We investigate the spectral behavior of nine Titan impact craters in order to constrain their composition. Past studies that have examined the chemical composition of impact craters on Titan have either used qualitative comparisons between craters or combined all craters into a single unit, rather than separating them by geographic location and/or degradation state. Here, we use Visual and Infrared Mapping Spectrometer (VIMS) data and a radiative transfer code to estimate the atmospheric contribution to the data, extract the surface albedos of the impact craters, and constrain their composition by using a library of candidate Titan materials, including essentially water ice, tholin, a dark component, and other possible ices at different grain sizes. Following a general characterization of the impact craters, we study two impact crater subunits, the "crater floor" and the "ejecta blanket". The results show that the equatorial dune craters - Selk, Ksa, Guabonito, and the crater on Santorini Facula - appear to be purely composed of organic material (mainly an unknown dark component). Titan's midlatitude plain craters - Afekan, Soi, and Forseti - along with Menrva and Sinlap, are enriched in water ice within an organic-based mixture. This follows the geographic pattern observed in our previous work with VIMS data, where the uppermost layers of the midlatitude alluvial fans, undifferentiated plains, and labyrinth terrains were found to consist of a mixture of organics and water ice, while the equatorial plains, hummocky terrains, and dunes were found to consist of a mixture of dark material and tholins. Furthermore, we found that the addition of some form of ice improves the fit in the ejecta spectra of Afekan and Sinlap craters. We find no indication for the presence of either NH3 or CO2 ice. Our main results agree with an existing Titan surface evolution scenario, wherein the impact cratering process produces a mixture of organic material and water ice, which is later "cleaned" through fluvial erosion in the midlatitude plains. This cleaning process does not appear to operate in the equatorial regions, which are quickly covered by a thin layer of sand sediment (with the exception of the freshest crater on Titan, Sinlap). Thus, it appears that active processes are working to shape the surface of Titan, and it remains a dynamic world in the present day.