Fluid atmospheric dispersion corrector
The effects of atmospheric dispersion on astronomical instruments are well known.1, 2 Extremely large telescopes, such as the Giant Magellan, the Thirty Meter Telescope and the European Extremely Large Telescope,3, 4require correction of this phenomenon, which results from the wavelength dependence of the refractive index of air. It causes the images of stars observed at typical elevations to appear extended by the order of a few arcseconds, as different wavelengths travel slightly different paths through the atmosphere. The apparent change in size is comparable with that of an image affected by atmospheric turbulence, and results in a lower image quality than that achievable with adaptive optics.
Traditional atmospheric dispersion correctors (ADCs) are made of two or more pairs of glass prisms. They are very bulky and expensive to build and need to be motorized to achieve correction at different altitudes. They require flint glasses, which inevitably introduce absorption losses, especially in the UV, and these losses scale with the size of the telescope. For multi-object spectroscopy, only small and discrete patches of the focal surface need to be corrected for atmospheric dispersion, so we can consider multiple small ADCs built into deployable units. However, the complexity of miniaturizing, motorizing, and controlling large numbers of individual correctors has encouraged us to look for passive solutions. To simplify the structure, Sorokin and coworkers proposed a fluid atmospheric dispersion corrector (FADC), which used a pair of fluids as the dispersion corrector.5 However, they did not present its design details or its performance.
An FADC uses a pair of immiscible liquids in a small glass container, placed very close to the telescope focal plane: see Figure 1. The liquids form two fluid prisms. Since it is placed normal to the telescope optical axis, the interface of the two fluids is kept horizontal by gravity. The apex of the fluid prisms automatically adapts to the telescope zenith angle. Potentially, independent deployable units, such as the Australian Astronomical Observatory's ‘Starbugs,’6 could each carry its own FADC, providing correction for each target object.
The fluids used in the FADC must have specific optical and physical properties. We select the refractive index of the fluids such that when the FADC is placed before the focus of the telescope, one wavelength (in this case 587.6nm) can be transmitted without refraction: see Figure 2(a). The two liquids have different dispersion properties and hence the fluid prisms can function dynamically when the telescope is rotated to observe different stars at various altitudes. In general, the dispersive properties of either organic or inorganic chemicals are poorly documented. We have developed a database of optical properties, including refractive indices and dispersion properties, for organic and inorganic chemicals.7 From this database, we identified two chemicals (anisole and 1-thioglycerol), whose properties we verified, and used them in an FADC to correct dispersion for a Cassegrain telescope. Their refractive indices are 1.517 and 1.524, respectively, at 587.6nm (the absorption line of helium and a common wavelength benchmark for astronomical optics). The chemicals' Abbe numbers—a measure of the materials' dispersion in relation to the refractive index—are 30.38 and 46.39, respectively, with the higher value indicating lower dispersion.
We performed an FADC on-sky demonstration using the 3.9m Anglo-Australian Telescope (AAT) at Siding Spring observatory.8 The colored CCD camera used is an IDS UI-2230SE-C, which has a 1024×768 unobstructed pixel array, allowing us to place the FADC close to the CCD chip. We mounted the FADC cell and the camera on the AAT's Cassegrain focus: see Figure 2(b). We carried out the on-sky demonstration during the telescope's twilight time, and therefore did not track one particular star at different zenith angles. Instead, we observed six different stars at different angles and produced Lucky Images, a form of speckle image generated using a high-speed camera with short exposure to reduce atmospheric effects and increase image resolution. We took 2000 frames of video (66 seconds at 30 frames per second), both with and without the FADC in place for each star. To analyze the data we took the best 10 frames of the video through the processing system AviStack. From these high-resolution Lucky Images we extracted the blue, green and red pixels and took the centroid of the star for each color. We calculated the dispersion length in arcsecs using a pixel scale of 0.031arcsec/pixel, derived from the AAT's f/8 plate scale. Figure 3 shows cropped Lucky Imaged stars at zenith angles of 7°, 33°, and 52°, without and with the FADC. Each image size is 200×200 pixels. They demonstrate that the FADC works very well to correct dispersion efficiently.
In summary, we have shown that the FADC can correct atmospheric dispersion passively with no moving parts. This concept shows potential as a good solution for next-generation extremely large telescopes. Although the FADC on-sky demonstration displayed corrective ability only for the visible wavelength range, this was limited by the spectral response of the colored CCD camera. We are now working to test the FADC across a wider wavelength range. We are seeking more fluids for application in different spectral ranges, and designing FADCs of varying size to meet the requirements of different systems.
Jessica Zheng is an instrument scientist whose main research interest is new technology development.