MSE Point Design V2 - cfht.hawaii.edu · ELT Extra Large Telescope ... LSST Large Synoptic Survey Telescope M1 Primary Mirror MSE Mauna Kea Explorer (formerly ngCFHT) PFS Subaru Prime

MSE Point Design Characteristics

02.03.01.001.DSN Version: B

Status: Released

2014-09-22

Prepared By: Name(s) and Signature(s) Organization Date David Crampton and Kei Szeto NRC-Herzberg

Approved By: Name and Signature Organization Date Rick Murowinski

MSE Project Office

17 September 2014

Maunakea Spectroscopic Explorer

MSE MSE Point Design Characteristics

Doc # : 02.03.01.001.DSN Date: 2014-09-22 Status: Rekeased Page: 2 of 13

Change Record

Version Date Affected

Section(s) Reason/Initiation/Remarks

A 2014-09-17 All First Release B 2014-09-22 Corrected references to WFMOS



Table of Contents

1 INTRODUCTION ........................................................................................................ 4 2 RELATED DOCUMENTS AND DRAWINGS .......................................................... 4

2.1 Relevant Documents............................................................................................... 4 2.2 References .............................................................................................................. 4 2.3 Abbreviations and Acronyms ................................................................................. 4

3 EXECUTIVE SUMMARY .......................................................................................... 5 4 GENERAL CONSIDERATIONS ................................................................................ 6

4.1 Constraints and Assumptions ................................................................................. 6 4.2 Future MSE Design Development.......................................................................... 6

5 TELESCOPE DESIGN................................................................................................. 7 5.1 Primary Mirror........................................................................................................ 7

5.1.1 Diameter (point design 10m)............................................................................ 7 5.1.2 f/ratio (point design f/1.8) ................................................................................ 7 5.1.3 Primary mirror concept (point design 10m segmented)................................... 8

5.2 Wide Field Corrector .............................................................................................. 8 5.2.1 Type (point design: refractive)......................................................................... 8 5.2.2 Field of View (point design: 1.5deg2, hexagonal)............................................ 8 5.2.3 ADC (point design: lateral-shift)...................................................................... 9

5.3 Telescope Structure (point design: Keck2 concept) ............................................... 9 5.4 Enclosure (point design: Calotte concept).............................................................. 9

6 INSTRUMENT DESIGN ............................................................................................. 9 6.1 Target pickoff (point design: WFMOS “Cobra” fibre positioner) ....................... 10

6.1.1 Number of targets (point design 3200) .......................................................... 10 6.2 Fibre system (point design: straightforward MOS system).................................. 10

6.2.1 Fibre system design (point design: similar to WFMOS without couplers).... 10 6.2.2 Fibre size (point design: 0.9 arcseconds) ....................................................... 10 6.2.3 Fibre output module (point design TBD)....................................................... 11

6.3 Spectrographs ....................................................................................................... 11 6.3.1 Spectral resolutions (point design R ~ 2000, 6500, 20000) ........................... 11 6.3.2 Wavelength range (point design: 370 – 1300nm) .......................................... 11 6.3.3 Spectrograph Concept (a triple-resolution spectrograph concept is discussed but no point design adopted) ...................................................................................... 11



1 Introduction

This report briefly summarizes the design parameters of the MSE system that emerged from the Feasibility Studies and provides a brief rationale for why they were selected.

2 Related Documents and Drawings 2.1 Relevant Documents

Reference Document title Date Document ID [RD-01] Feasibility Study Report for the Next Generation

CFHT: I. Science

[RD-02] Feasibility Study Report for the Next Generation CFHT: II. Technical

2.2 References

[1] Pazder, J. et al., The FRD and Transmission of the 270-m GRACES optical fiber link and a high numerical aperture fiber for Astronomy. SPIE 2014

[2] McConnachie, A. “Fibre diameter optimization for image quality conditions” Aug, 2014

[3] Ellis, R., et al., 2009, Wide-Field Fibre-Fed Optical Multi-Object Spectrometer WFMOS Study Summary

2.3 Abbreviations and Acronyms

ADC Atmospheric Dispersion Corrector CFHT Canada-France-Hawaii Telescope E-ELT European Extra Large Telescope ELT Extra Large Telescope FOV Field Of View FWHM Full Width at Half Maximum LSST Large Synoptic Survey Telescope M1 Primary Mirror MSE Mauna Kea Explorer (formerly ngCFHT) PFS Subaru Prime Focus Spectrograph PSF Point-Spread Function R Spectral resolution TMT Thirty Meter Telescope WFMOS Wide-Field Fibre-Fed Multi-Object Optical Spectrometer



3 Executive Summary

The initial MSE technical concept was designed to

a) meet the scientific requirements as effectively and efficiently as possible;

b) build upon the recent extensive developments for the ELT projects (especially primary mirror components, control systems, etc.) and modern wide field spectroscopic survey instruments (especially PFS), in order to minimize costs and schedule;

c) re-use the existing CFHT infrastructure (especially piers and support building).

More specifically, concepts developed for the TMT and WFMOS (that has since led to PFS) projects were relied upon to provide the bases for designs, costs and schedules since two of us (DC and KS) have been heavily involved in both projects for the past decade and thus have considerable insight into them. We also note that there is likely to be a heavy overlap in the partnerships for MSE, TMT (and, potentially, PFS).

The science imperatives that emerged from the science teams during the feasibility study phase can be summarized as:

• Must be able to efficiently obtain very large numbers (104 – 106) numbers of spectra with low (R = 2000), moderate (R ~ 6500) and high (R ~ 20,000) for faint (20 < g < 24) science targets over large areas of sky (103 – 104 deg2)

• Spectra should span the blue/optical wavelength region (0.37 – 1.3mu) • Complete optical wavelength coverage should be possible in a single observation at the

lower resolutions • Commission as soon as possible to be complementary with GAIA and planned imaging

surveys and be much more powerful than other spectroscopic facilities

Given the requirement to obtain large numbers of high R spectra simultaneously (and stack these efficiently on a detector), a fibre system that uses a positioner system to feed light from anywhere in the field through fibres to the linear slit(s) of the spectrograph(s) is the most feasible option. That immediately implies that the final f/ratio of the telescope optical system be > f/2.3 to efficiently couple the light into the fibres. To stay within the volume of the current CFHT enclosure and the mass of the current 3.6m, a 10m diameter telescope operated at prime focus is feasible, structurally and optically.

A major driver in the feasibility study was to unabashedly copy successful designs that would achieve the scientific goals rather than embark on significant development. programs. This approach leads to significant reductions in cost, schedule and risk. The selected point design MSE system is a fibre-fed WFMOS-like instrument [3], Keck-like telescope, and TMT-like Calotte enclosure.



Our findings indicate the performance of a telescope with a 10m aperture and a well-designed fibre transport system feeding efficient spectrograph(s) will meet the scientific requirements.

4 General Considerations

4.1 Constraints and Assumptions

For the purpose of the Feasibility Study, we selected the following constraints and assumptions to facilitate the technical work:

1. In consideration of the objectives of the Office of Mauna Kea Management Mauna Kea Science Reserve Master Plan, we have adopted the following redevelopment principles – a. We will not disturb the ground beyond what has already been done. b. We will stay within the current CFHT space envelope. c. We will minimize work at the summit by reusing the telescope and enclosure piers.

2. We maximize the telescope aperture size given the redevelopment principles with due consideration for technical risks.

a. We endeavour to build the biggest telescope that can be supported by the telescope pier and correspondingly an enclosure and the enclosure pier.

3. We minimize technical risks (and as well as programmatic risks, cost and schedule) by leveraging on existing designs of multi-object spectrographs, telescopes and enclosures.

4. We assume a segmented mirror telescope in order to exploit the mirror segment technologies, for production, support and alignment, developed for the ELT projects currently under construction.

a. For a monolithic mirror telescope, we recognize the current technology is limited to producing 8 m aperture having ~1/3 lower observation efficiency than a 10 m segmented primary.

5. We define a point design configuration in order to progress the programmatic development to estimate cost and schedule and therefore completing the feasibility study.

4.2 Future MSE Design Development

It is important to recognize the point design MSE configuration assumed in the feasibility study is a “vehicle” to illustrate feasibility and merely a “stepping stone” to facilitate future development. The point design is not intended for limiting considerations of alternative MSE concepts.



However, it is equally important that any future MSE design development must be approached as a system with interdependent choices between the spectrograph, telescope, enclosure and their supporting piers. In addition, the design space is limited by physical constraints such as mass and space and technical constraints such as maximum size of optics and fibre throughput etc.

5 Telescope Design

One, two and three mirror telescope optical designs were compared in [RD2], Section 3.1 and Appendix C. While two and three mirror concepts are more compact, potentially allowing a larger diameter primary, they are more complex and present other problems. Figure 9 of [RD2] and Figure 29 of Appendix C demonstrate that the throughput losses due to vignetting and extra reflections favour a simple, single mirror concept. A prime focus concept also facilitates direct calibration of the instrument from an on-axis location behind the primary mirror. The existence of several similar, successful, 8-10m telescopes with proven designs and costs is a significant factor.

The design that emerged is a single-mirror design with a wide field refractive corrector at prime focus compatible with a fibre-positioning unit. The design trade is summarized in [RD2], Section 3.1.5.

5.1 Primary Mirror

5.1.1 Diameter (point design 10m)

The requirement to take spectra of thousands of very faint targets (g > 24: [RD1]) implies that the raw light gathering power should be such that all exposures be less than about 4h – the exposure time estimator then demonstrates that diameters of about 8 – 10m are required. A much larger telescope will likely have difficulty fitting within the prescribed volume. Conversely, exposures shouldn’t be so short (< 15min) that the overall efficiency becomes very low due to operational overhead (slewing, acquisition, reconfiguration, etc.) – this would be an argument for a smaller diameter, wider field telescope given the interdependent considerations for mirror aperture and focal ratio, and the maximum of size refractive optics available.

5.1.2 f/ratio (point design f/1.8)

Although the primaries of many modern telescopes (e.g., TMT, E-ELT, LSST) are typically faster, approaching f/1, the final f/ratio of MSE should be > 2.3 in order to have efficient coupling of light into the fibres (Pazder [1]). The design and complexity (and hence throughput) of the wide field corrector is simpler (higher throughput) if the input and output f/ratios are similar. To quote John Pazder (optical designer): “the f/ratio of the primary should be as large as will fit within the enclosure, to make the overall optical design as simple as possible, with the best optical performance”. It is possible to use microlenses at the fibre input to increase the f/ratio (PFS does this) but this is not without impact – they have to be precisely centred and there will be some decrease in throughput. A better alternative is to use the new high numerical aperture fibres with a simple window in front.



5.1.3 Primary mirror concept (point design 10m segmented)

Keck demonstrated that segmented mirrors perform extremely well and weigh a fraction of that of a conventional single thick disk (cf. 8m telescopes like Gemini and others). Since the main function of the telescope structure is to support the primary mirror, the mass (and cost!) of the overall telescope scales with the mass of the primary. The 10m Keck telescope weighs the same (270 tons) as the 3.6m CFHT, hence the existing pier will support a telescope like Keck.

In preparation for the E-ELT and TMT, there has been a huge effort (with significant competition among companies and nations) to develop efficient and cost-effective production of 1.45 m (corner to corner) hexagonal segments, edge sensors, actuators and segment support assemblies. It is only logical to build upon this development for the MSE primary mirror.

A segmented primary mirror concept also provides the option of increasing or decreasing the overall diameter by adding segments (Figure 1), providing a true descope option. For example, a filled “10m” aperture would be comprised of 60 segments but the number could be reduced to 54 by removing the six outer most segments with minor impact on the science.

A monolithic borosilicate primary is presumably also a possibility but has the disadvantage of being limited to 8m diameter and having only one supplier could be problematic. In addition, fitting and operating an 8 m coating facility on-site would be challenging given the limited space and restricted access to the telescope pier.

5.2 Wide Field Corrector

5.2.1 Type (point design: refractive)

For the single-mirror telescope design, a “4+2” lens design with a simplified four-element refractive wide field corrector (WFC) like the Hyper-Suprime Cam wide field corrector on the Subaru telescope and a two-element atmospheric distortion corrector (ADC) appears to be the simplest and most cost effective option to deliver excellent telecentric images over a wide field with appropriate final f/ratio to couple into fibres. The WFC design is described in [RD2], Section 3.1.1.

5.2.2 Field of View (point design: 1.5deg2, hexagonal)

The diameters of the largest elements of a refractive corrector are limited by glass sizes that are available and practical to fabricate and mount. Our point design assumed a maximum glass diameter of 1.2m which results in a maximum 13% vignetting at the edge of the field. As reported in [RD2] Section 3.1.1, increasing the diameter to 1.35m would allow a field size of 1.5deg with no vignetting. However, diameters of up to 1.5m appear to be practical for glasses such as fused silica and so a larger field may be feasible.



Comment: a 1.5 square degree field requirement translates to a circular field with diameter of 1.38 deg – or to a hexagonal array that fits within a 1.5 degree circle. Thus, to first order, a “field of 1.5 deg” is equivalent to a “1.5 square degree field”.

A larger science field is possible with larger diameter glasses and/or by allowing some vignetting at the edges of the field. A full trade off study might be advisable to understand the scientific advantages and/or impact.

5.2.3 ADC (point design: lateral-shift)

Since MSE will obtain spectra at significant zenith angles with broad wavelength coverage (especially in the blue), an ADC is included. The proposed lateral-shift ADC performs well but will introduce a motion of the focal surface.

A study is required to further understand the design trades and the corresponding impacts on the overall science performance and operational model (if, say. observations have to be restricted by zenith angle). The inclusion of an ADC will, of course, reduce the throughput of the corrector (glass transmission and coatings) for all observations; it might limit the field size (if special glass pairs are required), and will increase cost. However, it will be necessary for some of the science programs.

5.3 Telescope Structure (point design: Keck2 concept)

A telescope structure similar to that used for the Keck telescope appears to meet the overall requirements and has the big advantage of being well understood by industry, in particular DSL (who built Keck2).

5.4 Enclosure (point design: Calotte concept) Extensive studies of enclosures for the TMT project demonstrate that the Calotte concept has the best structural and operational efficiencies to meeting the overall requirements, i.e. preserving the exceptional seeing on the site, and also the best match to the existing CFHT enclosure in size and mass than a conventional dome with a 10 m aperture opening. The cost, performance, and suitability of a Calotte style enclosure are well known, as are the procedures and costs for installation on site ([RD2], Section 5).

6 Instrument Design

The “instrument” is composed of three major components: the prime focus assembly (fibre positioner, acquisition system, calibration system, cable wraps, rotator, etc.), the fibre system that transports the light to the spectrographs, and the spectrographs.



6.1 Target pickoff (point design: WFMOS “Cobra” fibre positioner)

The WFMOS team (and subsequently, the PFS team) has invested a significant effort into the development of a fibre positioning system based on two stages of rotary motors which allow targets to be selected in overlapping circles giving 100% fill factor (any target can be selected by at least one fibre). In reality, the scientific ramifications of such a target selection system depends on the density of targets on the sky, clustering of objects on the sky, and the specific scientific goals, etc., and so studies are required to understand the optimal parameters (e.g., spacing, patrol region) for such a system. The rapid reconfiguration of the field enabled by this technology, and its relatively advanced technological development, were the main reasons for selecting this system as point design.

6.1.1 Number of targets (point design 3200)

The scientific requirements are for ~3000 spectra over a 1.5deg field [RD1]. For reference, the PFS system enables 2400 objects to be targeted (4 banks of 600 fibres) in a somewhat smaller field size.

6.2 Fibre system (point design: straightforward MOS system)

The possibility of including a few specialized fibres to feed dedicated spectrographs has been suggested to increase the versatility of MSE. Examples include an IFU bundle in the field or to feed an ultra high R spectrograph. These possibilities were not yet explored in the point design.

6.2.1 Fibre system design (point design: similar to WFMOS without couplers)

The point design concept for the whole upper end (comprising field selection and guiding, calibration, cable wraps, etc.) is similar to that originally proposed for WFMOS. It is recognized that there are several other positioning systems (e.g., MOONS, LAMOST, echidna, etc.). However, the WFMOS system has the advantage of being fully designed (and costed) for a similar telescope system.

Since MSE is a dedicated wide field spectroscopic facility, the fibre positioning system at prime focus does not have to be frequently installed on and removed from the telescope. Thus no fibre coupler (with attendant losses and calibration errors) is required – the fibre bundle can go straight to the spectrographs (located under the telescope in the pier lab).

6.2.2 Fibre size (point design: 0.9 arcseconds)

The optimum diameter for the fibres is a trade-off between maximizing the flux from the object and minimizing the sky background. The optimal size thus depends on the object size (seeing for point sources), object faintness relative to the sky and desired wavelength



range. Our analyses ([RD2], Section 3.2.1] indicate that the fibre diameter should be ~0.9” but this is contingent on both the median seeing expected and details of the anticipated science programs. Secondary reasons for choosing a small size are that if the PFS fibre system were to be cloned for MSE, the size would be 0.9” and a small input image makes it easier to achieve high spectral resolution. A more recent analysis by McConnachie [2] suggests that a fibre size corresponding to 1.0” is most appropriate for point sources with the currently expected seeing at the CFHT site. Further analyses need to be carried out to understand the optimal size for all sources, including faint galaxies.

Incidentally, a 120um fibre core corresponds to an aperture of 1.2 arcseconds for a 10m telescope with a f/2.1 focus.

6.2.3 Fibre output module (point design TBD)

It may be desirable to switch the fibre output to different spectrographs between spectral resolutions (e.g., low R, high R). This ability will become a requirement if, say, a PFS spectrograph design is selected for low R spectroscopy. It will also likely be necessary to accommodate other more specialized capabilities like IFU or very high R spectroscopy.

6.3 Spectrographs

6.3.1 Spectral resolutions (point design R ~ 2000, 6500, 20000) The consensus emerging from the feasibility studies [RD1] was that three spectral resolutions represent the best compromise for the majority of the science cases identified. However, scientific arguments have subsequently been made for even higher R for certain science. Given finite camera and detector sizes there is a trade-off between wavelength range and R. More scientific simulations and analyses should be carried out, accompanied by optical design studies to better understand the technical impact and costs and to provide reasonable constraints.

6.3.2 Wavelength range (point design: 370 – 1300nm) The findings emerging from [RD2], Section 3.3.1 are that simultaneous coverage of the full 370 – 1300nm wavelength range is required for the lowest resolution but more restricted wavelength intervals are probably adequate for the higher R modes. Given the constraints imposed by the spectrograph concept outlined in the Feasibility Study, the optimal ranges are:

• R 6500: 370-510nm & 770-910nm • R 20000: two * lambda/7 ranges in the 420-620nm region

As for the spectral resolution, more concepts, studies and tradeoffs should be considered by the expanded science team and technical teams.

6.3.3 Spectrograph Concept (a triple-resolution spectrograph concept is discussed but no point design adopted)

The science cases for the two lower spectral resolutions frequently require full wavelength coverage whereas more restricted ranges are more often required for high resolution science. Even so, the latter may require several orders to cover the desired wavelength range, in which



case the number of targets must be correspondingly reduced to fit all the spectra on the detector. It is thus difficult for single spectrographs to provide all of these options efficiently or effectively. However, if the high and low resolutions were provided by different spectrographs then some method of switching the fibre input between spectrographs would be required. A single “do-everything” triple-resolution concept was studied during the feasibility phase but the versatility is achieved through pupil slicing which then would result in decreased S/N in cases where detector noise is important. Another possibility for the lower resolutions would be to clone the PFS spectrographs – but then a new design for the R=20000 spectrographs is required, as well as the concept for switching the fibre input. Once updated scientific requirements for the spectrographs are available, this will guide future technical studies to explore alternate concepts and facilitate trade-offs. However, firm requirements for the initial spectrograph system should be defined as soon as possible in order that a solution (or solutions) be advanced in the next conceptual design stage.



Figure 1. 10m mirror with 60 segments. The diameters of the three circles are 9.4m, 10m and 11.2m