Search

The background noise between 1 and 1.8 microns in ground-based instruments is dominated by atmospheric emission from hydroxyl molecules. We have built and commissioned a new instrument, GNOSIS, which suppresses 103 OH doublets between 1.47 - 1.7 microns by a factor of ~1000 with a resolving power of ~10,000. We present the first results from the commissioning of GNOSIS using the IRIS2 spectrograph at the AAT. The combined throughput of the GNOSIS fore-optics, grating unit and relay optics is ~36 per cent, but this could be improved to ~46 per cent with a more optimal design. We measure strong suppression of the OH lines, confirming that OH suppression with fibre Bragg gratings will be a powerful technology for low resolution spectroscopy. The integrated OH suppressed background between 1.5 and 1.7 microns is reduced by a factor of 9 compared to a control spectrum using the same system without suppression. The potential of low resolution OH suppressed spectroscopy is illustrated with example observations. The GNOSIS background is dominated by detector dark current below 1.67 microns and by thermal emission above 1.67 microns. After subtracting these we detect an unidentified residual interline component of ~ 860 +/ 210 ph/s/m^2/micron/arcsec^2. This component is equally bright in the suppressed and control spectra. We have investigated the possible source of the interline component, but were unable to discriminate between a possible instrumental artifact and intrinsic atmospheric emission. Resolving the source of this emission is crucial for the design of fully optimised OH suppression spectrographs. The next generation OH suppression spectrograph will be focussed on resolving the source of the interline component, taking advantage of better optimisation for a FBG feed. We quantify the necessary improvements for an optimal OH suppressing fibre spectrograph design., Comment: Accepted for publication in MNRAS. 15 pages, 18 figures

A long-standing and profound problem in astronomy is the difficulty in obtaining deep near-infrared observations due to the extreme brightness and variability of the night sky at these wavelengths. A solution to this problem is crucial if we are to obtain the deepest possible observations of the early Universe since redshifted starlight from distant galaxies appears at these wavelengths. The atmospheric emission between 1000 nm and 1800 nm arises almost entirely from a forest of extremely bright, very narrow hydroxyl emission lines that varies on timescales of minutes. The astronomical community has long envisaged the prospect of selectively removing these lines, while retaining high throughput between the lines. Here we demonstrate such a filter for the first time, presenting results from the first on-sky tests. Its use on current 8m telescopes and future 30m telescopes will open up many new research avenues in the years to come., Comment: 15 pages, 5 figures. Nature Communications. Published 06 December 2011