PRIMA/FIRESS: Far-IR Observatory Concept

• PRIMA/FIRESS is a far-infrared observatory concept built around a 1.8 m telescope, enabling imaging, spectroscopy, and polarimetry in the 24–261 µm range.
• It integrates advanced instruments—PRIMAger for broad imaging and FIRESS for versatile spectroscopy—to study galaxy evolution, star formation, and cosmic dust.
• The mission employs innovative KID-based detectors with precise cryogenic performance, delivering high sensitivity and resolution for detailed astrophysical diagnostics.

PRIMA/FIRESS most commonly denotes the PRobe far-Infrared Mission for Astrophysics and its Far-Infrared Enhanced Survey Spectrometer, a Probe-class far-infrared observatory concept built around a 1.8 m cryogenic telescope. In current mission studies, PRIMA is described as covering 24–261/264 µm overall through the paired instruments PRIMAger and FIRESS, with FIRESS providing spectroscopy over 24–235 µm from low-resolution grating surveys to higher-resolution Fourier-transform operation (Dahal et al., 13 Nov 2025, Ciesla et al., 1 Sep 2025, Pontoppidan et al., 1 Sep 2025). The mission is framed as a background-limited platform for far-IR imaging, polarimetry, and spectroscopy, and about 75% of observing time is intended for General Observer programs (Dahal et al., 13 Nov 2025).

1. Mission definition and observatory context

PRIMA is presented as a far-infrared mission concept aimed at reopening the observational domain between the long-wavelength cutoff of JWST and the submillimeter regime of ALMA. The observatory is described as a 1.8 m telescope actively cooled to 4.5 K, with instrument stages at 1 K and detector stages near 120–125 mK, and its scientific scope is organized around the evolution of galactic ecosystems, the origins of planetary atmospheres, and the buildup of dust and metals over cosmic time (Ciesla et al., 1 Sep 2025, Dahal et al., 13 Nov 2025).

The payload consists of two main instruments. PRIMAger provides hyperspectral imaging from 24–84 µm and polarimetric imaging in four bands from 80 to 264 µm. FIRESS is the wide-band spectrometer, optimized for sensitive, multiplexed spectroscopy across 24–235 µm (Ciesla et al., 1 Sep 2025, Pontoppidan et al., 1 Sep 2025). The two instruments share the cryogenic chain and KID readout infrastructure, but they do not operate simultaneously; both remain cold while mission planning interleaves their observing programs (Ciesla et al., 1 Sep 2025).

Mission studies consistently position FIRESS as the spectroscopic counterpart to PRIMAger’s survey and imaging capability. PRIMAger supplies broad SED sampling, PAH-sensitive low-resolution imaging, and polarimetry, whereas FIRESS is designed for atomic, molecular, mineralogical, and kinematic diagnostics that require spectroscopic resolving power beyond the imaging modes (Ciesla et al., 1 Sep 2025, Moullet et al., 14 Nov 2025).

2. FIRESS architecture and operating modes

FIRESS is described as a long-slit grating spectrometer composed of four modules spanning the full 24–235 µm band, with a low-resolution dispersive mode and a higher-resolution mode implemented through a Fourier Transform Module inserted into the dispersed beam (Jiménez-Serra et al., 2 Sep 2025, Pontoppidan et al., 1 Sep 2025). In the low-resolution configuration, FIRESS delivers resolving power $R \sim 85{-}150$, often summarized as $R \sim 100$. In the high-resolution configuration, the resolving power scales approximately as

$$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$

giving $R \approx 20{,}500$ at 24 µm, $R \approx 4400$ at 112 µm, and $R \approx 2000$ at 235 µm (Pontoppidan et al., 1 Sep 2025).

FIRESS contains 672 spectral channels in total, with 336 channels accessible for a point source in a single exposure, so the full low-resolution band is covered in two spectral settings (Pontoppidan et al., 1 Sep 2025). Mapping is performed with a Beam Steering Mirror, which scans the slits across the sky while all four arrays are read continuously; low-resolution point-source spectroscopy uses beam steering for chopping, whereas high-resolution observations use the Fourier-transform configuration with nodding within the slit (Pontoppidan et al., 1 Sep 2025).

Mode	Spectral characteristics	Typical use
Low-resolution grating mode	24–235 µm, $R \sim 85{-}150$	Spectral mapping, surveys, fast point-source spectra
High-resolution FTM mode	24–235 µm, $R \approx 2000{-}20{,}500$	Disk gas lines, wind profiles, kinematics
Two-setting full-band survey	336 channels per setting	Full low-resolution spectral coverage

At wavelengths central to solid-state spectroscopy, the low-resolution mode is explicitly shown to be adequate for broad and narrow mineral bands. For example, smoothing sulfide opacities to $R>85$ retains the narrow FeS bands between 30–50 µm and fully captures the broad MgS band near 30 µm, while avoiding the prohibitive integration times that would accompany use of the high-resolution mode for such broad features (Jiménez-Serra et al., 2 Sep 2025).

3. Detectors, optics, and readout chain

PRIMA is designed to be background-limited over 24–261 µm, using roughly 11,000 kinetic inductance detectors (KIDs) operating at 120 mK (Dahal et al., 13 Nov 2025). Within FIRESS, the focal plane is implemented through KID arrays coupled by monolithic silicon lenslet arrays rather than bare pixels. Each FIRESS lenslet die contains 12 spatial positions × 84 spectral positions, for 1008 lenslets per die, on a 900 µm hexagonal pitch matched to the KID layout (Dahal et al., 13 Nov 2025).

The lenslet program reported for FIRESS addresses optical efficiency, spillover, and fabrication tolerances. The arrays are fabricated by grayscale lithography followed by DRIE, AR-coated with Parylene-C, and bonded to the KID arrays using Epo-Tek 301 epoxy and a flip-chip bonder (Dahal et al., 13 Nov 2025). The reported alignment performance is ≈3 µm pre-bond and ±0.5 µm post-bond, comfortably inside the ±10 µm FIRESS requirement; bond thicknesses of <1 µm for Band 1 and 1–4 µm for Band 4 meet the requirement that epoxy losses remain <5% (Dahal et al., 13 Nov 2025).

Warm readout electronics are shared between FIRESS and PRIMAger. The system is designed to multiplex more than 1000 detectors per chain over 2.5 GHz of instantaneous bandwidth, using 8 readout chains built around SpaceCube Mini v3.0 digital electronics, a radiation-tolerant Kintex KU060 FPGA, and custom 5 Gsps ADC/DAC hardware (Essinger-Hileman et al., 4 Dec 2025). The FIRESS resonators occupy the 0.4–2.4 GHz readout band, while PRIMAger uses 2.6–4.9 GHz; an RF board switches between the two instruments and provides filtering and gain conditioning (Essinger-Hileman et al., 4 Dec 2025).

Radiation hardness of FIRESS-style KIDs has also been studied directly. For a 5.3-year L2 mission, the total displacement damage dose in the aluminum inductors is estimated as

$$\mathrm{DDD}_{\mathrm{mission}} \approx 2.9\times10^7\ \mathrm{MeV\,g^{-1}},$$

and cryogenic alpha irradiation to 62% of that level produced no significant degradation in quasiparticle lifetime, resonant frequency, or internal quality factor (Kane et al., 30 Apr 2026). This suggests that cumulative displacement damage at L2 is not expected to be a limiting factor for FIRESS detector performance over mission lifetime (Kane et al., 30 Apr 2026).

4. Core science drivers and diagnostic content

The FIRESS science case is organized around three principal domains: origins of planetary atmospheres, co-evolution of galaxies and supermassive black holes, and buildup of heavy elements and dust (Pontoppidan et al., 1 Sep 2025). For protoplanetary disks, the spectrometer is explicitly optimized for the HD(1–0) line at 112 µm, dense H$R \sim 100$0O line forests, HDO at 234.8 µm, the CO rotational ladder, [O I] at 63 and 145 µm, [C II] at 158 µm, the 43 µm water-ice band, and the 69 µm forsterite feature. The requirement that HD at 112 µm retain a line-to-continuum ratio of at least 2.5% for disks above $R \sim 100$1 is one of the design drivers for the high-resolution mode near that wavelength (Pontoppidan et al., 1 Sep 2025).

For galaxy evolution, FIRESS is designed to recover the obscured side of star formation and black-hole accretion through mid- and far-IR line diagnostics. The low-resolution mode is used for PAH spectroscopy and lines such as [Ne II] 12.8 µm, [O IV] 25.9 µm, [Si II] 34.5 µm, [O I] 63 µm, [O III] 52/88 µm, [N III] 57 µm, [N II] 122/205 µm, and OH far-IR doublets, while the high-resolution mode targets P-Cygni profiles, blueshifted OH absorption, and high-ionization emission-line wings as feedback diagnostics (Fernández-Ontiveros et al., 8 Sep 2025, Pontoppidan et al., 1 Sep 2025). A simulated 200 arcmin$R \sim 100$2 blind FIRESS survey at cosmic noon is forecast to measure star formation and black-hole accretion rates for hundreds of galaxies out to $R \sim 100$3 (Fernández-Ontiveros et al., 8 Sep 2025).

The same spectral coverage underpins studies of heavily obscured nuclei and “HST-dark” galaxies. PRIMAger is used to identify such systems photometrically through the redshifted 9.8 µm silicate absorption band, while FIRESS low-resolution spectra at $R \sim 100$4 are expected to detect PAHs, ices, ionized and molecular gas in obscured nuclei out to $R \sim 100$5 (Donnan et al., 14 Mar 2025). For optically and near-IR dark galaxies at $R \sim 100$6, PRIMAger fills the 25–265 µm SED gap, while FIRESS is identified as the key instrument for diagnosing whether the power source is star formation or an AGN and for measuring the physical conditions of the ISM (Gruppioni et al., 2 Sep 2025).

Beyond the top-level drivers, FIRESS is repeatedly used as a general-purpose far-IR spectrometer in concrete case studies. One example is the detection of MgS and FeS solid-state absorption bands in dense-cloud sight lines, motivated by the sulfur depletion problem and by gas-phase NaS and MgS detections in the Galactic Center cloud G+0.693–0.027 (Jiménez-Serra et al., 2 Sep 2025). Another is the observation of stripped dust in cluster galaxies, where FIRESS maps [C II] 158 µm, [O I] 63 µm, and [N II] 122/205 µm in ram-pressure tails to constrain cooling, density, metallicity, and the relation between dust and multiphase gas (Boselli et al., 2 Sep 2025).

5. Observing strategies and representative performance

FIRESS observing is divided among low-resolution mapping, low-resolution pointed spectroscopy, and high-resolution targeted work. In low-resolution mapping mode, FIRESS is specified to reach 5σ line flux limits $R \sim 100$7 over 1 deg$R \sim 100$8 in 100 h, at $R \sim 100$9 (Pontoppidan et al., 1 Sep 2025). For full-band high-resolution spectroscopy over 51–210 µm, the projected line RMS in 1 h is $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$0, representing a survey-speed improvement of 3–4 orders of magnitude relative to Herschel/PACS full scans (Pontoppidan et al., 1 Sep 2025).

For high-redshift dusty galaxies, FIRESS low-resolution performance is often summarized using a 5σ, 1 h line sensitivity near $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$1 (Gruppioni et al., 2 Sep 2025). Under that assumption, a representative $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$2 HST-dark galaxy with $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$3 can yield detections of [Ne II] 12.8 µm, [Si II] 34.5 µm, [O I] 63 µm, [O III] 88 µm, and [O IV] 25.9 µm in about 2 h per source in low-resolution mode, while [Ne V] 14.3/24.3 µm requires a stronger AGN contribution (Gruppioni et al., 2 Sep 2025).

For blind spectroscopic surveys, a modeled 200 arcmin$$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$4 FIRESS program allocates 750 h total, including 640 h on source, and adopts a 5σ point-source sensitivity of $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$5 at 24 µm (Fernández-Ontiveros et al., 8 Sep 2025). The predicted yield is $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$6 galaxies detected in PAH 11.3 µm and/or [O III] 52 µm at $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$7, with roughly $$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$8 galaxies expected to have at least one robust SFR tracer (Fernández-Ontiveros et al., 8 Sep 2025).

A representative low-resolution mineralogy case is the FIRESS study of metal sulfides. The low-resolution 5σ point-source sensitivity requirement is

$$R(\lambda) \simeq 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$9

per resolution element, corresponding at 30 µm and $R \approx 20{,}500$0 to a continuum sensitivity of about 0.16 mJy in 1 h (Jiménez-Serra et al., 2 Sep 2025). Using Class 0/I protostars as background continua, the study finds that the broad MgS band at $R \approx 20{,}500$1 can be detected in 1 h for sources with $R \approx 20{,}500$2, while the narrower FeS bands in 20–50 µm require $R \approx 20{,}500$3 for $R \approx 20{,}500$4 in the same integration (Jiménez-Serra et al., 2 Sep 2025).

6. Confusion, calibration, and current limitations

A central systems-level issue for extragalactic PRIMA programs is confusion in PRIMAger imaging rather than FIRESS spectroscopy itself. The deblending study XID+stepwise shows that hyperspectral imaging can be pushed well below nominal confusion limits by propagating constraints from shorter to longer wavelengths. With Euclid-like positional priors, the method recovers fluxes to within 20% down to 0.2–0.7 mJy across 45–84 µm, corresponding to 1.3–3.4 times fainter than the confusion limit, and to 0.9, 2.5, 7.6, and 14.8 mJy at 92, 126, 183, and 235 µm, respectively, which are 3–5 times better than the confusion limit (Donnellan et al., 15 Dec 2025). A deeper Euclid-based prior catalogue plus weak ancillary 25 µm flux priors improves this further, reaching up to $R \approx 20{,}500$5 times fainter than the confusion limit at 96 µm (Donnellan et al., 15 Dec 2025). This strongly affects FIRESS target selection because dense, multi-channel PRIMAger SEDs remain available for IR-luminous galaxies at $R \approx 20{,}500$6 even several factors below classical confusion thresholds (Donnellan et al., 15 Dec 2025).

At the component level, FIRESS still has explicit development steps outstanding. The optimized Band 1 and Band 4 lenslet arrays have been fabricated and characterized, AR-coating and bonding recipes have been validated, and bonded lenslet–KID arrays are being prepared for cryogenic blackbody testing, but full cryogenic measurement of beam patterns, spectral response, cross-talk, and overall optical efficiency with operating KIDs remains future work, as do vibration tests in flight-like packaging and extension of the hex-corner design and multi-step AR coatings to the intermediate bands (Dahal et al., 13 Nov 2025). The mission itself is described as being in Phase A (Ciesla et al., 1 Sep 2025).

Science-specific caveats also remain. In the metal-sulfide case, modeled detectability depends on assumptions about sulfide abundance fractions and on laboratory optical constants for Mg$R \approx 20{,}500$7Fe$R \approx 20{,}500$8S and FeS; real mixtures may be more complex, and band confusion with other minerals or silicates requires detailed modeling (Jiménez-Serra et al., 2 Sep 2025). For cluster stripping studies, [C II] 158 µm is realistic in diffuse tails, whereas [O I] 63 µm and [N II] 122/205 µm are practical mainly in brighter or denser clumps (Boselli et al., 2 Sep 2025).

7. Historical nomenclature and acronym ambiguity

The acronym PRIMA has an older, unrelated usage in interferometry: “Phase-Referenced Imaging and Micro-arcsecond Astrometry” at the VLTI. In that context, PRIMA was a dual-feed upgrade whose central fringe sensor was the Fringe Sensor Unit (FSU), a K-band beam combiner using spatial phase modulation, a low-resolution $R \approx 20{,}500$9 spectrometer across K band, and phase/group-delay estimates sampled at rates up to 2 kHz (0909.1470). During commissioning, that FSU tracked fringes of stars as faint as $R \approx 4400$0 with the ATs and improved VLTI K-band sensitivity by more than one magnitude (0909.1470).

Later commissioning analyses emphasized that the interferometric PRIMA FSU suffered from non-linearities because it lacked real-time photometric correction and because its fringe encoding depended on polarization, so additional calibration and characterization were required before astrometric science operation (Sahlmann et al., 2010). This historical VLTI usage is distinct from the current far-infrared PRIMA mission concept and its FIRESS spectrometer (Pontoppidan et al., 1 Sep 2025).