ELT-HIRES the high resolution instrument for the ELT: optical design and instrument architecture

2018

Publication Year

2020-10-23T13:16:04Z

Acceptance in OA@INAF

ELT-HIRES the high resolution instrument for the ELT: optical design and

instrument architecture

Title

OLIVA, Ernesto; TOZZI, Andrea; FERRUZZI, Debora; RIVA, Marco; Genoni,

Matteo; et al.

Authors

10.1117/12.2309923

DOI

http://hdl.handle.net/20.500.12386/27967

Handle

PROCEEDINGS OF SPIE

Series

10702

Number

PROCEEDINGS OF SPIE

SPIEDigitalLibrary.org/conference-proceedings-of-spie

ELT-HIRES the high resolution

instrument for the ELT: optical design

and instrument architecture

Oliva, E., Tozzi, A., Ferruzzi, D., Riva, M., Genoni, M., et

al.

ELT-HIRES the high resolution instrument for the ELT: optical

design and instrument architecture

E. Oliva*

a

, A. Tozzi

a

, D. Ferruzzi

a

, M. Riva

b,

M. Genoni

b

, A. Marconi

a,c

, R. Maiolino

d

, L. Origlia

e

a

_{INAF-Arcetri Observatory, largo E. Fermi 5, I-50125 Firenze, Italy;}

b

INAF- Osservatorio Astronomico di Brera Via Bianchi 46, I-23807 Merate, Italy;

c

Dip. Fisica e Astronomia, Univ. di Firenze, via G. Sansone 1, I-50019, Sesto Fiorentino (FI), Italy;

d

Cavendish Laboratory University of Cambridge, 19 JJ Thomson Ave, Cambridge CB3 0HE, UK;

e

_{INAF-Osservatorio Astronomico di Bologna, via P. Gobetti 93/3, I-40129, Bologna, Italy}

ABSTRACT

The first generation of ELT instruments will include an optical-infrared High Resolution Spectrograph, conventionally indicated as ELT-HIRES. This paper describes the optical design and overall architecture of the instrument whose main capabilities can be summarized as follows.

– Fibers fed specrographs with resolving power R=100,000 (HR-modes), R=150,000 (UHR-modes) and R=20,000 (MR-modes).

– Simultaneous wavelength coverage from 400 nm to 1800 nm; extendable to 300-2400 nm. – Spectrometers with fixed configurations and without moving optical parts.

– Maximum size of entrance apertures ideal for seeing limited observations.

– Many observing modes, including seeing-limited single-object spectroscopy, multi-objects medium resolution spectroscopy and IFU observations with different spatial scales, down to the diffraction limit of the ELT telescope.

– Observing modes defined and selected in interfaces outside of the spectrographs.

– Modular design compatible with a minimal baseline that can be subsequently expanded and upgraded.

Keywords: Ground-based instruments, high resolution spectrographs, infrared spectrographs

1. INTRODUCTION

The European Extremely Large Telescope (ELT) will be the largest ground-based telescope at visible and infrared wavelengths. The flagship science cases supporting the successful ELT construction proposal were the detection of life signatures in Earth-like exo-planets and the direct detection of the cosmic expansion re-acceleration. It is no coincidence that both science cases require observations with a high-resolution spectrograph. When defining the ELT instrumentation, ESO commissioned two phase-A studies for high resolution spectrographs, namely CODEX1 _(covering the 370 nm – 710 nm wavelengths range) and SIMPLE2 (covering the 840 nm – 2400 nm wavelengths range). The studies, completed in 2010, demonstrated the importance of optical and near-IR high-resolution spectroscopy at the ELT. ESO thus decided to include a High-REsolution Spectrograph (HIRES) in the ELT instrumentation roadmap. At the same time, the CODEX and SIMPLE consortia realized the great scientific importance of covering the optical and near-infrared spectral ranges simultaneously. This marked the birth of the consortium that responded to the ESO call for the phase-A study of HIRES in March 2016. The study was successfully concluded in April 2018. More details on the HIRES project, science and consortium can be found in the dedicated paper3 presented at this conference.

- MR front -end Other front -ends J (L <200 m) Special IR fibers (L < ? ?? m) Cold IR spectrograph

r1

fiber to fiber = interface

-O Cold IR spectrograph K 1.44 -1.80 1.92 -2.40 HR and UHR front -end SCAO -IFU front -end interface Calibration Unit fiber to fiber interface z Y 0.55-0.67 0.67-0.83 0.83-0.95 0.95-1.13 I iZ-1 Z ELT HIRES UV libers (L «10 m) BB libers (L <10 m) fiber to fiber interface fiber to fiber 0.33 -0.40 0.40 -0.46 0.46 -0.55

2. OVERALL CONCEPT OF THE ELT-HIRES INSTRUMENT

The design consists of four spectrometer modules covering different wavelength bands; all fed by fibers (see Figure 1). Each spectrometer module is fed by a number of fiber bundles that correspond to the different observing modes. Inside the spectrometer each fiber bundle is connected to a linear array of micro-lenses that act as input slit. The input slits are mounted and fixed once forever.

The resolving power of a given bundle is determined by the core-size of the fibers. The baseline design (described in Section 4).includes 10 slits/bundles for R=100,000 (HR modes), 10 slits/bundles for R=150,000 (UHR modes) and 1 slit/bundle for R=20,000 (MR mode). Note that each spectrometer can host twice as many slits/bundles.

Different front-ends are used for seeing-limited HR/UHR observations, for SCAO-IFU observations and for MOS-MR observations. Each front-end includes dichroics that split the light to the spectrometer moduli. The active observing mode is selected using shutters in the front-ends and/or in the fiber-to-fiber interfaces.

Figure 1 Schematic layout of the HIRES spectrometer; the numbers on the right-hand side are wavelengths in microns.

Outside the spectrometer the fiber bundles terminate into the fiber-to-fiber interfaces that include the opto-mechanical devices coupling the light into external fiber bundles that are connected to the telescope front-end or to the calibration unit. The topology of this coupling defines the details of each observing mode that can be set-up, maintained, modified or added without touching the spectrometer modules. The parameters of the baseline observing modes are summarized in Table 1. It includes

– 4 modes optimized for throughput; i.e. with as few as possible optical elements in the front-end4 and in the fiber-to-fiber interfaces.

– 4 modes optimized for spectral accuracy; i.e. including in the frontend and in the fibertofiber inferfaces -opto-mechanical sub-systems that guarantee a uniform illumination of the fibers and of the spectrometer slit. These modes may result into some loss of efficiency, if necessary.

– 4 modes for spatial resolved spectroscopy; i.e. with the fiber-bundles organized for IFU observations in the 5

Main collimator: double pass parabola

Collimated beam 400x200 mm 1 Main disperser: R4 echelle grating 1.6m x 0.2m 0

Lcollimatorcollimator _{cross -dispersion}magnifier along

Collimated beams Collimated beams

280 x 140 mm 280 x 280 mm Transfer Anamorphic

collimator magnifier along cross- dispersion

U o

Camera and detector

Camera and detector

Camera and detector

J

Input slit Micro -lenses 10 arrays

Decker to select slit/mode

` 1

Fie- imager with anamorphic magnifier along main dispersion

Input slit image

r

HIRES

spectrometer module

Anamorphic

Camera and detector LinterfaceJ L interface_Jlinterface Jl interface interface .... other HR modes possible...

Similar modes with UHR fibers /slits

=I NI LI OI 1 _¡

"

O:

Functional scheme

ELT ELT ELT ELT

interface interface interface interface MR with HR1 with HR2 with HR3 with shutter shutter shutter shutter

N co x co 4) Q N CC x co 4) Q C`') CC x

Fibers to Fibers to Fibers to Fibers to fibers fibers fibers fibers optical optical optical optical

ELT interface UHR1 with shutter N CC Fibers to fibers nnti rai HR1 : HR2 : HR3 : HR4: HR5 : HR6: HR7: HR8 : H R9 : HR10 optimized optimized optimized optimized optimized optimized optimized optimized optimized : MOS 32 for RV accuracy for throughput

for IFU with zoom for polarimeter for ... for .... for ... for .... for ....

obj at AO- corrected focus

Table 1 Main parameters of the baseline observing modes

Name Type Resolving

power # of sky apertures Projected diameter of each fiber # of fibers per aperture Projected size of each aperture Simultaneous calibration B1 Throughput optimized 100,000 1 0.170” 64 Ø 1.36” No B2 Throughput optimized 100,000 2 0.170” 30 Ø 0.93” No B3 Throughput optimized 150,000 1 0.113” 96 Ø 1.11” No B4 Throughput optimized 150,000 2 0.113” 47 Ø 0.77” No

B5 Accuracy optimized 100,000 1 0.170” 56 Ø 1.27” Yes

B6 Accuracy optimized 100,000 2 0.170” 30 Ø 0.87” Yes

B7 Accuracy optimized 150,000 1 0.133” 96 Ø 1.06” Yes

B8 Accuracy optimized 150,000 2 0.133” 47 Ø 0.73” Yes

A1a Spatially resolved 100,000 1 0.010” 8 x 8 0.08” x 0.08” No

A1b Spatially resolved 100,000 1 0.100” 8 x 8 0.80” x 0.80” No

A2a Spatially resolved 150,000 1 0.010” 9 x 10 0.09” x 0.10” No

A2b Spatially resolved 150,000 1 0.100” 9 x 10 0.90” x 1.00” No

3. THE SPECTROMETER MODULES

The functional scheme of a spectrometer module is depicted in Figure 2.

100

0 8

o

'rn

60

40

C.)

20

BV

RT

ZYJ H

I I ! I I I I I I TTITT

0.4

0.5

0.6

0.7 0.8 0.9

1

1.2

Wavelength (sum)

rit

. -

1.6 1.8 2

2.4

All the spectrometer modules are equipped with a number of adjacent, parallel slits all with a length equal to 11 arcsec (sky projected angle). The slit-resolution product of the spectrometer is RS=17,000. Therefore, the R=100,000 slits are made of 64 fibers with projected diameter 0.170 arcsec, while the R=150,000 slits are made by 96 fibers with projected diameter of 0.113 arcsec. The R=20,000 slits are made by 10 fibers with a projected diameter of 0.85 arcsec.

The main spectrometers (BVRI and ZYJH) include four separate arms covering wavelengths that are adjusted to maximize the wavelength coverage and overlap with the regions of good telluric transmission (see Figure 3). The U and K modules are added to push the coverage toward the UV and infrared domains. The K module requires different type of fibers (see Figure 4).

Indeed, the spectral range covered by the Z arm (830-950 nm) would be much more efficiently covered using a deep-depleted CCD detector in a non-cryogenic module. The Z-arm is included in the IR cryogenic module solely to push to the wavelength coverage of the BVRI module to the shorter possible wavelength. An alternative configuration with a three-arms IR module (YJH) covering the 950-1800 nm range would yield a more natural, efficient, and cost-effective solution that however limits the wavelength coverage achievable in the blue using just two moduli.

o 0.4 0.5 0.6 0.7 0.8 0.9 1 1.2 1.6 1.8 2 2.4

Wavelength (µm)

E 20 W en O 40

/

z

ILIAJ 80 6 t¡LAIN L

YPicai

0.4 0.5 0.6 0.7 0.8 0.9 1 1.2 1.4, 1.6 1.8 2 2.4

Wavelength (sm)

h (meter o

-'

O 60 40 , .w ., /+h

/

Bro<<d-hand

V\1

Ñ tn E100 cn ,9 8

/

1

o

co u200

Te.ecc:ím/

ItZ

V(LIV RI

_{TrTT--T'l'TT 1TT1 TI"f"1"1"1"fT TTTTT'}

_TTTT

YJ

K

1

Figure 4 Transparency of different types of fibers and wavelength coverage of the spectrometer modules

The layout and rays-tracing of the ZYJH module are displayed in Figure 6 and Figure 7. Following the light path one finds the following elements.

• The input slit, consisting of linear arrays of micro-lenses as described in the Section 7. • A slit re-imager and anamorphic pupil expander that includes

o A lens-doublet that collimates the light from the slit.

o Two prisms that provide a 2x anamorphic magnification in the main dispersion direction. o A lens-doublet (identical to the collimator above) that refocuses the light.

o A wedge of fused-silica on the exit focal plane, needed to re-position the pupil image.

o Three flat mirrors with fold-angles adjusted to get a convenient mechanical positioning of the module. • The main collimator, consisting of an off-axis parabola used in double pass. The diameter of the collimated

beam is 400 mm in the main dispersion direction and 200 mm in the cross-dispersion direction. • The main disperser, consisting of a R4 (76o) echelle grating of 1600 x 200 mm.

• A wedged field lens of fused-silica, positioned at the intermediate focus. The power of the lens is used to correct the field curvature while the wedge is used to re-position the pupil image.

• A dichroic of fused-silica that reflects the “blue” light and transmits longer wavelengths. The exit face of the dichroic has a mild optical power to avoid ghosts.

• Three flat mirrors that are used to fold the light and position the arms at mechanically convenient positions. • Two identical transfer collimators, each consisting of an on-axis parabola. An even simpler solution with a

spherical mirror (as in ESPRESSO) could also be used; but the overall image quality is slightly worse.

• Two identical prisms of fused-silica each providing a 2x anamorphic magnification in the cross-disperser direction, thus re-obtaining a circular pupil (to the first order).

• Two wedges of fused-silica with a dichroic coating on the entrance face, used to split the light between the arms.

• Four cross-dispersers, consisting of transmission gratings working off-Littrow and providing an extra anamorphic magnification in the cross-dispersion direction.

• Four cameras each consisting of 2 lenses of fused-silica and 1 mirror (see Figure 10). This type of camera is the same as that developed for MOONS6 and currently under construction. The lenses are identical in all the cameras; this simplifies the manufacturing and reduces the costs.

• A detector of 61 mm x 61 mm (size of a H4RG 4k x 4k infrared array with 15 microns pixels). The detector is behind the center of the lens (see Figure 10). This causes an obscuration at the center of the pupil with vignetting that varies along each order as shown in Figure 11.

The same design can be conveniently adapted to the other spectrometer modules obtaining the layouts shown in Figure 8 and Figure 9.

The elements listed in Table 2 are identical in all the three modules, apart for the coatings of the optical elements and the groove spacing of the echelle gratings.

Table 2 List and description of the elements common to the three spectrometer moduli.

Element Description

Main collimator Off-axis parabola, f=4.0 m, D-parent=1.5 m

Main disperser Echelle R4 L=1.6 m W=0.2 m

Transfer collimators On-axis parabola, f=2.6 m, D=0.6 m

Field lens Fused-silica plano-convex lens, R=2.8 m, wedged 6o

Fold mirrors Various flats, same angles.

Anamorphic prisms Fused-silica prisms with aperture angle 35o

Dichroics Fused-silica; 1 plano-convex and 2 wedged

> _ E G e m Ú ö Ú CRDJ''`CRDH CRD Y Prism JH` Feld. 10 lchrolcJH Transfer collimator JH _Fold,

IField lens

Collimator

1

1m

Echelle

Slit re- imager and anamorphic pupil expander

To spectrometer -Exit slit Camera Fold "á F Fold 1

-

- - -'

Collimator Fold Input slit

Zoom of central region Slit re- imager and anamorphic pupil expander

;.. Fold Prisms ,...Collir EchelleR4(1.6mx0.2m)

)

5.5 m

E

nï

Figure 6 Layout and rays tracing of the ZYJH module

Fold 2 ^

Dichroic RI

Field lens

Zoom of central region

To spectrometer

Exit slit

Collimato

Fold 1

Input slit - Fold

r

Slit re- imager and anamorphic pupil expander

Collimator Echelle R4 (1.6 mx 0.2 m)

5.5 m

E

nï

The first order parameters of the optical design are summarized in Table 3.

Table 3 First order parameters of the optics; the values are the same for all modules unless otherwise specified.

Parameter Value

Slit length 11 arcsec (sky-projected angle); 41.6 mm at F/20

Slit width for R=100,000 0.17 arcsec (sky-projected angle); 0.64 mm at F/20 Input focal plane (before slit

re-imager)

F/20, flat, telecentric, 11 x 8 arc-sec (sky-projected angle) un-vignetted field of view

Focal plane at spectrometer slit F/10 (along dispersion) x F/20 (along cross-dispersion) Off-axis of echelle 0.9o (in the cross-dispersion direction)

Beam in main collimator 400 x 200 mm

Beam after transfer collimator 280 x 140 mm Beam after cross disperser 280 x 290 mm

Focal aperture of cameras F/1

Detector size 61 x 61 mm.

Detector type ZYJH: H4RG 4k2 detectors with 15 microns pixels.

BVRI: CCD 6k2 detectors with 10 microns pixels.

Detector L1, L2: Fused Silica L1

M1

Aspherics

150.00 MM 5.5 m E LO 5.5 m E

Figure 9 Layout and rays-tracing of the U module (left panel) and K module (right-hand panel).

Figure 10 Detailed view of the reflective camera. L1 and L2 are identical in the 4 arms of the ZYJH and BVRI spectrometers

09'T OL 'I 091 00'1 ,,,,,,,,, H lib NI ti %V\+. I

LI.i.iI

ii.I

00'1 901 01'1 98'0 98'0 08'0 99'0 ;El OCT ' S2'1 [1.\\\N1/4tL\1 (turf) inguatexela 09'0 1 ST 02'1

A

.&tiA\til\N,'

- I I - z _

\

\

.

\\4

\

\

\

\

C

\

1 ' - 94.-0 OL0

l.ilii

96'0 OS-0 29'0 6S'0 96'0 A

:i

-

'*rti\ti1\.iCU\t:

- I S90 09'0 SSO 99'0 46'0 26'0 04'0

a

04 G M 09 06 /1 M 0L 0ö 09 9 06

Figure 11 Effect of the central obscuration in the reflective cameras of the ZYJH and BVRI modules.

4. DISPERSERS AND SPECTRAL FORMAT

The parameters of the gratings and the spectral coverage of the modules and arms are summarized in Table 4. The corresponding spectral formats are displayed in Figure 12,

Figure

13

and Figure 14. These figures show the echellograms generated from the rays-tracing design.

The wavelength coverage of each arm is chosen to guarantee some overlap between adjacent spectral arms. The only exceptions are the J, H and K arms that are separated by regions of very poor telluric transmission (see Figure 3). The grooves spacing of the main dispersers (echelle R4 gratings) are chosen to maximize the overall spectral coverage; i.e. with the lowest order completely filling the detector. This yields values of grooves spacing different than those available in the current catalogs and, therefore, require the production of new master gratings. Note in particular that using the standard 31.6 gr/mm would result in a BVRI module with a spectral coverage reduced to 440 - 830 nm. The cross-dispersers are transmission gratings working in first order. Their size (340 x 380 mm), groove spacing and incidence angles are perfectly suited for the use of binary gratings lithographically etched into fused silica.

Table 4 Main parameters of the cross-disperser gratings.

Module Echelle R4 Arm Cross-disperser Orders Wavelength coverage (nm)

gr/m AOI ZYJH 16.00 gr/mm H 465 22.9o _{68 - 84 1435 - 1800} J 730 32.9o 90 - 107 1128 - 1356 Y 930 35.4o 108 - 127 950 - 1128 Z 1240 41.1o 127 - 146 827 - 955 BVRI 34.50 gr/mm I 1000 22.3o 68 - 84 666 - 833 R 1410 33.3o 84 - 102 549 - 674 V 1860 32.6o 102 - 121 483 - 555 B 2490 39.4o 121 - 140 400 - 467 U 67.00 gr/mm U 2240 27.2o _{72 - 87 330 - 405} K 11.95 gr/mm K 360 27.8o 72 - 86 1877 - 2405

H

10 mm 4 n] 107 108 105 104 103 102 101 100 2 99 v 98 9 Ó 97 96 95 94 93 92 91 90

J

10 mm E e o

íCi$RtúRRÑ$woro " n w.. o S 8

szapIp 148 145 144 143 142 141 140 139 198 137 136 7,135 0134 193 132 131 130 129 128 127 10 mm .fn! eye.g 428.9 9294 MO

932.6 es3.9 e3.3 .. ARIA

Bae.41:64:4 . _{. ftlt :9g1 AttM}

844.2114"*Itir _{-MO :::1Aliftl}

a.50.11.644

6'4

₀₄₇₄₁ 858 114644.¡38.81`88°.'f858..4. eg4.71_..14:¡,1 000.ojaas4 654 4 8852! 671.11 e7261 ' Vaal ' _PP./ 07481:8*:01 onAl 4014 '''''' ... 881 _{00341 88521.140.5). .. 696.0) ..} 06741iite1 80031 _{... 6946!} 8840).8694 ... . 8s55J QI 0) _ecos)

800.01 801Pil ewer wall 206.51:..004 ... 8074) ;sal 1004 914.21 918.71 9174 91 4:. 9211) 022.71 2.441 0.871 MI ... ... ... ...

_....

04.41 E48.51 940.11 94171 943.3J 942,4 9444r 9041 9741 5484:-.9094 ... ... ... ... ... ...

Figure 12 Echellograms and spectral format of the 4 arms of the ZYJH module; generated from the rays tracing. The green square represents the area of the 61x61 mm detector. Wavelengths are in nm.

I 10 mm 102 101 100 99 96 97 98 96 94 293 ,o 92 0 91 90 89 6a 87 88 85 84

R

_{10 mm} a4...54.4 560.1 M11.4 662Á 663,4 66141 63.0 .. l'ai 5xb 666.9 557.9 956.9...:.._5692I 59.31.. 4I{....fi01.4... 5824. ....603,51 594.51... 64.2'z 4.2 69py667:0 ....5:11... 569.21 '..,.¢71.01 70.7 671.Bj smelt 5I3.91 575.01 670.0 '. -;°;j 70.6 6T4AI 676.71...679:áI..._:!I...9B201r..::.1 :..6_..:...:666Aj .6B4.f1....91...657A1....666.II..::::.21_{... .} 9 6ÿ6;a7;I.._°1.._^e....iI..._.?I.::.rva4j

x.5I...8j e913j ówel:.'.-?ö?;ï1..::::..i.òj :"!#4.1. I r 6671I... 9m0 01.2 ...em0 9x. 9019

07.7 669Á /10.1I 9112j sls0].:;::.ëiäi[:::::::'älä:òj 14.3 516.9I 916AI 61101. 91,.9I.._..,!?0.6I_...,..0812j

...

921.1:sj 6Ba6I...6j...6,I::::::::äl:-:ëm>j 625.0 itëaj .eo el 99101 13a:I:::: .?:?:._:-:::;ië?;I

...

636.1...^I....^g!:e....6..m:!...äö:l::::::-«iäI:-::_...1

2,4...64,471 946.1] 9.9.51 94701...eNBL:::::,.äj...

20 .f.^.Ì:?1....SZ.äj 664.0I 656AI...::::ë.M1.ë.ë`l1':.:.:.:.:.6¢j

...7.4...SI...9/021 991.7I...:!I...7'`^:`1.:::::.::.

iiQlA 6e92 49a1 6995 6710 -... ....

zl:.::: .:::Il ..:..:: ...:::I::::::::..:::I:::.:::ë?e 6I.::.:.::..A7iöj

121 120 119 118 117 116 115 114 113 112 P., 111 7. 0110 109 108 107 106 105 104 103 102

V

10 mm 462.6 Ï95.9 9590 465.9 440.6 460.9 IM.I! 466.9 467.1 487.9 460.7 469.6 470.2 471.9] 476.Y 471.U'.171.8 476.6 479.4 ...4745,476.61 ::... 4742 476.Ó] 769 470.6j 477 41 478.21.".5,7219 479.2 476.11 479.9 lDal...!:I..._..466..AI:..fi9.Sl

462.6 '-159.e1 464.01 4049] 466.>I...!el..YB7_4]

466.6 _r..:.: 4B62I II 6L.._K0. 981:91

490.6 .o:6j 496.61 . g wa91 490.1I.. .4Y6:GI1 496.1 490.0y1y 49691...N7.1...9691._..49091: ..StAI

496.6 -600..1..60.:.I... Y]....?98:.1:::.90ä.0l: ..:ä9i91

609.9 .ew2]._g06:01..A.tL...OI....OI...p0?4f 609.6 -ó0i.41...910.41 01121 61221....6162j...fj14.II 619.1 ó,:]_.Ö16:01...81691...616.9]...4.I7:V1. ... ..k1&6] 617.9 6169]... 61991... 480.91... 621.71...ppy,jl...6i1.7] 662.7 622.7] 69491 62091 626.61 6219I ... iiiiël ...

sz7.e_4I 6969 63041 m1.eI...9së,?1...ä,ä+.ë1

592.6 . ëI 65491 6069I...yyp.ïl... ,.t1...t] 697.7.ës6f1.._999:à1 64091 641.9I:.:::.04Y;i1. ...

-1

.:. 642.9 el...b!^:01...I..:..?.!!.:I:.:..?.!9,?]1...1NO2I 646.1... 66141 662:::::.169:I:::::::.A4!;e. 140 139 136 137 138 136 134 133 132 131 0130 o129 l26 ln Ize 126 124 123 122 121

B

10 mm 220.2 4112.6. 408.7 M9.4 401 47.A 447.4682 4682

'"61

462p 4W ...

4118 41EE IILp 418A1 41{2 ..

`V

468.6 46... 410'E {06.

I

M6A ÖY.£ ,416A _411.8 .411

542;61

4119 {IQ6l

uv iïil' 418 1 6,3 il: .iïg!I..::ïï Pl.,,::iïëëll

L. I I rY1 1. ,.,A1

410.6 , .4i.1.2.1. .°..:1_u¢.al._:el...!qä:l1.:::.äl

47..0 KA.71 4664[ u6.il..{j{ï1..fY/.11:::xSSS

427.2 427.9I 42261 4aaI..!püÄl.. 4307J ::::.i.i:ël

4326 43121 Mt21 M24L...iäi;ïf"s1 493.8 47..61 4.6.1.61 0261 43261....!wT:=1 4272 M9.6I...4..'.M.A,.,.I¡T... N6i1.. .}bpll. :: 44U8N13I,wa1 l...M26I....1..:I-:::11 444.1 4M.61 4M4I....j>ci12...Ni1.c:...>1

447.8'"ïib.il .M21 4M21

_"'I

_{....2.i.21::.}

4612'. 11i:Ó1 482.61 M2e1...4eidl _{iÄ9.ë::'¡6pp}

..1.-46.6:4 _{1241 46+61.:}

"ip..C... xP Ö;:I...:::.

40861 469.41 41021 M1 61__. MI sl...1lYDI .

462.31^e!I47..61... :e1...N661...14S:eI.

ál

_,e,r

Figure 13 Echellograms and spectral format of the 4 arms of the BVRI module; generated from the rays tracing. The green square represents the area of the 61x61 mm detector. Wavelengths are in nm.

K

10 mm

U

_{10 mm}

Figure 14 Echellograms and spectral format of the U and K modules; generated from the rays tracing. The green square represents the area of the 61x61 mm detector. Wavelengths are in nm.

5. IMAGE QUALITY

The optical quality of the entire spectrometer can be conveniently evaluated in terms of the diameters with ensquared energies equal to 50%, 80% and 95%. These parameters are computed at different wavelengths, orders and spatial positions along the slit; thus covering the whole echellograms. The diameters of ensquared energies can be expressed in microns and in terms of the ratio between the diameter of encircled energies and the geometrical width of the slit image onto the detector at a given wavelength, order and position along the slit. The latter ratio provides direct information of the fraction of light falling inside the slit image. The results are displayed in Figure 15, Figure 16 and Figure 17. The values are remarkably good; for BVRI the optical quality is similar to that achieved in the optical design of ESPRESSO (see the comparison in Figure 18).

The image quality at the extreme wavelengths (K and U modules) is slightly deteriorated by the chromatic aberration introduced by the fused silica lenses in the cameras. Although better results could be in principle obtained using CaF2 elements; however this solution is risky - because it requires CaF2 crystals of size larger than those currently produced - and it would make the system much more expensive and difficult to handle.

0.5 l.0

B

95% . 80% 50% 0.48 0.50 0.52 0.54 0.58 95% 80% 50%. 0.5 0.56 0.60 0.65 0.70 0.76 0.80 Wavelength (ism) 45 30 0.42 0.44 0.48 0.48 0.50 0.52 0.54 0.58 30 15 0.55 0.60 0.65 0.70 0.75 0.80 Wavelength (tom) 95% 80% 50% 0.85 0.90 0.96 0.95 00 1.05 .10 1.15 1.20 1.26 1.30 1.35 1.50 1.80 Wavelength (ism) 1.70 1.80 45 30 0.85 0.90 0.95 0.95 1.00 1.05 1.10 30 15 1.15 1.20 1.25 1.30 1.35 1.50 1.80 1.70 1.80 Wavelength (p.m)

Figure 16 Diameter of encircled energies over the echellograms of the ZYJH spectrometer.

1.0 0.40 0.45 0.50 0.5 0.50 0.60 0.70 0.80 Wavelength (µm) 1.0 0.40 0.42 0.44 0.48 0.48 0.50 0.52 0.54 0.58 0.5 0.55 0.60 95% B07e 50% 0.65 0.70 Wavelength (µm) 0.75 0.80

Figure 18 Comparison between the image qualities of the BVRI module (left) and ESPRESSO (right). In both plots the diameters of the encircled energies are computed over the whole echellogram and are normalized to the geometrical width of the slit image, thus providing a direct comparison independent on the scale of the image on the detector.

6. ALTERNATIVE DESIGN WITH LENS-CAMERAS

An alternative design for the BVRI module with refractive cameras (i.e. cameras with only lenses) is shown in Figure 19. It uses larger detectors (92 mm x 92 mm) and slower beams at the focal plane (F/1.6). The lens-cameras are much larger (about 1 x 0.5 m) and heavier (about 200 kg) than the reflective cameras (see Figure 20). The lenses are made by fused silica, F2 and BK7 (or equivalent); i.e. the most popular and better controlled glasses that are normally available in large enough sizes and that have low internal absorption over the whole 400-1000 nm range. A minimum of six lenses with three aspheric surfaces are needed to obtain good images. The final optical quality is slightly worse than that achieved with reflective cameras (see Figure 21). The total thickness of glass is such that the overall throughput and image quality may be deteriorated by the internal absorption and non-homogeneities of the glasses. Moreover, the manufacturing and alignment errors of the many surfaces may introduce other aberrations. In practice, the efficiency that one formally gains avoiding the central obscuration could be partially lost by the practical problems related to such complex lens-cameras. Cost is another important drawback of lens-cameras.

Solutions with lenses are much more difficult (or even impossible) to achieve for the infrared modules because of the very fast focal ratio (F/1) imposed by the size of the IR detector and because of the very limited choice of glasses that are sufficiently transparent in the required spectral range and available in large size/thickness (in practice only fused-silica and CaF2). More details can be found in the trade-off analysis for the MOONS optics6.

0.48 0.50 0.52 0.54 0.58 0.55 0.60 0.85 2.70 0.75 0 80 Wavelength (µm) 0.40 0.42 0.44 0.46 0.40 0.50 0.52 0.54 0.58 0.55 0.80 0.65 0.70 0.75 0.80 Wavelength (am) Detector Aspherics 150.00 een 5 .5 m Zoom of al rag Exit Fold Pris Fold InFUtsln CollimatorIl

smre- andenanmd:n wm

w

,

5.5 m

N

C)

Figure 19 Comparison between the layouts of BVRI module with lens-cameras (leftl) and with reflective cameras (right-hand panel). Apart for the cameras, all the other elements are identical.

Figure 20 Layouts and rays-tracings of a reflective (left) and a lens camera. The detector in the lens camera is tilted.

Figure 21 Comparison between the image qualities of the BVRI module with reflective cameras (left) and with lens cameras (right-hand panel). In both plots the diameters of the encircled energies are computed over the complete echellograms and are normalized to the geometrical width of the slit image, thus providing a direct comparison

Name Ll (t,R in mm) L2 (t,R in mm) F0 fiber F0 lens 0 core (mm) D lens (mm)

Maxmum light losses for all fields and wavelengths

SRSL7 SLAMOS t =1.00 0=4.00 4% HR R1 flat R1 =0.510 F /3.5 F /20 0.113 0.646 R2 =0.510 K1 =-0.0 K2 =4.0 02 =2.407 SSS17 t=010 SLAM55 L =2.60 UHR R3 flat 01 =0.350 F/3.3 F /20 0.075 0.430 5% R2=0.350 K1=0.0 K2 =4.0 R2 =1.71 5SSL7 t =80 SLAWS 0-19.0 MR 01 flat 01 =2.80 F/3.5 F /20 0.570 3.40 2% R2 =2.00 K2 =3.0 K2 = -3.0 R2 =15.1

Main p rameters of micro -lenses (gl ed do blets( for the HIRES 25J11Sp ctrometer module. L3, L2: =thickness, R =radius of curvature, K conic constant. Fber is glu d to first (flat) surface of Ll.

FK fiber: nom hal aperture of light eam at the fibe beams are telecentric. FP lens: nominal aperture of light beam at the micm lens; beams aretelecentric. 41core: diameter of fiber core.

D lens: ion of the lens.

Includes light losses produced by blurring of images, by deviations from telecemricity, and by variations of the beam apertures. Maximum values among all fields and wavelengths; average values are smaller. UHR slits HR sits MR slit UHR mUrolenses HR microlenses MR mìcrolenses UHR mICrolenses HR microlenses UHR fibers HR fibers its R slits MR slit MR fibers 7. INPUT SLIT

The entrance of each spectrometer module consists of an optical system that re-images the input focal plane onto the virtual slit of the spectrometer (see Figure 6, Figure 7 and Figure 8). The re-imaging optics have an unvignetted field of view of 11 x 8 arcsec (sky projected angles). Therefore, they can be used to re-image the light from many parallel slits that can be separately illuminated through dedicated fiber bundles. Each fiber bundle corresponds to an observing mode and terminates in a linear array of micro-lenses, as shown in Figure 22.

Figure 22 CAD views of the arrays of micro-lenses and fibers at the input focal plane of a spectrometer module.

Each micro-lens consists of a lens doublet that produces a sharp, quasi-telecentric and chromatically corrected image of the pupil at F/20.

Figure

23

shows the layout and rays-tracing of the micro-lenses for the HR, UHR and MR modes of the ZYJH module. They are designed assuming that the intrinsic focal aperture of the fiber is F/3.5. The micro-lenses in the spectrometer have a rectangular shape. The same micro-lenses can be used in the front-end to receive the light from the telescope, and in the fiber-to-fiber interface to double-scramble the light of a whole fiber-bundle. The only differences are their external shape - that must be hexagonal - and organization on a quasi-circular field of view. Therefore, the large number of identical micro-lenses and fiber interfaces, that are needed to implement all the observing modes, can be most conveniently manufactured via an industrial process specifically set-up for this project.

8. COMPARISON WITH THE DESIGN OF OTHER HIGH RESOLUTION SPECTROMETERS

To evaluate the overall capabilities of a given spectrometer it is convenient to combine its parameters into a single figure of merit or “Spectrometer Equivalent Power” (S.E.P.). This quantity is proportional to the spectral resolving power (R), to the sky-projected area of the entrance aperture and to the number of spectral resolution elements that are simultaneously covered in the spectrum. Numerically:

A high resolution spectrometer with S.E.P. greater than unity can effectively operate under normal seeing conditions and acquire high resolution spectra covering a wide wavelengths range (one octave). Instead, instruments with S.E.P. much lower than unity require some improvement of the telescope PSF (i.e. AO correction) and/or must work over a reduced wavelengths range. For a given spectrometer, the value of the sky-projected aperture is proportional to (1/DTelescope)2. Consequently, modern spectrographs that are effectively operating on 4m class with S.E.P.~10 (e.g. HARPS, CARMENES) would have values of S.E.P.<0.1 if mounted on the ELT. Even replicating the most powerful spectrograph built so far – i.e. ESPRESSO – one would still fall short by a factor of two.

The above considerations are summarized in Table 5 that includes a few representative optical/IR HR spectrographs that are operating on, or are under construction for existing telescopes. The table also includes the designs of seeing-limited HR spectrometers for ELT’s, namely G-CLEF (GMT), MTHR (TMT), CODEX (first phase-A of ELT) and HIRES (this project). The entries “ESPRESSO*BVRI” and “CODEX*BVRI” are spectrographs similar to ESPRESSO and CODEX with parameters re-adjusted to get R=105 and shift the maximum wavelength to 830 nm.

Table 5 Parametric comparison of some high resolution spectrographs. Spectrometer name Parameters on ELT Resolving power Sky-projected aperture (sq.arcsec) Minimum wavelength (nm) Maximum wavelength (nm) Spectral coverage S.E.P. GIANO 50,000 0.03 950 2400 85% 0.01 NIRPS 100,000 0.02 980 1800 100% 0.02 HARPS 120,000 0.09 380 690 100% 0.09 CARMENES 100,000 0.10 520 960 100% 0.09 ESPRESSO 130,000 0.51 380 780 100% 0.69 ESPRESSO*BVRI 100,000 0.66 400 830 100% 0.69 G-CLEF7 (GMT) 130,000 0.37 350 950 100% 0.69 CODEX 130,000 1.6 370 720 100% 2.0 CODEX*BVRI 100,000 1.9 400 830 100% 2.0 HIRES-BVRI 100,000 1.9 400 830 100% 2.0 HIRES-ZYJH 100,000 1.9 830 1800 95% 2.0 MTHR8 (TMT) 100,000 4.8 310 1000 100% 8.1

To better understand the intrinsic characteristics and limits of seeing-limited HR spectrometer it is convenient to compare in some details the designs of ESPRESSO and of the instruments so far developed for ELTs.

Figure 24 shows a comparison between G-CLEFand ESPRESSO. The two spectrographs have the same value of S.E.P. and are remarkably similar. They both have a grating-echelle 1.2 meters long; and two cross disperser arms with F/2.5 cameras each feeding a detector of 92x92 mm2. The main differences are the length of the slit (longer in ESPRESSO) and the wavelength coverage (broader in CLEF). The two quantities compensate each-others, i.e. the design of G-CLEF gives higher priority to the wavelength coverage; and it accepts a smaller entrance aperture.

HIRES -ZYJH

> VINEM

v>

sIIP ₃ CRDJ

I

`

CRDY

CROZ Pnam,1114/4 CRDH

°''

'

Foltl, DichroicJH

,

Pñsm ZY

TrSflSl COI Irin ator JH _Fold-Flolom

collimalo

'Fold Slit image

EoHelle ZY 5.5 m E co N 3.6 m 4.0 m Blue evendku. Blue4anels Maul collimate,' -(anuFlpass) Bm 4

\

ÿ 41eensecml. Field lens -

--VLT ESPRESSO

114 e Ilene gratine

GMT G -CLEF

«1.C. 1.1 E ,....,... 3.7 m mm,

Figure 24 Layouts of the G-CLEF (left panel) and ESPRESSO spectrometers

Figure 25 shows a comparison between ESPRESSO and CODEX, the instrument proposed during the first phase-A of the ELT. The S.E.P. of CODEX is a factor of 2.9 larger because it has a longer echelle grating (1.6 m) and a much longer slit that requires 4 cross disperser arms. Another important difference is the main collimator: CODEX uses three aspheric mirrors, while ESPRESSO uses a single parabolic mirror. For this reason the collimator of CODEX is more compact.

Figure 25 Layouts of the CODEX (left panel) and ESPRESSO spectrometers

Figure 26 shows a comparison between ESPRESSO and HIRES. The latter has a larger S.E.P. because it has a longer echelle grating (1.6 m) and a longer slit that requires 4 cross disperser arms. Apart for the above differences, the overall layout of HIRES is similar to ESPRESSO; indeed HIRES could be called a “DOUBLE-ESPRESSO”.

F

2

D

2

A= 60 cine

_{total =}

S.E.P.

camera tel

1

39 in

\

Amax) (7]a-coverage ) (Fcamera) 2 ( Dtel l

2 Atotal = 60 cm2

R l

aperture 2

r

1og2 10P 1 sq. aresec ami. 100% 1 39 m

TMT-MTHR

HIRES -ZYJH 5.5 m

Figure 27 shows a comparison between the layouts of HIRES and MTHR, the concept design for the HR spectrograph of the TMT. MTHR achieves a much larger value of S.E.P. because it employs two echelle-gratings of 3.5 m. Its overall design is similar to a classical echelle spectrograph: clean, simple but very large.

Figure 27 Layouts of the HIRES (left panel) and MTHR spectrometers

9. GENERAL SCALING LAWS FOR SPECTROMETER DESIGN

The quantities that drive the size and the cost of a spectrometer are intrinsically related to the fundamental principle of throughput conservation (A*Omega = constant). The key parameters are as follows

i. The length of the main-dispersion grating (echelle) that determines the product between the spectral resolving power (R) and the sky-projected width (theta-slit) of the slit. Numerically:

ii. The total detector-area needed to collect the complete spectrum (Fcamera is the aperture focal ratio of each

camera)

iii. The number of separated arms/channels that are needed to cover the complete spectrum. This parameter is determined by the practical limits on the size of the beam at the cross-disperser gratings (DCRD) and on the

maximum range of angles that can be accepted by a camera. Moreover, the number of CRD-arms that can be coupled to a single echelle grating is limited by the need of creating a separate image of the pupil on each CRD

2nd half of slit )Cr.Disp)(Camera )Cr.Disp)(Camera 0.43-0.58 0.58-0.84

J

L 0.58-0.69 0.69-0.84 J CODEX -like s a of slit ó (Cr.Disp)(:::: ) L U ¡

\

I 0.43 -0.58 n 58-(1 s4 BB fibers (L <20 m) Double espresso (Cr.Disp)I \ \Camera) )Cr. DisnI r-Camera) 0.43 -0.50 0.50-0.58

In the above equation Nslits is the number of separated slit images created after the main collimator. This

parameter was introduced in the design of CODEX as an alternative solution to split in wavelengths the CRD channels. The conceptual difference is schematized in Figure 28.

The net result is that a given spectral coverage and aperture on sky can only be achieved using the same echelle grating, the same number of CRD arms and the same A*Omega for each camera.

Figure 28 Schematic comparison between slit-spitting and wavelength-splitting solutions in HR spectrometers

10. ACKNOWLEDGEMENTS

This work was financially supported by the Italian Institute of Astrophysics through the grant and “T-REX: Italian Technologies for ELT”

REFERENCES

[1] Pasquini, L.; Cristiani, S.; Garcia Lopez, R.; et al.; “CODEX”, SPIE, 7735E-2FP (2010)

[2] Origlia, L.; Oliva, E.; Maiolino, R.; et al..; “SIMPLE: a high-resolution near-infrared spectrometer for the E-ELT”, SPIE, 7735E-2BO (2010)

[3] Marconi, A.; Di Marcantonio, P.; Maiolino, R.; et al.; “ELT-HIRES, the high resolution spectrograph for the

ELT: results from the Phase A study”, SPIE, 10702-70 (2018); this conference

[4] Cabral, A.; Aliverti, M; Coehlo, J.; et al.; “ELT-HIRES the high resolution spectrograph for the ELT: the design

of the front end”, SPIE, 10702-358 (2018); this conference

[5] Tozzi, A.; Oliva, E.; Sanna, N.; et al.; “ELT-HIRES, the high resolution spectrograph for the ELT: the

IFU-SCAO module”, SPIE, 10702-319 (2018); this conference

[6] Oliva, E.; Delabre, B.;Tozzi, A.; et al.; “Toward the final optical design of MOONS, the Multi-Object Optical

and Near-Infrared spectrograph for the VLT”, SPIE, 9908-7RO (2016)

[7] Fürész, G.; Epps, H.; Barnes, S.; et al.; “The G-CLEF spectrograph optical design”, SPIE 9147-9GF (2014) [8] Vogt, S.; Rockosi, C.; Cowley, D.; “The MTHR Spectrometer”, TMT.INS.PRE.07.012.REL01 (2007)