Data-driven Prediction of HSQ Polymer Structure and Silicon Nanocrystal Photoluminescence

Synthesis of silicon nanocrystals through high-temperature pyrolysis of hydrogen silsesquioxane has emerged as a valuable approach for obtaining quantum-sized crystallites with controllable size and distinct photoluminescent maxima. Nevertheless, the use of hydrogen silsesquioxane has some notable disadvantages: poor shelf life, high cost, and limited supply. These issues motivate an exploration of alternative precursors. Recent studies have investigated silsesquioxane-like polymer precursors that are derived from various types of silanes (e.g., trichlorosilane, triethoxysilane), offering a cost-effective and tunable option for silicon nanocrystal synthesis. Here, we elucidate the relationship between the structure of the polymeric precursor and the photoluminescence of alkyl-passivated silicon nanocrystals by a statistical design of experiments technique called response surface methodology. Using this technique, we quantitatively model the relationship between precursor molar ratios, polymer structure (cage vs. network content), and photoluminescence quantum yield of alkyl-passivated silicon nanocrystals. We found that synthesis approaches to silsesquioxane polymers that use higher proportions of methanol and water in a trichlorosilane:water:methanol mixture result in larger amounts of network-type polymer structures, and that the network-type polymeric precursors yield silicon nanocrystals with higher photoluminescence quantum yields. While polymer structure strongly correlates with precursor composition, the correlation with the photoluminescence quantum yield is weaker, suggesting other factors play a key role in the photoluminescent properties of silicon nanocrystals.