Effect of Mechanical Stress on the Kinetics of Heterogeneous Electron

Aug 21, 2008 - File failed to load: https://cdn.mathjax.org/mathjax/contrib/a11y/accessibility-menu.js .... Present address: East Tennessee State Univ...
0 downloads 0 Views 576KB Size
Langmuir 2008, 24, 9941-9944

9941

Effect of Mechanical Stress on the Kinetics of Heterogeneous Electron Transfer Peng Sun,†,§ Zhen Liu,‡ Honghui Yu,*,‡ and Michael V. Mirkin*,† Department of Chemistry and Biochemistry, Queens CollegesCUNY, Flushing, New York 11367, and Department of Mechanical Engineering, City CollegesCUNY, New York, New York 10031 ReceiVed March 31, 2008. ReVised Manuscript ReceiVed June 1, 2008 The scanning electrochemical microscope (SECM) combined with a computerized tensile stage was employed to measure the kinetics of electron transfer (ET) reactions at stainless steel electrodes as a function of the applied mechanical stress. Reproducible current versus distance curves were obtained for different values of the tensile stress applied to a stainless steel (T-316) sample by using hexaammineruthenium as a redox mediator. The dependences of the extracted rate constant on substrate potential (i.e., Tafel plots, ln k versus E) were linear, in agreement with classical electrochemical theory. Possible origins of the stress effect on the ET rate and its implications for studies of stress corrosion cracking are discussed.

Introduction The interplay of the mechanical and electrochemical phenomena has been a subject of active research.1–9 For instance, the mechanical strain was induced by changes in the metal electrode potential (E), and its magnitude varied with E.1 The strain was attributed to the surface free-energy change with potential.1,2 The reverse is also true: mechanical stress can affect the contact potential.10 The rates of chemical processes, such as phase nucleation/growth and oxide formation, may also change significantly under the mechanical stress.9,11 A practically important case of the interplay between electrochemical reactions and mechanical phenomena is stress corrosion cracking (SCC), which is induced by tensile stress in a corrosive environment.12 Despite numerous experimental and theoretical studies of SCC, questions remain about the mechanisms of the formation and growth of a crack. Somewhat related mechanisms of pitting corrosion have been successfully studied by electrochemical scanning probes techniques13 including scanning electrochemical microscopy (SECM).14–18 The reports from White and Smyrl showed that pit nucleation occurred preferentially at microscopic precursor sites of high electrochemical activity.14,15 * Corresponding authors. E-mail: [email protected]; mmirkin@ qc.cuny.edu. † Queens CollegesCUNY. ‡ City CollegesCUNY. § Present address: East Tennessee State University. (1) Ibach, H.; Bach, C. E.; Giesen, M.; Grossmann, A. Surf. Sci. 1997, 375, 107. (2) Schmickler, W.; Leiva, E. J. Electroanal. Chem. 1998, 453, 61. (3) Guidelli, R. J. Electroanal. Chem. 1998, 453, 69. (4) Kiejna, A.; Pogosov, V. V. Phys. ReV. B 2000, 62, 10445. (5) Weissmu¨ller, J.; Viswanath, R. N.; Kramer, D.; Zimmer, P.; Wu¨rschum, R.; Gleiter, H. Science 2003, 300, 312. (6) Lu, B. T.; Chen, Z. K.; Luo, J. L.; Patchett, B. M.; Xu, Z. H. Electrochim. Acta 2005, 50, 1391. (7) Haiss, W. Rep. Prog. Phys. 2001, 64, 591. (8) Orlikowski, J.; Darowicki, K.; Arutunow, A.; Jurczak, W. J. Electroanal. Chem. 2005, 576, 277. (9) Kar, P.; Wang, K.; Liang, H. Electrochem. Solid State Lett. 2008, 11, C13. (10) Craig, P. P. Phys. ReV. Lett. 1969, 22, 700. (11) Barvosa-Carter, W.; Aziz, M. J.; Gray, L. J.; Kaplan, T. Phys. ReV. Lett. 1998, 81, 1445. (12) Arup, H., Parkins, R. N., Eds.; Stress Corrosion Research: Proceedings of the NATO AdVanced Study Institute on Stress Corrosion Research; NATO Advanced Study Institute Series: Series E, Applied Sciences, No. 30; Sijthoff & Noordhoff: Alphen aan den Rijn, The Netherlands, 1979. (13) Isaacs, H. S.; Kissel, G. J. Electrochem. Soc. 1972, 119, 1626.

We intend to explore the hypothesis that crack nucleation in SCC also occurs at the microscopic active sites, which can be detected by measuring local variations in surface redox reactivity. The first step in this direction reported here is to use SECM combined with a mechanical loading device (a miniloading stage; Figure 1A) to study the effect of mechanical deformation on the kinetics of an ET reaction. In SECM, a microscopic tip electrode is scanned over the substrate surface. A redox mediator species, O, contained in solution is reduced (or oxidized) at the tip (Figure 1B). The product of this reaction, R, diffuses to the substrate, where it can be reoxidized (or rereduced). This process produces an enhancement in the tip current, which depends on the rate of the mediator regeneration at the substrate and on the tip/substrate separation distance, d. A current versus distance (iT vs d) curve can be obtained by slowly moving the tip toward the substrate in the z direction (i.e., perpendicular to the substrate plane) and recording the current as a function of d. The heterogeneous rate constant of an ET reaction occurring at the substrate surface can be determined by fitting an iT versus d curve to the previously developed theory.19 Using our experimental setup, SECM measurements could be made while varying the mechanical load applied to the sample (Figure 1C). Stainless steel, which has previously been studied and has exhibited relatively stable electrochemical responses,15c,16a,17,18 was selected for these experiments. Because T316 stainless steel is used in implantable medical devices and is widely employed in corrosive surroundings (e.g., in seawater environments), its electrochemical properties and sensitivity to mechanical stress are important. Our first task is to show the possibility of reliable ET kinetic measurements at a stainless steel surface. Next, similar measurements will be performed with different values of tensile stress applied to a steel sample. (14) (a) Casillas, N.; Charlebois, S.; Smyrl, W. H.; White, H. S. J. Electrochem. Soc. 1993, 140, L142. (b) Casillas, N.; Charlebois, S.; Smyrl, W. H.; White, H. S. J. Electrochem. Soc. 1994, 141, 636. (c) Garfias-Mesias, L. F.; Alodan, M.; James, P. I.; Smyrl, W. H. J. Electrochem. Soc. 1998, 145, 2005. (15) (a) Basame, S. B.; White, H. S. J. Phys. Chem. B 1998, 102, 9812. (b) Basame, S. B.; White, H. S. Langmuir 1999, 15, 819. (c) Paik, C. H.; White, H. S.; Alkire, R. C. J. Electrochem. Soc. 2000, 147, 4120. (d) Serebrennikova, I.; Lee, S.; White, H. S. Faraday Discuss. 2002, 121, 199. (16) (a) Wipf, D. O. Colloid Surf., A 1994, 93, 251. (b) Still, J. W.; Wipf, D. O. J. Electrochem. Soc. 1997, 144, 2657. (17) Zhu, Y. Y.; Williams, D. E. J. Electrochem. Soc. 1997, 144, L43. (18) Lister, T. E.; Pinhero, P. J. Electrochem. Solid State Lett. 2002, 5, B33. (19) Wei, C.; Bard, A. J.; Mirkin, M. V. J. Phys. Chem. 1995, 99, 16033.

10.1021/la801009f CCC: $40.75  2008 American Chemical Society Published on Web 08/21/2008

9942 Langmuir, Vol. 24, No. 18, 2008

Letters

Figure 1. (A) Experimental setup for SECM studies of mechanical phenomena and (B) a scheme of the feedback mode SECM experiment. (C) Schematic drawing of the experimental system: (I) top view and (II) side view.

Experimental Section Chemicals. Hexaammineruthenium(III) chloride (Ru(NH3)6Cl3, 99%) was obtained from Strem Chemicals (Newburyport, MA). Potassium nitrate (Fisher Scientific) was recrystallized twice from water. Aqueous solutions were prepared from deionized water (MilliQ, Millipore Co.). Electrodes and Electrochemical Cells. Pt working electrodes (radius a ) 11.5 µm) with a radius of insulating glass surrounding the tip of rg ) 3a were prepared as described previously.20a Electrochemical measurements were performed in a four-electrode configuration, which included the working electrode (an SECM tip), a 0.25-mm-diameter Ag wire coated with AgCl serving as a reference, a counter electrode, and the SECM substrate (a metal sample), which also served as the bottom of the cell (Figure 1). T-316 steel sheets (0.5 mm thick from Goodfellow) were cut into small dog-boneshaped samples and polished with 0.3-µm alumina on a slowly rotating cloth. The cell was made by attaching a 7-mm glass tube to the middle of the sheet, which was mounted on the vibrationisolated horizontal SECM stage and attached to the minitensile tester. Instrumentation and Procedures. The hybrid experimental setup (Figure 1A) for mechanical/electrochemical measurements included a mechanical loading devicesa miniloading stage (Ernest F. Fullam, Inc., Latham, NY) and its data acquisition systemsthat was custom made and interfaced with the previously described SECM instrument.20b This instrument allowed us to obtain tip current versus distance curves, voltammograms, and maps of the substrate surface with or without applied mechanical load. The mechanical load applied by the Fullam miniloading stage (up to 1 kip, i.e., 4448 N) was gradually increased through a computer-controlled motor drive, and at each load level, the deformation was recorded using a Windowsbased MTEST Windows data acquisition system. After recording a current-distance curve at a specific value of the mechanical load, the tip was moved 50 µm away from the substrate, and the load was

increased to a new value. All experiments were carried out at room temperature (23 ( 2 °C).

Results and Discussion Initial experiments were carried out without the application of mechanical stress to check the possibility of accurate and reproducible ET kinetic measurements at the stainless steel/ aqueous solution interface. Cyclic voltammograms of 1 mM Ru(NH3)6Cl3 at the 8 mm2 T316 stainless steel electrode (Figure 2A) are reasonably well shaped, with a cathodic peak current directly proportional to the square root of the scan rate (V1/2) and the peak potential shifting with V, as one would expect for a kinetically controlled heterogeneous ET reaction. The ET kinetics at steel surfaces is much slower than those at Pt or Au electrodes, where the Ru(NH3)63+/2+ couple exhibits essentially reversible behavior. Cyclic voltammograms of Ru(NH3)6Cl3 obtained under the applied tensile stress (not shown) were similar to those in Figure 2A, and no significant effect of stress on the formal potential was found. A representative family of the SECM approach curves (iT-d) obtained with a small dog-bone-shaped steel sample used as a substrate is shown in Figure 2B. The mediator species were reduced at the 11.5-µm-radius Pt tip, and the product of this reaction was reoxidized at the steel surface

Ru(NH3)63+ + e- ) Ru(NH3)62+

(1)

Ru(NH3)62+ - e- ) Ru(NH3)63+

(2)

As the substrate potential gets more positive (from bottom to top), the oxidation rate increases and so does the tip current. The excellent fit between the experimental current-distance curves

Letters

Langmuir, Vol. 24, No. 18, 2008 9943

and the theory19 indicates that meaningful kinetic measurements of ET at the stainless steel substrate can be made by SECM. The rate constant of reaction 2 can be written as

kb ) k° exp[(1 - R)(E - E°′)F/RT]

(3)

where k° is the standard heterogeneous rate constant (cm/s), R is the cathodic transfer coefficient, E°′ is the formal potential of the Ru(NH3)63+/2+ redox couple, F is the Faraday constant, Ris the gas constant, and T is the temperature. kb was extracted from the iT-d curves and plotted in Figure 2C as a function of substrate potential (Tafel plot). The linearity and good reproducibility of the Tafel plots obtained from three different families of iT versus d curves and at different steel samples provide more evidence of the validity of our kinetic measurements. However, the kinetic parameters extracted from Figure 2C are somewhat surprising. The k° value, which can be evaluated by extrapolating the Tafel plot to E°′ ) -0.35 V versus Ag/AgCl, is ∼10-3 cm/s (i.e., ∼3 orders of magnitude lower than the rate constant values measured

Figure 3. (A) SECM current-distance curves for the reduction of Ru(NH3)63+ at a 11.5-µm-radius tip approaching a 0.5-mm-thick stainless steel substrate under the same potential (0.15 V vs Ag/AgCl) and different tensile stress (from top to bottom, in MPa): 0, 11, 18, and 27. Symbols are experimental data; solid lines are from the theory for finite substrate kinetics.19 The tip current is normalized by iT,∞ ) 3.58 nA. (B) Tafel plots obtained from SECM current vs distance curves at the above stress values. The solution contained 1 mM Ru(NH3)Cl3 and 0.2 M KNO3.

Figure 2. (A) Cyclic voltammograms and (B) SECM current vs distance curves of T316 stainless steel samples and (C) corresponding Tafel plots obtained with no mechanical load applied. The solution contained 1 mM Ru(NH3)Cl3 and 0.2 M KNO3. (A) The stainless steel working electrode area was 8 mm2. Scan rate, mV/s: 10 (green), 20 (yellow), 30 (pink), and 50 (blue). (B) An 11.5-µm-radius Pt tip was moved toward the substrate at an approach speed of 0.3 µm/s. From top to bottom, the substrate potentials (mV vs Ag/AgCl) are 200 (pink), 150 (orange), 100 (brown), 50 (blue), 0 (black), -50 (green), and -150 (red). Symbols are experimental data; solid lines represent the theory for finite substrate kinetics.19 The tip current is normalized by iT,∞ ) 3.58 nA. (C) Tafel plots obtained from three different families of SECM current vs distance curves.

for the Ru(NH3)63+/2+ redox couple at Pt and Au electrodes21–23). The transfer coefficient R ) 0.15 is also much smaller than expected for an uncomplicated outer-sphere ET. To our knowledge, no kinetic parameters of quasi-reversible ET reactions at stainless steel have yet been reported. A kinetic study of Ru(NH3)63+/2+ and several other outer-sphere redox couples on chromium yielded somewhat similar results (i.e., the rate constants were significantly lower than those measured at noble metal electrodes24). The transfer coefficient value decreased with increasing oxide layer thickness. It was suggested that both kinetic parameters were affected by the presence of a passive surface film. Most likely, a similar film is responsible for our present findings. The effect of tensile stress on the ET rate can be seen in Figure 3. In a series of approach curves obtained at a constant substrate potential (0.15 V versus Ag/AgCl; Figure 3A), the magnitude of positive feedback decreased systematically with the increase in applied mechanical load. Figure 3B shows the corresponding decrease in the rate constant of reaction 2 with increasing tensile stress. Tafel curves shifted downward in response to the increased value of the applied load, whereas their slope and therefore the charge-transfer coefficient (R) remained essentially constant. The SECM approach curves obtained at the same surface regions after the removal of the load yielded somewhat higher rate constants, but the kb values remained somewhat lower than those (20) (a) Laforge, F. O.; Kakiuchi, T.; Shigematsu, F.; Mirkin, M. V. Langmuir 2006, 22, 10705. (b) Shao, Y.; Mirkin, M. V.; Fish, G.; Kokotov, S.; Palanker, D.; Lewis, A. Anal. Chem. 1997, 69, 1627. (21) Beriet, C.; Pletcher, D. J. Electroanal. Chem. 1994, 375, 213. (22) Birkin, P. R.; Silva-Martinez, S. Anal. Chem. 1997, 69, 2055. (23) Sun, P.; Mirkin, M. V. Anal. Chem. 2006, 78, 6526. (24) Moffat, T. P.; Yang, H.; Fan, F.-R. F.; Bard, A. J. J. Electrochem. Soc. 1992, 139, 3158.

9944 Langmuir, Vol. 24, No. 18, 2008

Letters

measured before the sample was subjected to the tensile stress. T-316 is molybdenum bearing austenitic stainless steel whose surface is covered by a passive oxide layer.8 The samples used in our experiments were polycrystalline. In addition to changing the interatomic distances, the applied stress was likely to stretch the grain boundaries and induce changes in the surface oxide film. However, the observed change in the ET rate was not caused by the disruption of the oxide layer on the steel surface. If that were the case, then the appearance of a fresh (not passivated) steel surface under the mechanical stress would result in an increased rather than a decreased ET rate. Also, this effect could not be explained by the change in the formal potential of ET, which was found to be essentially independent of the applied stress. One can envision several ways in which the mechanical stress may influence the rate of the ET reaction at a metal electrode. The applied stress changes the work function10 and the potential of zero charge.25 The corresponding changes in the double-layer potential profile should influence the heterogeneous ET rate constant (Frumkin correction), according to eq 4 26

kt ° ) k° exp[-(R - z)Fφ2/RT]

(4)

where z is the charge on the reactant and φ2 is the potential at the outer Helmholtz plane. The above hypotheses could be checked by probing the effect of surface strain on the kinetics of the reverse ET reaction (i.e., the reduction of Ru(NH3)63+). However, it is hard to remove oxygen from the SECM cell attached to the loading stage, which is essential for working with air-sensitive Ru(NH3)62+ species. Our attempts to find another aqueous system, in which both oxidation and reduction reactions could be investigated at the stainless steel surface, have not yet been successful. Various species with redox potentials more positive than that of Ru(NH3)63+/2+ (e.g., ferrocenemethanol, Fe(H2O)63+/2+, and Fe(CN)63-/4-) apparently oxidize the steel surface, thus precluding meaningful kinetic measurements. This problem could be eliminated by employing noble metal substrates. However, most outer-sphere ET reactions at noble metal surfaces are too fast for measuring Tafel plots. Such metals (especially Au and Ag) have (25) Lipkowski, J.; Schmickler, W.; Kolb, D. M.; Parsons, R. J. Electroanal. Chem. 1998, 452, 193. (26) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; Wiley & Sons: New York, 2001; p 572.

much lower yielding strength than stainless steel so that any significant load applied to a sample would induce plastic deformation and expose fresh surface to solution. Additionally, destructive mechanical experiments using relatively large samples would require considerable amounts of a precious metal. Although the general trends discussed above are reproducible, the measured ET rates vary significantly from sample to sample and even between different points on the same sample surface. Kinetic experiments at stressed stainless steel are less reproducible than those carried out at noble metal electrodes because steel samples are microscopically heterogeneous and may differ in terms of grain size, the presence of inclusions and other defects, and variable oxide thickness. The nonuniformity of applied stress may further amplify the differences in local surface reactivity. Similarly to pitting corrosion,14,15 local variations in ET reactivity may point to precursor sites for the crack nucleation in SCC. Because surface defects, such as dislocations and particles, can generate very localized stress fields around them, an active site in SCC may be a stress concentration site. The possibility of correlation between the local changes in k° and crack nucleation is currently being investigated.

Conclusions We used the combination of the SECM with a computerized tensile stage to study the effect of mechanical stress on the kinetics of electron-transfer reactions at stainless steel (T-316) electrodes. The standard rate constant of Ru(NH3)62+ oxidation at the stainless steel substrate was several orders of magnitude lower than the values obtained for this reaction at noble metal electrodes. Linear Tafel plots were obtained from SECM current versus distance curves for different values of the applied mechanical load. The rate of the Ru(NH3)62+ oxidation decreased with increasing tensile stress. This effect may potentially be useful for understanding the mechanisms of nucleation phenomena in stress corrosion cracking. Future challenges include expanding the range of redox mediators suitable for such studies and the characterization of local changes in ET reactivity occurring in microscopic surface domains under mechanical stress. Acknowledgment. Support of this work by the CUNY Collaborative Incentive Research Grant and by the National Science Foundation (CHE-0645958) is gratefully acknowledged. LA801009F