Article Index

Provision of broad-band X-ray spectroscopy

The scientific objectives of THESEUS require the secure identification of sources types, in particular GRB triggers. The soft X-ray instrument is the primary source locator and has a high sensitivity to a wide variety of source types, as discussed above, as required to achieve the scientific goals.

This instrument will trigger on a large number of known sources which should not result in a spacecraft slew, but also other transient sources only some of which are GRBs. To reliably identify GRBs as well as spectroscopically characterise other sources and reduce the number of demanded slews requires the provision of a sensitive broad-band X-gamma-ray instrument well matched to the sensitivity of the soft X-ray instrument and with source location capabilities of a few arcmin in the X-ray energy band. To meet the scientific goals we require an instrument with the following characteristics:

  • extend the soft X-ray band of the imaging instrument up to the MeV band;

  • identify and localize with a few arcmin accuracy the GRBs by providing simultaneous triggers and by providing higher energy light curves and spectra to determine the luminosity of the GRB;

  • reduce the demanded number of spacecraft slews to observe with the IRT and act as a crucial filter to reduce soft X-ray trigger volume for ground/space telescope follow-up;

  • measure unbiased GRB/transient X- and gamma-ray spectra down to short time scales (ms time scales for the strongest events) to probe GRB physics

The proposed XGIS instrument provides the required sensitivity and bandpass. As can be seen in Figure 4, the sensitivity of the SXI and XGIS are well matched over the typical durations of GRBs (few to few tens of seconds). The XGIS will provide spectroscopy over 2-20000 keV with monitoring timescales down to milli-seconds. Despite advances during the Swift and Fermi era to identify and characterise GRB phenomenon requires study of the prompt emission. Planned future missions (e.g. SVOM, CALET/GBM, UFFO) do not provide the required combination of sensitivity and bandpass.

  • 4
  • 5

Figure 3. The left-hand panel shows the GRASP (FOVxEff. Area) of the THESEUS/SXI in the soft X-ray energy band compared to XMM-Newtin and eROSITA . The GRASP of X-ray monitrs on-board MAXI and ASTROSAT are also show for completeness, even though these are not focusing and their sensitivity for a given eff. area is substantially sworse than that of focusing teelscopes . The leap in monitoring / syervey of the soft X-ray sky allowed by THESEUS/SXI is outstanding. The right panel shows The THESEUS/ XGIS fills the parameter space in the top-left corner of the right-hand panel where other instruments have either too high an X-ray threshold or too low effective area, and will still provide 1000-1550 cm2 effective area up to several MeV. 

  • 6
Figure 4. Sensitivity of the SXI (black curves) and XGS (red) vs. integration time. The solid curves assume a source column density of 5x1020 cm-2 (i.e. well out of the Galactic plane and very little intrinsic absorption). The dotted curves assume a source column density of 1022 cm-2 (significant intrinsic absorption). The black dots are the peak fluxes for Swift BAT GRBs plotted against T90/2. The flux in the soft band 0.3-10 keV was estimated using the T90 BAT spectral fit including the absorption from the XRT spectral fit. The red dots are those GRBs for which T90/2 is less than 1 second. The green dots are the initial fluxes and times since trigger at the start of the Swift XRT GRB light-curves. The horizontal lines indicate the duration of the first time bin in the XRT light-curve. The various shaded regions illustrate variability and flux regions for different types of transients and variable sources.


  • 7
  • 8

Figure 5. Left: Time scales and luminosities of soft X-ray transients. Right: typical light curve of an X-Ray Flash, a GRB emitting only in the X-ray energy range (Amati et al. 2004). THESEUS will provide new results on all of these types of transient.


Optical-IR follow-up 

The scientific goals of THESEUS require the following on-board capabilities for an optical/near IR telescope (IRT) to follow-up GRBs after a demanded spacecraft slew:

  1. Identify and localize the GRBs found by the SXI and XGIS to arcsecond accuracy in the visible and near IR domain (0.7-1.8 μ);

  2. Autonomously determine the photometric redshift of GRBs for z>4 and provide redshift upper limits for those at lower redshift;

  3. Provide precise spectroscopic redshift measure for bright GRBs, together with limits on the intrinsic NH and metallicity for the majority of GRBs at z >4

The requirement number 1 is justified by the fact that the goal of the THESEUS mission is to study the Universe at redshift beyond z=6 in order to study the epoch of reionization. CMB experiments suggest that reionization was underway at z~9, while it appears to be completed by z~6.5. The question is whether massive stars could sustain a largely reionized Universe at z=6-9, and beyond. GRB afterglow spectra are power laws and, due to the dots are those GRBs for which T90/2 is less than 1 second. The green dots are the initial fluxes and times since trigger at the start of the Swift XRT GRB light-curves. The horizontal lines indicate the duration of the first time bin in the XRT light-curve. The various shaded regions illustrate variability and flux regions for different types of transients and variable sources.

Lyman-alpha drop-off (i.e. the Lyman alpha absorption within the GRB host galaxies and intervening IGM), a very attenuated signal is expected at wavelengths shorter than the Lyman alpha break, providing an unmistakable feature. Due to cosmological expansion the Ly-alpha wavelength (1216 Å = 0.126 μ) moves along the energy band, and in order to measure GRB redshifts between z=4 and z=10 the telescope detector has to be sensitive in the 0.7 to 1.8 μm range. GRBs with redshifts below 4 can be used as comparison to evaluate how massive stars evolve along the history of the Universe, and they can be easily followed up from ground. It is for high redshift GRBs that a NIR telescope in space really takes advantage of the absence of background due to the atmosphere. The field of view of the IRT telescope shall be of about 15x15 arcmin, given that the SXI will provide error boxes which are of about 2 arcmin radius.

  • 9
  • 10

Figure 6. Left: a simulated IRT high resolution (R=500) spectrum for a GRB at z=6.3 observed at 1 hour post trigger assuming a GRB similar to GRB 050904. The spectrum has host log(NH)=21 and neutral fraction Fx=0.5 (and metallicity 0.1solar). The two models are: Red: log(NH)=21.3, Fx=0 Green: log(NH)=20.3, Fx=1. The IRT spectra provide accurate redshifts. Right: simulated IRT low resolution (R=20) spectra as a function of redshift for a GRB at the limiting magnitude AB mag 20.8 at z=10, and by assuming a 20 minute exposure.The underlying (noise-free) model spectra in each case are shown as smooth, dashed lines. Even for difficult cases the low-res spectroscopy should provide redshifts to a few percent precision or better. For many applications this is fine - e.g. star formation rate evolution.


For requirement number 2 the telescope will be operated in low resolution mode (R~10-20), and the Ly-alpha drop-off will be searched for. A fit of the sources’ low resolution spectra, done on-board, should be capable of identifying high-redshift candidates. If we focus on GRBs at redshift larger than 6 the detector shall be optimized in terms of QE the 0.8-1.5 μm wavelength range. To obtain reliable results the detector QE shall be known within 10-20%.

The requirement number 3 deals with the high-resolution spectra (R~500). A resolution of the order of R~500 is good enough to identify the main absorption lines in bright GRB NIR afterglow spectra. Such detections would (i) enable a more precise measure of the GRB redshift, that goes beyond the “mere” detection of the Lyman-alpha drop off achievable with lower resolution spectra and (ii) help in discriminating between highly-extinguished and high-redshift (z>5) events. Besides the redshift measure, with IR high resolution spectroscopy it will be possible to derive limits on element relative abundances and metallicity (together with a measure of NH for the brightest events, obtained by fitting the red wing of the Ly-alpha). As discussed above (Sect. “Exploring the early Universe with GRBs”), GRBs are powerful tools to study the early Universe. With IRT spectroscopy we will able to promptly identify the high-redhift GRBs. Such information will be vital to optimize ground-based follow-up (with the large aperture facilities that will be available in 2028, like E-ELT), aimed at precise optical/IR spectroscopic studies for a detailed characterization of the GRB environment throught the measure of chemical abundances, metallicity, SFR.

The sensitivity of the SXI triggering system in the 0.3-5 keV energy band probes a fluence range of 10-8 and 10-9 erg cm-2. Based on Ghirlanda et al. (2015) the proposed THESEUS IRT with a limiting sensitivity of 20.6 mag in the H (1.6+/-0.15 μm) filter is expected to detect all the GRB counterparts in imaging and low-resolution spectroscopy, if pointed early after the GRB trigger, in a 300 s exposure, and for the large majority of GRBs high-resolution spectra can be taken even 1-2 hours after the GRB (first or second spacecraft orbit) with the 19th magnitude sensitivity IRT. The annual rates for on-board spectroscopy are shown below. THESEUS out-performs Swift by about an order of magnitude at all redshifts and by more at the highest redshifts. Using the IRT to follow-up the SXI and XGIS will identify the highest priority high-redshift targets for the early Universe science goals.

In Figure 2, we show the rate of GRBs whose redshift will be spectroscopically determined by THESEUS on-board as a function of redshift. For the other GRBs detected by the SXI/XGIS, the IRT will provide a location and a redshift limit and thus provide a redshift estimate for the entire sample detected on-board. The cumulative distribution represents the rate (number of GRBs per year) that can be detected by THESEUS (red solid filled region). The width of the distribution accounts for the uncertainties of the population synthesis code adopted. For comparison, the rate of detection of GRBs by Swift is shown (blue line). This rate is derived from the actual population of GRBs detected by Swift and with measured redshift multiplied by a factor 3. Indeed, approximately only 1/3 of the GRBs detected by Swift have their redshift measured. The upper axis shows the age of the Universe. The detection and redshift-estimate annual rates expected form THESEUS are also summarized in Table 3.

Table 3. THESEUS yearly detection and on-board redshift measurements rates. Note that photometric redshifts are possible only at z >~5 , when the Lyman “dropout” or “break” gets inside the IRT band



z > 5

z > 8





Photometric z




Spectroscopic z