Fibre systems for future astronomy: anomalous wavelength–temperature effects

C. L. Poppett 0 J. R. Allington-Smith 0 0 Centre for Advanced Instrumentation, Physics Department, Durham University , South Road, Durham DH1 3LE A B S T R A C T Focal ratio degradation is an important property of optical fibres that determines the design and cost of instruments using fibres. Motivated by the importance of fibres in feeding instruments on Extremely Large Telescopes, the need for cryogenic-cooling to reduce thermal background and the desire for broad-band performance, we have studied the dependency of focal ratio degradation (FRD) on both temperature and wavelength. This shows a small but significant reduction in performance when cooled as expected from previous work. We also find an increase in FRD with wavelength broadly consistent with theory at room temperature but this dependency reverses in sign when the fibres are cooled to 77 K, contrary to existing theory. We parameterize the wavelength dependency by an ad hoc extension to an existing model but it is clear that existing theory does not provide a good description of the operation of fibres in astronomical systems. This unexpected behaviour, which may relate to frozen-in stress from the manufacturing process, will need to be taken into account when designing future fibre systems. 1 I N T R O D U C T I O N Focal ratio degradation, a violation of Etendue conservation equivalent to an increase in entropy and thus a loss of information content, can create problems for astronomical instruments which use optical fibres (Heacox 1986; Clayton 1989; Carrasco & Parry 1994; Parry 2006) such as multi-object and integral-field multiplexed spectroscopy to route light from the telescope focus to the entrance aperture (pseudo-slit) of a wide-field spectrograph. To match the scale of the astronomical images requires fibres of diameters typically between 50 and 500 m, which are thus multimode at the wavelengths of current interest (0.3 < < 2.5 m). The increased speed of the beam exiting the fibres results in decreased throughput and/or reduced spectral resolution; or increased cost if both spectral resolution and throughput are to be maintained (Parry 2006). This penalty is often well compensated by their versatility in transporting light from the telescope focus to remote instruments and reformatting it for multiplexed spectroscopy. The enormous sizes projected for instruments for the next generation of telescopes with apertures between 20 and 40 m (ELTs: Extremely Large Telescopes) and their stringent stability requirements mean that most instruments must be mounted on fixed platforms and not ride on the telescope as it tracks (Allington-Smith 2007). This, and the increasing importance of multiplexed spectroscopy, means that fibres are of increasing importance to modern astronomy. The use of photonic crystal fibres has been realized in niche areas such as interferometry and is being explored in other areas relevant to astronomical spectroscopy (Corbett et al. 2005), including the removal of OH emission lines from the night sky in the near-infrared (Bland-Hawthorn, Englund & Edvell 2005). For spectroscopy they offer the potential of efficient coupling into single-mode fibres which would remove problems such as FRD and modal noise which currently affect high-resolution spectroscopy. However, for the present, we need to optimize the performance obtainable with step-index multimode fibres made from fused silica. The increase in telescope aperture means that increasing numbers of targets are highly cosmologically redshifted so that diagnostic ionic bound-state transitions are observed in the near-infrared. To reduce the thermal background, the instrument must be cooled to 100200 K while the telescope focus stays at 260290 K; so we require the fibres to tolerate a difference of 100200 K along their length (1100 m). A further requirement is for broad-band performance which will allow the same fibres to feed separate wavelengthoptimized spectrographs. Previous studies of the dependency of FRD on temperature (Lee, Haynes & Skeen 2001) concluded that fibres remain flexible at 80 K temperature and, after a brief equilibration period, show only a slight degradation in performance. This is true of fibres mounted without encapsulation. For the more realistic case where fibres are mounted in ferrules and matrices of holes using adhesives, a variety of effects are seen depending on the thermal mismatch between fibre and mount. However, this can be ameliorated by designing the system to eliminate differences in thermal properties. Several authors have studied the wavelength dependence of FRD at room temperature in the visible regime (Carrasco & Parry 1994; Schmoll, Roth & Laux 2003; Oliveira, de Oliveira & dos Santos 2005) and found it to be small or unmeasurable. These results are broadly consistent with the picture of fibres where light is scattered by small defects (microbends) presumably introduced during the drawing process and modelled as a modal diffusion process by Gloge (1972). Any wavelength dependency in this process, depending on the typical size and number density of the defects and, possibly, on their internal properties and contrast with the surrounding material may manifest itself as a dependency of FRD on wavelength. We have examined the dependency of FRD on both wavelength and temperature for the first time. Although we confirm the reality of both dependencies, we find that the effects are not independent, with the sign of the wavelength dependency varying with temperature. 2 D E S C R I P T I O N O F E X P E R I M E N T 2.1 Apparatus The equipment is illustrated in Fig. 1. Incoherent broad-band visible light is injected into the fibre in a filled cone such that the far-field distribution of specific intensity with angle is constant for angles in < < in which is set by the adjustable iris placed in the collimated input beam to determine the focal ratio of the input beam. The light source is a pinhole and diffuser illuminated by a tungsten lamp. Light exiting the fibre is intercepted by a thermoelectrically cooled SBIG ST-7 CCD from which the distribution of specific intensity with angle (i.e. the far-field pattern) is measured. A laser beam defines the optical axis to which all components were aligned. A viewing system allows the end of the fibre to be examined in situ by means of a beam splitter and microscope. The input beam can be set within the range 2.13 < Fin < 42.5 and the output beam is limited to Fout > 1.8. Light entering the fibre is first passed through a bandpass filter. The position of the fibre is adjusted as required for the change in focus caused by the different thicknesses and refractive indices of the different filters used, although this should not affect the angular distribution striking the fibre face. The complete fibre core was Name FMOS-a FMOS-b TEIFU Core Diameter ( m) Cladding Buffer 485 10 illuminated. The fibres used in the experiment were of doped stepin (...truncated)