The following specs has been used by our clients for the to coat the silicon surface with an epi-layer.
The silicon wafers we purchase from you are used as substrates for epitaxial coating of silicon from our silicon gases (i.e. silane, disilane, and dichlorosilane). These epi coatings are deposited onto the wafers, then analyzed via Fourier Transform Photoluminescence spectrometry for acceptors and donors. Acceptors and donors are Group III and V elements phosphorus, boron, arsenic, aluminum, and in rare cases gallium, which give the silicon (normally an insulator) its semiconducting properties. As we test for the above elements, we need the substrate wafers to be largely free of them. The FTPL analyses are done down to the single-digit ppta level, so ultrahigh purity throughout this analysis is of the utmost importance. This is because the ultra-thin epitaxial layers (i.e. <100 microns) allow the substrate to be somewhat "visible" through the epitaxial coating. Hence, doped wafers will distort the analysis/measurement, and make the epitaxy (which we use to quantify these dopants in our silicon gases) appear to be more contaminated that it actually is. The FTPL measurement is what we use to certify the acceptor and donor levels to our customers, on our COAs for our products.
Having bismuth or indium doped wafers would be a huge plus for us, as the major feature in our FTPL spectra is the large silicon two-phonon band. If the substrate were doped with bismuth or indium, it would make the two-phonon band appear smaller to the spectrometer, effectively removing the large two-phonon band from the substrate spectrum, while not interfering with the spectrum of the epitaxial layer. However, we have so far been unable to find any such doped wafers. Even gallium-doped wafers would possibly be valuable in this regard, although the gallium peak is much closer to the no-phonon bands of phosphorus arsenic and aluminum that we use to quantify these analytes.
I hope this information is useful to you. If this is unclear, I have some spectra marked up with the features I have described that I can e-mail to you that might make things a bit more clear.