Extended Data Fig. 3: Schematic illustrations of the light paths in conventional spectrophotometry and SFP method.

a, Light paths in conventional UV-VIS-NIR spectrophotometry for non-PL (left) and PL coolers (right). This configuration works for non-PL coolers as all the reflected light (at the wavelength of \(\lambda\)) is correctly collected by the detector. While for PL coolers, both reflected light (at the wavelength of \(\lambda\)) and re-emitted light (at the wavelength of \({\lambda }^{{\prime} }\)) are redirected to the detector. Due to the lack of monochromators in the integrating sphere, the re-emitted light will be misinterpreted as having the same wavelength as the reflected light (that is, \(\lambda\)) by the detector, resulting in misinterpretation of the contribution of PL emissions. b, Light paths in Step 1 of SFP method for non-PL (left) and PL coolers (right). For non-PL coolers, this configuration works similarly to conventional spectrophotometry. While for PL coolers, the added filter will reject re-emitted light (at the wavelength of \({\lambda }^{{\prime} }\)). Only reflected light (at the wavelength of \(\lambda\)) will be collected by the detector, leading to exact spectral reflectance. c, Light paths in Step 2 of SFP method. The background photon numbers are measured (left) as reference, followed by the measurement of PL coolers (right). Either reflected light or re-emitted light will be correctly collected by the detector array at corresponding wavelengths.