Mono- and disilicic acids were stabilized by uncharged polyethylene glycols (PEGs) in silica-supersaturated solutions (the starting solution contained 500 ppm/8.3 mM sodium orthosilicate, Na2SiO3 center dot 5H(2)O, expressed as SiO2) at pH = 7, most likely by hydrogen bonding between the silanol groups and -CH2-CH2-O ether moieties. The stabilization was monitored by measuring molybdate-reactive silica and also by a combination of liquid- and solid-state Si-29 NMR spectroscopy. It depends on PEG concentration (20-100 ppm) and molecular weight (1550-20 000 Da). Two narrow Si-29 NMR signals characteristic for monosilicic acid (Q(0)) and disilicic acid (Q(1)) can be observed in Si-29 NMR spectra of solutions containing PEG 10000 with intensities distinctly higher than the control, that is, in the absence of PEG. Silica-containing precipitates are observed in the presence of PEG, in contrast to the gel formed in the absence of PEG. These precipitates exhibit similar degrees of silica polycondensation as found in the gel as can be seen from the Si-29 MAS NMR spectra. However, the D-2 HETCOR spectra show different H-1 NMR signal shifts: The signal due to H-bonded SiOH/H2O, which is found at 6 ppm in the control, is shifted to similar to 7 ppm in the PEG-containing precipitate. This indicates the formation of slightly stronger H-bonds than in the control sample, most likely between PEG and the silica species. The presence of PEG in these precipitates is unequivocally proven by C-13 CP MAS NMR spectroscopy. The C-13 signal of PEG significantly shifts and is much narrower in the precipitates as compared to the pristine PEG, indicating that PEG is embedded into the silica or at least bound to its surface (or both), and not phase separated. FT-IR spectra corroborate the above arguments. The H-bonding between silanol and ethereal 0 perturbs the band positions attributed to vibrations involving the 0 atom. This work may invoke an alternative way to envision silica species stabilization (prior to biosilica formation) in diatoms by investigating possible scenarios of uncharged biomacromolecules playing a role in biosilica synthesis.