Multivariate Statistical Analysis of Two-Dimensional Metal Cluster Arrays Grown in Vitro on a Bacterial Surface Layer

Abstract

Regular nanocluster arrays were fabricated by chemical reduction of Pd complexes in the presence of the two-dimensional crystalline surface layer protein from the bacterium Sporosarcina ureae. The size of Pd clusters grown on the protein, its occupation properties, and the local cluster distribution on the protein surface were analyzed by electron microscopy, correlation averaging, and multivariate statistical analysis. The Pd clusters of 1.2-1.5 nm in size are preferably located in and close to pores and gaps of the protein framework in seven cluster sites of three different types per unit cell. The sites are almost completely occupied. The clusters appear strongly directed by the protein template but show positional variability within the main cluster sites. This variability is an inherent property of the Pd clusters and cannot be explained by statistical error with positional analysis. Binding studies with Pd complexes prior to reduction suggested that a high number of potential nucleation sites are formed on the protein surface, facilitating the initial growth of clusters at variable positions. A model for heterogeneous nucleation and cluster development on protein surfaces is proposed and compared to the mechanism of reductive metallization of DNA investigated previously.