Wavelength calibration from 1 − 5µm for the CRIRES+ high

Wavelength calibration from 1 − 5µm for the CRIRES+

high-resolution spectrograph at the VLT

U. Seemann

*,a

, G. Anglada-Escud´e

, D. Baade

, P. Bristow

, R. J. Dorn

, R. Follert

, D.

Gojak

, J. Grunhut

, A. P. Hatzes

, U. Heiter

, D. I. Ives

, P. Jeep

, Y. Jung

, H.-U. K¨auﬂ

F. Kerber

, B. Klein

, J.-L. Lizon

, M. Lockhart

, T. L¨owinger

, T. Marquart

, E. Oliva

, J.

Pauﬁqu´e

, N. Piskunov

, E. Pozna

, A. Reiners

, A. Smette

, J. Smoker

, E. Stempels

, E.

Valenti

Georg-August Universit¨at G¨ottingen, Institut f¨ur Astrophysik, Friedrich-Hund-Platz 1,

D-37077 G¨o ttingen, Germany;

European Southern Observatory, Karl Schwarzschild Str. 2, D-85748 Garching, Germany;

Th¨uringer Landessternwarte Tautenburg, Sternwarte 5, D-07778 Tautenburg, Germany;

Uppsala Universitet, Department for Physics and Astronomy, Box 515, 751 20 Uppsala,

Sweden;

INAF-Arcetri Observatory, Largo E. Fermi 5, I-50125 Firenze, Italy

ABSTRACT

CRIRES at the VLT is one of the few adaptive optics enabled instruments that oﬀer a resolving power of 10

from 1 − 5 µm. An instrument upgrade (CRIRES+) is prop osed to implement cross-dispersion capa bilities ,

spectro-polarimetry modes, a new detector mosaic, and a new gas absorption cell. CRIRES+ will boost the

simultaneous wavelength coverage of the current instrument (∼ λ/70 in a single-order) by a factor of & 10

in the cross- dispe rsed conﬁguration, while still retaining a 10 arcsec slit suitable for long-slit sp e ctroscopy.

CRIRES+ dramatically enhances the instrument’s observing eﬃciency, and opens new scientiﬁc opportunities.

These include high-precision radial-velocity studies on the 3 m/s level to characterize extra-solar planets and their

athmospheres, which de mand for specialized, highly accurate wavelength calibr ation techniques. In this paper,

we present a newly developed absor ption gas-cell to enable high-precision wavelength calibration for CRIRES+.

We also discuss the strategies and developments to cover the full operational spectral range (1−5 µm), employing

hollow-catho de emission lamps, Fabry-P´erot ´etalons, and absorption gas-cells.

Keywords: astronomica l instrumentation, infrared spectrographs, high spectral resolution, upgrade of existing

VLT instruments, infrared spectrometers, wavelength calibration, radial-veloc ities, extra-solar planets

1. INTRODUCTION

The CRyogenic high-resolution InfraRed Echelle Sp ectrograph (CRIRES

) is installed on UT1 (Antu) of ESO’s

Very Large Telescope (VLT), and is in operation since 2006. The spectrograph features a nominal resolving

power of R = 10

(using a 0.2

′′

slit) over the entire 1 − 5 µm near-infrared (NIR) wavelength domain. CRIRES

is equipped with a MACAO

adaptive optics module to enhance the energy concentration within the 0.2 × 30

arcsec slit. A single order is dispersed on a fo cal plane mosaic of 4 1k x 1k InSb Aladdin detectors,

resulting in

a simultano us wavelength coverage of ab out λ/70 (a t 1µm) to λ/50 (at 5µm).

The CRIRES upgrade project aims to increas e the obser ving eﬃciency of the instrument, transfo rming it

into a cross-dispersed echelle spec trograph with new capabilities, termed “CRIRES+”. The major changes

implemented by the upgrade comprise a cross-dispersion unit in place of the existing prism pre-disperser, a

new focal plane array of 3 HAWAII-2RG 2k×2k IR detectors, introduction of a new spectro-polarimeter, and

additional means of wavelength calibration to enable high-precision velocimetry using a gas-absorption cell. The

Further author information:

* e-mail: [email protected] ingen.de

 

Figure 1. Optical layout of the upgraded CRIRES, termed “CRIRES+”. The instrument itself is located in a cryostat (dark

shaded area), interfaced by a dichroic entrance window. The upgraded optics

1, 2

consist of cross-dispersion capabilities (via

a grating wheel), camera, and a new 3-chip detector mosaic. The adaptive optics module, the instrument derotator, and

calibration devices are comprised in a warm part between the telescope and the spectrograph (light grey area). Wavelength

calibration sources are inserted into the optical b eam just after the telescope. This is where the new absorption gas-cell

is mounted (red cylinder), as well as feeds for emission line lamps, ´etalons, and a polarimeter.

Figure 2. ThAr catalog

(top) and corresponding histogram (bottom) of lines used for the standard CRIRES(+) wavelength

calibration. The density of availabe lines rapidly decreases towards longer wavelengths, rend ering ThAr unsuitable as a

calibration source for the entire NIR range. ThAr lines greatly vary in intensity, so that eﬀectively only a subset can be

used.

CRIRES+ project and its main components are further described elsewhere in these proceedings.

1, 2, 6, 7

Figure 1

shows an overview of the new optical layout of CRIRES+ and its ma in components.

This paper focuses on the new absorption gas-cell for CRIRES+, a nd outlines the wavelength calibration strat-

egy for CRIRES+ comprised of various refer ence sources. It is organized as follows: §2 describes the wavelength

calibration in the NIR and the curr ently used techniques with CRIRES. In §3, we present a newly developed

gas-cell for precisio n ca libration with CRIRES+, and discuss its design, construction, and characterization. The

concept for wavelength calibrators with CRIRES+ is described in §4. Concluding remarks are pr ovided in §5.

2. WAVELENGTH CALIBRATION IN THE NIR

High-resolutio n spectroscopy demands for precise wavelength c alibrators. The well established technique in the

optical domain using commercially available hollow-cathode lamps (HCLs) such as ThAr provides unreso lved

emission lines even at the high spectr al resolving power of 100 000 oﬀered by CRIRES. ThAr as the default

wavelength referenc e for optical spectrographs can provide a line position accuracy

of 16 − 82 m/s and up to

∼ 10 m/s with carefully selected lines

in the 400 − 700 nm range for single lines. In a stable environment, ThAr

is thus suitable for high-precision applications such as doppler measurements on the m/s scale using large echelle

formats. Blended lines and large intensity diﬀerences between the metal lines and the lines originating from the

ﬁll gas can be handled by careful selection and masking during calibration, however, two major problems exist

with HCLs. First, HCLs are intrinsically unstable as their individual line intensities vary fro m lamp to lamp, and

are a function of age, pressure, and operating current. These eﬀects degrade the stability as a reference source.

Over their lifetime, the gas pres sure inside a HCL decreases, resulting in wavelength changes of the gas line s.

The metal lines change in intensities as cathode material is sputtered oﬀ, and the geometr y of the cathode thus

depe nds on lamp age. Accordingly, the illumination of the instrument by the HCL varies over time. Secondly,

Figure 3. Absorption spectrum of the CRIRES N

O gas-cell

in the near-infrared K–L bands. N

O oﬀers a number of

transition bands redwards of 2µm as a wavelength reference in selected, small wavelength regions.

ThAr does not exhibit a homogenous density of spectral features in terms of wavelength. While the line density

is high in the optical (λ < 1µm), atomic transitions become infrequent at longer wavelengths. In the NIR, the

line dens ity of ThAr decreases so that it is not suitable to provide a wavelength so lution for CRIRES beyond

∼ 2.5µm.

Nevertheless, a ThAr HCL is a C RIRES general purpose wavelength refere nc e from 0.95 − 2.5 µ m,

and yields a level of ∼ 50 m/s precision for the line positions. Fig. 2 shows the ThAr emission lines for calibrating

CRIRES+ spectra.

As with all NIR spectrographs, wavelength solutions for CRIRE S at longer wavelengths rely on sky emission

lines, telluric absorption lines, and absorption lines by specializ e d gases enclosed in reference gas cells. While the

hydroxyl night-time airglow (OH) is present over a wide wavelength region in the NIR

and well characterized,

at R = 10

these feature s are mostly too weak and too infrequent to provide a generally reliable wavelength

standard for C RIRES(+). Similarly, molecules prese nt in earth’s athmosphere provide a multitude of absorption

lines ac ross the NIR, dominated by the strong a bsorber H

O, among o thers, who se transition ba nds deﬁne the

boundaries of the band passes. The presence and strength of telluric calibrators, however, are strongly site-

depe ndent and prone to variations on diﬀerent timescales from hours to years. Thus, a detailed understanding

and modelling of the atmospheric conditions, based on simultaneously measured sounding data , is required, as

eg. high-altitude winds induce signiﬁcant uncertainties (≤ 25m/s) on the line positions.

Telluric molecules such

as CO

can be used with CRIRES and have demons trated radial velocity precisions o f 5 − 10 m/s,

but are

restricted to narrow wavelength intervals.

The beneﬁts of the wavelength calibration technique using telluric features can be mimicked by exploiting

ro-vibr ation lines from a selected mo le cular gas. Such a captive gas (or a mixture of several) is then kept

under well controlled conditions (inside a gas-cell vessel), and this vessel is brought into the lig ht entering the

spectrograph at the time of observations (analogous to the iodine method

in the o ptical domain). T he eﬀect

of such a gas-cell is thus similar to the imprint of mole c ular features from the atmosphere, on the actual target

spectrum itself, but for selected (non-telluric) species and without the inﬂuences of the perturbing dynamical

processes in the atmosphere. By using a reference gas in the optical path shared with the light coming from the

astronomical target, waveleng th calibratio n expo sures can be recorded simultanously (the gas-cell is illuminated

by the target), or oﬄine with a broadband calibration lamp. A detailed knowledge of the r eference gas spe ctrum

is gene rally required, and precise line positions be known from laboratory mea surements if absolute calibration

is desired. The gaseous absorption lines can then be used to track all instrumental instabilities and drifts that

follow the gas-cell in the optical chain. CRIRES has o ﬀered two such absorption gas cells for simultanous and

oﬄine ca libration. A CO cell to cover the CO bands around 2.1µm in the K- band, and a N

O g as-cell (Fig. 3)

with usable absorption features in selected rang es between 2.6 − 4.7µm.

While the latter appro ach is a viable option in the L- and M-bands, the range covered by N

O is very limited.

In the K-band, N

O transition bands are rather weak, while the pro minent (and wider) bands in L and M are

easily saturated (and thus do not facilita te determination of accurate line positions nor proﬁles).

The la ck of ca libration sources in the NIR comparable to those in the visua l regime poses a major challenge

to the exploitation o f these wavelength regions using high-resolution spectros copy. One o f the goals of the

CRIRES+ project is to improve the current wavelength calibration situation. In an on-going eﬀort, we seek

to provide enhanced wavelength reference so urces for CRIRES+, both for general purpose calibration (of the

facility instrument modes) and for specialized applications (high-precision radial velocities in the NIR).

3. THE CRIRES+ PRECISION RADIAL VELOCITY GAS-CELL

The gas-ce ll wavelength calibration method is the only technique that provides a wavelength reference information

and enables monitoring of instrumental eﬀects such as drifts in the dispersion re lation and changes in slit

illumination simultaneously with the astronomical observation, both in time and co-located on the detec tor. This

means that the wavelength so lution is recorded at the very same time and under the insta ntanous conditions

as the stellar light is rec orded. Within the instr ument, the exact same optical pa th is used fo r both the source

of wavelength reference and the celestial object, s o that the spe c tral features of both contributions end up on

the same pix els in the focal plane. When modelled properly,

15, 16

this approach a llows for very high precision

(relative) wavelength calibration and long-term repeatability as ne eded for radia l velocimetry work on the m/s

level.

For HARPS-type spectrogr aphs, a high degree of illumination stability can be achieved by appropriately

feeding the instrument via optical ﬁber s. As CRIRES is not a ﬁber-fed spectrograph, it is important to provide

a stable slit-illumination (aided by the adaptive optics), or to monitor the instrumental proﬁle variations during

the observation, by the use of the gas-cell’s s pectral lines.

For high-prec ision wavelength calibration with CRIRES+, we have developed a new NIR absorption gas-cell

primarily aimed at exoplanet precision Doppler experiments. The design considerations of the reference gas for

CRIRES+ are summarize d as follows. The ﬁlling gas must provide a set of dense , homogenous a nd sharp spectral

lines over a wide wavelength range, as to provide a high density of information usable to construct the dispersion

relation. At the same time, the radial velocity information content

15, 17, 18

of the stellar spectrum under study

must match the anticipated RV precision. This, in turn, is subject to a multitude of parameters, a mong them

attainable SNR of the target obse rvations, availability of spectral features, ie. spectral type , wavelength range

and instrumental (spectral) resolution. One of the science drivers

of CRIRES+ is the search of super-Earths

around low-mass stars, using the RV technique. To accomplish a ∼ 3 m/s level of RV precision for M-s tars with

CRIRES+ in cross-dispersion, we identiﬁed a suitable reference gas that fulﬁlls the requirements in the K-band.

Our choice of ﬁlling gas for the CRIRE S+ short-path gas -cell is a compound of three molecular species. The

mixture is tailore d to provide an optimum wavelength coverage in the H and K-bands, with an emphasis on a core

region in the K-band. The tr ansmission spectrum is generated by acethylene (C

), ammonia (NH

), and an

isotopologue of methane (

), and draws from the experience of a previous NH

gas-cell for CRIRES,

and

experiments with methane isotopo logues on IRTF/CSHELL.

Methane-13 (

) features the carbon atom

with an additional neutron co mpared to naturally abundant methane (

). The latter is a major contributor

to the telluric transmission spectrum, and could thus not be disentangled from a captive methane-12 held under

controlled conditions (ie. in the gas-cell). However, methane-13 has a diﬀerent reduced mass due to the heavier

nucleus, responsible for shifting its rotational transitions signiﬁcantly. Methane-13 is a stable isotopologue and is

easily disting uished from standard methane-12. An important point is also that methane has well characterized

absorption features with high frequency precision.

20, 21

The partial pressur es of the three molecular species ar e

listed in Table 1. We have optimized these for the CRIRES+ conﬁguration (R = 10

, ≥ 200 nm continous

wavelength coverage in the K-band, available path-length) to allow for the best “aspect ratio”, ie. the optimum

line-depth (∼ 80%) while retaining the smallest p ossible line-width (unresolved at R = 10

3.1 Design and construction

The gas-cell is designed to utilize the maximum available clearance in front of the CRIRES+ instrument cryostat.

This renders the length of the cell vessel to be no longer than ≈ 18 cm (constrained by the telescope ﬂange and

the instrument derotator), and hence also deﬁnes the optical path-length through the reference gas. The gas-cell

module is composed of a mounting structure and the actual gas-cell, with the latter be ing an all-glass cylinder

ﬁlled with the c aptive gas mixture. The primary design driver for the cell vessel are a) a long term, gas-tight

sealing (10 years), and b) a high optical eﬃciency. Experiments have shown that standard v iton se als and

glued windows cannot provide the long-term stability, and were therefore discarded. Instead, we chose the same

Figure 4. Filled gas-cell cylinder with fritted q uartz windows, containing a mixture of

, NH

, C

. The “glas nose”

on the top middle of the cell is a remainder of the ﬁlling stem used to insert the gases.

material (quartz) for both the cylinder and windows, which allows for fritting them together with a glass mold to

form a sea mless, homogenous joint. The optically polished low-OH quartz glass windows are anti-reﬂection coated

for the K-band, and mutually tilted with respect to each other so that internal fringing eﬀects are minimized.

Filling the cell with the appropriate partial gas pressure s is done after evacuating the cylinder to below 10

−3

mbar

using a glass stem attached to the cylinder body. The mixture of gases is then achieved by subsequently ﬂooding

the cell with the individual species , at the desired partial pressures. Figure 4 shows the sealed and ﬁlled gas-c ell

cylinder.

Our tests over a period of four months have proven that the cell is well sealed, and no change in gas pressure

could be detected. Four glas-air interfaces by the two w indows inevitably result in a reduced throughput of

the instrument. We measured the total trans mis sion of the complete cell to b e above 90% in the K-band, a nd

around 85% in the Y– H bands (on average, including windows and gas ﬁlling). This only marginally reduces the

eﬃciency of CRIRES+.

3.2 Gas-cell mount

CRIRES(+) features a calibration slide close to the VLT Nasmyth foca l plane, before the light enters the derotato r

and the adaptive optics module. This movable stage holds two slots for gas- cells with s hort optical path-length

that can be inserted directly into the optical beam. The spectral features of the re ference gas is thus imprinted

onto the stellar light immediately after leaving the telescope. To host the sealed glas vessel in a deﬁned and

reproducible location and orientation with respect to the telescope beam, we have designed a spe c ialized mount

for the gas-cell (see Fig ure 5 ). The glass c orpus is ﬁxed in two rings, where four rubber padded screws e ach

Table 1. Molecular species and their partial pressures for the CRIRES+ precision radial velocity gas-cell. The total

pressure of the ﬁlled cell is ∼ 300 mbar, with a path-length of 18 cm. The fourth column d enotes the average line-width

of t he corresponding species, as measured from high-resolution spectra. In the last column, t gives t he percentage of

wavelength coverage with usable absorption features deep er than 10% from the gas mixture, for each of the J HK b an d

range deﬁnitions.

Molecule Pressure line-width t

[mbar] [km/s]

[%]

Acethylene

∼ 60 1.791 ± 0.010 K: ∼ 76%

Ammonia

∼ 90 1.978 ± 0.011 H: ∼ 67%

Methane

∼ 150 1.797 ± 0.022

J: < 10%

Figure 5. Design of the CRIRES+ short gas-cell mount. The fragile glass cylinder is held in a stress-free conﬁguration.

Side view (left) and cut front view (right). Parts are labelled as follows: protective aluminum sh ells (1-2), gas-cell cylinder

(3), mounting posts and ring ﬁxtures (4, 8), straylight removal masks (5-6), dust protection caps (7), stress-free alignment

(9-11), and locking mechanism (12-14).

hold and align the cylinder almost free of stress. This proved especia lly important during shipping and handling

to prevent cracks o r breakage of the glass and seals, and ensures long-term endurance under changing ambient

conditions. Two aluminum half-shells further protect the cell mechanically and provide shielding from external

straylight.

3.3 Laboratory FTS measurements

A prototype gas-cell has been succ e ssfully built and ﬁlled. We have started to characterize its transmission

spectrum by means of a Fourier transform spectrometer (FTS). Spectra of the gas-cell were recor de d with

a Bruker IFS1 25HR device at a resolution between 0.04 − 0.0065 cm

−1

(corresponding to resolving powers of

250 000−1 538 000 at 1 µm). The ce ll is measured at an ambient pressure < 0.1 mbar to supress the athmospheric

contributions by water vapor etc. within the instrument. A blackbody-like continuum lamp is used to illuminate

the gas-cell. Normalized spectra of the H and K-bands are on display in Figures 6 and 7, respectively.

In the H-band (1.47 − 1.83µm), the strongest absorption feature s in the cell’s s pec trum (Figure 6, bottom

panel) originate from acethylene in two very deep (∼ 90%) bands around 1.53µm. Similarly prominent contribu-

tions in the neighbouring range (1.5µm) are a set of ammonia lines, partly mixed with acethylene. The largest

wavelength range in this band, starting at 1.63µm, is covered by deep (≤ 80%) line systems fr om methane-13,

with minor contributions by ammonia. Overall, the line coverage of the three ga ses in the H-band is quite

complete, with only a small gap around 1.6µm. The upper panels in Figure 6 show FTS spectra of the individual

gases ammonia, acethylene, and methane-13. These are measured sepa rately under diﬀerent ambient conditions,

and include some water vapor contributions from a normal athmosphere during the experiment (in the case

of NH

and C

). Note that the individual species have lower gas pre ssures than pres ent in the prototype

gas-cell (Table 1). For guidance, the top panel in Figure 6 gives the telluric spectrum of the same FTS setup as

measured with the gas-cell, and represents the ambient air from the optical path between so urce and detector.

Note that this telluric spectrum is much weaker and cleaner than a corres ponding spe ctrum measured through

Earth’s athmosphere with the VLT, due to the much smaller pa th-length. The shaded areas in Figure 6 denote

wavelength ranges o utside the H-band deﬁnitions .

Our FTS measurements of the K-ba nd (1.95 − 2.4µm) are plotted in Figure 7. This is the wavelength range

for which the gas pressures of ammonia and methane-13 are designed to provide the optimum line depths.

Acethylene (third panel from top) has only very weak bands in this range and does not add signiﬁcantly to

Figure 6. Laboratory FTS measurements of the CRIRES+ gas-cell species in the H-band. Plotted from top to bottom are

the ambient athmosphere (mostly contributions by wator-vapor, as present during the gas-cell measurement), ammonia

and acethylene gas (both includ ing residual water vapor in the setup), methane-13, and the compound gas-cell mixture

as measured from the ﬁlled CRIRES+ cell.

Figure 7. K-band measurements of the ﬁlled CRIRES+ gas-cell and its individual species. See Figure 6.

Figure 8. K-band regions with transition bands by ammonia (top, blue), acethylene (middle, pink), and methane-13

(bottom, green) as measured with the ﬁlled CRIRES+ cell. Also shown in detail are single line proﬁles of the corresponding

molecule (right panels). In all three regions, some smaller lines from the other two molecules may be present, respectively,

as the wavelength regions of their absorption features partly overlap. The line proﬁles are mostly resolved at the resolution

of these FTS measurements (R = 1 538 000), yet for CRIRES+ (R = 100 000) will q ualify as largely unresolved.

the spectrum of the mixture (botto m panel). The major constituent in the K-band is methane-13, achieving

large line-depths of mor e than 95% in the high-resolution FTS spectra. The central part of the band is further

populated by str ong ammonia lines, w hich also has features at around 2.0µm and be low. In the K-band, the

coverage with reference gas lines is complete but for a ∼ 30nm range at 2.1µm. In total, more than 75% of the

entire K-band a re po pulated with suitable gas lines deeper than 10% by any of the three reference gas species.

The prototype CRIRES+ gas-cell demonstrates the power of the selected gas mixture in the H and K-bands.

For the remaining operating wavelength domains of CRIRES+, other wavelength reference sources are fo reseen

(Section 4).

To achieve the demanding RV precision of CRIRES+ in the K-band, the spectra l lines exploited for wavelength

calibration must enable a highly accurate measurement of their line positions. This immediately translates into

sharp, unresolved referenc e lines at the given resolving power (R = 10

) of CRIRES+. Our gas-cell mixture is

designed such that the line-widths of all three molecules is smaller than 3 km/s, corresponding to the resolution

element of CRIRES+. We measured the line-widths in the FTS spectra with the highest attainable reso lution

for all three gases in selected, strong absorption bands. At ∆λ = 0.0065 cm

−1

(R = 1 538 000) these lines are well

discernible for

, NH

, and C

(though numerous are blended, particularly in

). Line-widths are

typica lly closely distributed around 2 km/s (cf. Table 1), with very similar average widths of 1 .797±0.022 km s

−1

1.978 ± 0.011 km s

−1

, and 1.79 1 ± 0.010 km s

−1

(for

, NH

, and C

, respectively). Figure 8 shows these

transmission features in narrow re gions of the K-band separately for all three constituents, and also illustrates

corres ponding typical line proﬁles.

4. WAVELENGTH CALIBRATION STRATEGY FOR CRIRES+

CRIRES+ will be a facility instr ument at the VLT, addressing a large variety of scientiﬁc questions. Appropriate

wavelength calibration sour ces must b e provided for the entire working range of the spectrograph (0.95− 5.3µm).

Moreover, wavelength ca libration needs to support all observing modes of CRIRES+. In combination with

the limited wavelength coverage of the individual sources, this leads to a numb er of constraints the CRIRES+

calibration plan must match:

1. The ex isting HCL techniques as well as the gas-cells (including the new CRIRES+ c ell) cannot operate in

polarimetry mode. This is a re sult of the opto -mechanical layout, where all calibra tion devices are located

on a movable stage in front of the Nasmyth focus. At the same time, the pola rimeter unit

is also located

on the same stage, and thus cannot be inserted at the same time as the gas-cells or the HCL ﬁber -feeds.

2. The ThAr HCL is only suitable for the Y–K bands (see Section 2).

3. The available and new gas -cells lack bro adband coverage in, or do not operate in the Y–J and L–M bands.

4. The available techniques in the L–M bands are not satisfactory (Section 2).

We address these issues by expanding the suite of calibration sources for CRIRES+. The concept for calibrators

comprises the following changes and additions. First, the ThAr HCL will be superseded by an UNe HCL.

The characteristics of UNe have recently been studied, and prove to be very co mparable to ThAr. A major

advantage of UNe is that it provides signiﬁcantly more lines (three to ﬁve times) in the NIR, and thus increases

the suitability of the HCL method for CRIRES+ in the J, H, and K-bands. UNe catalogs

22–24

are now ava ilable

with line lists covering the Y–L bands. These HCLs are readily available and their line density is suitable for

CRIRES+ as a general purpose calibrator up to (including) the K-band, where the tra nsmission of commercially

availabe UNe bulbs, due to the quartz window, in combination with standard optical ﬁbers breaks down. The

implementation of the UNe HCL in CRIRES+ (aided by the cross-dispersed spe c tral format) will signiﬁcantly

improve the dispers ion s olution.

Secondly, to enable wavelength c alibration for polarimetric observations, an additional reference source with

high stability cover age is considered. Etalons provide a homogenously spa ced set of lines, where the line sharpness

(“ﬁnesse”) and contrast can b e controlled by the quality and reﬂectivity of the reﬂective coating. At the sa me

time, the peak-to- peak distance (free spe ctral range) is a function of the cavity length. This allows to tune the

free spectral range to the resolution of the instrument. For CRIRES+, we a re developing stabilized Fa bry-P´erot

´etalons to provide suitable oﬄine ca libration lines over the entire NIR range. Two etalon units are studied, where

their usa ble wavelength range is driven by the availability of suitable broadba nd cavity coatings. Although a small

ﬁnesse (∼ 10) satisﬁes a level of 5 m/s calibration precision, a broadband performance over several NIR bands (eg.

YJ and HK) is not easily achieved. Furthermore, challenges include achieving the necessar y long-term stability

of such a unit. The s tability is directly linked to the temper ature stability of the etalon, and thus de mands for

an actively controlled environment inside a vacuum tank. For CRIRES+, ﬁber-fed ´etalons are under study. This

poses stringent requirements on the etalon illumination and its stability by the optical ﬁbers. A quasi-continuum

laser-driven light source (LDLS) is used to provide the necessary white-light to the etalon. These FP ´etalons are

compatible with the CRIRES+ po larimetric modes when the ca libration light is fed into the optical path before

the movable stage, where the polarimeter is mounted on. This ca n be accomplished by exploiting the existing

calibration unit outside of the telescope beam. The output light of that unit is redire c ted into the beam by a

folding mirror, before the movable stage. To obtain an absolute wavelength solution with the etalon, the UNe

HCL will be employed as a reference. This is because the line-position of the etalo n peaks is not known a priori

(albeit their relative pos itions ar e), but can be assigned by comparison with a HCL line. The development of

an etalon for CRIRES+ will signiﬁcantly enhance its dispersion solution across the entire spectral format, by

providing homogenously and densely distributed reference lines. With a dedicated and optimized etalon unit,

this technique may also overcome the lack of calibrators in the L and M-bands.

In a third approach, we are also s eeking to further advance the gas-cell desig n and to expand the spectral

coverage beyond the H and K-bands. Absorption gas-cells with an increased optica l path-le ngth to several meters

are under study. Such a cell could be ope rated with ﬁlling gases at much lower pressures, yielding signiﬁcantly

reduced line-widths for gasous spe cies that at short path-lengths a) provide unsuitably broad spectral lines, or

b) have too weak abso rption features due to small absorption coeﬃcients.

The CRIRES+ suite of wavelength reference sources als o c omprises the existing N

O gas-cell for use in

selected L&M wavelength ranges. Also supported are the established methods of using sky lines and telluric

lines to construct the dispersion solution in wavelength domains where applicable (cf. Section 2).

5. CONCLUSION

The CRIRES+ project will upgrade VLT/CRIRES into a cross-dispersed, high-eﬃciency NIR spectrograph at a

resolving power of 10

. New capabilities such as high-precisio n radial velocity measurements o n low-mass stars

and po larimetric observa tions call for new and improved wavelength calibration techniques across the operating

domain from 1 − 5µm.

To satisfy the corresponding req uirements, we have develop e d a ne w CRIRES+ absorption gas-cell speciﬁcally

tailored to enable RV experiments with M-stars on the 3 m/s level in the K-band. We have designed, built, and

characterized a prototype gas-cell for CRIRES+ ﬁlled with a compound of ammonia, acethylene, and methane-

13 molecular spec ies. This cell provides a forest of absorption lines in the H and K-bands for simultaneous

wavelength calibration, evidenced in high-reso lution FTS spectra. Our data show that the population of lines

covers a high fraction of the ba ndpasses, and that the gas line- w idths are smaller than can be resolved with

CRIRES+, thus providing high-precision wavelength markers.

We have also outlined the wavelength calibration strategy and our de velopments to expand the suite of

frequency calibration sour ces for CRIRES+. This includes an improved UNe-HCL over ThAr, the new gas-cell

for precision applications, and the design of Fabr y-P´erot ´etalons to cover the Y–M bands. The set of advanced

wavelength calibrators will enable CRIRES+ to e xplore the NIR regime in unprecedented detail.

ACKNOWLEDGMENTS

The CRIRES+ co nsortium acknowledges funding from the German federal ministry of education and research

(BMBF) and the Wallenberg founda tion, Sweden.

REFERENCES

[1] O liva, Ernesto et al., “Concept and optical design of the cross-disperser module for CRIRES+,” Proceedings

of the SPIE, 9147-289 (2014).

[2] Lizon, Jean-Louis et al., “Opto-mechanical design o f a new cross dispersion unit for the CRIRES+ high

resolution spectrograph for the VLT,” Proceedings of the SPIE, 9147-290 (2014).

[3] K ¨auﬂ, H.-U., Ballester , P., Biereichel, P., Delabre, B., Donaldson, R., Dorn, R., Fe drigo, E., Finger, G.,

Fischer, G., Franz a, F., Gojak, D., Huster, G., Jung, Y., Lizon, J.-L., Mehrgan, L., Meyer, M., Moor wood,

A., Pirard, J.-F., Pauﬁque, J., Pozna, E., Siebenmorgen, R., Silber, A., Stegmeier, J., and Wegerer, S.,

“CRIRES: a high-resolution infrared spectrograph for ESO’s VLT,” in [Society of Photo-Optical Instrumen-

tation Engineers (SPIE) Conference Series], Moorwood, A. F. M. and Iye, M., eds., Society of Photo-Optical

Instrumentation Engineers (SPIE) Conference Series 5492, 1218–1227 (Sept. 2004).

[4] Pauﬁque, J., Biereichel, P., Donaldson, R., Delabre, B., Fedrigo, E., Franza, F., Gigan, P., Gojak, D.,

Hubin, N. N., Kasper, M. E., Kaeuﬂ, H.-U., Lizon, J.-L., Oberti, S., Pirard, J.-F., Pozna, E., Santos, J.,

and Stroebe le , S., “MACAO-CRIRES: a step toward high-resolution spectroscopy,” in [Advancements in

Adaptive Optics], Bonaccini Calia, D., Elle rbroek, B. L., and Ragazzoni, R., eds., Society of Photo-Optical

Instrumentation Engineers (SPIE) Conference Series 5490, 216–227 (Oct. 2004 ).

[5] Dorn, R. J., Finger, G., Huster, G., Kaeuﬂ, H.-U., Lizon, J.-L., Mehrgan, L., Meyer, M., Pirard, J.-F.,

Silber, A., Stegmeier, J., and Moorwood, A. F. M., “ T he CRIRES InSb megapixel focal plane array detector

mosaic,” in [Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series], Garnett, J. D.

and Beletic, J. W., eds., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series

5499, 510–517 (Sept. 2004).

[6] Follert, Roman et al., “CRIRES+: a cross-dispersed high resolution infrared spectro graph for the ESO

VLT,” Proceedings of the SPIE, 9147-44 (2014).

[7] Lockhart, Matthew et al., “Novel infrared polarimeter for the ESO CRIRES+ instrument,” Proceedings of

the SPIE, 9147-329 (201 4).

[8] K e rber, F., Nave, G., Sansonetti, C. J., and Bristow, P., “From laboratory to the sky: Th-Ar wavelength

standards for the cryogenic infrared echelle spectrograph (CRIRES),” Physica Scripta Volume T 134, 014007

(May 2009).

[9] Palmer, B. A. and Engleman, R., [Atlas of the Thorium spectrum] (1983).

[10] Lovis, C. and Pepe, F., “A new list of thorium and argon spectral lines in the visible,” A&A 468, 1115–1121

(June 2007).

[11] Kerber, F., Nave, G., and Sansonetti, C . J., “The Spectrum of Th-Ar Hollow Cathode Lamps in the 691-5804

nm region: Establis hing Wavelength Standards for the Calibration of Infrared Spectrographs,” ApJS 178,

374–381 (Oct. 2008 ).

[12] Rousselot, P., Lidman, C., Cuby, J.-G., More els, G., and Monnet, G., “Night-sky spectral atlas of OH

emission lines in the ne ar-infrared,” A&A 354, 1134–1150 (Feb. 2000).

[13] Abrams, M. C., Davis, S. P., Rao, M. L. P., Engleman, Jr., R., and Brault, J. W., “High-resolution Fourier

transform spectroscopy of the Meinel system of OH,” ApJS 93, 351–3 95 (July 1994).

[14] Fig ue ira, P., Pepe, F., Melo, C. H. F., Santos, N. C., Lovis, C., Mayor , M., Queloz, D., Smette, A., and

Udry, S., “Radial velocities with CRIRES. Pushing pre cision down to 5-10 m/s,” A&A 511, A55 (Feb.

2010).

[15] Butler, R. P., Marcy, G. W., Williams, E., McCarthy, C., Dosanjh, P., and Vogt, S. S., “Attaining Doppler

Precisio n of 3 M s-1,” PASP 108, 500 (June 1996).

[16] Be an, J. L., Seifahrt, A., Hartman, H., Nilsson, H., Wiedemann, G., Reiners, A., Dreizler, S., a nd Henry,

T. J., “The CRIRES Search for Planets Around the Lowest-mass Stars. I. High-precision Near-infrared

Radial Velocities with an Ammonia Gas Cell,” ApJ 713, 410–422 (Apr. 2010).

[17] Bouchy, F., Pepe, F., and Queloz, D., “Fundamental photon noise limit to r adial velocity measurements,”

A&A 374, 733–739 (Aug. 2001).

[18] Reiners, A., Bea n, J. L., Huber, K. F., Dreizler, S., Seifahrt, A., and Czesla, S., “Detecting Planets Around

Very Low Ma ss Stars with the Radial Velocity Method,” ApJ 710, 432–443 (Feb. 2010).

[19] Anglada-Escud´e, G., Plavchan, P., Mills, S., Gao, P., Garc´ıa-Berr´ıos, E ., Lewis , N. S., Sung, K ., Ciardi, D.,

Beichman, C., Brinkworth, C., Johnson, J., Davison, C., White, R., and Pr ato, L., “Design and Construction

of Absorption Cells for Precision Radial Velocities in the K Band Using Methane Isotopologues,” PASP 124,

586–597 (June 2012).

[20] Nai-Cheng, S., Yao-Xiang, W., Yi-Min, S., Cheng-Ya ng, L., Xue-Bin, Z., and Chu, W., “Iodine and Methane

Stabilized He-Ne Lasers as Wavelength Standards,” in [Precision Measurement and Fundamental Constants],

Taylo r, B. N. and Phillips, W. D., eds., 77 (Aug. 1981).

[21] Boudon, V., Champion, J.-P., Gabard, T., Lo¨ete, M., Coustenis, A., de Bergh, C., B´ezard, B., Lellouch,

E., Drossart, P., Hirtzig, M., Negr˜ao, A., and Griﬃth, C. A., “Methane in Titan’s atmosphere: from

fundamental spectroscopy to planetology,” Europhysics News 40, 17–20 (July 2009).

[22] Redman, S. L., Lawler, J. E., Nave, G., Ramsey, L. W., and Mahadevan, S., “The Infrared Spectrum of

Uranium Hollow Cathode Lamps from 850 nm to 4000 nm: Wavenumbers and Line Identiﬁcations from

Fourier Transfo rm Spectra,” ApJS 195, 24 (Aug. 20 11).

[23] Redman, S. L., Ycas, G. G., Terrien, R., Mahadevan, S., Ra msey, L. W., Bender, C. F., Osterman, S. N.,

Diddams, S. A., Quinlan, F., Lawler, J. E., and Nave, G., “A High-resolutio n Atlas of Uranium-Neon in the

H Band,” ApJS 199, 2 (Mar. 2012).

[24] Sarmiento, Luis-Fernando, Reiners, Ansg ar, Seemann, Ulf et al., “Chara cterizing UNe hollow cathode lamps

at NIR wavelengths for the CARMENES survey,” Proceedings of the SPIE, 9147-194 (201 4).