Instrument Performance

This page describes the performance of the upgraded CRIRES, currently installed at UT3. An overview for the original (pre-upgrade, pre-August 2014) CRIRES can be in the following link.

Sensitivities

Table 1 lists preliminary values for the maximum S/N values in the stellar continuum and the maximum flux on the detector, both derived from commissioning data. They were determined for the spectro-photometric standard star Pi.02 Ori (H=4.2; spectral type A1V) using a 0.4 arcsec slit, adaptive optics under a Turbulence category of 30% (seeing 0.8 arcsec and coherence time 4.1 ms) and an airmass of 1.3. The DIT and NDIT were 15 s and 3, respectively.

 

Table 1. Sensitivities for Pi.02 Ori  
Band
max S/N
peak flux on detector [ADU]
Y (946-1120 nm)
570 11000
J (1116-1362 nm)
530
10000
H (1423-1769 nm)
480
8300
K (1972-2624 nm)
350
5200
L (2869-4188 nm)
- 9600 1)
M (3583-5300* nm)
(* detector cutoff)
- DIT=15s: saturation;
DIT=5s: ~20000s 1)

Please note that IR observations are sensitive to the precipitable water vapor (PWV) in the Earth's atmosphere. Observations in the L and M bands should only be requested for PWV values <= 2.5 mm.

1) Given that in the L and M bands the thermal background is the dominating flux contributor, the maximum integration times (DITs) should be limited. Particularly, in the M-band the DITs should not be longer as 10s and 20s, respectively for the 0.4" and 0.2" wide slit. Longer DITs will result in saturation and can lead to persistence effects in the detectors that could affect the data quality of subsequent observations.

 

Slit Viewer Characteristics

For guiding on target (Target == SV guide star), the slit viewer guide star needs to be relatively bright as only a small fraction of the light is reflected by the slit jaws to the slit viewer detector. For guiding off target, the star can be fainter as no light is "swallowed" by the slit. The following H-band magnitude limits have been established during commissioning:

  on-slit guiding
off-slit guiding
AO NGS through 0.2" slit
H =  0 .. 15 H =  0 .. 16
AO NGS through 0.4" slit H =  0 .. 14.5 H =  0 .. 16
NoAO mode H =  -2 .. 15 H =  -2 .. 16

Please note that to ensure proper guiding, no other object should be closer than 2" to the Slit Viewer guide star. However, this restriction does not apply if the SV guide star is brighter by at least 3 magnitudes in the H-band than any close-by object.

For faint SV guide stars (H>13), a turbulence category of <= 30% should be requested.

 

Adaptive Optics Correction

The Adaptive Optics (MACAO - Multi-Applications Curvature Adaptive optics) is used to concentrate the light of the target into the slit and thereby significantly enhance the signal-to-noise ratio of the observations. MACAO is equipped with a wavefront sensor (Avalanche Photo Diodes) sensitive in the R band. Since the flux on these diodes is limited to 1 million counts to not damage the devices, stars brighter than R=0.2 mag cannot be used as AO guide stars. Stars fainter than R =15 mag will not result in any improvement. Good correction under very good atmospheric conditions (turbulence category <= 30%) can be obtained for faint AO stars (R >=12 mag) at an airmass <= 1.4. Note that the AO mode cannot be used with THICK clouds.

The quality of the AO correction degrades with increasing airmass; we strongly recommend limiting the use of AO to airmass values up to 1.4, as otherwise the AO loop may become unstable and the AO correction poor.

The field selector allows the selection of the AO star within a 25 arcsec on the nominal position of the slit center. However, if the AO star lies at more than 10 arcsec away from the science target, then the AO correction degrades and very good atmospheric conditions, AO stars brighter than R=12 mag as well as low airmass (<=1.4) are required.

Science detector characteristics

Typical values for the main characteristics of the science detectors are listed in Table 2.

Table 2. Typical detector characteristics
Read-out noise in e- (rms) for Det1, Det2, Det3
up to 12 e- for DITs < 50s,
6e- for DITs >= 50s
Read-out mode
Sample up the Ramp
Dark current  (e-/second/px)
< 0.003 
Gain (e-/ADU) for Det1, Det2, Det3
 2.15, 2.19, 2.00
Saturation level (e-) for all detectors
37,000 ADU
Non linearity <1% below 6000 ADU/px
>5% above 18,000 ADU/px
>10% above 29,000 ADU/px

Users can select predefined detector integration times (DITs) of 1.427, 2, 3, 5, 7, 10, 15, 20, 30, 45, 60, 75, 90, 120, 150, 180, 240, 300, 450, 600 and 900 seconds. The simple reason for this setting is to avoid extensive overheads for the morning calibrations.

 

Metrology

The CRIRES metrology system ensures that the spectrum falls on the same pixel. It is operational in all wavelength settings and ensures repeatability with subpixel precision (<0.2 px in dispersion). Without metrology, the spectrum can be shifted by 5-10 pixel in dispersion, thereby affecting the wavelength calibration. All morning calibrations are taken with the metrology.

 

Radial velocity precision

The short gas cell (SGC) provides a stable long-term wavelength reference in the H and K bands.

For a S/N of 150 per spectral pixel in the spectral continuum, an RV precision of 4 m/s has been attained in the lab by employing the short gas cell (SGC) as a simultaneous wavelength calibrator in the K-band with the 0.2” slit. The best RV precision that has been attained with CRIRES on sky is 2.6 m s-1 for GJ588 (K=4.76 mag) with DIT x NDIT = 60 s and a S/N ratio of ~280 in the brightest regions of the stellar continuum. For longer exposure times, the RV precision will be degraded because of grating drifts.

Users should consider to make use of the telluric absorption lines of Earth’s atmosphere as a simultaneous wavelength reference. In most CRIRES wavelength settings, these lines will anyways be imprinted on the science data. Figueira et al. (2010; A&A, 515, 106) demonstrated that telluric lines are intrinsically stable down to 10 m/s (rms). For information on how to model the telluric lines we refer to the paper by Seifahrt et al. (2010; A&A, 524, 11).

Observations without any simultaneous wavelength calibrator can result in an RV precision on the order of a few times 100 m/s due to grating drifts.

 

Schematics of the lightpath in CRIRES