Pushing ELT’s sensitivity through PCA background subtraction

The thermal infrared is an important wavelength range for a wide variety of astronomical observations. At these wavelengths, the large ground-based telescopes such as the ELTs reach higher angular resolution than space-based telescopes. However, their sensitivity is limited by the high thermal background due to both photon noise and imperfect removal of background structures from the sky and warm telescope optics. Developing more effective methods for the removal of spatially and temporally variable background structure is paramount for unlocking the full potential of existing and future large and extremely large ground-based telescopes operating at thermal-infrared wavelength. As recent studies have shown, the Principal-Component-Analysis (PCA) method can significantly improve the background subtraction compared to the more common method of subtracting the mean image from dedicated background exposure. We developed and optimized background subtraction routine based on PCA. We used imaging data taken in nulling-interferometric mode from the Large Binocular Telescope Interferometer (LBTI) in the N band (11 μm). The LBTI connects the two aperture of the LBT to form the largest single-mount telescope in the word. This make the LBTI the best place to pioneer data reduction methods and procedures for future ELTs. We present a comparison of classical (mean) and PCA background subtraction for both aperture photometry (mean retrieval and RMS) and high contrast imaging (contrast curves). The PCA background subtraction allows for improvement factors of two to three for both observing techniques. This will allow for improving the sensitivity of suitable, existing data by the same factors. More importantly, it will reduce by a factor five to ten the extremely valuable integration time required to detect faint sources with future ELTs