GaP and its ternary relatives such as AlGaP and InGaP, are commonly used in devices ranging from LEDs to solar cells. The development of these materials for nonlinear optical devices, however, is more recent. Growth and processing of GaP need to be specifically tailored for orientation-patterning, which is a process by which one can control the two crystallographic orientations of a polar, III-V material on a nonpolar, Si substrate. This technique involves growth by molecular beam epitaxy (MBE), lithography and etching, and regrowth. For epitaxial regrowth, the GaP surface after etching and photoresist removal is extremely important. In this study, we investigated the surface preparation of GaP for MBE regrowth by AFM to study surface morphology, XPS to detect the presence of remaining oxide or contaminants on the surface, and cross-section TEM to study the film regrowth quality. 200-nm GaP films on GaP substrates were used. By using standard films for these studies, we could isolate any processing and regrowth issues to these steps, rather than problems related to GaP on Si growth. The samples underwent various times of surface cleaning by O2 plasma etching: 0, 1, 5, and 10 min. AFM images before and after O2 cleaning revealed smooth surfaces with rms roughness values of ~0.2 nm. The XPS scans before and after the surface cleaning steps revealed several differences in the oxide created by the O2 plasma clean and native oxide including different ratios of the P2p and O1s peak areas and a 1-eV chemical shift in the Ga3d peak. Regrowth on a 30-sec plasma-cleaned sample resulted in areas of poor crystalline quality as seen by cross-section TEM. This is likely due to residual plasma-induced oxide that could not be completely thermally desorbed in the MBE chamber. The need for an additional surface cleaning step prior to regrowth was necessary to ensure that the plasma-induced oxide was fully removed and replaced by the native oxide to protect the surface yet could be thermally desorbed under UHV. O2 plasma-cleaned samples were etched with either dilute hydrofluoric acid (HF) or hydrochloric acid (HCl). A rough surface resulting from the HF etch and smooth surface from the HCl etch was seen by AFM, yet no difference between the surfaces was detected by XPS. GaP and GaP/Si samples were processed and prepared for regrowth by a 1-min O2 plasma clean and HCl etching, then regrown by MBE. Cross-section TEM will be presented to show the regrowth quality. Using a combination of AFM, XPS, and TEM, surface preparation steps were optimized to ensure high-quality regrowth by MBE.