Abstract
The ability to accurately measure polarization changes at the part-per-million (ppm) level enables a wide variety of astronomical and defense-related programs, as light reflected by any medium tends to impart polarization. For instance, multi-wavelength, optical studies of Jupiter-sized planets around other stars may probe their atmospheres and eventually surfaces, even though the measurements are contaminated by 100,000 times as many stellar photons as planetary ones. Additionally, our pilot study of resident space objects suggests that potentially game-changing results may occur if polarization is calibrated to a high degree. We have developed the POLISH2 aperture-integrated, ground-based optical polarimeter at the Lick Observatory Shane 3-m telescope using two photoelastic modulators instead of conventional, rotating waveplates. This instrument measures AC-coupled intensity variations from 10-100 kHz due to the resonant modulators. Measurements of both strongly (0.1%-1%, or 1,000-10,000 ppm) and weakly polarized calibrator stars (0.0001%-0.001%, or 1-10 ppm) are necessary to calibrate for instrumental orientation and telescope-induced polarization, respectively. Since stellar polarization is due to scattering of starlight by intervening dust particles, stellar distance is linearly proportional to stellar polarization. While strongly polarized stars abound in astronomical stellar catalogs, such catalogs are incomplete below 0.01%-0.1% polarization. This is because 1) conventional polarimeters are limited to ~0.01% accuracy due to linear to circular polarization conversion in waveplates, and 2) unpolarized stars, necessarily bright, tend to saturate on conventional waveplate / CCD instruments. POLISH2, being designed to observe Venus and the Crab Pulsar (apparent visual magnitude ranges between -4 and 16, or a factor of 108 in dynamic range) with photon-limited sensitivity, is uniquely suited to measure stellar polarization to the ppm level. We have used the Lick Observatory 1-m telescope to perform a survey of bright stars to identify nearly unpolarized calibrator stars.
Laboratory calibration of POLISH2 is quite difficult, as the linear (LP) and circular polarization (CP) of even a bare, incandescent bulb with no reimaging optics is found to be LP = 0.67520% +/- 0.00054% and CP = -0.08643 +/- 0.00028%. Propagating the bulb through two collimating and focusing lenses, we measure LP = 3.0319 +/- 0.0032% and CP = -0.36029 +/- 0.00033%. Thus, propagation of light through lenses introduces both linear and circular polarization due to static stress frozen into the lens via the annealing process. Even linear film polarizers are found to be non-ideal, as they appear to introduce circular polarization with 2-3% amplitude depending on the rotational orientation of the linear polarizer. We find that integrating spheres reduce bulb polarization to LP = 0.02%-0.05% (200-500 ppm) and CP = 0.001%-0.002% (10-20 ppm), with larger spheres producing weaker polarization. Finally, cavity blackbodies are found to present the weakest linear polarization, with LP ~ 0.01% and |CP| ~ 0.01% (100 ppm) for two separate blackbodies. These measurements show that the best sources for ultra-high accuracy polarimetric calibration are stars themselves.
Calibration of an Ultra-High Accuracy Polarimeter at the Part-Per-Million Level
The ability to accurately measure polarization changes at the part-per-million (ppm) level enables a wide variety of astronomical and defense-related programs, as light reflected by any medium tends to impart polarization. For instance, multi-wavelength, optical studies of Jupiter-sized planets around other stars may probe their atmospheres and eventually surfaces, even though the measurements are contaminated by 100,000 times as many stellar photons as planetary ones. Additionally, our pilot study of resident space objects suggests that potentially game-changing results may occur if polarization is calibrated to a high degree. We have developed the POLISH2 aperture-integrated, ground-based optical polarimeter at the Lick Observatory Shane 3-m telescope using two photoelastic modulators instead of conventional, rotating waveplates. This instrument measures AC-coupled intensity variations from 10-100 kHz due to the resonant modulators. Measurements of both strongly (0.1%-1%, or 1,000-10,000 ppm) and weakly polarized calibrator stars (0.0001%-0.001%, or 1-10 ppm) are necessary to calibrate for instrumental orientation and telescope-induced polarization, respectively. Since stellar polarization is due to scattering of starlight by intervening dust particles, stellar distance is linearly proportional to stellar polarization. While strongly polarized stars abound in astronomical stellar catalogs, such catalogs are incomplete below 0.01%-0.1% polarization. This is because 1) conventional polarimeters are limited to ~0.01% accuracy due to linear to circular polarization conversion in waveplates, and 2) unpolarized stars, necessarily bright, tend to saturate on conventional waveplate / CCD instruments. POLISH2, being designed to observe Venus and the Crab Pulsar (apparent visual magnitude ranges between -4 and 16, or a factor of 108 in dynamic range) with photon-limited sensitivity, is uniquely suited to measure stellar polarization to the ppm level. We have used the Lick Observatory 1-m telescope to perform a survey of bright stars to identify nearly unpolarized calibrator stars.
Laboratory calibration of POLISH2 is quite difficult, as the linear (LP) and circular polarization (CP) of even a bare, incandescent bulb with no reimaging optics is found to be LP = 0.67520% +/- 0.00054% and CP = -0.08643 +/- 0.00028%. Propagating the bulb through two collimating and focusing lenses, we measure LP = 3.0319 +/- 0.0032% and CP = -0.36029 +/- 0.00033%. Thus, propagation of light through lenses introduces both linear and circular polarization due to static stress frozen into the lens via the annealing process. Even linear film polarizers are found to be non-ideal, as they appear to introduce circular polarization with 2-3% amplitude depending on the rotational orientation of the linear polarizer. We find that integrating spheres reduce bulb polarization to LP = 0.02%-0.05% (200-500 ppm) and CP = 0.001%-0.002% (10-20 ppm), with larger spheres producing weaker polarization. Finally, cavity blackbodies are found to present the weakest linear polarization, with LP ~ 0.01% and |CP| ~ 0.01% (100 ppm) for two separate blackbodies. These measurements show that the best sources for ultra-high accuracy polarimetric calibration are stars themselves.