S1 The Journal of Materials Chemistry Microscopic vs macroscopic structural evolution of SiO 2 sols and gels employing a tailor-made fluorescent reporter dye Brita Unger, Knut Rurack, Ralf Müller, Harald Jancke and Ute Resch-Genger Electronic Supplementary Information S1 Correction of fluorescence spectra. Fluorescence spectra obtained for TEOS-derived sols and gels can be strongly distorted by Raman scattering from the matrix. To account for this problem, we investigated a series of undoped samples prepared and treated in the same way as the dye-doped samples. Additionally, measurements with crossed polarizers were performed. After background subtraction, all the spectra were corrected for the spectral responsivity of the instrument. To minimize potential errors from residual background signals, every corrected spectrum was fitted to a single or a progression of lognormal functions. The band maxima given in the text and ESI always correspond to the results of the fitted spectra (reproducibility of the fits: ± 1 nm). S2 1–II and aggregation. Based on the following considerations, aggregation does not seem to be the effect triggering the appearance of type II fluorescence in sols and gels doped with 1. As mentioned in the text, even when saturated, ethanolic solutions of 1 never showed any other bands than the respective corresponding dilute solutions. For 1, especially the large and flexible crown unit should prevent spontaneous aggregation. Moreover, since the crown unit does only bind to Hg(II) ions S1 —which are not present under the conditions used here—cation-assisted dimerization via the crown units sandwiching the metal ion is also very unlikely. Furthermore, the similarity of the spectral data found in the protonation studies (see 3.4.2) disfavor aggregation. This is stressed by a comparison of the red-shifted spectra of 1–II with those of closely related styryl dyes such as 4. S2 Additionally, the occurrence or absence of the red-shifted components in the case of 1 as compared to 2 is well-reflected by their different proton or hydrogen-bonding sensitivities. For π-stacking and aggregation, the exchange of a dimethylamino for a methoxy group is unlikely to have such a dramatic effect.