Optical attenuators extend dynamic range but alter angular response of planar ultraviolet-C dosimeters

A challenge for sensors used in ultraviolet-C (UV-C) decontamination protocols of N95 respirators is validation that the entire N95 surface receives the minimum acceptable dose. Photochromic indicators (PCIs) can accurately measure UV-C dose on nonplanar surfaces, but often saturate below doses required to decontaminate porous, multilayered textiles such as N95s. Here, we investigate the use of optical attenuators to extend PCI dynamic range while maintaining a near-ideal angular response - critical for accurate measurements when UV-C is uncollimated. Through an analytical model, we show that tuning attenuator refractive index, attenuation coefficient, and thickness can extend dynamic range, but compromises ideal angular response unless the attenuator is an ideal diffuser. To demonstrate this tradeoff empirically, we pair PCIs with model specular (floated borosilicate) and diffuse (polytetrafluoroethylene) attenuators, characterize the angular response, and evaluate on-N95 UV-C dose measurement accuracy of each PCI-attenuator stack in a UV-C decontamination system. While both borosilicate and polytetrafluoroethylene increase PCI dynamic range >4$\times$, both attenuators introduce angle-dependent transmittance, which causes location-dependent underestimation of UV-C dose. The PCI-borosilicate and PCI-polytetrafluoroethylene stacks underreport true on-N95 dose by 1) 14.7% and 3.6%, respectively, on a surface near-normal to the array of source lamps, and 2) 40.8% and 19.8%, respectively, on a steeply sloped location. Overall, we demonstrate that while planar optical attenuators can increase PCI dynamic range, verification of near-ideal angular response is critical for accurate UV-C dose measurement.