An improved, self-contained gas injection system for the divertor material evaluation system (DiMES) on DIII-D has been employed for in situ study of chemical erosion in the tokamak divertor environment. To minimize perturbation to local plasma, the Mark II porous plug injector (PPI) releases methane through a porous graphite surface at the outer strike point at a rate precisely controlled by a micro-orifice flow restrictor to be approximately equal as that predicted for intrinsic chemical sputtering. Effective photon efficiencies resulting from CH4 are found to be 58 ± 12 in an attached divertor (ne ∼ 1.5 × 1013/cm3, Te ∼ 25 eV, Tsurf ∼ 450 K), and 94 ± 20 in a semi-detached cold divertor (ne ∼ 6.0 × 1013/cm3, Te ∼ 2-3 eV, Tsurf ∼ 350 K). These values are significantly more than previous measurements in similar plasma conditions, indicating the importance of the injection rate and local re-erosion for the integrity of this analysis. The contribution of chemical versus physical sputtering to the source of C+ at the target is assessed through simultaneous measurement of CII line, and CD plus CH-band emissions during release of CH4 from the PPI, then compared with that seen in intrinsic sputtering.