Voltage-dependent first-principles simulation of insertion of chloride ions into Al/Al$_2$O$_3$ interfaces using the Quantum Continuum Approximation

TL;DR

This paper introduces the Quantum Continuum Approximation (QCA), a novel method combining DFT and electrostatic models to predict chloride ion insertion into aluminum oxide interfaces at various voltages, aiding corrosion understanding.

Contribution

QCA enables first-principles voltage predictions for electrochemical interfaces by reducing computational cost and coupling atomistic and continuum models.

Findings

Chloride insertion favorable at voltages > -0.2 V in systems with grain boundaries

QCA predicts corrosion onset consistent with experimental observations

Method reduces DFT atom count by an order of magnitude

Abstract

Experiments have shown that pitting corrosion can develop in aluminum surfaces at potentials $> -0.5$ V relative to the standard hydrogen electrode (SHE). Until recently, the onset of pitting corrosion in aluminum has not been rigorously explored at an atomistic scale because of the difficulty of incorporating a voltage into density functional theory (DFT) calculations. We introduce the Quantum Continuum Approximation (QCA) which self-consistently couples explicit DFT calculations of the metal-insulator and insulator-solution interfaces to continuum Poisson-Boltzmann electrostatic distributions describing the bulk of the insulating region. By decreasing the number of atoms necessary to explicitly simulate with DFT by an order of magnitude, QCA makes the first-principles prediction of the voltage of realistic electrochemical interfaces feasible. After developing this technique, we apply QCA to predict the formation energy of chloride atoms inserting into oxygen vacancies in Al(111)/$\alpha$-Al$_2$O$_3$ (0001) interfaces as a function of applied voltage. We predict that chloride insertion is only favorable in systems with a grain boundary in the Al$_2$O$_3$ for voltages $> -0.2$ V (SHE). Our results roughly agree with the experimentally demonstrated onset of corrosion, demonstrating QCA's utility in modeling realistic electrochemical systems at reasonable computational cost.