2. Mission Description

This chapter is a brief introduction to the satellite and its instruments, and is intended as a simplified guide for the proposer. Reading it thoroughly should provide the reader with the necessary information to understand the capabilities of the instruments at a level sufficient to prepare the feasibility section of an Astro-E2 proposal.

Astro-E2 is scheduled to launch from the Uchinoura Space Center (USC) on a Japanese M-V rocket in February 2005. After launch, the satellite with undergo an initial checkout phase lasting approximately 4 weeks, including instrument turn-on and initial calibration. Science observations will begin immediately after this initial phase and will focus on using the XRS for as long as the cryogenic system lasts, which is expected to be $\sim 2.5$years.

Figure 2.1: The 96 minute Astro-E2 orbit.

Astro-E2 is in many ways similar to ASCA in terms of its orbit, the pointing and tracking capabilities of the spacecraft, and uses the same station (USC) for uplink and downlink. Therefore, the operational constraints for Astro-E2 are also similar to those of ASCA. Astro-E2 will be placed in a near-Earth orbit, with inclination of 31 degrees, orbit altitude of roughly 550 km, and orbital period of about 96 minutes. The maximum slew rate of the spacecraft is $12$degrees/min, and settling to the final attitude will take about 10 minutes. The normal mode of operations will have the spacecraft pointing in a single direction for at least 1/4 day (10 ksec; but see ``raster-scanning'' below). With this constraint, most targets will be occulted by the Earth for about one third of each orbit, but some objects near the orbital poles can be observed nearly continuously. The current projection is that the observing efficiency of the satellite will be about 45%.

Figure 2.2: [Left] Schematic picture of the bottom of the Astro-E2 satellite. [Right] A side view of the instrument and telescopes on Astro-E2.

2.1 A Brief Introduction to Astro-E2

The scientific payload of Astro-E2 (Fig. 2.2) consists of three types of co-aligned scientific instruments. High-resolution X-ray spectroscopy will be carried out with the X-ray Spectrometer (XRS). The XRS is an X-ray calorimeter array capable of precisely measuring the energy of an incoming X-ray photon (between 0.3-12 keV) by detecting the temperature rise resulting from the absorption of the photon. This instrument uses consumable cryogens, and it is expected that its lifetime will be $\sim 2.5$ years. The observatory also features four X-ray sensitive imaging CCD cameras (X-ray Imaging Spectrometers, or XISs), two front-illuminated (FI; energy range 0.4-12 keV) and two back-illuminated (BI; energy range 0.2-12 keV), capable of moderate energy resolution. The XRS and each XIS are located in the focal planes of five dedicated X-ray telescopes. The third type of instrument is the non-imaging, collimated Hard X-ray Detector (HXD), which extends the bandpass of the observatory to much higher energies with its 10-600 keV bandpass.

Table 2.1: Overview of Astro-E2 capabilities

S/C	Orbit Altitude	550 km
	Orbital Period	96 minutes
	Observing Efficiency	$\sim 45$%
XRT	Focal length	4.75 m (XRT-I), 4.50 m (XRT-S)
	Field of View	$17'$ at 1.5 keV
		$13'$ at 8 keV
	Plate scale	0.724 arcmin/mm (XRT-I), 0.764 arcmin/mm (XRT-S)
	Effective Area	440 cm$^2$ at 1.5 keV
		250 cm$^2$ at 8 keV
	Angular Resolution	$2'$ (HPD)
XRS	Field of View	$2.9'\times2.9'$
	Bandpass	0.3-12 keV
	Pixel Grid	6 $\times$ 6; 30 active
	Pixel size	0.624 mm$\times$0.624 mm
	Pixel spacing	0.64 mm$\times$0.64 mm
	Energy Resolution	6.5 eV (FWHM) at 6 keV
	Lifetime	$\sim 2.5$ years
	Effective Area	190 cm$^2$ at 1.5 keV
	(incl. XRT-S)	100 cm$^2$ at 8 keV
	Absolute Line Centroiding	$\lesssim 2$eV
	Absolute Time Resolution	100 $\mu$s
	Average Telemetry limit	5 counts/s/pixel
	Background rate	$\sim 2$counts/pixel/day
XIS	Field of View	$18'\times18'$
	Bandpass	0.2-12 keV
	Pixel grid	1024$\times$1024
	Pixel size	24 $\mu$m$\times$ 24 $\mu$m
	Energy Resolution	$\sim130$eV at 6 keV
	Effective Area	340 cm$^2$ (FI), 390 cm$^2$ (BI) at 1.5 keV
	(incl XRT-I)	150 cm$^2$ (FI), 100 cm$^2$ (BI) at 8 keV
	Time Resolution	8 s (Normal mode), 7.8 ms (P-Sum mode)
HXD	Field of View	$4.5^{\circ}\times4.5^{\circ}$ ($\gtrsim 100$keV)
	Field of View	$34' \times 34'$ ($\lesssim 100$ keV)
	Bandpass	10 - 600 keV
	- PIN	10 - 60 keV
	- GSO	30 - 600 keV
	Energy Resolution (PIN)	$\sim 3.0$keV (FWHM)
	Energy Resolution (GSO)	$7.6 / \sqrt{E_{MeV}}$ % (FWHM)
	Time Resolution	61 $\mu$s

Figure 2.3: Effective area as a function of photon energy of the combined XRT + XRS system, for all 30 pixels. The upper (solid) curve shows the effective area with the open position of the filter wheel; the lower (dashed) curve shows it with the 300 $\mu$m Be filter.

Figure 2.4: XIS Effective area of one XRT + XIS system, for both the FI and BI chips.

All of the instruments on Astro-E2 operate simultaneously. Each of the five co-aligned XRTs features an X-ray mirror with an angular resolution (expressed as Half-Power Diameter, or HPD) of $\sim 2'$ (cf. Fig.2.6). Figure 2.3 shows the total effective area of the XRS+XRT, which includes features due to the elemental composition of the XRS and XRT. K-shell absorption edges from the oxygen (0.54 keV), aluminum (1.56 keV), and nickel (8.33 keV) in the blocking filters can be seen, as well as a number of weak M-shell features between 2-3 keV arising from the gold in the XRT. Of course, similar features exist in the total effective area of the XIS+XRT (for both FI and BI CCDs), seen in Figure 2.4, but the energy resolution of the XRS makes these more significant if they coincide with an astronomical feature.

The XRS includes a $6 \times 6$ array of individual detector elements (pixels), of which 30 are active^2.1, arranged as illustrated on Fig.2.5. There is also one calibration pixel used for gain tracking, set away from the main array and illuminated by a collimated $^{55}$Fe source (see § 6.2). As of this writing, the final assembly of the XRT for the XRS (the XRT-S) is in progress and its HPD has not yet been measured, but it is expected to be between 1.6 and 2.0 arcminutes. Gravity will also slightly distort the ground-based HPD measurements (cf. Fig.2.6) for all the XRTs in as-yet unknown ways. If the HPD is $2'$ (a conservative estimate), an on-axis point source will put $\sim 64$% of its flux onto the XRS array, which is $\sim 2.9' \times 2.9'$ on the sky.

Figure 2.5: The XRS calorimeter array, showing the layout of the pixels. Each pixel has a dimension of 0.64 mm $\times$ 0.64 mm, corresponding to a sky projection of $\sim 0.49' \times 0.49'$; the entire X-ray sensitive portion of the array projects a solid angle of $\sim 2.9' \times 2.9'$ on the sky. The pixel numbers are determined by the geometry of the detector assembly; the marked but unlabeled pixel is inactive as of March 2004. In this Figure, the array is shown projected onto the sky.

The four XISs (cf. Fig.7.2) are true imagers, and have a larger field of view ( $\sim 18' \times 18'$), but significantly lower spectral resolution than the XRS ($\sim130$eV at 6 keV; see Fig.2.9). The XISs do not require cryogenic cooling, and should continue to operate well after the cryogenic system has been exhausted.

Figure 2.6: The Encircled Energy Function (EEF) showing the fractional energy within a given radius for one quadrant of the XRT-I telescopes on Astro-E2 at 4.5 and 8.0 keV. The XRT-S EEF is expected to be similar.

The HXD (cf. Fig.8.1) is a non-imaging instrument with an effective area of $\sim300$cm$^{2}$, featuring a compound-eye configuration and an extremely low background. It dramatically extends the bandpass of the mission with its nominal sensitivity over the 10 - 600 keV band (cf. Fig.2.7). The HXD consists of two types of sensors: 2 mm thick silicon PIN diodes sensitive over 10 - 60 keV, and GSO crystal scintillators placed behind the PIN diodes covering 30 - 600 keV. The HXD field of view is actively collimated to $4.5^\circ \times 4.5^\circ$ by the well-shaped BGO scintillators, which, in combination with the GSO scintillators, are arranged in the so-called phoswich configuration. At energies below $\sim 100$ keV, an additional passive collimation further reduces the field of view to $34' \times 34'$. The energy resolution is $\sim$ 3.0 keV (FWHM) for the PIN diodes, and $7.6 / \sqrt{E}$ % (FWHM) for the scintillators (where $E$ is energy in MeV). The HXD time resolution for both sensors is 61 $\mu$s. While the HXD is intended mainly to explore the faintest hard X-ray sources, it can also tolerate very bright sources up to $\sim 10$ Crab.

Figure 2.7: Total effective area of the HXD detectors, PIN and GSO, as a function of energy.

Although this document emphasizes the XRS, each observation will contain data from all instruments since they are co-aligned and operate simultaneously.

2.2 Operational Constraints of Astro-E2

2.2.1 Raster-scanning

While a substantial change of the spacecraft attitude (more than a few degrees) is time-consuming, small offsets can be made efficiently. Such offsets would require that no new stars need to be acquired by the star trackers, and thus are limited to $\sim$ 1 to $2^{\circ}$. This is particularly useful for studies of extended sources larger than a few arcmin with the XRS (which is a unique tool for high resolution spectroscopy of extended sources). The minimum total time is still 10 ksec, and each point in the raster scan must be observed for at least 3 ksec. Raster-scanning observations should be specified on the RPS form, with the details to be entered in the ``remarks'' field. If the raster-scan pattern is complex or unusual, PIs are encouraged to contact the Astro-E2 team at ISAS/JAXA or the NASA Astro-E2 GOF for assistance before submitting a proposal.

2.2.2 Cooling system constraints

Another operational constraint for the XRS arises from the operation of the refrigerator for the XRS. To maintain the base temperature of the array at 60 mK, the instrument uses an Adiabatic Demagnetization Refrigerator (ADR), which has to be recharged periodically, about once every day. The recharge will take about 65 minutes. The current plan is to recharge when the spacecraft is slewing, or when a target of an observation is in the Earth's shadow, but this may change as a result of the on-orbit verification of the satellite.

2.2.3 Telemetry rates

Astro-E2 carries a 6 Gbit data recorder. Data will be downlinked to USC at a rate of 4 Mbps for a total of 2 Gbits per pass, up to 5 times a day. This allows a maximum of 10 Gbits of data to be obtained per day, but fewer passes may be available to Astro-E2 as it will share the use of USC ground station with other ISAS satellites^2.2. Data can be recorded at 4 different rates: Super-High (524 kbps), High (262 kbps), Medium (131 kbps), and Low (33 kbps). The recording rate will be changed frequently throughout an observation, according to a sequence that will be determined by the operations team at ISAS. This is to optimize the selection of the data rates and the usage of the data recorder, taking into account the expected count rates supplied by the proposers. Thus an accurate estimation of the count rates is important for the optimization of the mission operation. We emphasize that proposers cannot arbitrarily choose the data recording rate.

We expect that on-source data will usually be recorded at High (during contact orbits, during which the satellite passes over USC) or Medium (during remote orbits, without USC passes) data rate. The Low rate will be used mostly for times of Earth occultations and SAA passages, as it is probable that the background rates in the XIS and HXD exceed their telemetry allocation limit at Low data rate. The XRS telemetry allocation is a constant 160 ct/s regardless of the total data recording rate, while the telemetry limits for the XIS are presented in Chapter 7. Unless otherwise specified, the XIS data mode will be chosen for each data recording rate used to prevent telemetry saturation, based on the count rate supplied by the proposer.

2.3 Features and Capabilities of the XRS

The truly new and unique aspect of the Astro-E2 mission will be the availability of the XRS to obtain high-resolution X-ray spectra of a wide variety of celestial sources. The XRS differs from a grating spectrometer in two important ways. First, it is a non-dispersive detector so it can measure high-resolution spectra of extended sources as well as point source. Secondly, since the XRS energy resolution is nearly constant the resolving power $E/\Delta E$ increases with energy rather than decreasing.

2.3.1 Design and Performance

The XRS absorbs the energy of each photon $E_{\rm ph}$ in a pixel (one of the 30 active elements of the XRS detector), which is then converted into heat (see Fig.6.3). This results in a change of the temperature of the pixel $\Delta T$, which is roughly $\Delta T = E_{\rm ph} / C$, where $C$ is the heat capacity of the pixel. As discussed in Chapter 6, maintaining the sensitivity of the detector requires keeping its operating temperature to a minimum; at the operating temperature, the typical $C$ of all constituents of an XRS pixel is $\sim 10^{-6}$ erg/K. With this, a 1 keV photon with $E_{\rm ph}$ of $1.6 \times 10^{-9}$ erg produces $\Delta T$ of $\sim 1.6$ mK. The energy resolution of the XRS is thus related to the accuracy with which the above $\Delta T$ can be measured. The pre-launch energy resolution $\Delta E$ of the flight detector pixels is 6.5 eV (FWHM) at 6 keV. With adequate counts, the centroid energy of monochromatic photons can be determined to $\lesssim 2$ eV; this is likely to be limited by the systematics of the instrument, and specifically, the temperature stability of the cooling system, but should be aided by the on-board calibration sources (cf. §6.2). While at low energies, X-ray spectrometers such as the LETG on Chandra and RGS on XMM-Newton have better energy resolution, in the crucial Fe K region (6 - 7 keV), the energy resolution of the XRS is superior (by $\sim5 \times$) to the Chandra HETG (see Figs.2.8 & 2.9). Note that there are essentially no user-specified parameters for the XRS, except for the setting of the Filter Wheel (cf. §6.2), which is external to the instrument itself.

2.3.2 Unique attributes

While both XMM-Newton and Chandra have high resolution spectroscopy capabilities, there are several aspects where the Astro-E2 XRS will provide unique data. First, both XMM-Newton and Chandra rely on dispersive spectroscopy using diffraction gratings, and unambiguous spectra can be obtained only for point or point-like sources with angular extent (or small, well-defined structure) which is smaller than or comparable to the PSF of the telescope. There are many classes of diffuse celestial X-ray emitters (such as supernova remnants and clusters of galaxies) with a larger angular size, and these simply cannot be effectively studied with gratings. The XRS, on the other hand, is a non-dispersive instrument, and can study these sources and so it will open a new, unique window on the X-ray Universe. Secondly, the XRS has superior resolution at high energies compared to existing X-ray telescopes. The XMM-Newton gratings have essentially no effective area above $\sim 2$ keV and above $\sim 3$ keV, the energy resolution of the Chandra gratings is inferior to that of the XRS. Therefore, the high resolution spectroscopy with the XRS lends itself well towards answering such questions as:

• What is the composition and temperature of hot cosmic plasma (as inferred from the relative intensities of emission lines)?
• What is the velocity field of line-emitting material (as inferred from Doppler shifts of emission lines such as Fe K$\alpha$)?
• What is the nature and kinematics of hot, tenuous absorbing material in the line of sight to an astrophysical source (from profiles of absorption lines)?

2.3.3 Characteristics

By comparison with some other instruments (such as proportional counters) the time to detect and measure the energy of an X-ray photon with the XRS is relatively long-on the order of 100 milliseconds (although the arrival time can be measured to a few tens of microseconds). As a result, the throughput of events whose energy can be determined with high resolution will drop as the counting rate increases beyond $\sim3.5$ photons s$^{-1}$per pixel. The on-board signal processing (see §6.3) distinguishes ``high resolution'' events (entirely uncontaminated by temporally neighboring events), ``medium resolution'' events (slightly contaminated events, with $\Delta E_{\rm mid} \sim 1.2 \times \Delta E_{\rm hi}$), and ``low resolution'' events, with $\Delta E_{\rm low} \sim 5 \times \Delta E_{\rm hi}$. Note that this limitation affects each pixel independently, since all the channels are processed separately: even when the count rate on one pixel may be excessive, another pixel, receiving a lower photon rate, may be unaffected. As a result, considering the point spread function of the XRT, it is possible that the XRS pixels close to the nominal source position may be strongly affected, while the pixels farther away from the peak may yield mostly ``high resolution'' and ``medium resolution'' events. The relative fractions and event rates are illustrated in Figures. 2.10 & 2.11, which assume a Poisson distribution of arrival times between events from a point source.

The relative fraction of event types will also be affected by cosmic rays impinging on the XRS. Cosmic rays can heat the entire array, leading to false events on every pixel. The rate will have to be measured on orbit, but a conservative estimate is $\sim 0.5$count/s. The effect of these cosmic rays will be to decrease the number of high resolution events, as it puts a lower floor on the count rate in any pixel even for faint sources. Since the Hi-res record length is 0.16 s, the fraction of time affected by cosmic rays is 0.5 counts/s $\times 0.16$s = 8%. In earlier X-ray detectors such as CCDs or proportional counters, this would be considered unusable ``deadtime.'' In the XRS, cosmic rays do not cause true deadtime, but instead migrate events toward lower resolution.

For the studies of very bright sources, there is a filter wheel located in the X-ray beam between the mirror and the detector, capable of reducing the photon flux incident on the detector. This filter wheel (see § 6.2) has three different filter elements; OPEN, BE, and ND, with optional $^{55}$Fe calibration sources that illuminate the detector. BE is a soft X-ray cutoff filter containing 300$\mu$m of beryllium; its effect can be seen in Figure 2.3. ND is a neutral density filter that is partially opaque with energy-independent transmission of 10%. It is important to stress that except for the Crab nebula, there are no known extended sources such as clusters of galaxies or supernova remnants that would cause enough pulse pileup to warrant the use of any filters in the filter wheel. We thus expect that the filter wheel should be necessary only for bright Galactic binary sources (with flux in excess of $\sim 50$ milliCrab).

Figure 2.8: Comparison of the effective area of the XRS against XMM-Newton's RGSs and Chandra's grating instruments.

Figure 2.9: Comparison of the energy resolution of the XRS and XIS against XMM-Newton's RGSs and Chandra's grating instruments.

The XRS has a relatively modest field of view, covering only $\sim 2.9' \times 2.9'$. For sources with larger angular extent, multiple pointings may be required, increasing the integration time needed to map a large diffuse region. In addition, for point sources, the energy resolution of the XMM-Newton RGS or Chandra LETG is superior below $\sim 2$ keV.

2.3.4 Summary

The XRS excels for observations such as:

• Spectroscopy at energies $E > 2$ keV where the highest spectral resolution is required, and in particular, in the Fe K line region.

• All high spectral resolution studies of extended sources (modulo the surface brightness; very faint and/or large sources may require inordinate integration times in multiple pointings to take advantage of the energy resolution of the XRS, and are better suited to study with instruments subtending a larger solid angle on the sky).

The XRS is less appropriate for:

• Studies requiring primarily high spatial resolution: the resolution of Chandra is about 100 times better, while the XMM-Newton is about 10 times better.

• Studies of very large ($>$ a degree), diffuse objects with low surface brightness: collecting a sufficient number of photons for a meaningful study may require unreasonable integration times.

• Soft X-ray spectroscopy of point sources at $E < 1.0$ keV. The XMM-Newton RGSs and Chandra LETG are more appropriate, as the Astro-E2 XRS has worse spectral resolution and a lower effective area than these.

• Spectroscopy of broad ($>$ 200 eV) spectral features. For this, the XMM-Newton EPIC and Chandra ACIS detectors are superior, as they provide greater collecting area.

• Rapid variability studies on millisecond time scales: for bright sources, the counting rate limitation requires the use of filters, resulting in the loss of counts. For such studies, the RXTE or XMM-Newton may be more appropriate. Nonetheless, the XRS does have good time resolution, and may be an excellent instrument for temporally resolved spectroscopy (pulsars).

Figure 2.10: The fraction of X-ray photons processed into the ``low-res,'' ``mid-res,'' and ``high-res'' events by the XRS for a point source with a power-law spectrum ($\Gamma =2.0$) between 0.5-10 keV. The mid-res counts are separated into primary events, which have nearly the same resolution as the hi-res events, and secondary events (marked with an ``S'') which have somewhat lower resolution. Secondary events occur when the event time is within 142 ms of an earlier pulse (see also Fig.6.5).

Figure 2.11: Similar to Fig.2.10, but showing the counting rate as a function of source flux. This figure includes the effects of the point-spread function of the mirror.

2.4 Calibration Plan

Observations for scientific instrument calibration are to start $\sim$3 weeks after the launch. Those to calibrate the functions essential for achieving the mission goals have the highest priorities, and hence, all of them will be carried out by the end of Phase-1A. After that, long term variation of the functions will have been monitored every half a year or so. The list of all the calibration targets during Phase-1A can be found on the Astro-E2 websites listed in Appendix C.

Table 2.2 summarizes the calibration items of all scientific instruments and their expected accuracy.

Table 2.2: Error Budgets of Scientific Instrument Calibrations

S/S	Calibration Item	Requirement	Goal
XRT	absolute effective area (2-10 keV in XIS)	5%	5%
	XRT-S to XRT-I effective area cross-calibration	5%	5%
	vignetting for all XIS f.o.v.	2%	2%
	on-axis PSF (for all integration radii from 1$'$-6$'$)	1%	1%
	off-axis PSF (as on-axis but for all XIS f.o.v.)	2%	2%
	optical axis position in XRS	$< 1'$	$< 1'$
	optical axis position in XIS	$< 0.2'$	$< 0.2'$
XRS	energy scale	2 eV	1 eV
	line spread function (FWHM)	1 eV	1 eV
	effective area	10%$^{\rm a}$	5%$^{\rm a}$
	absolute timing	300 $\mu$s	100 $\mu$s
	relative timing	80 $\mu$s	10 $\mu$s
XIS	Q.E. at 10 keV	5%$^{\rm b}$	5%$^{\rm b}$
	energy scale	0.1%$^{\rm b}$	0.1%$^{\rm b}$
	resolution (FWHM) at 5.9 keV	1%$^{\rm b}$	1%$^{\rm b}$
	Q.E. vs position	5%$^{\rm b}$	5%$^{\rm b}$
	OBF integrity	broken/unbroken	broken/unbroken
HXD	absolute effective area	20%	5%
	relative effective area	10%	5%
	vignetting	10%	5%
	background modelling$^{\rm c}$	$\sim$10%$^{\rm c}$	5%$^{\rm c}$
	absolute timing	300 $\mu$s	100 $\mu$s
	relative timing	10$^{-8}$	10$^{-10}$
	GBST absolute timing	100 ms	15 ms

Note $\cdots$ All the values quoted are preliminary.
a: Including the errors associated with the XRT.
b: All values of the XIS are based on the ground calibration data and very preliminary. See the text in more detail.
c: Study continues. The HXD team will prepare some documents on HXD BGD estimation for updated results by the start of AO-1 observations.

Note that the values listed are those required from the scientific purpose and ultimate goals which are possible to be realized on the basis of the instrument design, and are not the results of real measurements. Since the scientific instruments of Astro-E2 are very sensitive to the environment, it scarcely makes sense to list the values of the ground-based calibration. We will add a column to the table representing the results from inflight calibration until the start of AO-1 observations (2005 September). Please regard the current values in the table as providing rough idea on the order of calibration accuracy in orbit.

• Note from the XIS team -- All the values about the XIS in the table are estimated from the very preliminary results of the ground calibrations of the FI sensors. The goals of the BI sensor calibrations should be almost the same. Furthermore, since the XIS will suffer the radiation damage, it is very difficult to predict how the characteristics of the XIS will change even in the first half year after the launch. The XIS team will update these values as ground or flight calibrations progress.
• Note from the HXD team -- It is very difficult to estimate the background levels on the ground. The HXD team will continue study on them, and will prepare some documents on HXD BGD estimation until the start of AO-1 observations.