Anal. Chem. 2005, 77, 2601-2607
Microelectrode Array Microscopy: Investigation of Dynamic Behavior of Localized Corrosion at Type 304 Stainless Steel Surfaces Tedd E. Lister* and Patrick J. Pinhero
Idaho National Laboratory, P.O. Box 1625, Idaho Falls, Idaho 83415-2218
Scanning electrochemical microscopy (SECM) and a recently developed microelectrode array microscope have been used to study localized corrosion and electrontransfer characteristics of native oxide layers of type 304 stainless steels. The I-/I3- redox couple was employed as a mediator and allowed sensitive detection of oxide breakdown events. In solutions containing I-, a signal at the microelectrode was observed on type 304 stainless steel surfaces at active pitting corrosion sites. Under conditions where pitting corrosion occurs, SECM was used to track the temporal characteristics of the reaction in a spatial manner. However, because of the time required to create an image, much of the temporal information was not obtained. To improve the temporal resolution of the measurement, microelectrode array microscopy (MEAM) was developed as a parallel method of performing SECM. The demonstration shown reveals the potential of MEAM for analysis of surface chemistry on temporal and spatial domains. Native oxide films formed on metal surfaces are crucial to their stability. The technological importance of this phenomenon is demonstrated through the continued research activity and interest in corrosion science. Thus, understanding processes that control passive film stability and failure is an important scientific endeavor. The stability of technically important alloys in various environments has been cataloged for pure metals and technologically important alloys.1 Fundamental studies have taken a different approach by attempting to better understand the issues and factors influencing the stability of the oxide layer. Understanding the fundamental properties of passive films on metals may lead to the development of materials and coatings with superior corrosion resistance. The difficulty in the study of oxide films lies in the small dimensions of the film and their nonuniform properties. These nonuniformities (multiple phases, inclusions, defects, and contaminants) control the reactivity of the surface although they may comprise only a small percentage of the surface area. Very little is known about the evolution of oxide films during exposure to chemical environments. Heterogeneities in the oxide film can lead * To whom correspondence should be addressed. Phone: (208) 526-4320. Fax: (208) 526-4822. E-mail:
[email protected]. (1) Craig, B. D., Anderson, D. B., Eds. Handbook of Corrosion Data, 2nd ed.; ASM International: Materials Park, OH, 1995. 10.1021/ac0485785 CCC: $30.25 Published on Web 03/10/2005
© 2005 American Chemical Society
to nonuniform corrosion processes, i.e., localized corrosion. Localized corrosion is a particularly difficult problem to study due to the relatively small area of the oxide film that is involved. Despite the small surface area affected, once the passive film is broken, significant amounts of material can be removed, leading to degradation of physical and mechanical properties of the material. Unfortunately, most traditional surface analysis methods are applicable to postmortem studies only and thus may only represent the state after emersion of the sample. In situ techniques offer a powerful way to study oxide films as they provide the ability to investigate processes while they occur (for reviews of various in situ techniques, see ref 2). Scanning probe microscopy (SPM) techniques have shown great promise in studying immersed surfaces.2 SPM techniques have been used as a tool for analyzing topological features at high resolution in the study of metal dissolution and oxidation.3-5 SPM techniques have also been used to image localized dissolution events.6,7 The high spatial resolution imposes limits to the area of a surface that can be measured in a reasonable time. Thus, heterogeneous processes or kinetically fast processes such as localized corrosion can be difficult to study using single probe designs. Multiplexing designs for SPM are beginning to emerge for application in data storage8-10 which could improve the data acquisition speed for imaging applications. Scanning electrochemical microscopy (SECM) (for a review see ref 11) is a method that is amenable to the use of an array methodology as arrays of electrodes can be addressed without the requirement of individual feedback for each element, a requirement for most SPM techniques. SECM has been utilized in studying the electronic properties of oxide films and corrosion processes.12-17 While noble metals (2) Wieckowski, A., Ed. Interfacial Electrochemistry, Theory, Experiment, and Application; Marcel Dekker: New York, 1999. (3) Suggs, D. W.; Bard, A. J. J. Am. Chem. Soc. 1994, 116, 10725. (4) LaGraff, J. R.; Gewirth, A. A. Surf. Sci. 1995, 326, L461. (5) Nanjo, H.; Newman, R. C.; Sanada, N. Appl. Surf. Sci. 1997, 121, 253. (6) Williford, R. E.; Windisch, C. F., Jr.; Jones, R. H. Mater. Sci. Eng. 2000, 54, A288. (7) Femenia, M.; Pan, J.; Leygraf, C. J. Electrochem. Soc. 2002, 149, B187. (8) Bullen, D.; Wang, X. F.; Zou, J.; Chung, S. W.; Mirkin, C. A.; Liu, C. J. Microelectromech. Syst. 2004, 13, 594. (9) Despont, M.; Drechsler, U.; Yu, R.; Pogge, H. B.; Vettiger, P. J. Microelectromech. Syst. 2004, 13, 895. (10) Pantazi, A.; Lantz, M. A.; Cherubini, G.; Pozidis, H.; Eleftheriou E. Nanotechnology 2004, 15, S612. (11) Mirkin, M. V.; Horrocks, B. R. Anal. Chim. Acta 2000, 406, 119. (12) Serebrennikova, I.; White, H. S. Electrochem. Solid State Lett. 2001, 4, B4. (13) Fushimi, K.; Okawa, T.; Azumi, K.; Seo, M. J. Electroanal. Chem. 2000, 147, 524.
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005 2601
(i.e., Au and Pt), having a bare or an extremely thin oxide layer (atomically thick), show relatively uniform reactivity across the surface (contamination not withstanding) at low overpotentials for facile electrochemical reactions, metals covered by thicker oxide layers (such as chromium oxides) show blocking behavior.18 Oxide films have shown activity at higher overpotentials, where electron transfer is observed at localized sites on the surface which has been attributed to defects in the oxide film. Native oxide films on pure Ta, Ti, and Al displayed specific sites of electron transfer.12-17 In some cases, these defect sites have been correlated to regions that were susceptible to pitting corrosion. These sites can be considered analogous to circuit elements, representing low impedance in the equivalent circuit analysis of electrochemical impedance spectroscopy data,19-21 and thus present the most likely points for corrosion activity. SECM can provide specific information about the location of defects and the relative magnitude of current flowing through them as a function of potential. Other SECM studies have focused on the study of surfaces undergoing localized corrosion. SECM has been used to image corrosion activity at stainless steel surfaces where the oxidizable products of the corrosion process were detected.22 SECM has also been used to initiate isolated corrosion processes by generation of corrosive species at the tip.22,23 Cathodic corrosion activity of Fe- and Cu-rich inclusions was examined on Al 2024.24,25 The dissolution of sulfide particles in type 304 SS (SS ) stainless steel) was imaged using iodide as a mediator species.26,27 Localized corrosion of type 304 SS has been studied using SECM and interesting temporal behavior observed.28-30 In the these experiments, the images captured by SECM displayed the limitations of the time required to capture the image using a single microelectrode. Temporal information is crucial to understanding the effect of chemistry and microstructure on localized corrosion. Attempts to understand these effects have been made using planar electrode arrays composed of the material of interest, where specific elements are controlled (corroded) and others passively observed to determine spatial-chemical effects.31,32 These experiments have shown that sites of localized corrosion influence the localized corrosion of surrounding regions. In light of these (14) Basame, S. B.; White, H. S. Anal. Chem. 1999, 71, 3166. (15) Basame, S. B.; White, H. S. J. Phys. Chem. 1995, 99, 16430. (16) Casillas, N.; Charlebois, S.; White, H. S.; Smyrl, W. H. J. Electrochem. Soc. 1994, 141, 636. (17) Casillas, N.; Snyder, S. R.; Smyrl, W. H.; White, H. S. J. Phys. Chem. 1991, 95, 7002. (18) Lee, C.; Bard, A. J. Anal. Chem. 1990, 62, 1906. (19) Wenger, F.; Cheriet, S.; Talhi, B.; Galland, J. Corros. Sci. 1997, 39, 1239. (20) Annergren, L.; Thierry, D.; F. Zou, F. J. Electrochem. Soc. 1997, 144, 1208. (21) Philippe, L. V. S.; Walter, G. W.; Lyon, S. B. J. Electrochem. Soc. 2003, 150, B111. (22) Wipf, D. O. Colloid Surf., A 1994, 93, 251. (23) Still, J. W.; Wipf; D. O. J. Electrochem. Soc. 1997, 144, 2657. (24) Seegmiller, J. C.; Buttry, D. A. J. Electrochem. Soc. 2003, 150, B413. (25) Seegmiller, J. C.; Bazito, R. C.; Buttry, D. A. Electrochem. Solid State Lett. 2004, 7, B1. (26) Paik, C. H.; White, H. S.; Alkire, R. C. J. Electrochem. Soc. 2000, 147, 4120. (27) Lister, T. E.; Pinhero, P. J. Electrochim. Acta 2003, 48, 2371. (28) Lister, T. E.; Pinhero, P. J. Electrochem. Solid State Lett. 2002, 5, B33. (29) Lister, T. E.; Mizia, R. E.; Pinhero, P. J. NACE Corros. 2003, Paper No. 03379. (30) Lister, T. E.; Pinhero, P. J. Proc. Electrochem. Soc. 2002, 2002-24, 368. (31) Lunt, T. T.; Brusamarello, V.; Scully, J. R.; Hudson, J. L. Electrochem. Solid State Lett. 2000, 3, 271. (32) Lunt, T. T.; Scully, J. R.; Brusamarello, V.; Mikhailov, A. S.; Hudson, J. L. J. Electrochem. Soc. 2002, 149, B163.
2602
Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
experiments, a method of capturing the dynamics of localized corrosion is needed to further understand these processes. As localized corrosion processes at stainless steels have been captured via SECM, the use of many microelectrodes simultaneously capturing data was envisioned.30 Electrode arrays have been used in various applications such as environmental sensing,33 neurological studies,34,35 and also localized corrosion processes.31,32 The incorporation of an electrode array into a traditional SECM instrument scheme results in what has been called the microelectrode array microscope.36 The experiments outlined below detail the development of the microelectrode array microscope (MEAM), designed to address both spatial and temporal domains simultaneously. The promising results observed with SECM, also shown below, led to interest in devising a method to improve the temporal resolution of the experiment. A parallel method using multiple electrodes would reduce the time required to form an image of activity across the surface. MEAM couples a microelectrode array with electronics capable of making electrochemical measurements in a potentiostatic fashion for each array element. The experimental apparatus employs an array with 100 microelectrodes spaced 400 µm apart in a square (10 × 10) arrangement. The distance between the array elements limits the spatial resolution of the measurement compared to that of SECM, but improves the potential temporal resolution (0.1 s/image) by a factor of about 30000 (compared to that of previous SECM measurements of similar areas examined). The work is presented here to demonstrate the application of MEAM as an analytical tool in the study of corrosion processes and to point out the potential for use in other studies. EXPERIMENTAL SECTION Scanning Electrochemical Microscopy. The SECM instrument used in these experiments was based on previous designs.16 Control of the microelectrode was performed using inchwormtype piezo motors (Burleigh) with optical encoders to allow reproducible motion. The motors were mounted onto a three-axis stage to provide three-dimensional motion. The encoded (reproducible) resolution was 20 nm. A bipotentiostat (Topometrix) was used to independently control the potential of the sample and microelectrode versus a common reference electrode. The system was computer controlled using programs written in LabVIEW (National Instruments) language. The sample cell was placed on a holder having three thumbscrews arranged in a tripod configuration. A video microscope was used to aid microelectrode positioning and sample leveling using the thumbscrews. Carbon fiber microelectrodes (8 µm diameter) were constructed using an established method.37 The microelectrode was used as an aid in sample leveling, by checking the sample to tip distance across the surface and repositioning the sample accordingly. The sample to tip distance was set by slowly approaching the surface until (33) Tercier, M. L.; Buffle, J. Anal. Chem. 1996, 68, 3678. (34) Strong, T. D.; Cantor, H. C.; Brown, R. B. Sens. Actuators, A 2001, 91, 357. (35) James, C. D.; Spence, A. J. H.; Dowell-Mesfin, N. M.; Hussain, R. J.; Smith, K. L.; Craighead, H. G.; Isaacson, M. S.; Shain, W.; Turner, J. N. IEEE Trans. Biomed. Eng. 2004, 51, 1640. (36) Lister, T. E.; Glenn, A. W.; Pinhero, P. J. NACE Corros. 2004, Paper No. 04445. (37) Potje-Kamloth, K.; Janata, J.; Josowicz, M. Ber. Bunsen-Ges. Phys. Chem. 1989, 93, 1480.
Figure 1. Diagram of the MEAM instrument.
contact and then retracting to a specific distance (5 or 10 µm) above the sample surface.15 As the shaft is flexible, contact of the tip to the surface does not damage the sample. Images were acquired by moving the probe in a raster-type motion. Line scans were made using a step/acquire scheme where the step size was 10 µm and the data acquisition time was 20 ms. Line scans were acquired in the same scan direction. The SECM images shown in this paper were 3000 µm × 3000 µm in area, created using 100 line scans spaced 30 µm apart. The scan rate was calculated to be 175 µm/s. Microelectrode Array Microscopy. A diagram of the MEAM apparatus is shown in Figure 1. The MEAM instrument uses an array with 100 microelectrodes spaced 400 µm apart in a square (10 × 10) arrangement (Acute 100 electrode array, Cyberkinetics Inc., Foxborough, MA). The array effectively analyzes a 4000 µm × 4000 µm area. The microelectrodes were made of doped-Si shafts that come to a sharp point. The shaft design was employed, because more readily available and easier to fabricate planar arrays would lead to diffusion and solution resistivity issues in the MEAM configuration. The final 30 µm of the tips was coated with Pt. The remainder of the tip was coated with an insulating coating of paralene. The electrode array was mounted on the SECM motor stage for positioning above the sample but held stationary during the experiment. Each microelectrode was individually connected and controlled by the multichannel microelectrode analyzer (MMA) and packaged software (model 900, Scribner Associates) that potentiostatically controls up to 100 electrodes. MEAM images are displayed using the MMA software graphics, where each square (pixel) represents a microelectrode response shaded by the magnitude of the response and positioned in the correct spatial orientation. When used, the sample potential was controlled by a separate chassis-isolated potentiostat (Omni 90, Cypress Systems). Separate reference and auxiliary electrodes were used for the array elements and the sample. Electrochemical Cells. Two electrochemical cells were used in the experiments. The first cell, used for SECM measurements only, was constructed from PTFE and quartz and held a volume of approximately 60 mL. The samples (5.6 mm disk) were sealed
to a Teflon sample holder using FEP Teflon shrink tubing, such that only the finished surface was exposed to solution. The sample holder screwed into the cell base. The second electrochemical cell was used for MEAM experiments. This cell was constructed of Nylon and held approximately 200 mL of solution during the tests. Samples were potted in epoxy (Epo Kwick, Buehler) and press fit into the bottom of the cell. The sample for MEAM was a 6 mm diameter disk, so the array only measures activity at a portion of the sample. Pt auxiliary electrodes were used in all cells. Experimental Details. Type 304 stainless steel (304 SS) samples were cut from 6 mm diameter rods (Eagle Alloys). The samples had the following composition (wt %): C, 0.013; Mn, 1.70; Si, 0.42; P, 0.029; S, 0.026; Cr, 18.32; Ni, 8.50; Mo, 0.50; Cu, 0.44; Co, 0.13; N, 0.076; Fe, balance. These samples were polished to a final 800-grit finish using SiC paper (Beuhler). After the samples were loaded into the cell, they were rinsed sequentially in acetone, ethanol, and deionized water. Solutions were made using 18 MΩ‚cm water (Nanopure Diamond UV/UF, Barnstead). Chemicals were all ACS grade or better (Alpha Aesar). All experiments in this paper were performed at laboratory ambient temperature, which ranged from 22 to 25 °C. Postexperiment Analysis. Samples were analyzed following the corrosion experiment using light optical microscopy (LOM) and in some cases using scanning electron microscopy (SEM). An environmental scanning electron microscope (Philips model XL 30) was used for SEM analysis. RESULTS AND DISCUSSION Electrochemistry of Austenitic Steels in Dilute KI Solutions. Experiments to demonstrate the ability of SECM to detect oxide film breakdown were carried out in 10 mM KI. This solution acts as both a mildly corrosive solution and a redox mediator (I-/ I2). Similar experiments were discussed elsewhere.28 Cyclic voltammetry of 304 SS in 10 mM KI solution is shown in Figure 2A. The positive sweep from the open circuit potential displays a passive current until the potential is positive of the oxidation potential for I- (approximately 0.4 V vs Ag/AgCl). A peak centered at approximately 0.65 V overlaps with a broad increase in current to the reversal potential. On the return sweep, a hysteresis is observed, indicating surface activation, usually a sign that pitting corrosion sites have been initiated and are growing. A broad reduction feature was observed on the reverse sweep centered at -0.05 V. The cyclic voltammetry response is interpreted as follows. The oxidation peak in the forward sweep is attributed to oxidation of I- ions superimposed on oxide growth and dissolution processes. Repeated cycling results in attenuation in the overall current.29 This phenomenon is attributed to oxide film growth, where oxide film thickening increases the barrier to electrontransfer processes. Despite the overall decrease in current in cyclic voltammetry experiments, the oxide film on 304 SS is susceptible to localized breakdown as evidenced by current measurements at fixed potentials. Figure 2B shows a chronoamperometry curve for a fresh 304 SS sample polarized from open circuit (near 0 V) to 0.60 V for many hours. The erratic current trace is typical of stable pitting corrosion, which is confirmed by microscopic examination of the sample surface following the experiment. The conditions (solution composition and potential) in these studies were chosen such that the sample was on the edge of stability Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
2603
Figure 3. SECM images (3000 µm × 3000 µm area, Z-scale 0 to -25 nA) of 304 SS in 100 mM KI. The potential of the sample was stepped to 0.5 V just prior to the initiation of imaging. Each image was taken of the same region of the surface and required 39 min to acquire. Some sites in images 3, 4, and 6 are greater than the maximum scale value to allow weaker sites to be observed in the plots. Figure 2. Electrochemistry of 304 SS in 10 mM KI: (A) cyclic voltammogram at 10 mV/s and (B) chronoamperometry curve where the potential was stepped from 0 to 0.60 V at time 0 s.
during the experiment. Pitting corrosion events are likely initiated at manganese sulfide inclusions, a known mechanism for pitting corrosion of stainless steels.38 SECM Studies of Localized Corrosion. The use of SECM in detecting sites of oxide film breakdown will be described below. This technique provides information about localized behavior, albeit with poor temporal resolution. To improve the temporal resolution of the experiment, the MEAM instrument was developed as will be described later in the paper. To perform SECM imaging of the surface, the sample is leveled with respect to the X-Y tip travel to allow the tip to cover an area of 3000 µm × 3000 µm at a 10 µm tip-sample distance without touching the surface. This provides a large enough area to capture several pitting corrosion sites. The sample is polarized from open circuit conditions typically near 0 V to the breakdown region, positive of 0.40 V. Imaging is initiated following the potential step with the microelectrode potential at -0.2 V to collect oxidized iodide products (I2 or I3-). Previous work has associated the sites of activity observed on 304 SS with pitting corrosion sites.28-30 This was based on careful observation of sites of activity followed by postmortem examination of the surface in that region. LOM and SEM analysis confirmed that the pitting corrosion is associated with the observed activity. In the process of studying the breakdown of oxide films on stainless steel, it was realized (38) Eklund, G. S. J. Electrochem. Soc. 1974, 121, 467.
2604 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005
that the technique could be used to observe the evolution of corrosion on surfaces by capturing many SECM images over the same region of the surface. These studies can be performed by positioning the tip above the place of interest on the surface or by capturing several images of the same area of the surface. A paper outlining this approach was published.28 In those experiments, the evolution of sites of electron transfer evolved in a dramatic fashion. The responses observed were explained as a result of three processes: (1) sites occur where growth or a change in the composition of the oxide film effectively turns off delocalized electron transfer across the oxide film, (2) weak (
