Soft X-ray Telescope

Soft X-ray Telescope (SXT) onboard ASTROSAT is sensitive to soft X-rays in the energy range of 0.3 − 8 keV. Here are basic characteristics of SXT.

Telescope 2.0 m focal length
Telescope Mirrors Conical Shells
Telescope PSF 3- 4 arcmin
Detector e2V CCD-22
Pixel Size 40μm × 40μm
Pixel Scale 4.12” per pixel
Useful Image Area 600 pixel × 600 pixel
Field of View ~40’
PSF 2.0' FWHM
10.0' HPD
Position accuracy 2.0’
Energy Range 0.3-8keV
Energy Resolution 90 eV at 1.5 keV
136 eV at 5.9 keV
Time resolution ~ 2.4 s (full frame)
~ 0.278 s (150x150 centered pixel frame)
Effective Area ~90 cm 2 at 1.5 keV
Sensitivity (obs. time) ~10 -13 ergs cm-2 s-1 (5 σ) (20000 s)

SXT Components

Telescope :

The telescope consists of a tubular structure housing the assembly of X-ray reflecting mirrors and other components. There is a “Charge Coupled Device” (CCD) camera at the common focus of all the mirrors in order to image the cosmic sources.


Basic Components

The basic components of the telescope are given below (see Fig. 1).

  1. A Mirror Assembly.
  2. A Focal Plane Camera Assembly housing a cooled CCD.
  3. A deployable cover/door at the top end of the telescope that covers the optical elements on the ground and protects them from contamination. This was deployed ~2 weeks after launch, in a one-time operation, and is perched at an angle of 256°.
  4. A “Thermal Baffle” placed between the mirror assembly and the telescope door. All parts are made of anodized aluminum alloy 6061 T6. The function of the thermal baffle is to protect the telescope from the Sun, and to provide a base for mounting the heaters to maintain the optics within a certain specified range of temperatures, and to block the unwanted area of the optics. The sun avoidance angle with the thermal baffle is ~45o.
  5. A “Forward tube” made up of Composite Fiber Reinforcement Plastic (CFRP). The Forward tube extends from the bottom of the “Top Lid” and it covers the thermal baffle assembly and 1α section (see below) of the Optics assembly.
  6. A metallic ring (Ring 1) provides an interface between the Rear tube 1 and the middle flange of the optics from the bottom side. Another ring (Ring 2) provides an interface between the Forward tube and the Rear tube 1 from the top size.
  7. Rear tube-1 made up of CFRP is a hollow cylinder of diameter 343 mm ID and 347.8 mm OD and extends from Ring-1 to the “Deck Interface Ring” (DIR). It houses 3α (see below) optics while the Forward tube houses the 1α assembly (see below).
  8. “Deck Interface Ring” (DIR) is made up of Al alloy 6061 and is used to assemble rear tube-1 and rear tube-2 of the telescope. DIR has 9 nos. of holes of size M8 to provide interface between the payload and the top deck of the satellite.
  9. Rear tube-2 made up of CFRP is a hollow stepped cylinder with a top portion thicker than the bottom portion to provide stiffness. Rear tube-2 extends from the DIR to the CCD interface ring.
  10. CCD interface ring is provided to align the CCD Camera with the Rear tube-2 tubular structure to the desired accuracy. This is made up of aluminum alloy 6061.


Figure 1: Left: Structure and various components of the Soft X-ray Telescope from outside (simulated picture). Top: The optical modules with gold-coated mirrors (upper: 1 α; lower: 3 α; actual photo).


Figure 2: The principle of the Wolter I optics using ray diagram. The green lines show the conical approximations of the paraboloidal and hyperboloidal mirrors.


The Mirror Assembly for X-ray reflection

The mirror assembly of the X-ray telescope in the SXT consists of a set of coaxial and con-focal shells of conical mirrors approximating paraboloidal and hyperboloidal shapes and arranged behind each other. This geometrical arrangement is known as Wolter I optics (Fig. 2). X-rays are first reflected by an internally reflecting paraboloidal (1α section) mirror and then reflected to the prime focus of the telescope by an internally reflecting hyperboloid (3α section) mirror. At grazing incidence, the active region of the mirror is just a thin annulus giving a small collecting area even for a large diameter mirror. Thus nesting of Wolter I shells is incorporated to improve the filling factor of the circle defined by the outermost shell. Higher nesting is achieved by using shells made of very thin foils but figured in a conical approximation to Wolter I optics. SXT has 40 complete shells of mirrors assembled quadrants wise (320 mirrors) for 1α and 3α mirrors (Fig. 1). The focal length of the telescope is 2000 mm, constrained by the available space in the launch vehicle flaring. Each mirror is made of aluminum (thickness ~0.2 mm) covered with gold on the reflecting side. The length of each mirror is 100 mm. The radius of the outermost shell is 130 mm, while that of the innermost shell is 65 mm. The on-axis FWHM and half-power diameter (HPD) of the point spread function (PSF) in the focal plane is ~2’ and ~10' respectively. The on-axis effective area of the telescope is about 90 cm2 at 1.5 keV (Fig. 3).

Figure 3: Telescope effective area (including CCD Quantum Efficiency (isolated and bi-pixel events 1-4) and the absorption by the optical blocking filter as measured) vs. energy for on-axis.


Focal Plane Camera Assembly (FPCA) :

The primary instrument in the Focal Plane Camera Assembly (FPCA; see Fig. 4) of the SXT is a special CCD, which is the focal plane imager. The health and the operational conditions of the CCD require vacuum, low temperature and protection from the optical light and energetic protons. Therefore, components such as thermo-electric cooler (TEC), optical blocking filter, proton shield, etc. have been included in FPCA. The operation of the CCD also requires the processing electronics, and the calibration of the CCD + processing electronics requires calibration sources.

Figure 4: Focal Plane Camera Assembly (FPCA) and its components.


Focal Plane Camera Assembly Devices and Components
  1. Charge Coupled Device CCD-22:
    It is a MOS device built by the e2V Technologies Inc., UK, for the European Photon Imaging Camera (EPIC) onboard the XMM-Newton observatory and supplied to the University of Leicester and has the following characteristics (also see Fig 5):
    1. A three-phase frame transfer.
    2. Open electrode structure for useful band pass of 0.2 to 10 keV.
    3. Operating area of 610 × 602 array of 40 micron by 40 micron pixels including over-scan.
    4. The storage region is a 600 × 602 array of 38 x 12 micron pitch.
  2. Figure 5: Schematic Diagram of the CCD22.


  3. Calibration sources:
    Four individual 55Fe radioactive calibration sources are provided in the camera for in-flight calibration at energies of ~5.9keV (Mn). These illuminate the corners of the CCD outside the Field Of View (FOV). A fifth source under the door facing the CCD is no longer available after the deployment of the door. See Fig 6.
  4. Figure 6: The CCD illuminated by 4 internal calibration sources, shown at corners.


  5. Optical Blocking Filter:
    A thin filter is installed in front of the CCD to block visible light. The filter consists of a single fixed polyimide film 1840 A° thick coated with 488 A° of Aluminum on one side. The typical optical transmission of the filter is less than 5 × 10-3 (similar to the XMM-Newton thin filter).The filter design provides ~ 7 magnitude of optical extinction over the visible band. For the SWIFT XRT with a PSF of ~15" a 6th magnitude star gives an optical loading of ~ few e- per pixel, at which point the quality of the X-ray data begins to be affected. For the SXT with a ~7-8 times larger PSF and a 2 times larger (angular) pixel the safe optical limit should be closer to a ~4 magnitude star. This limit is best set using a bright non-X-ray emitting star. The X-ray transmission of the filter is shown in Fig. 7.
  6. Figure 7: The transmission of X-rays through the optical blocking filter.


  7. Operating Temperature of the CCD:
    The X-ray CCD detector is cooled to 191οK (-82o C) by a thermo-electric cooler (TEC) and a radiator assembly during its operation for low dark current and to reduce sensitivity to radiation damage. The TEC is coupled to a cold finger connected to a “Heat Pipe” unit which is connected to a radiator plate designed by ISRO for effective heat dissipation and which provides a maximum temperature of −40°C at the junction between the heat pipe and the camera cryostat.
  8. Proton Shield:
    The camera is provided with a proton shield to minimize any damage to the device due to energetic protons. The degradation of energy resolution is, therefore expected to be contained during the 5 years lifetime of the AstroSat mission. Due to the orbit of the AstroSat being more benign than the orbit of SWIFT, the mass and size of the proton shield was significantly reduced from the original SWIFT design.
  9. Readout Nodes:
    Provision has been made for readout from either left node or the right node using node with two individual pre-amplifiers.

  10. Figure 8: CCD Quantum Efficiency points (refer to the subsection "CCD Quantum Efficiency Curves”).



CCD Quantum Efficiency (QE) Curves

The flight device has been calibrated across the energy range giving the measured QE values for iso events, for events up to 4 pixels in size, and for “all” events, i.e. including those with 5 or more active pixels. The results of these calibration tests are shown in Fig. 8.




Charge Transfer Inefficiency (CTI) and Pile-up effects

The Charge Transfer Inefficiency (CTI) has been measured along both the serial and parallel readouts using the four corner sources. The serial and parallel CTIs thus measured are shown as a function of energy in Fig. 9. The CCD camera door is always open and remains open during the data readout. Therefore, if the source is too strong or the PSF is too sharp then multiple photons can fall on the same pixel during the 2.4 s that it takes to readout the CCD, leading to a pile-up and affecting spectra, light curves and images. The amount of pile-up depends on the PSF of the telescope and will need to be corrected for when the source count rate at the maximum of the PSF becomes ~0.5 cts/pixel/frame. The measurements of the PSF suggest that it will not become an issue in the PC mode for sources fainter than ~200 mCrab. For Chandra, XMM-Newton and Swift in imaging modes pile-up occurs for sources of intensity ~0.1, 1 and 5 mCrab respectively. The pile-up effect depends on the sharpness of the PSF and the intensity of the source. The exact measurement of this effect is complicated due to corrections for the attitude variations that are still being implemented, and it will be provided soon. At present, if a source is producing (or expected to produce) < 40 counts , it is safe to assume no pile-up for the PC mode. This limiting count rate will be ~8.6 times larger in the FW mode. Our current analysis shows that in order to get rid of pile-up for ultra-bright sources like Crab, Cyg X-1 etc., a 1 arcmin central removal of the PSF is enough.


Figure 9: The serial and parallel CTI as functions of energy.


Data Modes and Telemetry :

Data from the SXT CCD are stored on-board, and then sent to the ground station once in each orbit around the Earth. The SXT on-board memory quota is 280 Megabytes per orbit (~ 98 minutes). This puts serious constraints on the data modes and how the data are packaged. There are four data modes and one separate mode for only housekeeping. In each mode, data are packed in 2 Kbyte (2K) segments. So the SXT memory per orbit is filled with approximately 143360 2K blocks.


The various data modes are "Photon Counting" (PC) mode, "Fast Windowed Photon Counting" (FW) mode, "Bias Map" (BM) mode and "Calibration" (Cal) mode (CM). There is also the "House Keeping" (HK) mode for health parameters of the electronics system. In the PC mode, data from the entire CCD (i.e., 600 X 600 pixels) are collected, provided these are above a specified threshold energy (set through a tele-command by the SXT team and it can be between 100 - 200 eV). Currently the default value set is 105 eV that is 4 above the noise peak. It is, however, recommended that low energy threshold to be used during analysis should be >=200 eV. Moreover, data from a maximum of 36000 pixels only can be transmitted in this mode. The read-out time of this mode is ~2.4 s. In the FW mode, a 150 X 150 pixel window is used in the center of the CCD. The read-out time of this mode is ~278 ms. Data in the FW mode are also thresholded as in the PC mode. The Cal and bias modes are used to check the calibration of SXT and the zero point for the noise level. The read-out time of the Cal mode is ~2.4 s. Data from 4 corner windows of the CCD with a size of 80x80 pixels (under 4 X-ray radioactive sources) and a central window with a size of 100x100 pixels are transmitted with zero threshold in the CM. BM mode is a separate mode in which the entire CCD frame is sent with zero threshold. Incrementally addressed 60 rows per CCD frame along with their co-ordinates are sent in this mode.  On the ground, these rows of each incrementally addressed CCD frame are mapped to generate an individual CCD frame. It takes 24 seconds of data to generate one completely mapped CCD frame. The HK mode is operated only when there is a failure of both LBT telemetry channels (main and redundant). When HK mode data command is uploaded, LBT data information in the form of HK data are sent in 2K data package. Hence only one frame will be generated and pushed in currently operating mode data package. Spectral information is available under all 4 modes: PC, FW, CM and BM.


Data (Level 1) in each mode are packaged in 2K blocks. However, the content of a 2K block is not the same for all the modes. For the PC and FW modes (i.e., the science data), only the channel number above the pre-selected threshold is stored in the 2K block along with the pixel coordinate and the CCD frame identification. These data are stored in the 15th to 2042nd bytes of the total of 2048 bytes of a 2K block. Three bytes are required to store each of (a) CCD frame identification, (b) CCD row number of the pixel and threshold value and (c) CCD column number of the pixel and the channel number. The bytes 1-14 (header) and 2043-2048 (footer) store the 2K block number, mode information, on-board time, window location and numbers to check the validity of the 2K block.


Operation Procedure

Normal operations began after the SXT door was opened. The CCD is currently maintaining its operational temperature with an accuracy of ±2 degs. SXT is pointed towards a celestial source only after all the observational constraints are met, and the FPCA is normally operated in the PC mode with full CCD readout unless specified otherwise. The calibration data from the four corner point sources are part of the observations but unlike in the CM these have a threshold already applied. Thus normal data have full energy resolution and time resolution of 2.4 sec.



Observational Constraints :

There are several pointing constraints on the SXT observations, primarily to protect the CCD, the optical blocking filter above the CCD, and the mirror coating. The most important one is the Sun avoidance angle (> 45 degree) and is absolutely essential for the safety of the SXT. The other constraints that can affect the data quality are the Earth limb / bright Earth avoidance angle, the RAM direction avoidance angle and the Moon avoidance angle. The RAM avoidance of >12 degree is applied by the mission operations for all the observations as it can also affect the mirrors. The bright Earth avoidance angle of >110 degree is used in the SXT pipeline while converting level1 data to level 2 data products.



In-flight Performance :

The first six months of observations with AstroSat were dedicated for performance verification (PV) observations, followed by a six-month long guaranteed time (GT) observation phase. In a typical orbit, parts of the SXT data are not usable due to SAA passage, eclipse of the source by the Earth and by the bright Earth viewing that floods the available memory allocation. The net observing efficiency of the SXT varies from source to source but on the average it is about 25%. The SXT has observed many X-ray sources including PKS 2155-304, Tycho SNR, 1E 0102-72.3 - an SNR, AB Dor - an active sun-like star, A1795 - a cluster of galaxies, and other AGN, X-ray binaries, etc. An example of Tycho spectrum is given in Fig. 10. A comparison of the 1E 0102-72.3 spectrum with the IACHEC model is also shown in Fig. 11. The spectral comparison with IACHEC model confirms the spectral calibration of SXT. SXT was pointed such that 1E 0102.2-7219 and PKS 2155-304 were incident on different parts of the CCD to determine the bore-sight of the telescope and the vignetting in the SXT at different off-axis angles. Since PKS 2155-304 had shown strong variability during our monitoring, we have mostly relied on the use of the supernova remnant 1E 0102.2-7219 in the Small Magellanic cloud which, however, required long observations, as the source is very weak in the SXT. This source emits mostly soft X-rays and is seen in 0.3 - 3 keV energy band.

Figure 10: The SXT sprectrum of Tycho.


Figure 11: The X-ray spectrum of 1E0102-72.3 as fitted with the IACHEC model derived from several X-ray observatories carrying a CCD camera in the focal plane of a telescope. The SXT spectrum was extracted from a radius of 10 arcmin. No significant contributions from the closest (<2arcm) XRB are noticed.


Figure 12: The point spread function for on-axis SXT observations of the blazar 1ES 1959+650 (exposure: 71 ks; 0.3-7.0 keV) made by using equal distance method. The X-axis shows the distance from the bore sight (arc sec), whereas, Y-axis gives the normalized counts. The best-fit double King function model, as well as the two King function components, are shown.


The results from these observations showed that the bore sight of the SXT is close to center of the foV at the CCD detector coordinates X=302±7 and Y=285±7 in pixels. The vignetting of the telescope or the projected area as a function of off-axis angle was also determined. The on-axis point spread function (PSF) in the focal plane is well characterized by a double King function (see Fig. 12). The two King functions have core radii of 54±8 and 680 (-200, +400) arcsec respectively with the broader King function having ~8% of the intensity compared to the narrower King function. The broader King model is indicative of the misalignment and the scattering in the mirrors. The half light (intensity) radius is ~70 arcsec, the half encircled energy radius is ~7-8 arcmin. Care must be taken to include as much of the encircled energy as possible while extracting a spectrum, and then use the corresponding response for the telescope area function from those provided. For very bright sources, the user may have to include a radius as large as 18 arcmins to get all the photons and then use background from a deep field with no detectable objects. The deep field background images to be used for image analysis of extended sources like clusters etc. During the last two years of AstroSat, the CCD gain of SXT has been changed by 30-40 eV and the users are requested to apply it externally (using gain command) when fitting the spectrum. One should confirm it by matching the RMF (by shifting the gain) with gold absorption edge, seen clearly in the spectral data. A description of the in-flight performance can be found in


Singh, K. P., et al., “Soft X-ray Focusing Telescope aboard AstroSat: Early Results”, Current Science, in press, 2017.
Singh, K. P., et al., “The Soft X-Ray Focusing Telescope aboard AstroSat and its Post-launch Scientific Capabilities”, JAA, in press, 2017.
Singh, K. P., et al., “In-orbit performance of SXT aboard AstroSat”, Proceedings of the SPIE, 9905, 2016.


The numbers corresponding to the most recent pipeline updates (coordinator tool) are being worked out - a preliminary report can be obtained from the SXT-POC on request (email: astrosat_sxt@tifr.res.in).