OPTIMOS is the only E-ELT instrument exploring a large field of view. The fibre/positioner approach provides the advantage of avoiding flexure issues when invoking such a large physical field of view (2m x 2m). Another noticeable advantage is that fibres can be positioned at any available location over the largest field of view of the telescope.
Astrophysical sources have many different apparent sizes on the sky, ranging from unresolved stars or QSOs to ≥ 10 arcsec extended sources, even at high redshift (e.g. galactic haloes, Ly alpha blobs). For unresolved sources or Mono-Objects (MO) the aperture has been optimised to be 0.9 arcsec (see RD13). Other apertures are matching the size of distant galaxies (Medium IFU (MI) = 1.8x2.9 arcsec2) and extended sources (Large IFU (LI) = 7.8x13.5 arcsec2). The sampling of the IFU is settled to 0.3 arcsec for a seeing limited instrument covering the UV to the NIR.
OPTIMOS-EVE is the only E-ELT instrument that includes medium to high spectral resolution (R~5,000 – 30,000) so that many scientific programs (see section 5.1) are feasible that cannot be done by JWST or by other first-generation E-ELT instruments under study. A spectral resolution larger than 5000 is mandatory in the NIR to provide enough spectral regions that are only limited by sky background and not affected by strong OH skylines. The covered spectral ranges are all adjusted to the OPTIMOS Science Cases (see Table 1 in RD13). NIR spectroscopy is strongly affected by OH skylines and many science cases require to work in between the skylines, implying a minimal spectral resolution (Fig. 2). For an emission line with an observed FWHM of 200-300km/s (typical values for z> 6 galaxies, see Kashikawa et al. 2006), the problem is even more complex because the fraction of spectral windows free of skylines decreases rapidly. It may reach ~ 25% at R=5000. Such a resolution is a good compromise because it also significantly helps to remove unresolved skylines when they coincide with the source emission (see RD13).
Figure 2 (Left) Night sky in the visual and the near-infrared; notice that OPTIMOS-EVE does not extend beyond 1.6 micron; (Right) Fraction of sky background limited regions in the J and H band, as a function of spectral resolution.
Sky correction
Sky correction is a crucial issue especially for the detection of faint sources that are in reach of the E-ELT. It is often believed that fibre-fed spectrographs are less efficient to perform robust sky corrections, mostly because of the possible different throughput of individual fibres and the difficulty to sample the sky very close to the object. These can be overcome by the fact that we can model the sky variations by placing (sky and MO) fibres in the overall field of view, by the possibility to calibrate the fibre throughput or to use beam-switching observations (see RD13). With the use of the OPTIMOS-EVE simulator (see RD13) we have tested the above assumptions and verified to which accuracy the sky correction can be performed. We have used a model of the sky variations over the OPTIMOS-EVE FOV (over a diameter of 7 arcmin), together with existing FLAMES data, to investigate the reliability of sky subtraction in a variety of observing scenarios (RD4, see also RD13). From these simulations, it has been shown that the accuracy of the sky removal can easily reach 1% of the sky signal, with a number of sky fibres that depends on the amplitude of the sky fluctuations (see Figure 3 and RD13). A significantly better value (0.3-0.4%) is expected for the Medium IFU (MI), which design includes 4 sky-fibres allocated to each MI (see small red squares in Figure 3).
Figure 3 (Left) Fraction of sky residuals (to the sky signal) as a function of the number of allocated sky fibres (MO mode); in the visible and near-IR, the average sky variation is 10% and 20%, respectively (see RD13); (Right) fraction of the sky residual (see the color-coded ruler, in %) for the MI mode. It assumes that the target (a very distant galaxy) illuminates a fraction of the IFU and that sky is sampled from other IFU pixels as well as from the four sky fibres surrounding the IFU (see RD13). Size is indicated in pixels.
Armed with such results, we have performed simulations of faint (Science Case 5) or very faint (Science Case 3) sources that are the most demanding regarding the sky correction. These show that for a m(AB)=25 galaxy, 10 hrs exposure may help to reach a sky residual that is less than 1% of the source continuum, which is sufficient to recover the absorption lines from the IGM (case 5). For a much fainter source, 40 hrs exposure is enough to recover the Ly alpha emission (EW above 20 A) of several hundreds of Ly Alpha Emitters (LAEs) or a m(AB) of 28 (see Figure 4). This warrants the strategy delineated in RD4 to observe a large number of very distant LAE candidates with the MO mode and to re-observe them with the MI to better study their properties (the 30 MIs are able to detect galaxies down to the 30th magnitude, see RD13 for a complete description).
Figure 4 On the left, two recovered Ly alpha lines after 40hrs of observation using the MO mode; the input line (red curve) has an energy of 10-19 erg/s/cm2 for a m(AB)=28 galaxy at z=8.8; S/N are 8 and 4 respectively. Two sets of skylines are visible at 1184-1187nm and 1193-1195nm. On the right, same conditions but after a full correction of the sky in the MI, using surrounding sky fibres and also sky sampled by surrounding pixels of the 1.8x2.9 arcsec2 IFU (GLAO, FWHM=0.45 arcsec); under these conditions the S/N reaches 16; strong sky lines on either sides of the spectra demonstrate the need to work at moderate spectral resolution (R=5000).
Multiplex
The effective multiplex of OPTIMOS-EVE depends on the science observations and can reach a maximal value of 240 for programs not very demanding in sky correction (bright targets). For the faintest targets we have tested several configurations for which from one third to one half of the fibres have to be allocated to the sky, either by using a staring or a beam switch mode (see RD4). These numbers are well matched to the expected LAEs density at very high redshifts (see RD2, Science Case 3). For other Science Cases, the OPTIMOS-EVE multiplex is compliant with the Top Level Requirements (see RD3).
Overall efficiency
The throughput of the whole instrument (including detectors) is provided in RD5. It averages to 26% in the R=5000 mode, for the mono-object fibres, including an estimated 67% throughput of the fibre system. On the other hand the comparison of the overall efficiencies of different instrument concepts depends on each science case goal. For the most demanding one (very distant and faint galaxies, Science Case 3), the detection of the Lyman alpha line depends on 7 different factors, including (1) the spectrograph efficiency, (2) the sky correction efficiency, (3) the aperture losses, (4) flexure effects if the instrument is mounted to the telescope, (5) the spectral resolution that ensures sufficiently large spectral regions not affected by OH lines, (6) the effective multiplex of the instrument and its consistency with the actual number density of targets and (7) the overall field of view that can be attained in one observational set-up. A fibre-fed spectrograph is likely less efficient with respect to item (1) and, depending on the actual number density of very high-z sources, for item (6). For the factor (2) OPTIMO-EVE is probably less efficient in its MO mode, while the MI mode may allow a better sky correction for very compact objects (sky is sampled in all directions surrounding the object). Because astrophysical sources do not have a rectangular shape oriented along the numerous slits of a multiple object spectrograph, OPTIMOS-EVE takes the advantage regarding item (3) and is the best option for items (4), (5) and (7).
OPTIMOS-EVE provides a number of different observing modes that allow the same spectrograph to sample the telescope plane in different ways. This concept allows for the best use to be made of the prevailing conditions, and enables the implementation of a wide range of science goals with one single instrument. This modularity also ensures that OPTIMOS-EVE will be well matched to the development path of the telescope, and can be adapted to new scientific topics that could be prioritized after the completion of the E-ELT.
The OPTical Multi Object Spectrograph 'Extreme Visual Explorer' for the European - Extremely Large Telescope - Consortium partners: GEPI, NOVA, RAL, NBI, INAF, AIP, LSW, IAG, LNA, ON
