Numerical optimization of integrating cavities for diffraction-limited millimeter-wave bolometer arrays
Far-infrared to millimeter-wave bolometers designed to make astronomical observations are typically encased in integrating cavities at the termination of feedhorns or Winston cones. This photometer combination maximizes absorption of radiation, enables the absorber area to be minimized, and controls the directivity of absorption, thereby reducing susceptibility to stray light. In the next decade, arrays of hundreds of silicon nitride micromesh bolometers with planar architectures will be used in ground-based, suborbital, and orbital platforms for astronomy. The optimization of integrating cavity designs is required for achieving the highest possible sensitivity for these arrays. We report numerical simulations of the electromagnetic fields in integrating cavities with an infinite plane-parallel geometry formed by a solid reflecting backshort and the back surface of a feedhorn array block. Performance of this architecture for the bolometer array camera (Bolocam) for cosmology at a frequency of 214 GHz is investigated. We explore the sensitivity of absorption efficiency to absorber impedance and backshort location and the magnitude of leakage from cavities. The simulations are compared with experimental data from a room-temperature scale model and with the performance of Bolocam at a temperature of 300 mK. The main results of the simulations for Bolocam-type cavities are that (1) monochromatic absorptions as high as 95% are achievable with <1% cross talk between neighboring cavities, (2) the optimum absorber impedances are 400 Ω/sq, but with a broad maximum from ~150 to ~700 Ω/sq, and (3) maximum absorption is achieved with absorber diameters ≥1.5λ. Good general agreement between the simulations and the experiments was found.
Additional Information© 2002 Optical Society of America. Received 9 February 2001; revised manuscript received 7 August 2001. We thank Jonas Zmuidzinas for helpful discussions and Sunil Golwala for critical comments on the manuscript. J. Glenn and A. E. Lange acknowledge support from National Science Foundation (NSF) grant AST-9618798. A. Lee was supported by NSF grant AST-9120005.
Published - GLEao02.pdf