XRISM Resolve Calorimeter Array

• Resolve is a soft X-ray imaging spectrometer that employs a 36-pixel micro-calorimeter array operating at 50 mK to deliver high energy resolution (<7 eV) over the 0.3–12 keV range.
• Its design, inherited from Hitomi/SXS, integrates optimal filtering, detailed calibration frameworks, and advanced cryogenic cooling systems to ensure high spectral fidelity and imaging capability.
• Ground tests and early observations, such as those of Mrk 509, highlight its ability to resolve complex spectral features and separate narrow and broad emission components.

Searching arXiv for recent and primary sources on XRISM/Resolve. Resolve is the soft X-ray imaging spectrometer on the X-Ray Imaging and Spectroscopy Mission (XRISM), and its focal plane is a 36-pixel X-ray micro-calorimeter array operated at 50 mK. It is the recovery-mission successor to the Hitomi/ASTRO-H Soft X-Ray Spectrometer (SXS), with mostly the same design and the same intended in-flight performance. In this configuration, the array covers a $3' \times 3'$ field of view and is designed to achieve an energy resolution of better than 7 eV over the 0.3 -- 12 keV energy range for more than 3 years in orbit (Sato et al., 2023).

1. Mission role and instrument identity

Resolve is the spectrometer carried by XRISM for non-dispersive, high-resolution X-ray spectroscopy. The instrument is explicitly described as the recovery version of Hitomi/SXS: XRISM is equipped with the Resolve spectrometer, which has mostly the same design as SXS and is expected to have the same in-flight performance (Sato et al., 2023).

This continuity is central to the identity of the calorimeter array. The retained elements include the 36-pixel X-ray micro-calorimeter array, 50 mK operating temperature, 0.3 -- 12 keV energy coverage, the $3' \times 3'$ field of view, the general readout architecture, event grading, optimal filtering, the general cryogenic-chain philosophy using liquid helium, mechanical coolers, and adiabatic demagnetization refrigerators (ADRs), and the same anti-coincidence detector concept (Sato et al., 2023). The principal Resolve-specific changes discussed in the source concern the surrounding instrument and mission robustness rather than the calorimeter principle itself. In particular, a new cylindrical aperture baffle was added around the gate-valve area to reduce micro-meteoroid and orbital debris risk and to help against Earth albedo optical light leak and atomic oxygen (Sato et al., 2023).

Hitomi/SXS achieved the energy resolution of $\sim$5 eV in orbit, but it was lost after only a month of operation due to the loss of spacecraft attitude control. A plausible implication is that Resolve inherits a detector architecture with demonstrated in-orbit spectral performance but was configured within a mission framework shaped by that earlier failure mode (Sato et al., 2023).

2. Detector array architecture and operating principle

The Resolve calorimeter array is a 36-pixel detector arranged as $6 \times 6$ pixels at an 832 ${\rm \mu m}$ pitch (Sato et al., 2023). One upper-left corner position outside the telescope field of view is a calibration pixel, so the hardware is a 36-position array while the on-sky field is defined by the mirror illumination (Sato et al., 2023).

Each pixel consists of an ion-implanted Si thermistor with an HgTe absorber (Sato et al., 2023). The operating principle is calorimetric: an incident X-ray is absorbed in a low-heat-capacity absorber, converting photon energy into heat, with temperature rise approximately

$\Delta T = E / C,$

where $E$ is the photon energy and $C$ is the sensor heat capacity (Sato et al., 2023). The thermal pulse decays on a timescale

$\tau \sim C/G,$

where $G$ is the thermal conductance to the heat sink (Sato et al., 2023). The thermistor senses the transient temperature rise as a resistance change. Because the detector measures deposited energy thermally rather than by charge collection, the spectral resolution is largely independent of photon energy across the bandpass (Sato et al., 2023).

The energy-resolution scaling is given as

$3' \times 3'$0

where $3' \times 3'$1 is Boltzmann’s constant, $3' \times 3'$2 is absorber temperature, and $3' \times 3'$3 is the logarithmic thermometer sensitivity (Sato et al., 2023). The thermometer operates in the variable-range-hopping regime, with resistance described by

$3' \times 3'$4

although the source notes that ion-implanted Si often deviates from this exact form (Sato et al., 2023). For the SXS heritage detector, the thermistor sensitivity at operating temperature under bias was $3' \times 3'$5 (Sato et al., 2023).

The absorber material is HgTe, chosen for small heat capacity and large stopping power for X-rays (Sato et al., 2023). The paper also quotes, for SXS heritage, designed quantum efficiency at 6 keV and filling factor both greater than 95\%, though it does not separately restate those exact numbers specifically for Resolve (Sato et al., 2023). This suggests a close continuity between the inherited focal-plane design and the XRISM implementation.

3. Cryogenic environment and supporting detector systems

Resolve operates at 50 mK because the required energy resolution of $3' \times 3'$6 eV (FWHM) at 6 keV demands an ultra-low thermal-noise environment (Sato et al., 2023). The cryogenic chain combines 30 L liquid helium, a 2-stage ADR for detector cooling, an additional stage-3 ADR for cryogen-free operation after helium depletion, a $3' \times 3'$7He Joule-Thomson cooler, and multiple 2-stage Stirling coolers (Sato et al., 2023).

The cooling architecture is organized so that the 2-stage Stirling coolers cool shields to about $3' \times 3'$8 K, the Joule-Thomson cooler cools the innermost shield to about $3' \times 3'$9 K, liquid helium provides a heat sink for the low-temperature stage, and the 2-stage ADR cools the calorimeter thermal sink and detector array to 50 mK (Sato et al., 2023). In cryogen mode, stage-1 recycle occurs about every 44 hours in orbit, recycle takes about an hour, and detector-cooling duty cycle is $\sim$0 (Sato et al., 2023). In cryogen-free mode, stage-1 recycles about every 16 hours in orbit, each recycle takes about an hour, and detector-cooling duty cycle is again $\sim$1 (Sato et al., 2023).

Ground tests give more specific performance figures. The cryogen-mode hold time is about 37.8 hours at liquid-helium temperature typically about 1.24 K, with operational duty cycle $\sim$2; the cryogen-free hold time is about 16.7 hours with duty cycle $\sim$3 (Sato et al., 2023). The temperature stability requirement is 2.5 $\sim$4K RMS over 10 minutes, and achieved stability was 0.6 $\sim$5K RMS average in cryogen mode and 0.7 $\sim$6K RMS in cryogen-free mode (Sato et al., 2023).

The focal-plane assembly includes a rear anti-coincidence detector behind the calorimeter array. This anti-co detector is a silicon ionization detector, specifically 1 cm$\sim$7 $\sim$8 0.5 mm high-purity silicon configured as a p-i-n diode and covering a larger area than the micro-calorimeter array (Sato et al., 2023). A plausible implication is that the array is embedded within a broader background-rejection and thermal-control system rather than functioning as an isolated sensor plane.

4. Readout chain, event processing, and calibration framework

Because the calorimeter thermistors have high electrical impedance, the design uses a JFET source follower to convert the detector’s high impedance to a lower output impedance, reducing sensitivity to microphonics through cable capacitance (Sato et al., 2023). Signals from the calorimeter pixels and anti-co detector are amplified and digitized in the X-ray amplifier BOX, with ADC sampling at 12.5 kHz (Sato et al., 2023).

The pulse shape processor performs event triggering, pulse detection, grade assignment, optimal filtering, and pulse-height estimation (Sato et al., 2023). Optimal filtering uses a pixel-dependent template based on the average pulse and noise spectrum, and pulse height is found by cross-correlation of waveform and template (Sato et al., 2023). Resolve launch templates were generated on the ground from instrument-level test data using the flight dewar and flight electronics, and high-frequency weight was removed with a cut-off frequency of 366 Hz (Sato et al., 2023).

Event grades are subdivided into Hp, Mp, Ms, Lp, and Ls (Sato et al., 2023). The high-resolution pulse record length is 1024 samples (81.92 ms), the medium-resolution pulse record length is 256 samples (20.48 ms), and the key pulse-separation intervals used for event grading are 18.32 ms and 70.72 ms (Sato et al., 2023). Event timing is interpolated to 1/16 sample, corresponding to 5 $\sim$9sec resolution, while the absolute timing accuracy requirement is 1 ms (Sato et al., 2023).

The offline timing calibration relation is

$6 \times 6$0

For SXS and Resolve, only coefficient $6 \times 6$1 is used, with $6 \times 6$2 (Sato et al., 2023).

Calibration is distributed across three source classes: a calibration pixel with a collimated $6 \times 6$3Fe source, a $6 \times 6$4Fe source on the filter wheel, and a modulated X-ray source (MXS) (Sato et al., 2023). The calibration pixel provides continuous illumination and ongoing monitoring of gain scale and line spread function; the filter-wheel source provides supplemental full-array calibration with Mn fluorescent lines; and the MXS illuminates all pixels in pulsed operation, allowing gain tracking during observations (Sato et al., 2023). Ground calibration files are prepared for 49 mK, 50 mK, and 51 mK operation temperatures for each channel (Sato et al., 2023). The gain-drift tool is rslgain, and pulse height to energy conversion uses rslpha2pi (Sato et al., 2023).

5. Performance specifications and response model

The mission-level requirement is energy resolution of better than 7 eV over the 0.3 -- 12 keV energy range, with orbital operation for more than 3 years (Sato et al., 2023). The direct heritage benchmark is Hitomi/SXS, which achieved 5 eV (FWHM) at 6 keV in orbit (Sato et al., 2023). For Resolve itself, the source does not claim in-flight performance in the 2023 overview paper, but it states that in the March 2022 instrument-level ground test the energy resolution for all the pixels met the requirement with a margin for high and medium resolution grade, and the absolute energy scale over 0.3 -- 9 keV also met the requirement with margin (Sato et al., 2023).

Several time-domain and throughput quantities are specified. The micro-calorimeter pulse falling time is 3.5 msec, the anti-co pulse falling time is 0.15 msec, and the anti-co dead time is $6 \times 6$5 msec (Sato et al., 2023). Resolve has the requirement of processing up to 200 s$6 \times 6$6 array$6 \times 6$7, including spurious events, without event losses (Sato et al., 2023). At higher count rates, events can be discarded without processing, losses are recorded as pseudo-events, the fraction of best-grade events is reduced, and pile-up and electrical cross-talk degrade spectral performance (Sato et al., 2023).

The Gaussian-core energy-resolution model is given in the paper as

$6 \times 6$8

as printed in the source (Sato et al., 2023). The line spread function consists of a Gaussian core and an extended line spread function including an exponential tail with e-folding of $6 \times 6$912 eV, electron-loss continuum, and escape peaks (Sato et al., 2023). The Gaussian core is pixel-dependent and environment-dependent, while the extended line spread function is described by common parameters representing absorber-related loss mechanisms (Sato et al., 2023).

The principal quantitative characteristics are succinctly summarized below.

Quantity	Value
Array format	36 pixels, ${\rm \mu m}$0
Pixel pitch	832 ${\rm \mu m}$1
Operating temperature	50 mK
Field of view	${\rm \mu m}$2
Energy range	0.3 -- 12 keV
Required energy resolution	Better than 7 eV

These values define the calorimeter array as a cryogenic, imaging, non-dispersive spectrometer optimized for soft X-ray line diagnostics rather than broadband calorimetry in the high-energy-physics sense (Sato et al., 2023).

6. Resolve in scientific operation

Resolve’s scientific role is to provide high-resolution imaging spectroscopy across the soft X-ray band, and the data supplied include an early observational demonstration in the 2 -- 12 keV spectrum of the Seyfert 1 galaxy Mrk 509 (Dadina et al., 4 May 2026). In that study, the instrument is described as having ${\rm \mu m}$3 eV spectral resolution at 6 keV in the Introduction and ${\rm \mu m}$4 eV at the ${\rm \mu m}$5Fe reference energy in the reduction section (Dadina et al., 4 May 2026).

The observation shows how the array’s resolution can separate narrow and broad Fe-K components that are blended in lower-resolution spectroscopy. The XRISM/Resolve spectrum reveals a narrow FeK${\rm \mu m}$6 core resolved with ${\rm \mu m}$7 eV (${\rm \mu m}$8 km/s) and a broader component with ${\rm \mu m}$9 eV (Dadina et al., 4 May 2026). It also finds tentative evidence (3.6$\Delta T = E / C,$0) for an ionized absorber, with the data suggesting infall at $\Delta T = E / C,$1 km/s and a location within a few thousands gravitational radii (Dadina et al., 4 May 2026).

This use case is instrumentally important because it illustrates the specific advantage of the Resolve calorimeter array: it combines throughput and line-resolving capability in an energy band where narrow emission, broad reflection signatures, and absorption features coexist. The secure instrumental result in the supplied source is that Resolve can resolve a narrow Fe K$\Delta T = E / C,$2 core in a bright Seyfert nucleus and separate it from additional broad emission structure (Dadina et al., 4 May 2026). A plausible implication is that the array’s scientific value lies as much in decomposition of complex line-rich spectra as in the nominal FWHM figure itself.

7. Significance, limitations, and technical interpretation

Resolve is a micro-calorimeter array rather than a dispersive spectrometer or a conventional scintillation calorimeter. Its defining characteristics are the 36-pixel HgTe-plus-ion-implanted-Si focal plane, 50 mK operation, optimal-filter pulse processing, and a calibration framework built around continuous gain tracking and line-shape control (Sato et al., 2023). The design is explicitly intended to deliver spectral resolution that is largely independent of photon energy across the bandpass, unlike dispersive spectrometers (Sato et al., 2023).

Several limitations are also explicit in the supplied sources. The 2023 overview does not provide a single exact Resolve in-flight energy-resolution number analogous to the SXS 5 eV result; it states ground-test compliance and expected equivalence to SXS (Sato et al., 2023). High count rate reduces the fraction of best-grade events, and in ground tests electrical cross-talk degraded resolution by a few eV in high-rate conditions (Sato et al., 2023). The line spread function is not purely Gaussian, and practical response modeling must include the Gaussian core, exponential tail, electron-loss continuum, and escape peaks (Sato et al., 2023). In the Mrk 509 observation, some interpretations, notably the proposed inflowing absorber, remain tentative and model-dependent even though the line-resolving capability itself is directly demonstrated (Dadina et al., 4 May 2026).

Taken together, the supplied material defines Resolve as a high-resolution cryogenic X-ray calorimeter array whose architecture is inherited from Hitomi/SXS and whose operational design couples millikelvin detector physics, optimal digital pulse processing, and multi-channel gain calibration. Its scientific significance follows from that systems integration: a 36-pixel array at 50 mK can provide better than 7 eV spectroscopy over 0.3 -- 12 keV while retaining imaging capability over a $\Delta T = E / C,$3 field of view, and this capability is already sufficient to resolve astrophysical line structures at the $\Delta T = E / C,$4 eV scale in orbit-class observations (Sato et al., 2023, Dadina et al., 4 May 2026).