Characterizing the next instrument for Gemini
The Gemini Observatory is considering the next instrument to include in their long-range plan. The US members of the Gemini Science and Technology Advisory Committee (GSTAC) hereby solicit your input. The US members of the GSTAC are Henry Roe (Lowell, hroe@lowell.edu), Tom Matheson (NOAO, matheson@noao.edu), Alice Shapley (UCLA, aes@astro.ucla.edu), and Nathan Smith (Arizona, nathans@as.arizona.edu).
Large-scale surveys of the US community have shown that the greatest demand is for general purpose (so-called workhorse) spectrographs. These surveys include the ALTAIR report, the CURRENTS survey, and the Ground-Based OIR System Roadmap Committee report. One of the highly desired instruments from these surveys, the Gemini high-resolution optical spectrograph (GHOS), is currently in development. Another capability with high demand is a spectrograph with moderate resolution and wide wavelength coverage. We seek your concise thoughts about the scientific problems that can be addressed by such an instrument and implied desired instrument capabilities
This is not a survey where the greatest number of votes will decide the design of the instrument. Rather, it is a method to obtain scientific guidance for a white paper that the US GSTAC members will present at the next GSTAC meeting. We intend to keep the community engaged as the white paper develops.
As a strawman concept for discussion, consider a high-throughput, single-object instrument that simultaneously covers as much of the near-UV to near-IR range as possible at moderate spectral resolution (R ~ 5000).
There is a diverse set of science cases that could be addressed with such an instrument. Examples include:
- Targets of Opportunity – Gamma-ray bursts occur over a wide range of redshifts so the optimal spectral coverage is not known a priori. Current and future time-domain surveys will produce large numbers of transients at unknown redshifts and with unknown characteristics, so wide-wavelength coverage is essential.
- Late-type stars and brown dwarfs – Simultaneous overage from the red portion of the visible spectrum through the IR is important for characterizing these objects.
- Metal-poor stars – Even moderate resolution spectra can yield fundamental stellar atmosphere parameters. For carbon-enhanced metal-poor stars, the addition of the IR provides access to molecular CO bands and thus oxygen abundances.
- High-redshift galaxies – Both high-throughput in the blue and extensive IR coverage would enable thorough emission-line analysis of these objects. High-throughput in the IR would enable detection of primordial galaxies.
Naturally, various technical and financial considerations may constrain the design and hence capabilities of such a spectrograph. Obvious technical tradeoffs include:
-
Reduce wavelength coverage
- Blue cutoff at 400 nanometers
- Red cutoff in H band (~ 1800 nanometers)
- 360 to 1100 nanometers (OUV optimized, silicon detector only)
- 800 to 2500 nanometers (near-IR optimized)
-
Reduce spectral resolution
- R~3500 beyond 1000 nanometers
- R~3000, or lower below 1000 nanometers
-
Fiber-fed, bench-mounted spectrograph
- Wholly fiber-fed
- Hybrid with some wavelength coverage diverted through fiber
The silver coatings used on the Gemini mirrors dramatically reduce transmission below 400 nanometers. If UV-to-blue wavelength coverage is essential for this instrument, then a future change to a more UV-reflective coating such as aluminum may be considered.
We would like to hear your views of this instrument. Comments on this page are now closed, but you may still email comments to us-gstac@noao.edu. A brief (100 – 300 words) description of the science you would like to do, along with specifications that you would need in terms of wavelength range and resolution. A description of how the various compromises might affect your science would also be useful.
Input from the entire international Gemini user community is welcome and would be greatly appreciated as we work together towards a strong Gemini.
On behalf of the entire Gemini community, thank you for your time and assistance.
Last updated or reviewed August 20, 2012.
NOAO is the national center for ground-based nighttime astronomy in the United States and is operated by the Association of Universities for Research in Astronomy (AURA). under cooperative agreement with the National Science Foundation. If you would like information about solar astronomy, visit the National Solar Observatory. If you would like information about radio astronomy, visit the National Radio Astronomy Observatory.
Comments (13)
GREAT idea! Consider also Time-critical SolSys observations
The correlation of optical and NIR features helps to better constrain vairous species on solid and gaseous solar-system bodies. Numerous examples exist where a full range of NIR and optical wavelength coverage will determine the actual chemical constituent rather than merely limit the range of species. Surfaces can vary compositionally, and so too may the spectral features as regions may rotate out of view or outbursts of activity may diminish or grow. More problems arrise when most bodies are too faint or the signals are too weak to spectroscopically detect these features, whether broad or narrow, especially on small bodies (asteroids, comets, and SDOs/TNOs/Centaurs), and/or very distant objects (outer solar system dwarf planets or Neptune and Uranus). This would be a big step forward, coupling the spectrograph with a large-aperture 8-meter telescope, where the instrumental resolution is at a more moderate resolution, and hence higher signal-to-noise per element, appropriate to maximizing the yeild of the observations of such bodies.
An additional thought would be to make this instrument capable of second time-scale, or sub-second, sampling, for time-critical events, such as occultations, spacecraft encounter events (think Deep Impact), and satellite mutual events and transits.
Very Best Regards!
-James Bauer
Consider slightly higher resolution for Transient followup?
Dear Gemini instrument team,
With the advent of Pan-Starrs, PTF, and soon the SST, the need for good spectroscopic followup will be nearly overwhelming. It is therefore very refreshing to see this new spectrograph come available. The European community has access to the X-shooter and UVES spectrographs on the VLT, which combine wide wavelength coverage with high resolution. This has served them very well in followup observations for supernovae, and gamma-ray bursts. Our community lacks this sort of higher resolution, and so I am hopeful that in the proposed instrument the resolving power of at least R ~ 5000 be preserved, and perhaps even increased some, to enable resolution of absorption lines in the ISM and IGM for newly discovered transients, and studies of the evolution of lines within supernova remnants in nearby galaxies. Better still would be to have a resolution comparable to the Keck/ESI spectrograph with the range of 10,000 < R < 15,000. Perhaps if the spectrograph could be configured dynamically between R ~5000 and R ~10,000 it would enable a lot more science and more detailed characterization of transients and ISM features.
Best wishes,
Bryan Penprase
Spectrograph
One of the great successes at Gemini has been the effective use of queue-scheduled time to obtain rapid and effective response to transient events. Coupling that style of operation to an exceptional spectrograph would be a powerful scientific system. It would be great for supernova science, where the IR is the frontier.
I can think of two points that might affect the chosen design.
One is that the objects of interest are sometimes very faint, so that exquisite sky subtraction is essential. If someone has done this with fibers, I would like to see the evidence.
The other is that the energy distributions of the objects will determine whether simultaneous observations at all wavelengths are going to be effective. There's less point in having super-wide wavelength coverage if the reality is that your exposure time is set by the poorest-performing part of the system. You'll just pick a reasonable exposure time and live with the noise in the UV or in the IR, as the case may be.
Bob Kirshner
missing the boat...
The comment that this survey "isn't meant to be a vote" is interesting. Big telescopes aren't just for seeing faint objects -- they're unique for collecting the photons needed to do sensitive polarimetry! Sensitive polarimetry is needed to characterize, and even visualize scattering regions. Polarized line profiles, optical pumping, and the quantum mechanics of Raman scattering are now being used as nimble tools for everything from double-differential photometry of extrasolar planets, resolving the 3-d magnetic structure of stellar atmospheres, to describing the isotropy and structure of the radiation environment in ubiquitous scattering regions. We're far from days when Thomson scattering and unpolarized line scattering were our primary tools for describing just a few Halpha polarized line profiles that barely break the 1% sensitivity level. Big telescope (8m or more) are begging for spectropolarimetric capability -- and not as an add-on afterthought. We need to reach 0.1% full Stokes sensitivity in resolved line spectra. Build a spectropolarimeter with even R=30000, that beats the systematics of EspaDons and has full Mueller calibration and stability characteristics on an 8m telescope and you'll have built not just an instrument with "convenient" features -- but a unique instrument with ground-breaking scientific capabilities that we haven't yet begun to explore.
Include spectropolarimetric capability
Whatever the final configuration adopted I would urge that the spectrograph include spectropolarimetric capability. This would not compromise the performance of the spectrograph. A lot of our understanding of active galactic nuclei and the nature of supernovae explosions has come from spectropolarimetry.
Martin Gaskell
polarimetric capability would be highly desireable
I fully concur with Kuhn and Gaskell. Having a high accuracy polarimetric capability would be a major asset for a variety of astrophysical problems, from the Solar System to high z. It would also make the instrument very competitive vis-à-vis XShooter, which lacks such capability.
This probably means leaving the instrument at the straight-through port.
Antonio Mario Magalhaes
(1) spectropolarimetry; (2) small bodies
I'd like to agree strongly with those already arguing for spectropolarimetry and for Solar System small body capabilitity.
(1) SPECTROPOLARIMETRY: It can be argued, as Omer Blaes has done, for example, that a handful of papers on the spectropolarimetry of quasars and active galactic nuclei have done more to advance that science than thousands of papers on, e.g., emission lines - (a) the 'Unified Scheme' of flattened obscuring regions in AGNs was established via highly polarized broad emission lines; (b) the solution to basic Broad Absorption Line (BAL) quasar questions (saturated? Yes; in all quasars? Yes) from polarimetry of the BAL troughs; (c) the rotation of the broad emission line region, from systematic Position Angle changes with velocity in H-alpha; (d) the geometry of the X-ray 'warm absorbers' (WA) from optical continuum polarization of WAs vs non-WAs.
But, polarimetry is photon hungry. The polarizations found are usually a few percent, equivalent to using a 1-meter telescope in unpolarized light. The above results came mainly from 4-meter class telescopes. (exeption: BALs were done most spectacularly on Keck.) Only the brightest lines in the brightest objects have been accessible to date. The change in established quasar structure with black hole mass, accretion rate, or even with simply broad emission line width or emission line ionization, could not be investigated. Making spectropolarimetry available on Gemini would be a major step forward. If built in to the spectrograph design the additional complexity is minimal.
(2) SMALL SOLAR SYSTEM BODIES: Optical-NIR (0.5-2,.2micron) spectra are the workhorse tool for characterizing asteroid surface composition. Despite some 9000 near-Earth asteroids [NEAs] now being known, the number with spectral types from spectra is only ~1000, the bulk coming from the 3.5m IRTF. Asteroids need only low resolution (R~50-100). However, rotation periods of minutes to hours suggest that getting a spectrum in a fraction of a period (say ~1/10th) could enable differential surface composition measurements. In situ measurements of a handful of asteroids suggests non-uniform surface composition. Hence jsing a larger telescope, such as Gemini, would be a direct benefit. [Co-adding phased spectra are of course possible using smaller telesecopes, but 'wobble' of some significant fraction of the asteroids means that the summed spectra are of different regions and not strictly summable by phase.]
Queue-scheduling and modest time-resolution (~1 minute) would be needed for theses studies.
Many of the newly found NEAs are small and faint. Employing an effective OH suppression system in the NIR would be a major enhancement for these objects.
Good idea - keep it simple
I think this is going to be a widely used instrument - as evidenced by the success of X-Shooter on VLT and GNIRS a few years ago when it was the only instrument in the world (on an 8m class telescope) giving simultaneous J+H+K coverage. I would argue against adding complications, such as a high resolution mode or spectropolarimetry.
For high redshift galaxies, emission lines are not the only (and perhaps not the most exciting) thing to go after. Multi-object spectrographs such as MOSFIRE are well equipped to do emission line surveys of many objects at once, but they won't contribute much to continuum studies. Galaxies at z>1.5 bright enough for continuum / absorption line studies will rarely number more than 1 per MOSFIRE field, which means the new spectrograph will be the instrument of choice for studies of stellar kinematics, detecting Balmer breaks at z>2, Lyman breaks at z>7, etc. (For stellar kinematics having a wide wavelength coverage also directly benefits the measurement)
On the technical side, the key will be to be sky-background limited in between the sky lines. This has not been achieved by any near-IR spectrograph to date, including X-Shooter which turns out to have an ill-understood problem with non-Gaussian detector noise. MOSFIRE may actually come close, although it's too early to say for sure. This has implications for the resolution: too high resolution will imply too long (> a few minutes) integration times to beat the detector noise. On the other hand, too low resolution and the wings of the sky lines will dominate the noise everywhere in the J and H bands. A resolution of 4000-5000 is probably ideal - IF the detector noise is both low and well behaved, there are no thermal leaks, etc.
As far as descoping goes, I think the instrument as proposed is just about right. If forced to identify the core capability, I would probably say it is crucial to retain optimal, sky-limited performance at 0.8-1.6 microns at a resolution of 4000-5000.
Pieter van Dokkum
High redshift QSOs
Metal lines detected in absorption towards high z (5>z>7) QSOs offer an independent measure of the state of the IGM, its metallicity and the integrated star formation history of the Universe at the end of the epoch of Reionization. Many (50-100) new bright QSOs are expected to be discovered with future surveys including Pan-Starrs and Skymapper. The strongest absorption features include CIV, CII, SiV, SII and NV, the rest wavelength of these ions, including Lyman-alpha, range from 122 to 155 nm. Thus, for a redshift range of 5 to 7, wavelength coverage of 720 - 1249 nm would be required.
I agree with Pieter van Dokkum on a number of points.
1. Keep it simple
2. Best descoping option is a sky-limited performance at 0.8-1.6 microns with R~5000.
X-shooter has been a highly sought-after instrument, however it suffers from a loss of sensitivity around the Optical/Near-IR dichroic at 1 micron, right at the 'sweet spot' for the above science. Magellan's FIRE currently leads the way on solving this problem.
Emma Ryan-Weber
Spectroscopy of Icy Dwarf Planets
A moderate resolution (R ~ 5000), optical-near ir (0.4 - 2.2 micron) spectrograph on Gemini would be big boost to my research program. In particular, I would use it to detect new ice bands in spectra of icy dwarf planets as well as observe currently known ice bands to better characterize the surfaces of these objects. Moderate spectral resolution, wide spectral coverage, and a large aperture telescope are essential for my research program.
high-z galaxies and quasars
I totally agree with Pieter van Dokkum and Emma Ryan-Weber that this instrument is great for high-z galaxies and 
quasars studies. Here I just want to mention that K-band (up to 2.4 microns) coverage is crucial for these studies.
1. high-z galaxies
K band will cover Halpha and [NII] lines for galaxies at 2.0<z<2.6 and the Hbeta and [OIII] doublet for galaxies at 3.1<z<3.8. These optical diagnostic lines are crucial to determine the systemic redshift and to understand the physical properties in these galaxies, including metallicities, star formation rates, AGN activities etc.
2.high-z quasars.
K band will cover the MgII emission line for quasar at z>6. MgII line is an important line to determine the supermassive black hole mass. Compared to the CIV line, it is not affected by the outflow of black holes, giving us a better constraint on the mass of these supermassive black hole at the early universe.
I play the GMOS and GNIRS ITC a little bit. I assume that we need to achieve S/N of ten per resolution element on the rest-frame UV (observed-frame optical) continuum to study the ISM absorption lines and 5 sigma detection on the rest-frame optical (observed-frame near-IR) nebular line (e.g. Hbeta), and I find the exposure time is comparable for these two instruments for a given high-z galaxy. At least for high-z galaxy studies, the concern from Robert Kirshner won't be an issue.
Best Regards,
Fuyan Bian
WISE followup
The Wide-field Infrared Survey Explorer has found numerous brown dwarf and ultra-luminous galaxy candidates. R=500-1000 spectroscopy gives useful followup on both classes of object. For the ULIRGs the velocity dispersion broadens the lines, and for BDs the spectral features are molecular bands with intrinsic widths.
Covering a wide-band in one shot is very convenient and the source density is low enough that multi-object spectroscopy is not too important.
Infrared guiding is useful, with a choice of bands. Cool BDs are not bright in the K band.
Giant planet aerosols
Particularly on Jupiter, it has been a challenge to determine the composition of aerosols responsible for the different shades of red/brown coloration. There is a lack of consensus regarding the altitude levels of these chromophores --- are they in the clouds, or the featureless overlying hazes?
Aerosol altitudes are typically retrieved from infrared observations, while color is obviously measured in the visible. It would be very valuable to obtain a complete spectrum of both regions at the same time; the 0.89-micron methane feature could be used for altitude discrimination. Problems with prior studies include low spectral resolution (like with filtered imaging) and asynchronous visible and IR observations (not good because the clouds can evolve between observations).
Spatial resolution
Global spectra have already been retrieved. But properties like aerosol color, cloud heights, thicknesses, etc. are highly spatially variable, so we need spatially resolved spectra to make new advances. It is not clear from this page whether the spectrograph will offer this capability, perhaps using a long (~50" for Jupiter) slit.
Bright object accommodation
At visible wavelengths, Jupiter is very bright. A concern with high-throughput instruments is that the detector must be able to observe bright targets without saturating (e.g., with short integration times).
-- Mike Wong