Computing with Inorganic Cellular Frameworks and Nets

The enormous potential of parallel computing has led to the first prototype devices being constructed. However, all the examples to date rely on complicated chemical and/or physical manipulations, and hence do not lend themselves to the kind of widespread investigation necessary to advance the field. This article presents a new paradigm for parallel computing: the use of solid, single crystalline materials as cellular automata suggesting the idea of the “Crystal Computer,” now possible due to a new class of crystalline cellular materials that undergo single-crystal-to-single-crystal (SC-SC) oxidation and reduction (REDOX) reactions. Two avenues are proposed for investigation: reversible single-crystal to single-crystal electronic transformations and solid-state spin transfer within spin-crossover complexes. Both schemes allow computation to occur in three dimensions, within cheap and easy to assemble materials and using commonplace techniques for input and readout. and white cells (the input to the system), patterns can be created out of apparent disorder by repeatedly applying the over-riding laws to each successive “generation” of cells. These patterns can be viewed as the output of a computational sequence by the cellular array, and it has been shown that CAs of this type are Universal Turing Machines (Turing, 1937; Berlekamp et al., 2003) capable of computing anything that can be computed algorithmically. Recent practical examples of functioning CA based on chemical systems include automata capable of modeling cancer cell proliferation (Bandyopadhyay et al., 2010) and playing simple “games” against a human opponent (Pei et al., 2010). However, these surface-confined DOI: 10.4018/jnmc.2011010103 International Journal of Nanotechnology and Molecular Computation, 3(1), 24-34, January-March 2011 25 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. systems are yet to exploit the inherently three dimensional nature of chemical interactions fully and as a consequence the computing power of such 2-D devices may be limited by these geometrical constraints. Specifically, creating real-world CA capable of acting as true Universal Turing Machines remains an enormous challenge. These extant real-world systems also rely on time and labor-intensive techniques for data input and read-out, such as scanning-tunnel microscopy (Bandyopadhyay et al., 2010) and/or “wet chemical” manipulations (Pei et al., 2010). Herein, we propose an approach to cellular automata based on single crystal-single crystal (SC-SC) transformations driven by optical and/ or electrical impulses supplied to the crystal by simple and readily available equipment. Two possible systems will be expounded, both with the potential for three-dimensional computing using very simple apparatus under standard “bench-top” conditions (e.g. room temperature and pressure). The first relates to work performed in the Cronin group on reversible electron uptake and loss leading to SC-SC transformations in crystals of composition [(C4 H10NO)40(W72M12O268Si7)]·48H2O (M = Mn III or CoIII) (Ritchie et al., 2008; Thiel et al., 2009; Thiel et al., 2010). The second approach relies on the phenomenon of spin cross-over within certain crystalline materials (Gütlich et al., 2000; Turner & Schultz, 2001; Halcrow, 2008; Bodenthin et al., 2009). CRYSTAL COMPUTING WITH POLYOXOMETALATES Crystals of the form [(C4H10NO)40(W72M12O268 Si7)]·48H2O (M = Mn III) (1ox) self-assemble from a solution of the well-studied tungsten polyoxometalate [γ-SiW10O36] 8and a simple manganese(II) salt and hence are straightforward to prepare. The structure of these crystals comprises an infinite array of 3and 4-connected Keggin polyanions (Long et al., 2010), where each three connected unit is surrounded by three neighboring clusters in a trigonal-planar fashion, and each 4-connected unit features four nearest neighbors located on the vertices of a distorted tetrahedron (Figure 1). The linkages between the clusters in Figure 1 are [W-O-Mn] bridges, which can be considered as “bolts” holding together the tetravacant {SiW8O36} (tetrahedral) and trivacant {SiW9O37} (trigonal-planar) building blocks. These bridges have no preferred orientation, and so statistically half of the linkages are between an Mn on a tetravacant cluster and a W on a trivacant cluster and the other half are between a W on a tetravacant cluster and an Mn on a trigonal cluster. We can thus begin to generate some simplified schemes representing the molecular structure within these crystals which will demonstrate their potential as CA grids (Figure 2). Figure 2a shows a 2-D slice through a crystal of form 1ox, with tetrahedral cross-secFigure 1. Photograph of crystals of 1ox with schematic showing how triangular (green) and tetrahedral (red) units link up in three dimensions to produce an infinite 3,4-connected framework 9 more pages are available in the full version of this document, which may be purchased using the "Add to Cart" button on the publisher's webpage: www.igi-global.com/article/crystal-computer-computinginorganic-cellular/54342