XRISM: X-ray Imaging & Spectroscopy Mission

• XRISM is an advanced X-ray observatory that combines non-dispersive microcalorimeter spectroscopy with wide-field CCD imaging to probe cosmic plasma dynamics.
• The mission employs complementary instruments—Resolve for precise high-resolution spectroscopy and Xtend for contextual imaging—to study galaxy clusters, black holes, and diffuse sources.
• XRISM’s open-data operational model integrates automated pipelines and guest-observer support, accelerating astrophysical discoveries through methodical calibration and public access.

The X-ray Imaging and Spectroscopy Mission (XRISM) is Japan’s seventh X-ray astronomical observatory and a JAXA–NASA mission with ESA participation, conceived as the recovery of the high-resolution soft X-ray spectroscopy capability lost after Hitomi. Launched on September 7, 2023, XRISM is built around the joint use of non-dispersive microcalorimeter spectroscopy and wide-field CCD imaging to probe material circulation and energy transfer in cosmic plasmas, clarify the nature of cosmic structures and objects, and restore a broad community platform for precision X-ray line diagnostics (Mori et al., 2024, Team, 2020).

1. Mission identity and scientific objectives

XRISM was designed to restore high-resolution X-ray spectroscopy as a core observatory capability rather than as a narrow technology demonstration. Its stated scientific goals include revealing the structure formation and evolution of galaxy clusters, understanding the circulation history of baryons, investigating the transport and circulation of energy in the Universe, and enabling new science with high-resolution X-ray spectroscopy (Team, 2022). In the mission white paper, these goals are recast in terms of measuring line centroids, widths, and relative intensities precisely enough to probe gas dynamics, plasma conditions, chemical enrichment, and compact-object accretion physics across sources ranging from comets and the interstellar medium to black holes, supernova remnants, clusters of galaxies, and the warm-hot intergalactic medium (Team, 2020).

A defining methodological feature is that XRISM performs non-dispersive spectroscopy. This matters particularly for extended and morphologically complex sources, because unlike grating spectrometers it preserves spatial information while resolving emission and absorption features. The mission documentation emphasizes that Resolve provides spectral resolution 20–40 times better than CCD instruments on Chandra, XMM-Newton, and Suzaku, while Xtend supplies the larger-scale imaging and contextual spectroscopy needed to interpret the much smaller calorimeter field (Team, 2020).

The mission was also architected as a public observatory. Its science-operations design explicitly separates mission operations from science operations, formalizes guest-observer support, and treats software, calibration, and data distribution as first-class mission products rather than auxiliary tasks. That operational model was derived from lessons identified from ASCA, Suzaku, and Hitomi, including the need to start science-operations preparation early, avoid non-public “animal software,” and maintain clear boundaries among scientists, engineers, and instrument teams (Terada et al., 2021).

2. Observatory configuration and instrument complement

XRISM carries two soft X-ray telescopes with a shared basic mirror design and a common focal length of $5.6\,\mathrm{m}$: Resolve, the high-resolution spectroscopy telescope, and Xtend, the wide-field imaging telescope (Mori et al., 2024). Both are fed by X-ray Mirror Assemblies based on a thin-foil-nested, conically approximated Wolter-I design. Each mirror assembly contains 203 nested shells; the mirror system has a diameter of $45\,\mathrm{cm}$, and the quick reference summarizes an HPD of $1.3\,\mathrm{arcmin}$ for XRISM/XMA, improved over Suzaku/XRT-I (Team, 2022).

Resolve is the flagship spectrometer. It is a $6\times6$ pixel X-ray microcalorimeter array operating at about $50\,\mathrm{mK}$, covering $0.3$–$12\,\mathrm{keV}$ with a $2.9' \times 2.9'$ field of view and a mission energy-resolution target of $\leq 7\,\mathrm{eV}$ FWHM at $6\,\mathrm{keV}$ (Team, 2022). Xtend is the complementary soft X-ray imaging spectrometer. It combines the XMA with the Soft X-ray Imager (SXI), an X-ray CCD camera, to provide a $45\,\mathrm{cm}$0 field of view over $45\,\mathrm{cm}$1–$45\,\mathrm{cm}$2 (Mori et al., 2024). The mission documents repeatedly frame the pair as deliberately complementary: Resolve delivers unprecedented high-resolution spectroscopy over a small field, whereas Xtend supplies broad spatial context, background characterization, source finding, and mapping of diffuse emission (Mori et al., 2023).

This complementarity is not merely geometric. Resolve is the instrument for resolving narrow line complexes, bulk velocities, turbulence, ionization structure, and line broadening; Xtend is the instrument that contains Resolve’s field within a much larger imaging region, monitors contaminating sources, and provides the wide-field grasp required for diffuse targets and serendipitous transients (Mori et al., 2024). A plausible implication is that XRISM’s scientific identity is defined less by either instrument in isolation than by the combined use of narrow-field calorimetry and contextual imaging spectroscopy.

3. Xtend: detector architecture, commissioning, and in-orbit performance

Xtend consists of the X-ray Mirror Assembly and the Soft X-ray Imager. The SXI uses four P-channel, back-illuminated CCDs with a $45\,\mathrm{cm}$3 thick depletion layer, arranged in a $45\,\mathrm{cm}$4 grid (Mori et al., 2024). Each CCD has a $45\,\mathrm{cm}$5 physical-pixel format and is read out as $45\,\mathrm{cm}$6 logical pixels via on-chip $45\,\mathrm{cm}$7 binning, giving a pixel scale of $45\,\mathrm{cm}$8 (Mori et al., 2024). The CCDs are cooled by a single-stage Stirling cooler to $45\,\mathrm{cm}$9 in nominal operation, while pre-launch thermal-vacuum tests also demonstrated operation at $1.3\,\mathrm{arcmin}$0 (Mori et al., 2024).

The pre-flight design retained Hitomi/SXI heritage but introduced specific changes aimed at two known failure modes: optical leakage and charge-transfer degradation. The CCD incident surface was given a $1.3\,\mathrm{arcmin}$1 aluminum coat to improve optical and infrared blocking, and a notch structure was introduced to reduce CTI; the contamination-blocking filter and baked components were designed to suppress molecular deposition, with the estimated contamination on CCD surfaces kept below $1.3\,\mathrm{arcmin}$2 over three years (Noda et al., 12 Feb 2025). Ground testing with the full flight-model configuration showed $1.3\,\mathrm{arcmin}$3Fe energy resolution of $1.3\,\mathrm{arcmin}$4–$1.3\,\mathrm{arcmin}$5 FWHM at $1.3\,\mathrm{arcmin}$6, optical leakage below $1.3\,\mathrm{arcmin}$7 when CCD and filter suppression were combined, CCD quantum efficiency of 0.957 at $1.3\,\mathrm{arcmin}$8 and 0.993 at $1.3\,\mathrm{arcmin}$9, and filter transmissivity of 0.857 at $6\times6$0 and 0.892 at $6\times6$1 (Noda et al., 12 Feb 2025).

Commissioning followed a staged activation after launch. The initial operation of SXI began on October 17, 2023, about 40 days after launch, with electronics powered on first, cooling initiated on October 20, and the CCDs stabilized at $6\times6$2 after a small adjustment from the originally planned $6\times6$3 (Suzuki et al., 2024). The more-than-one-month initial operation verified raw telemetry integrity, observation modes, automated and time-tagged procedures, South Atlantic Anomaly handling, and dark-level management. All four nominal modes were exercised: full-window, $6\times6$4-window, $6\times6$5-s burst, and $6\times6$6-window+burst; all use a 4 s frame cycle, with exposures of 3.96 s for full-window, 0.46 s for $6\times6$7-window, and 0.06 s for the burst modes (Suzuki et al., 2024).

The in-orbit performance of Xtend was reported as fully consistent with ground measurements and compliant with mission requirements. The measured XMA half-power diameter and effective area were consistent with ground calibration, the in-orbit energy resolution was around $6\times6$8 and thus satisfied the beginning-of-life requirement of $6\times6$9, the non-X-ray background was low, stable, and clean, and there was no apparent contamination buildup more than half a year after launch (Mori et al., 2024). The first-light image of Abell 2319 showed that the full $50\,\mathrm{mK}$0 imaging region was functioning for X-ray detection except for known bad columns identified in ground tests (Mori et al., 2024).

Xtend also developed contingency mitigation against a specific CCD artifact discovered during 2020/2021 ground cooling tests: anomalous charge intrusion from outside the imaging area. That effect produced saturated “Icicle,” “Mountain,” and “Field” structures with pulse heights pinned to 4095 ch. A new CCD driving technique was then developed in which $50\,\mathrm{mK}$1 is maximized and $50\,\mathrm{mK}$2/$50\,\mathrm{mK}$3 minimized during exposure, with only brief $50\,\mathrm{mK}$4 returns to normal potentials every 80 logical lines for charge injection. In non-flight-model reproduction tests this suppressed the anomaly, and flight-model thermal-vacuum tests at $50\,\mathrm{mK}$5 showed count maps and Mn-K$50\,\mathrm{mK}$6 spectra consistent with normal operation (Noda et al., 9 Mar 2025).

Operationally, Xtend is optimized for wide-field soft X-ray imaging spectroscopy rather than for bright-source timing. Its ground-calibration effective area was approximately $50\,\mathrm{mK}$7 at $50\,\mathrm{mK}$8 and $50\,\mathrm{mK}$9 at $0.3$0, its frame cycle is 4 s, and charge injection is applied every 80 logical rows to suppress charge-transfer inefficiency (Mori et al., 2024). This combination of wide field, low and clean background, relatively sharp PSF, and high-energy response is what underlies its stated suitability for diffuse X-ray sources, supernova remnants, galaxy cluster outskirts, solar wind charge exchange emission, and serendipitous transients (Mori et al., 2024).

4. Resolve: calorimeter performance, background control, and timing fidelity

Resolve is the enabling instrument for XRISM’s high-resolution spectroscopy, and several in-orbit characterization studies have refined how its nominal performance is realized in practice. The on-orbit energy resolution is about $0.3$1 at $0.3$2, exceeding the original mission target of $0.3$3 FWHM (Mochizuki et al., 11 Jan 2025). The instrument’s low-background requirement is $0.3$4 over 0.3–12.0 keV, equivalent to about one background event per spectral bin per 100 ks exposure, and its event screening is therefore a major part of the performance budget (Mochizuki et al., 11 Jan 2025).

The screening framework optimized 19 items in three classes: event pulse shape, relative arrival times among multiple events, and good time intervals (Mochizuki et al., 11 Jan 2025). Using 650 ks of in-orbit non-X-ray background data from October 2023 to February 2024, the initial screening produced a background rate of $0.3$5, meeting the mission requirement, while additional screening based on correlations among pulse-shape properties reduced the background further to $0.3$6, with the strongest improvement below $0.3$7 (Mochizuki et al., 11 Jan 2025). This is particularly relevant for faint diffuse targets, where the mission’s scientific advantage depends as much on background cleanliness as on intrinsic calorimeter resolution.

Bright-source performance introduces a different class of systematics. In orbit, high-count-rate effects were characterized with Crab offset observations and a GX 13+1 test case. The analysis identified three main effects: exposure-time loss from PSP overflow, energy-scale shifts at 6 keV, and energy-resolution degradation from electrical cross talk between neighboring pixels (Mizumoto et al., 7 Jun 2025). At low count rates the energy offset at 6 keV is about $0.3$8, attributed mainly to orbital variation in Resolve electronics combined with sparse fiducial sampling; at higher rates the offset becomes progressively negative owing to local heating by incoming X-ray flux (Mizumoto et al., 7 Jun 2025). The degradation in FWHM at 6 keV before cross-talk screening is described by

$0.3$9

with $12\,\mathrm{keV}$0, $12\,\mathrm{keV}$1, and $12\,\mathrm{keV}$2, where $12\,\mathrm{keV}$3 is the pixel-GTI-corrected count rate and $12\,\mathrm{keV}$4 is the FWHM in eV. After applying a nearest-neighbor coincidence “cross-talk cut,” the slope becomes $12\,\mathrm{keV}$5 and the baseline is restored to $12\,\mathrm{keV}$6 (Mizumoto et al., 7 Jun 2025). For precision velocity work at the level of tens of $12\,\mathrm{keV}$7, the paper explicitly warns that pixel-dependent eV-scale offsets must be accounted for (Mizumoto et al., 7 Jun 2025).

XRISM’s timing system was separately designed to meet a mission requirement of 1.0 ms absolute timing accuracy at $12\,\mathrm{keV}$8 (Terada et al., 18 Mar 2025). The bus+ground component was allocated a $12\,\mathrm{keV}$9 absolute budget, verified on the ground and then tested in orbit with PSR B1937+21 and simultaneous Crab pulsar observations with NICER and NuSTAR (Terada et al., 18 Mar 2025). The commissioning pulsar test found bus+ground timing jitter below $2.9' \times 2.9'$0, and the Crab campaign showed the main pulse aligned with NICER and NuSTAR within $2.9' \times 2.9'$1 (Terada et al., 18 Mar 2025). This places XRISM’s bus+ground timing performance comfortably within its allocated budget, with the remaining mission-level timing assessment depending on the Resolve instrument timing calibration itself.

5. Science operations, pipeline architecture, and analysis ecosystem

XRISM’s science operations were designed as an independent system rather than an extension of instrument development. The Science Operations Team is defined around four responsibilities: guest-observer program and data distribution, distribution of analysis software and the calibration database, guest-observer support, and performance verification and optimization activities (Terada et al., 2021). Operationally, responsibilities are distributed across SOC at JAXA, SDC at NASA, and ESAC at ESA, while mission operations remain separate under the Mission Operations Team (Terada et al., 2021). This structure was an explicit response to lessons from previous Japanese X-ray missions regarding software delays, informal support arrangements, and the need to avoid hidden internal tools (Terada et al., 2021).

The data-processing chain is two-stage. Telemetry generated onboard as SMCP messages is downlinked as CCSDS space packets, stored in SIRIUS, converted by the JAXA pre-pipeline into Raw Packet Telemetry FITS files and then First FITS Files, and finally calibrated by the NASA pipeline into Second FITS Files and cleaned-event products (Eguchi et al., 24 Jul 2025). The mission documentation stresses that no data are lost during PPL or PL; later stages populate FITS tables with calibrated values rather than discarding information (Eguchi et al., 24 Jul 2025). Scientists receive SFFs and ready-to-analyze cleaned-event FITS files, while the intermediate RPT and FFF products remain internal (Eguchi et al., 24 Jul 2025).

At the SDC, the pipeline is implemented as a daemon-driven automated system spanning Japan and the United States. The pre-fetch daemon triggers xDTS transfer from ISAS to GSFC, fetch validates and prepares data, stream executes the pipeline script, post_proc validates outputs, pre_archive packages products, archive transfers them to archives in both countries, and deleteit cleans the post-processing area (Doyle et al., 2022). Processing uses HEASoft plus XRISM-specific tasks; the 2022 pipeline paper counted 62 tasks in total, including Resolve, Xtend, multi-instrument, and general utilities, and emphasized preview products such as spectra, light curves, and images for rapid inspection (Doyle et al., 2022).

As reprocessing demand increased after commissioning and the performance-verification period, the JAXA pre-pipeline was ported to the TOKI-RURI HPC system using Singularity containers and ext3 working disk images. For 161 OBSIDs, this produced a reported speedup of 33 relative to the regular virtual-machine environment, with zero failed OBSIDs after the ext3-based solution to the 32-bit inode problem was adopted (Eguchi et al., 24 Jul 2025). The practical significance is operational rather than merely computational: central reprocessing campaigns that would otherwise take weeks become manageable within mission-required timescales (Eguchi et al., 24 Jul 2025).

The broader software ecology also includes quick-look scientific tools. “XSLIDE” (Braun et al., 2022) is a graphical interface for rapid inspection of Resolve spectra that approximates direct spectral reconstruction by treating the RMF as effectively diagonal and the ARF as locally constant. It supports rebinning, continuum fitting, automatic line finding via a piecewise statistics method or continuous wavelet transform, line identification using AtomDB 3.0.9 and XSTAR-generated databases, approximate diagnostics such as Ly$2.9' \times 2.9'$2/Ly$2.9' \times 2.9'$3, $2.9' \times 2.9'$4, $2.9' \times 2.9'$5, and K$2.9' \times 2.9'$6/K$2.9' \times 2.9'$7 ratios, and export to formats usable in XSPEC (Braun et al., 2022). Its role is not precision inference but fast triage of line-rich spectra, which is particularly valuable for a mission whose scientific leverage lies in identifying fine spectral structure early and efficiently.

6. Scientific reach and early observational results

XRISM’s early literature already shows the intended division of labor between Resolve and Xtend. In solar physics, a by-product observing mode during day-Earth occultation has become scientifically productive: using data from October 2023 to November 2024, Xtend measured Mg, Si, S, Ar, Ca, and Fe abundances in reflected solar-flare X-rays, finding an inverse-FIP pattern and flare-class-dependent trends in which Si, S, and Ar decrease with increasing flare magnitude जबकि Ca shows the opposite trend; Resolve then separated Rayleigh- and Compton-scattered Fe XXIV/XXV lines from neutral or low-ionized Fe-K$2.9' \times 2.9'$8, with the neutral/low-ionized Fe-K$2.9' \times 2.9'$9 equivalent width anti-correlated with hard X-ray flux with best-fit slope $\leq 7\,\mathrm{eV}$0 (Suzuki et al., 5 Sep 2025). This demonstrates that the mission’s “by-product” data can support high-statistics CCD spectroscopy and high-resolution calorimeter spectroscopy in the Fe-K band.

Xtend’s wide field and low background have also opened a parameter space in faint diffuse emission around compact Galactic sources. In September 2024, XRISM detected extended X-ray emission around the PeVatron microquasar V4641 Sgr with $\leq 7\,\mathrm{eV}$1 significance from the radial-profile analysis and $\leq 7\,\mathrm{eV}$2 from the spectral analysis, finding a characteristic Gaussian width of $\leq 7\,\mathrm{eV}$3 arcmin, corresponding to $\leq 7\,\mathrm{eV}$4 pc at $\leq 7\,\mathrm{eV}$5 (Suzuki et al., 2024). The paper interprets the result either as synchrotron emission requiring an enhanced magnetic field of order $\leq 7\,\mathrm{eV}$6 or a suppressed diffusion coefficient of order $\leq 7\,\mathrm{eV}$7 at 100 TeV, or as thermal plasma from a jet termination shock with $\leq 7\,\mathrm{eV}$8 and an implied jet luminosity of $\leq 7\,\mathrm{eV}$9 (Suzuki et al., 2024). Independently of the preferred physical model, the detection exemplifies Xtend’s stated niche in faint, wide-field structures.

In accreting compact objects, Resolve has begun to separate physical regimes that were blended in pre-XRISM spectroscopy. The review “X-ray Spectroscopy of Disk Winds in Black Hole X-ray Binaries” (Shidatsu et al., 30 Jan 2026) argues that XRISM has already distinguished true blueshifted winds from hot static atmospheres, obscuring outer-disk plasmas, and unusual emission-line systems. The clearest early case is 4U 1630$6\,\mathrm{keV}$0472, where Resolve resolved Fe XXV He$6\,\mathrm{keV}$1 and Fe XXVI Ly$6\,\mathrm{keV}$2 absorption profiles and found a persistent component with line-of-sight velocity $6\,\mathrm{keV}$3, interpreted not as an escaping wind but as a hot, bound disk atmosphere (Shidatsu et al., 30 Jan 2026). That result addresses a long-standing ambiguity in black-hole wind studies: whether Fe-K absorption implies outflow by default or can instead arise in a non-escaping atmosphere.

The mission’s sensitivity to line kinematics is also evident in neutron-star binaries. In Cyg X-2, Resolve measured broad Fe-K emission lines from the accretion disk corona with velocity dispersion $6\,\mathrm{keV}$4, argued for at least two distinct ADC regions, and marginally detected a blueshifted absorption feature interpretable as Fe XXVI or Fe XXV with outflow velocity $6\,\mathrm{keV}$5 or $6\,\mathrm{keV}$6, respectively, implying a mass-loss rate of order $6\,\mathrm{keV}$7 (Mizumoto et al., 29 Nov 2025). The paper explicitly compares this with narrower XRISM Fe-K lines in GX 340+0 and with earlier Chandra measurements, using the broader Cyg X-2 lines to argue for a more turbulent and dynamically active ADC (Mizumoto et al., 29 Nov 2025).

Taken together, these results show XRISM functioning in the mode for which it was designed. Resolve is resolving Fe-K structure, measuring centroids and widths at the eV level, and separating multi-zone plasma components; Xtend is supplying large-grasp imaging spectroscopy, diffuse-emission sensitivity, and source context over a $6\,\mathrm{keV}$8 field. The mission papers do not present this as a finished calibration state—extensive calibration studies are repeatedly said to remain in progress—but they already show that the observatory is operating as a technically mature, scientifically general, and methodologically distinctive platform for high-resolution X-ray spectroscopy combined with wide-field imaging (Mori et al., 2024).