Electrical properties of Mg-doped p-type GaN grown by metalorganic chemical vapor deposition have been investigated by Hall effect and conductivity measurements. Metastability and persistent photoconductivity effects have been observed in GaN. It was found that at low temperatures, it takes several hours for the free hole concentration to reach its equilibrium value in the dark as well as in the photoexcited state, implying a bistable nature of impurities in p-type GaN. Temperature dependence of these behaviors have been studied, from which the energy barrier for free hole capture by ionized impurities as well as between the metastable and the stable states of neutral impurities have been obtained.

High-purity aluminum samples were implanted with 35 keV Cl ϩ then polarized in both Cl Ϫ -containing and Cl Ϫ -free electrolytes in order to ascertain corrosion behavior as a function of Cl Ϫ content in the oxide. Implant fluence between 5 ϫ 10 15 and 2 ϫ 10 16 Cl ϩ cm Ϫ2 resulted in little or no localized attack. Implant fluences of 3 ϫ 10 16 and 5 ϫ 10 16 Cl ϩ cm Ϫ2 resulted in significant pitting in a Cl Ϫ -free electrolyte with the severity scaling as a function of implant fluence. The low variability in the pitting behavior of the 5 ϫ 10 16 Cl ϩ cm Ϫ2 sample suggests that this implant dosage results in a critical Cl Ϫ concentration in the oxide for pit nucleation. The passive current density (i pass ) decreased with increasing implant fluence. A space-charge effect is proposed to account for this phenomenon, although effects from defect interactions and possible oxide thickening are still under consideration.

The pitting potential of pure aluminum thin films in 50 mM K 2 SO 4 was measured as a function of implanted Cl fluence. Samples were implanted with 35 keV Cl + at room temperature using fluences from 2.25 ϫ 10 16 to 3.25 ϫ 10 16 ions cm −2 in increments of 0.25 ϫ 10 16 . An empirical relationship between pitting potential and fluence was found which suggests a critical Cl concentration in the oxide is necessary for pit initiation. No correlation between pitting potential and the measured Cl concentration or distribution in the metal was found.A key aspect to understanding oxide destabilization and subsequent pitting of passive metals is to elucidate the role of aggressive anions, particularly Cl − . Although numerous theories have been proposed in which fundamental physical and chemical processes for the interaction of Cl − and passivating oxides are described, 1-9 the exact mechanism which causes oxide breakdown has yet to be determined. Some of the information that is lacking to date is the concentration of Cl and its specific location ͑i.e., surface, bulk, or interfacial͒ in the oxide that results in failure.The interaction of aqueous chloride and passive oxides has been studied by a variety of methods. Kolics et al. used radioactive tracers to study passive film breakdown as a function of chloride adsorption incorporation. In a comparative study of sulfate and chloride adsorption on 99.999% aluminum they showed that a significant portion of adsorbed chloride is incorporated into the passive film and the irreversibility of adsorption is enhanced by film growth. 10 Their findings support that chloride anion incorporation rather than adsorption alone contributes to initiation of the pitting process. Berzins et al. 11 measured localized chloride adsorption on 99.5% aluminum and found a quantitative relationship between Cl adsorption and NaCl solution concentration between 1 and 10 −5 M. The concentration of chloride within the oxide was not determined, however, and pitting did not occur at all adsorption sites. Therefore, the relationship between localized chloride adsorption and pitting was not clear. Bockris and Minevski found a linear dependence of the breakdown potential on the log of concentration for aqueous chloride levels between 10 −4 and 10 −2 M. 12 In another set of studies, ion implantation was used to study the migration of halide species in anodic alumina. Brown and Mackintosh and Skeldon et al. found that during film growth ͑i.e., under an electric field͒, implanted chloride ions migrate inward, toward the aluminum substrate. 13,14 These findings were supported by Shimuzu et al. in another migration study which used a different source of chloride ions incorporated into oxide films. 15 Shimuzu's reported chloride migration rates were in excellent agreement with values obtained by using implanted chloride, indicating that ionimplantation damage had little effect on the ionic transport processes of chloride. While the mobility of chloride in aluminum oxide has therefore been established, its relations...

Fluctuation electron microscopy studies have been performed on several aluminum oxides exposed to different electrochemical conditions. Little is known about amorphous aluminum oxide structures and their relationship with their passivation behaviors. Corrosion studies have shown that exposure of aluminum oxide films to Cl ions in solution reduces the oxide's passivity, and this results in the onset of pitting corrosion. The physical changes that occur in the oxide as a result of Cl exposure have not been previously identified due to the difficulty in investigating the structure of this amorphous material. Fluctuation microscopy is a new electron microscopy technique that is able to detect the presence of medium range order structures in amorphous systems. In this paper, we will report fluctuation microscopy results on amorphous aluminum oxides that have been exposed to Cl ions in solution and compare them with oxides that have seen no electrolyte exposure or that have been exposed to electrolytes that do not contain Cl-,such as SO42- containing electrolytes. We will also compare the Cl-exposed oxides with oxides that have been implanted with Cl ions. The differences in pitting behaviors for these oxidesare consistent with our previous speculation on the effect of medium range order on the passivation behavior of aluminum oxides grown using ozone.

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 Å͒. 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 both͒ on 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. Pitting corrosion can be described in terms of four distinct and consecutive stages: ͑i͒ interactions at the oxide/solution interface, ͑ii͒ interactions within the passive film, ͑iii͒ metastable pitting, and ͑iv͒ stable 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 nature 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 following: 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 dissolution or thinning 5 ; and according to the point defect model ͑PDM͒, 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 potential͒. Measurements of average chloride levels in the oxide, however, may not be representative of the local distribution 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 different approach to the problem, allowing the Cl concentration and distribution to be specified t...