Do zeolite precursor species really exist in aqueous synthesis media?

The authors of a recent report in this journal (Houssin, et al., Phys. Chem. Chem. Phys., 2003, 5, 3518, ref. 29) claim that a tetrapropylammonium (TPA) silicate mixture with molar composition 4.41 TPAOH : 10 SiO2 : 117 H2O contains high concentrations of two silicate oligomers, the prismatic double five-ring and a novel pentacyclic dodecamer. The latter species is derivative of the framework structure of silicalite-1, a MFI-type zeolite that spontaneously crystallizes from this system, and, indeed, the authors declare it to be a "nanoprecursor'' in the TPA-mediated growth of silicalite-1. Using two-dimensional Si-29 COSY NMR spectroscopy to examine an equivalent mixture enriched in the Si-29 isotope, we show that the published peak assignments are erroneous. The signal assigned to the double five-ring corresponds, in reality, to the well-characterized pentacyclic octamer, while the three signals assigned to the pentacyclic dodecamer arise from three completely separate species. We find no evidence, therefore, to support the suggestion that silicalite-1 is formed by sequential condensation of precursor species any more complex than the simple orthosilicate anion.