels at saturation are comparable to those of the solutions where contact adsorption occurs. For these anions, both the surface and the anion remain predominately hydrated throughout the charging potential region.18,20As a result, the closest approach of either sulfate or perchlorate ions to the surface is at least two layers of water molecules.l8 The only potential-dependent change occurring at the aqueous interface is water dipole reorientation. Simple calculations using Millers rule34 demonstrate that the contribution of the water should be small relative to the nonlinear properties of the surface. Further support for the electric field effect at the surface is provided by inspection of the relative intensities of the signals a t the anodic potentials when the S H level approaches constancy. The observed decreased SH signal with increased halide ion size is exactly opposite to that expected if the nonlinear polarizability of the ions dominated the SH production. The results presented demonstrate that SHG is sensitive to anion adsorption a t the interface but in a manner consistent with the charging of the electrode surface rather than the intrinsic nonlinear polarizability of the adsorbate species. A t a given potential there will exist a charge on the metal electrode, qm, and an equal but opposite charge in the solution, qs. By Gauss's law, for a flat conducting metal with the electric field normal to the surface, the electric field should be proportional to the charge density, qm, at the surface. The second harmonic signal should therefore show quadratic dependence with surface charge.16 When the SH-potential profiles were analyzed in this manner, the derived charge density vs. potential curves are qualitatively similar to electrochemical measurement^.'^ Using these SH profiles as a means of measuring the potential dependence of the charge density at the surface, it is possible to obtain thermodynamic values for the adsorption process. For KBr studied over the concentration range 1-100 mM, using K2S04to maintain 0.1 M ionic strength, a Frumkin isotherm analysis of the data gave a free energy of adsorption for Br- of -113 f 3 kJ/mol at -0.8 V. For this preliminary analysis, the fractional cov(29) Bloembergen, N.; Chang, R. K.; Jha, S. S.; Lee, C. H. Phys. Reu. 1968, 174, 813. (30) Johnson, P. B.; Christy, R. W. Phys. Reu. E 1972,6,4370. (31) Terhune, R. W.; Maker, P. D.; Savage, T. M. Phys. Reu. Lett. 1962,8, 404. (32) Wang, C. Phys. Reu. 1969, 178, 178. (33) Brown, C. Phys. Reu. 1969, 178, 178. (34) Miller, R. C. Appl. Phys. Lett. 1964, 5, 17.
Langmuir, Vol. 2, No. 2, 1986 137 erage was estimated to be the ratio of [ 1 ( 2 ~ ) ]at~ -0.8 /~ V and [1(20)]~/~ at saturation coverage.35 The reported differential capacitance value for Br- on silver has been reported as -114 kJ/molZ1at a coverage of 15 pC/cm2. More detailed studies are currently in progress, using a wider concentration range and a more accurate HurwitzParsons a n a l y s i ~ to ~ ~determine ,~~ more accurately the fractional coverage. It is interesting to compare these results with electroreflectance (ER) studies in which the reflectivity of the metal surface has been monitored as a function of surface charge, wavelength, and optical polarization.3g42 For all of the different silver single-crystal surface preparations, a potential step to positive values relative to the PZC decreases the number of electrons at the surface, leading to a decrease in reflectivity. As with SHG, a clear distinction can be observed between the optical properties of various crystal faces when immersed in nonadsorbing electrolytes. Although considerable controversy exists regarding the nature of the excitation, surface charge density is agreed to play an important role. In early studies, the ER effect was described in terms of the free electron model in which the change in the free electron concentration at the metal surface with electrode potential is considered to be the main cause for the reflectance ~ h a n g e . ~Other ~ , ~ studies ~ suggest that the pronounced potential-dependent features that are observed are caused by electromodulation of optical transitions involving surface state^.^^,^^ These surface states are presumed to be sensitive to excess surface charge. C. Molecular Adsorption. The adsorption of pyridine on all of the smooth silver surfaces is clearly detectable by SHG. Differential capacitance measurements indicate that pyridine adsorption occurs in this region near the PZC.45 SH signals attributed to pyridine adsorption have been previously observed on electrochemically roughened surfaces.12 Electrochemical roughening is not a prerequisite for the observation of pyridine reported here. Although the pyridine adsorption must be responsible for the observed increase in SH signal in the double layer charging region, it is not evident from these studies what part of the interface is the major contributor. In the most direct sense, the nonlinear polarizability of the adsorbed layer of pyridine can cause the observed SH behavior. It is also true that the presence of an adsorbed layer at the surface influences the strength of the dc electric field in the double layer. Hence, there exists the possibility that, as with the simple electrolytes, the distortion of the electric field induced in the metal by the adsorbed pyridine is responsible for the increase in SH signal. The relatively low signal level from the pyridine relative to the anion adsorption region would favor the latter mechanism, but further experiments are necessary to verify this claim. (35) The SH background signal present at the PZC, [ 1 ( 2 ~ ) ~ ] ~was '*, subtracted from each value. (36) Hurwitz, H.0. J . Electroanal. Chem. 1965, I O , 35. (37) Weaver, M. J.; Anson, F. C. J . Electroanal. Chem. 1975,65,737. (38) See, for example: McIntrye, J. D. E. In "Advances in Electrochemistry and Electrochemical Engineering"; Muller, R. H., Ed.; WileyInterscience: New York, 1973; Vol. 9, p 61. (39) Kolb, D. M.; Kotz, R. Surf. Sci. 1977, 64, 96. (40) Kofman, R.; Garrigos, R.; Duteil, L.; Cheyssac, P. J. Electroanal. Chem. 1983, 150, 253. (41) McIntyre, J. D. E. Surf.Sci. 1973, 37, 658. (42) McIntyre, J. D. E.; Aspnes, D. E. Surf. Sci. 1971, 24, 417. (43) Ho, K. M.; Harmon, B. N.; Liu, S. H. Phys. Reu. Lett. 1980,44, 1531. (44) Kolb, D. M.; Boeck, W.; Ho, K. M.; Liu, S. H. Phys. Reu. Lett., 1981, 47, 1921. (45) Fleishmann, M.; Robinson, J.; Waser, R. J . Electroanal. Chem. 1981, 117, 257.
138 Langmuir, Vol. 2, No. 2, 1986
D. Effect of Surface Roughness. Slight electrochemical roughening of the surface structure has a distinct effect on the SH profile. This is evident in Figure 4 in which the SH signal after the ORC is enhanced relative to the initial level. A closer analysis of this phenomena is shown in Figure 5 for controlled amounts of oxidative and reductive changes in the surface. With increased oxidation of the surface, the resulting signal is greater in overall level, but the dominant factor is the growth at negative potentials. This signal is even further enhanced by the addition of silver ions which reduce inhomogeneously on the roughened surface as shown in curve of Figure 5. Differential capacitance studies of electrochemically roughened surfaces have previously been measured with similar electrolytes in order to understand the influence of surface roughening on the double layer structure.& For comparable degress of roughening, only a moderate increase in actual surface area was found. In addition, changes in average surface concentration for the adsorbing electrolytes were minor. However, roughening did induce noticeable changes in the morphology of the capacitance-potential curves, which were traced to alterations in the surface crystallite structure. The lack of evidence for the roughening process resulting in a significant increase in anion adsorption or surface area suggests that the increased SH signal observed here must be related to a change in the optical properties of the original surface. It is well-known that similar types of electrochemically induced changes in the surface optical properties are operative in enhancing SERS signals.47 The optical absorptivity of the metal surface shows a similar enhancement and potential dependence upon electrochemical roughening.26These studies were performed with photoacoustic detection and will be discussed in a later publication. The acoustic studies demonstrate the commonly held belief that as with other optical processes, the variation in the metal surface dielectric properties, i.e., absorbance, is closely related to the optical surface enhancements. This increased optical absorbance and S H production is most likely a result of a rise in activation of surface plasmon (SP) excitation by the rough metal surface. Although surface plasmons are generally not excited on a prefectly flat surface by a transverse electromagnetic waves, surface irregularities due to roughening procedures should facilitate the momentum conservation and allow optical excitation. The plasmon resonance frequency of a particular roughness feature should be different from that of the smooth surface and dependent upon the geometry of the surface irregularity. For silver surface, SP excitation appears near 3.5 eV whereas the SP plasma frequency of a spherical well-isolated silver particle appears in the near-UV depending upon the size of the particle.4s On a roughened surface, the nonspherical shapes and dipole coupling of these surface particles result in a shift of the plasma frequencies to lower energies at wavelengths pertinent to the studies presented here. Electroreflectance studies have demonstrated this effect to be true, showing the appearance of additional features in the reflectance spectra at lower frequencies as a result of continued r o u g h n e ~ s .This ~ ~ increased efficiency in optical coupling to the surface plasmons should increase the local field strength at these small surface irregularities, resulting in (46) Hupp, J. T.; Larkin, D.; Weaver, M. J. Surf. Sci. 1983,125, 429. (47) Chang, R. K., Furtak, T. E., Ed. “Surface Enhanced Raman Scattering”; Plenum Press: New York, 1982. (48) Glass, A. M.; Wokaun, A.; Heritage, J. P.; Bergman, J. G.; Liao, P. F.; Olson, D. H. Phys. Reu. B 1981, 24, 4906. (49) Kotz, R.: Kolb, D. M.; Sass, J. K. ,Surf. Sci. 1977, 69. 359.
Richmond
an enhancement of both the incoming and outgoing optical fields. Such a mechanism appears to be general for many optical processes at surfaces including hyper-Raman ~ c a t t e r i n gl,u~m ~ i n e s c e n ~ e ,and ~ ~ ,SHG ~ ~ from bare meta l ~ . ” , Wokaun ~~ et al.52have explicitly demonstrated the dependence of the SH enhancement of the resonant excitation of a localized surface plasmon. The fact that a similar enhancement is not significant when silver ions are added to the smooth surface suggests that the formation of the enhancing microscopic nodules at the surface is largely aided by reduction onto bumps and pits of the rough surface. It appears from these studies that the newly measured SH behavior from the rough surface biased between the limits of the double layer charging region is a reflection of the potential dependence of the optical dielectric constant of the surface plasmons. This potential dependence in the dielectric constants t and 2~ at the silver surface is a consequence of the change in electron density with applied field. A further demonstration of this effect in shown in the studies of Figure 6 in which the surface charge density, and consequently the electric field, has been varied with the choice of electrolyte. Although the minima in the profiles of the roughened electrodes are consistently shifted positive to the smooth surfaces, the sensitivity to the different adsorptive properties of the electrolytes in the double layer is retained after roughening. This parabolic SH potential dependence from the roughened surface is similar to the type of profile observed by Corn et a1.16J7 in their studies of surface-plasmon-enhanced second harmonic generation at thin-film silver electrodes. Optical coupling with the surface plasmons in the thin film was effected by using the technique of attenuated total reflectance. Slight roughening of the surface also has a dramatic effect on the signal resulting from pyridine adsorption at the surface. As mentioned earlier, the signal from pyridine is observed between -1.3 and -0.6 V on all of the surfaces. When pyridine is added to a solution containing an electrode that has been electrochemically cycled and held at -1.3 V for several minutes, a strong signal at negative potentials results which has a potential dependence that mimics the behavior of the potential dependence of the bare silver surface. This signal is reasonably reproducible when scanned reversibly between the potential limits. When silver ions are added to this pyridine-containing solution, a tremendous enhancement in SH signal is observed a t the negative potentials. The SH profile tends to grow as the potential is scanned between the two limits. Pyridine ion is known to stabilize the positively charged silver ions at certain potentials, but the potential dependence of the SH behavior is completely different from the hysteresis in the pyridine-silver signal measured with SERS. The increasing signal with scan could possibly be due to increased stabilization of the silver ions by the adsorbed pyridine. E. SHG during an ORC. At potentials where electrochemical modification is occurring, a growth in signal is observed in Figure 4 as the silver is oxidized and subsequently reduced. On the basis of the observations discussed in the last section, the question that arises is (50) Murphy, D. V.; Von Raben, K. U.; Chang, R. K.: Dorain, P. B. Chem. Phys. Lett. 1982, 85, 43. (51) Harstein, A.; Kirtley, J. R.; Tsang, J. C. Phys. Reu. Lett. 1982, 45, 201. ( 5 2 ) U‘okaun, A.; Bergman, J. G.; Heritage, J. P.; Glass, A. M.; Liao, P. F.; Olson, D. H. Phys. Reu. E 1981,24, 2. (53) Bergman, J. G.; Chemla, D. S.; Liao, P. F.; Glass, A. M.: Pinczuk, A.; Hart, R. M.; Olson, D. H. Opt. Lett. 1981, 6 , 33.
Langmuir 1986,2, 139-146
139
whether this increased SH signal is generated by the oxidatively formed AgC10, on the surface or by the silver surface alone which is actively being roughened. Due to the inversion symmetry of the crystalline form of AgClO,, no signal would be expected from this complex. In addition, silver perchlorate is relatively soluble and should not exist to a significant extent on the silver surface. As a comparison, the signal during the formation of AgC10, has been found to be a factor of 5-10 lower than that generated by silver halides a t the surface.54 The alternative explanation that the roughening silver surface contributes significantly to this signal would account for the fact that the SH response tends to peak beyond the +0.5 V where AgC10, should be undergoing reduction. These effects have been studied in more detail by photoacoustic methods and will be discussed in a later publication.26
of the electrical double layer structure that dominates the SH production in these aqueous nonabsorbing electrolytes is the charged silver metal surface. A strong correlation has been demonstrated between the potential dependence of the SH response and the changes in excess surface charge density as measured by electrochemical methods. The reproducibility and simplicity of the method offers considerable potential for future studies of kinetic processes at the electrochemical interface. These studies also demonstrate how SH measurements in the double layer charging region can be significantly altered by electrochemical roughening of the surface. The observed SH enhancement and altered potential dependence resulting from surface roughness is attributed to local field enhancements created by the local surfaceplasmon excitation and corona effects in the rough surface protrusions.
Conclusions In these studies, SHG has been shown to be extremely sensitive to adsorption and desorption of ions and molecules at smooth silver electrode surfaces. The component
Acknowledgment. Financial support from Research Corporation, the donors of the Petroleum Research Fund, administered by the American Chemical Society, and the National Science Foundation (CHE-8513008) is gratefully acknowledged. Registry No. Ag, 7440-22-4;KBr, 7758-02-3;K,SO,, 7778-80-5; KClO,, 7778-74-7; KI, 7681-11-0.
(54) Richmond, G. L., unpublished results.
Interactions of Human Milk Lipase with Sodium Taurocholate and Other Surfactants? Charmian J. O’Connor* and Peter Walde Chemistry Department, University of Auckland, Private Bag, Auckland, New Zealand Received February 1, 1985. I n Final Form: October 10, 1985 The effects of Triton X-100, Brij 56, sodium dodecyl sulfate (SDS),sodium bis(2-ethylhexyl) sulfosuccinate (AOT), hexadecyltrimethylammonium bromide (CTAB), dodecylammonium propionate (DAP), and egg yolk lecithin on the hydrolysis of 4-nitrophenyl propionate catalyzed by human breast milk lipase have been compared, in both the absence and the presence of 2 mM sodium taurocholate. All reactions were carried out in 0.1 M Tris-HC1buffer, pH 7.5, at 298 K. The nonionic surfactants that were tested stimulate the hydrolysis, but the ionic surfactants, with the exception of the bile salt, behave as inactivators (except at very low surfactant concentration) of the esterase activity. Measurements of the fluorescent and circular dichroism spectra indicate that the inhibition caused by SDS does not arise from the denaturation of the enzyme. Mixed micelles of bile salt and ionic surfactants are less inhibitory than are micelles of the ionic surfactants. Mechanisms for these reactions are discussed which consider the cmc of the surfactants and the interactions between the surfactants and the enzyme, the surfactants and the substrate, and the enzyme and the substrate. In general, the data seem well fitted by the “surface-as-cofactor” theory.
Introduction Small molecules can significantly alter protein structure by interacting preferentially with some of the charged sites or by forming more powerful hydrophobic bonds than exist in the structural protein molecule. Since surfactants possess hydrophobic and hydrophilic regions of known properties and chemical constitution, studies on the effects of amphiphilies on protein stability and activity can provide insight into the structure of proteins. It has long been known that surfactants may precipitate or form complexes with, or denature proteins a t low concentrations. Surfactants produce conformational changes in proteins at these low concentrations, combining with native proteins in multiple equilibria (Le., many molecules per molecule + T h i s is part 16 in a series. P a r t
15: ref 7.
of protein) as the protein unfolds, and the protein then binds more surfactant molecules as more binding sites become exp0sed.l Binding to high affinity sites does not necessarily result in denaturation a t low surfactant concentration.2 In recent years, this present research group has made extensive investigation^^-^ of the bile salt stimulated est(1) Lapanic, S. “Physicochemical Aspects of Protein Denaturation“; Wiley: New York, 1978; Chapter 3. (2) Gennis, R. R.; Jonas, A. Annu. Rev. Biophys. Bioeng. 1977,6,195. (3) O’Connor, C. J.; Wallace, R. G. Eur. J . Biochem. 1984, 141, 379. (4) O’Connor, C. J. and Wallace, R. G. J. Colloid Interface Sci. 1984, 539-547. (5) OConnor, C. J.; Wallace, R. G. Pediatr. Gastroenterol. Nutr. 1985, 4 , 240, 446, 587. (6) OConnor, C. J.; Wallace, R. G. FEES Lett. 1984, 170, 375. (7) OConnor, C. J.; Wallace, R. G. In “Proceedings of 5th International Symposium on Surfactants in Solution”; Mittal, K. L., Botherel, P., Eds.; Plenum Press: New York, in press.
0743-7463f 86f 2402-0139$0~50f 0 0 1986 American Chemical Society
