Article Index

The InfraRed Telescope (IRT)

The InfraRed Telescope (IRT) on board THESEUS is designed in order to identify, localize and study the transients and especially the afterglows of the GRBs detected by the Soft X-ray Imager (SXI) and the X and Gamma Imaging Spectrometer (XGIS).

The telescope (optics and tube assembly) can be made of SiC, a material that has been used in other space missions (such as Gaia, Herschel, Sentinel 2 and SPICA study). Simulations using a 0.7 m aperture Cassegrain space borne NIR telescope (with a 0.23 m secondary mirror and a 10x10 arc min imaging flied of view), using a space qualified Teledyne Hawaii-2RG (2048x2048 pixels) HgCdTe detector (18 μm/pixels, resulting in 0.3 arcsec/pix plate scale) show that for a 20.6 (H, AB) point like source in a for a 300 s integration time one could expect a SNR of ~5 for a point source. The telescope sensitivity is limited by the platform jitter. Taking the latter into account (1 arc sec jitter over 10 s, at 3 σ level), we foresee to limit the image integration time to a maximum of 10 seconds per frame in order to correct for the jitter, and hence such short integration times will induce a high Read-out Noise (RoN, see below) degrading in turn the IRT sensitivity. In addition, due to the APE capability of the platform (2 arc minutes), the high resolution spectroscopy mode cannot make use of a fine slit, and a slit-less mode over a 5x5 arc min area of the detector will be implemented (similarly to what is done for the WFC3 on board the Hubble Space Telescope), with the idea of making use of the rest of the image to locate bright sources in order to correct the frames a posteriori for the telescope jitter. This is possible thanks to the selectable number of outputs (up to 32) of the Hawaii 2RG detector. The same goal could also be obtained by making use of the information provided by payload the high precision star trackers mounted on the IRT.

Hence the maximum limiting resolution that can be achieved by such a system for spectroscopy is limited to R~500 for a sensitivity limit of about 17.5 (H, AB) considering a total integration time of 1800 s. The IRT expected performances are summarized in Table 7.

In order to achieve such performances (i.e. in conditions such that thermal background represents less than 20% of sky background) the telescope needs to be cooled at 240 (±3) K , and this can be achieved by passive means. Concerning the IRT camera, the optics box needs to be cooled to 190 (+/- 5) K and the IR detector itself to 95 (+/-10) K: this allows the detector dark current to be kept at an acceptable level. The Hawaii-2RG provides a high quantum efficiency (80-90%) over a wide energy range, since substrate-removed HgCdTe can simultaneously detect visible and infrared light, enabling spectrographs to use a single focal plane assembly for Visible-IR sensitivity. The maximum wavelength cut-off (50% of quantum efficiency) varies from 1.8 (H band) to 2.5 (K band) μm as a function of operation temperature detector thickness, and Cadmium fraction of the detector.

In the simulations we used a conservative value of 0.1 e-/s/pixel for dark current at 100 K and 10 e- for the RoN At higher temperatures the dark current increases exponentially (already 1.0 e-/s/pixel at 110 K). The cooling of the detector at these low temperatures can hardly be achieved with a passive system in a low Earth orbit such as the one foreseen for THESEUS, due to the irradiation of the radiators of the infrared flux by the Earth atmosphere. A TRL 5 cooling solution for space applications is represented by the use of a Miniature Pulse Tube Cooler (MPTC). Such devices are available e.g. at Air Liquide Advanced Technologies. Air Liquide has studied in detail an engineering model of MPTC for the CNES MICROCARB mission, and is planning to have a full space qualified version (TRL 8) by 2018. Preliminary thermal studies making use of the Earth fluxes computed for the SVOM mission, which will have a similar orbit in terms of altitude but with a higher inclination (30°), indicate that use of such a device coupled to a 0.3 m2 radiator, which periodically faces the Earth atmosphere, would allow to reach the desired temperatures. Due to the tight link between the thermal aspects at optical and instrument level we propose that ESA takes the responsibility delivering the IRT telescope and its associated thermal system (radiators, etc.) including the IRT instrument cooling system (for the camera and the detector).

In order to keep the camera design as simple as possible (i.e. avoiding to implement too many mechanisms, like tip-tilting mirrors, moving slits etc.), we could implement a design with an intermediate focal plane making the interface between the telescope provided by ESA/industry and the IRT instrument provided by the consortium, as shown in the block diagram in Figure 13. The focal plane instrument is composed by a spectral wheel and a filter wheel in which the ZYJH filters, a prism and a volume phase holographic (VPH) grating will be mounted, in order to provide the expected scientific product (imaging, low and high-resolution spectra of GRB afterglows and other transients).


  • 22
  • 23
Figure 13. The IRT Telescope sketch (left) and optical diagram (right).


Specifications of the entire system are given in Table 3. The mechanical envelope of IRT is a cylinder with 80 cm diameter and 180 cm height. A sun-shield is placed on top of the telescope baffle for IRT straylight protection. The thermal hardware is composed by a pulse tube cooling the Detector and FEE electronics and a set of thermal straps extracting the heat from the electronic boxes and camera optics coupled to a radiator located on the spacecraft structure. The overall telescope mass is 112.6 kg and the total power supply is 95W.


Telescope type:


Primary & Secondary size:

700 mm & 230 mm


SiC (for both optics and optical tube assembly)

Detector type:

Teledyne Hawaii-2RG 2048 x 2048 pixels (18 μm each)

Imaging plate scale


Field of view:

10’ x 10’

10’ x 10’

5’ x 5’

Resolution (λ/Δλ):

2-3 (imaging)

20 (low-res)

500 (high-res), goal 1000

Sensitivity (AB mag):

H = 20.6 (300s)

H = 18.5 (300s)

H = 17.5 (1800s)




VPH grating

Wavelength range (μm):

0.7-1.8 (imaging)

0.7-1.8 (low-res)

0.7-1.8 (high-res, TBC)

Total envelope size (mm):

800 Ø x 1800

Power (W):

115 (50 W for thermal control)

Mass (kg):


Table 3. IRT specifications


IRT Observing sequence

1) The IRT will observe the GRB error box in imaging mode as soon as the satellite is stabilized within 1 arc sec. Three initial frames in the ZJH-bands will be taken (10s each, goal 19 AB 5 σ sensitivity limit in H) to establish the astrometry and determine the detected sources colours.

2) IRT will enter the spectroscopy mode (Low Resolution Spectra, LRS) for a total integration time of 5 minutes (expected 5 σ sensitivity limit in H 18.5 (AB)).

3) Sources with peculiar colours and/or variability (such as GRB afterglows) should have been pinpointed while the low-res spectra were obtained and IRT will take a deeper (20 mag sensitivity limit (AB)) H-band image for a total of 60s. These images will be then added/subtracted on board in order to identify bright variable sources with one of them possibly matching one of the peculiar colour ones. NIR catalogues will also be used in order to exclude known sources from the GRB candidates.

  1. In case a peculiar colour source or/and bright (< 17.5 H (AB)) variable source is found in the imaging step, the IRT computes its redshift (a numerical value if 5<z<10 or an upper limit z<5) from the low resolution spectra obtained at point 1) and determines its position. Both the position and redshift estimate will be sent to ground for follow-up observations. The derived position will then be used in order to ask the satellite to slew to it so that the source is places in the in the high resolution part of the detector plane (see below) where the slit-less high resolution mode spectra are acquired. Following the slew, the IRT enters the High Resolution Spectra (HRS) mode where it shall acquire at least three spectra of the source (for a total exposure time of 1800s) covering the 0.7-1.8 μm range. Then it goes back to imaging mode (H-band) for at least another 1800s (TBC). Note that while acquiring the spectra, continuous imaging is performed on the rest of the detector, see Fig 3.16. This will allow to the on board software to correct the astrometry of the individual frames for satellite drift and jitter and allow a final correct reconstruction of the spectra by limiting the blurring effects.

  2. In case that a faint (> 17.5 H (AB)) variable source is found, IRT computes its redshift from the low resolution spectra, determines its position and sends both information to the ground (as for 3a). In this case IRT does not ask for a slew to the platform and stays in imaging mode for a 3600s time interval to establish the GRB photometric light curve (covering any possible flaring) and leading the light curve to be known with an accuracy of <5 %.


  • 19
  • 18

Figure 14. Left: the IRT focal plane division. The blue area (10’x10’) is used for imaging and low resolution spectra. The orange area (5’x5’) is used for high-resolution slit-less spectra. The size of the high-resolution spectral area is limited by the satellite pointing capabilities. Right: Teledyne Hawaii 2RG detector at ESA Payload Technology Validation section during the tests for the NISP instrument on board the EUCLID mission (Credits ESA)


IRT specific calibration needs

The IRT detector single pixel and intra-pixel response will be characterized using a dedicated optical bench in the lab at CEA-Irfu/SAp on the IRT camera qualification model. A similar activity is on-going at CEA in the framework of the EUCLID NISP instrument. The overall instrument calibration can be obtained on ground using the IRT camera flight model connected to a telescope simulator. The most accurate and final calibration of the IRT will be obtained in flight by observing a number of known calibrating sources, and by means of LED lights installed within the camera that uniformly illuminate the detector, as done on the NISP camera on the EUCLID mission.


IRT Telemetry requirements
The telemetry requirements for the different observing modes described above are reported in Table 4. Transmitting full frames is impractical except in the rarest of cases, so in imaging and LRS mode tiles centered on identified sources of interest are chosen for transmission: 64 x 64 pixels in imaging mode, 128 x 64 in LRS mode.  In both modes, tiles around comparison objects are also likely to be necessary.  In HRS mode, a fixed window of 2048 x 64 is used. Clearly, full frame image transfer must be exceptional and tiled mode somewhat restricted. The figures shown are for all observations, so calibration data is included.


Mode and tilesize


Integration (s)

Volume (Mbit/hr)

Full frame (2048 x 2048) [8 MB]


Combined 60


Image tile (64 x 64) [8 KB]


5 sources @ 10s + 10 ref. objs @ 60s


LRS tile (128 x 128) [16 KB]


10 sources @ 60s (6x10s combined)


HRS image (1024 x 1024) ) [2 MB]


1800 (30x60s combined exposures)


 Table 4. IRT data rate.