Effect of Implanted Cl and Deposited Oxides on the Pitting Behavior of Aluminum L. M. Serna, z C. M. Johnson, F. D. Wall, * and J. C. Barbour Sandia National Laboratories, Albuquerque, New Mexico 87185-0888, USA Insight into the influence of Cl on the pitting behavior of aluminum has been gained using a combination of ion implantation and oxide deposition. High-purity Al thin-film samples were implanted with 35 keV Cl + followed by plasma deposition of an aluminum oxide Al 2 O 3 . The pitting potential of unimplanted areas in 50 mM NaCl increased with increasing deposited oxide thickness 0, 80, 140 Å. For implanted areas, polarization in 50 mM NaCl resulted in pitting that was insensitive to oxide thickness and implant fluence above a critical level. By comparison, polarization in 50 mM K 2 SO 4 resulted in pitting which exhibited a dependence on both deposited oxide thickness and implant fluence. Finally, the effect of the Cl source solution, implantation, or bothon the pitting behavior of Al with a deposited oxide was examined. These results are used to support the hypothesis that both a critical chlorine distribution and oxide modification are contributing factors to pit nucleation. © 2005 The Electrochemical Society. DOI: 10.1149/1.1924308All rights reserved. Manuscript submitted August 3, 2004; revised manuscript received January 5, 2005. Available electronically June 7, 2005. Pitting corrosion can be described in terms of four distinct and consecutive stages: iinteractions at the oxide/solution interface, iiinteractions within the passive film, iiimetastable pitting, and ivstable pit growth. 1 The first two stages have been the subject of intense research efforts yet remain the least understood. The inherent difficulties in studying pit nucleation arise due to the localized na- ture of the attack and the short time scale associated with nucleation. Furthermore, while there are many proposed theories which seek to explain the breakdown of passive films, 2-10 the role of chloride in the oxide destabilization and pit nucleation process remains unclear. The prominent mechanisms describing the interaction of aggressive anions with the passive film include but are not limited to the fol- lowing: Cl - penetration of the oxide resulting in film breakdown at the metal/oxide interface 2 ; competitive anion O 2- vs. Cl - adsorption 3,4 ; complex ion formation leading to localized film dis- solution or thinning 5 ; and according to the point defect model PDM, increased cation diffusion resulting in void formation at the metal/oxide interface. 6 The first step in the penetration of chloride into the oxide film is adsorption of ions followed by absorption and migration. Several authors have studied the distribution of chloride at the oxide/solution interface 11,12 and within the oxide 12,13 under electrochemical conditions relevant to pit nucleation i.e., below and at the pitting potential. Measurements of average chloride levels in the oxide, however, may not be representative of the local distribu- tion associated with pit nucleation in the metal. Recently, ion implantation techniques have been used to study the pitting behavior of aluminum. 14,15 For clarity, in this paper the aqueous species is referred to as chloride, Cl - , and the implanted species is chlorine, Cl. The use of ion implantation provides a dif- ferent approach to the problem, allowing the Cl concentration and distribution to be specified then linked to the observed pitting be- havior. In addition, ion implantation effectively bypasses the trans- port steps leading to pit nucleation by preloading the oxide with Cl. Testing implanted samples in Cl - and Cl - -free electrolytes allowed for the pitting process to be investigated with and without Cl - inter- actions at the oxide/solution interface and possibly without Cl - transport. It was previously demonstrated that implantation of Cl into pure aluminum above a critical fluence 2 10 16 cm -2 ren- ders the material susceptible to pitting, even in the absence of aque- ous Cl - . 14 Because implantation resulted in the presence of Cl in both the native oxide and the underlying metal, the locations where Cl interacts with the system to initiate breakdown i.e., at the oxide/solution interface, in the oxide, or at the metal/oxide interface could not be unambiguously determined. In the present study, a combination of ion implantation and oxide deposition is used to better define the distribution of Cl in the Al/Al 2 O 3 system. Three objectives were set out for this work. First, to determine the pitting behavior of Al with a deposited oxide ex- posed to aqueous Cl - . Second, to establish the impact of implanted Cl on the pitting behavior of a sample capped with a Cl-free depos- ited oxide in a nonhalogen electrolyte. In the above two objectives, there is only one available source of destabilizing species Clwhich are differentiated by their origin in the system. The final objective is to determine the pitting behavior of a system with a combination of both Cl sources in an effort to determine possible controlling or contributing mechanisms for the role of Cl in pit nucleation. Experimental Sample preparation and analysis.— A sample structure where a Cl-free deposited oxide caps an implanted Al electrode schemati- cally shown in Fig. 1was created using a series of steps. Thin-film samples were generated by depositing 2000 Å of aluminum 99.9999%onto a SiO 2 /Si wafer using electron-beam evaporation. A native oxide formed upon exposure of the sample to air. Next, samples were implanted at room temperature with 35 keV Cl + to fluences of 3 10 16 and 5 10 16 cm -2 through 0.5 cm 2 apertures. The spatial distribution of implanted Cl was calculated using TRIM-92 software 16 and is shown for the stated fluences in Fig. 2. For an implant energy of 35 keV, the Cl concentration in the native oxide approximately the first 50 Å of the sampleis approximately 1-3 atom %. The peak concentration of Cl is dependent on the im- plant fluence and is located in the metal. Finally, an aluminum oxide was deposited over the entire surface of the sample implanted and unimplanted areas. This was done by exposing the sample to electron-beam-evaporated aluminum and an electron cyclotron reso- nance ECRoxygen plasma simultaneously, resulting in the depo- sition of an Al 2 O 3 layer that mimics a pure passive oxide. Oxide * Electrochemical Society Active Member. z E-mail: lmserna@sandia.gov Figure 1. Schematic of structure for samples investigated in this study show- ing approximate locations of implanted chlorine Cl. Journal of The Electrochemical Society, 152 7B244-B249 2005 0013-4651/2005/1527/B244/6/$7.00 © The Electrochemical Society, Inc. B244 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 54.237.14.251 Downloaded on 2016-10-25 to IP