3-2. Comprehensive experimental and computational screening of magnesium corrosion inhibitors

S.V. Lamaka¹*, C. Feiler¹, T. Wurger, Di Mei¹, B. Vaghefinazari¹,

R. H. Meissner, M.L. Zheludkevich^1,2

1. Magnesium Innovation Centre - MagIC, Helmholtz-Zentrum Geesthacht (HZG),Germany

2. Institute for Materials Science, Faculty of Engineering, Kiel University, Germany

Abstract: This work presents the results of a systematic search for magnesium corrosion inhibitors by traditional experimental and computational machine learning methods .

The extensive database gathered the influence of a large number of organic/inorganic compounds on behaviour of Mg materials in aqueous electrolytes. A number of new corrosion inhibitors were discovered with efficiency exceeding that of chromate. Inhibiting efficiency of multiple compounds was found to be dependent on the alloy specification.

At the same time, the field of potential magnesium inhibitors is vast, containing several millions of individual compounds. This makes it impossible to test all promising compounds experimentally and collect detailed atomistic understanding of the inhibition mechanisms for each of them.

Instead, Machine Learning approach employing Artificial Neural Network was developed (combining experimental, machine learning, DFT calculations and molecular dynamics tools) to enable in silico prediction of inhibition (and dissolution promoting) properties. Experimental validation shows high potential of computational methods for rapid, precise and cost-effective search of corrosion inhibitors urgently needed to substitute the carcinogenic chromates .

Concomitantly, a property-structure landscape based on the results of hydrogen evolution measurements was created. It vividly demonstrated the relationship between corrosion inhibition efficiency and corresponding molecular structure of magnesium corrosion inhibitors. After creating a high-dimensional similarity measure with the SOAP-REMatch kernel between 74 tested compounds, the similarity matrix was reduced to a two-dimensional visualization with sketch-map, providing a reference to qualitatively predict the inhibition behaviour of yet to be tested molecules .