FIRESS: Far-Infrared Enhanced Survey Spectrometer

• FIRESS is a far-infrared spectrometer covering 24–235 µm, designed for survey-scale mapping and high-resolution follow-up on the PRIMA observatory.
• It utilizes four slit-fed grating modules and a Fourier Transform Module, coupled with advanced kinetic inductance detectors, to achieve broad spectral coverage and exceptional sensitivity.
• The instrument supports diverse science drivers—including planetary atmospheres, galaxy evolution, and interstellar medium studies—by delivering orders-of-magnitude improvements in sensitivity and mapping speed.

The Far-Infrared Enhanced Survey Spectrometer (FIRESS) is the primary spectroscopic workhorse on PRIMA, a cryogenic 1.8-m space telescope, and is designed to provide low-resolution grating spectroscopy over 24–235 µm in both point-source and mapping modes together with a high-resolution mode implemented through an internal Fourier Transform Module. Its formal science drivers are the origins of planetary atmospheres, the co-evolution of galaxies and supermassive black holes, and the buildup of heavy elements in the Universe, but the instrument is also conceived as a general-purpose far-infrared spectrometer for a dominant General Observer program (Pontoppidan et al., 1 Sep 2025).

1. Mission setting and instrument identity

FIRESS is one of the two primary instruments on PRIMA, alongside PRIMAger. Within that payload, FIRESS is the wide-band survey spectrometer: it disperses the telescope beam onto large kinetic-inductance-detector arrays, provides broad instantaneous spectral coverage, and supplies the mapping capability needed for large-area surveys of far-infrared lines and continuum (Dahal et al., 13 Nov 2025). PRIMA itself is described as a cryogenic 1.8-m observatory designed around astrophysical-background-limited observations, with mission-wide coverage extending across the far-infrared; FIRESS occupies the 24–235 µm spectroscopic portion of that parameter space (M. et al., 2 Sep 2025).

The instrument’s stated role is broader than a single flagship program. In the current PRIMA General Observer science book, two-thirds of the 76 science cases request FIRESS, either alone or together with PRIMAger, and 75 % of the mission lifetime is reserved for General Observer programs (Pontoppidan et al., 1 Sep 2025). This combination of dedicated key-science drivers and broad community use explains the architectural emphasis on simultaneous wavelength coverage, multiple observing modes, and high multiplexing.

The mission papers present FIRESS as a direct response to long-standing limitations of earlier far-infrared spectrometers. Relative to ISO-LWS, Herschel-PACS, and SOFIA instruments, the combination of cold optics, large spectral multiplex, and background-limited detector arrays is intended to shift FIR spectroscopy from narrow, source-by-source programs to survey-scale observing (Pontoppidan et al., 1 Sep 2025).

2. Optical architecture and observing modes

FIRESS spans 24–235 µm with four slit-fed grating modules. Each module couples to a 24 (spatial) by 84 (spectral) pixel detector array, and all four arrays are read out simultaneously (M. et al., 2 Sep 2025).

Band	Spectral range (µm)	Resolving power $R$
1	24–43	90–150
2	42–76	85–120
3	74–134	90–125
4	130–235	95–130

In the low-resolution mode, the four bands are logarithmically spaced and slightly overlapping, with pixel sizes of 7.6″, 7.6″, 12.7″, and 22.9″ in Bands 1–4, respectively (Yoon et al., 2 Sep 2025). The telescope beam is divided by dichroics into four spectral channels, each feeding its own long slit and grating. Bands 1 and 3 are coaligned on the sky, as are Bands 2 and 4, with the two slit pairs separated by about 3.3 arcmin; accordingly, a point source can be placed on Bands 1+3 or 2+4 simultaneously, but not on all four bands at once (M. et al., 2 Sep 2025).

The focal-plane geometry is hexagonal rather than rectangular. Because adjacent spatial rows are offset by half a pixel in the dispersive direction, chopping a point source between complementary slit rows yields half-pixel spectral sampling. The beam-steering mirror is therefore integral not only to background subtraction but also to recovery of the intrinsic grating resolution (M. et al., 2 Sep 2025).

High-resolution spectroscopy is provided by a post-dispersed Fourier Transform Module inserted ahead of the grating optics. The module is a polarizing Martin–Puplett interferometer with an optical path difference range from $-20$ mm to $+260$ mm, a carriage travel of 35 cm, and a field of view of 4.5′. Its resolving power scales approximately as

$$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$

giving $R\approx 4400$ at 112 µm, $R\approx 20{,}500$ at 24 µm, and $R\gtrsim 2000$ at 235 µm; the maximum optical path difference can also be reduced to trade resolution for sensitivity, as in the $R\sim 900$ configuration discussed for OH outflow work (Pontoppidan et al., 1 Sep 2025).

3. Detector system and focal-plane implementation

FIRESS uses aluminum kinetic inductance detectors operated at 120 mK. The current performance model adopts an intrinsic detector noise-equivalent power of approximately $0.9\times10^{-19}\,\mathrm{W\,Hz^{-1/2}}$, with thermal quasiparticles negligible at that temperature and detector time constants of order 1 ms (M. et al., 2 Sep 2025). In the radiation-hardness study, the FIRESS arrays are described more specifically as lumped-element KIDs with a niobium interdigitated capacitor and a thin-film aluminum inductor/absorber of volume $V\sim15~\mu\mathrm{m}^3$, for which the background-limited scaling is stated as $-20$0 (Kane et al., 30 Apr 2026).

Each FIRESS focal plane consists of two 1008-pixel subarrays, yielding 2016 pixels per spectrometer module. A representative array has roughly 1008 resonators distributed over a 500–2500 MHz readout band, with resonances approximately equally spaced in logarithmic frequency and typical loaded quality factor $-20$1 (Kane et al., 30 Apr 2026). The instrument uses eight readout circuits in total, one per subarray, and achieves frame rates near 9.5 kHz at the focal plane before downsampling to science sample rates set by detector bandwidth and the mission’s data-rate constraint (M. et al., 2 Sep 2025).

Optical coupling from the fore-optics to the absorbers is provided by monolithic silicon lenslet arrays. Each lenslet array has 1008 pixels arranged as 12 spatial by 84 spectral, is hexagonally packed with 900 µm pitch, and is fabricated as a monolithic silicon die. For Bands 1 and 4, the diffraction-limited spot diameters at the detector plane are approximately 26 µm and 61 µm, respectively, smaller than the corresponding absorber diameters of 70 and 115 µm. The lenslets are anti-reflection coated with Parylene-C, with $-20$2, and are bonded to the KID arrays with Epo-Tek 301 epoxy; measured bond thicknesses are $-20$3 µm for Band 1 and 1–4 µm for Band 4, consistent with the requirement that epoxy-layer loss remain $-20$4 (Dahal et al., 13 Nov 2025).

Radiation robustness at the Sun–Earth L2 point has been assessed directly for FIRESS KIDs. The radiation study calculates a 5.3-year mission displacement-damage dose of $-20$5, dominated by solar protons, and irradiates an array to a median $-20$6, or about 62 % of that mission dose, under fully cryogenic conditions. Within measurement uncertainty, no systematic degradation is found in quasiparticle lifetime, resonant frequency, or internal quality factor, and the inferred impact on the FIRESS NEP requirement of $-20$7 is negligible (Kane et al., 30 Apr 2026).

4. Science drivers

The first formal science driver is the origins of planetary atmospheres. In this program, HD $-20$8 at 112 µm functions as a direct tracer of disk gas mass, while multiple water lines spanning a wide range of excitation trace the radial distribution of volatile oxygen. The instrument must cover HD 112 µm, the ortho-H$-20$9O ground-state $+260$0 line at 179.53 µm, HDO $+260$1 at 234.8 µm, low-$+260$2 CO such as $+260$3 at 217 µm, and the H$+260$4 S(0) line at 28.2 µm. Disk modeling ties the HD requirement directly to spectral resolution: to maintain a line-to-continuum contrast $+260$5 for disks more massive than $+260$6, the instrument needs $+260$7, which is one of the principal reasons for the high-resolution mode (Pontoppidan et al., 1 Sep 2025).

The second major driver is the co-evolution of galaxies and supermassive black holes. In low-resolution blind surveys, FIRESS is designed to measure star-formation and black-hole accretion rates in the same systems through rest-frame mid- and far-infrared diagnostics such as PAH 11.3 µm, [O III] 52 µm, and [O IV] 25.9 µm. A simulated 200 arcmin$+260$8 blind spectroscopic survey with 640 h on source, or about 750 h including mapping geometry and overheads, yields approximately 1000 galaxies detected at $+260$9 in at least one of PAH 11.3 µm and/or [O III] 52 µm, with AGN identified through [O IV] 25.9 µm. In the same framework, 20–50 % of the galaxies are expected to host active nuclei detectable in [O IV], and pointed follow-up of 75 dusty star-forming galaxies in three redshift bins is proposed to measure N/O and O/H using [N III] 57 µm, [O III] 52 µm, and other mid-infrared lines (Fernández-Ontiveros et al., 8 Sep 2025).

A third major extragalactic program is high-redshift PAH spectroscopy. Simulations of the FIRESS low-resolution mode indicate that PAH emission from galaxies with $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$0 is detectable to $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$1 with integration times of order 10 h, while galaxies with $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$2 can be observed at the end of reionization, $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$3, and possibly beyond. The same study concludes that PRIMA observations of PAH emission are $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$4 times more efficient at detecting galaxies than VLA observations of CO(1–0) for galaxies with the same infrared luminosity, and that FIRESS can serve as a PAH mapping instrument for protoclusters at $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$5 (Yoon et al., 2 Sep 2025).

FIRESS is also explicitly motivated by targeted ISM problems outside the core flagship programs. In the “missing interstellar oxygen” science case, the key observables are the broad crystalline water-ice bands at 44 and 62 µm, which can constrain crystalline ice grains up to $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$6m or even larger sizes. Because the features are broad, low spectral resolution of $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$7 is sufficient, but relative calibration better than 1 % is required, and the 40–70 µm region must include the [O I] 63 µm cooling line as well as the ice bands. In that context, a sensitive FIR spectrograph represented by PRIMA/FIRESS is described as indispensable (Onaka et al., 2 Sep 2025).

5. Sensitivity, calibration, and operational performance

The current FIRESS performance model provides both conservative General Observer guidelines and wavelength-dependent current best estimates. For unresolved lines in low-resolution point-source mode, the recommended planning value is a $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$8 minimum detectable line flux of $$R_{\rm FTM}(\lambda)\approx 4400 \times \frac{112\,\mu\mathrm{m}}{\lambda},$$9 in 1 h, with the usual background-limited $R\approx 4400$0 integration-time scaling (M. et al., 2 Sep 2025). For continuum work in the low-resolution PAH simulations, the adopted $R\approx 4400$1, 1 h, point-source continuum sensitivities after $R\approx 4400$2 binning are 64 µJy in Band 1, 125 µJy in Band 2, 196 µJy in Band 3, and 193 µJy in Band 4 (Yoon et al., 2 Sep 2025).

The high-resolution mode has a different noise model because the Fourier-transform architecture integrates photon background over the width of each grating channel. In the GO guidelines, the $R\approx 4400$3 minimum detectable line flux tends to $R\approx 4400$4 in 1 h for very faint sources, with an additional term that scales with source continuum flux density $R\approx 4400$5 and as $R\approx 4400$6 for brighter continua (M. et al., 2 Sep 2025). A representative disk example gives an HD 112 µm line flux of $R\approx 4400$7 detected at S/N = 10 in 10 h per setting at $R\approx 4400$8 (Pontoppidan et al., 1 Sep 2025).

Survey performance is a central design metric. The science-driver paper states that FIRESS low-resolution mapping can scan 1 square degree to $R\approx 4400$9 line sensitivity better than $R\approx 20{,}500$0 in 100 h at $R\approx 20{,}500$1 (Pontoppidan et al., 1 Sep 2025). The GO mapping guideline is more granular: to reach $R\approx 20{,}500$2 over 100 sq. arcmin requires 800 h in the 24–75 µm region and $R\approx 20{,}500$3 between 75 and 235 µm, with the time scaling linearly with area and as depth$R\approx 20{,}500$4 (M. et al., 2 Sep 2025).

Operationally, low-resolution point-source observations use beam-steering-mirror chopping along the slit, with the current concept adopting approximately 5 Hz chopping. High-resolution FTM spectroscopy fixes the BSM and scans the interferometer so that the optical modulation from each grating channel lands in an audio-frequency band where KID $R\approx 20{,}500$5 performance is favorable; for an OPD scan speed of $R\approx 20{,}500$6, a 112 µm channel modulates near 9 Hz (M. et al., 2 Sep 2025). Some science cases impose unusually stringent calibration demands. The crystalline-ice program, for example, requires relative calibration better than 1 % because the expected 44 µm excess can be only 0.4–2 % above the continuum (Onaka et al., 2 Sep 2025).

6. Broader significance, heritage, and design constraints

FIRESS is situated within a lineage of far-infrared spectrometers, but its intended performance is substantially beyond earlier survey instruments. Compared with Herschel-PACS full spectral scans over 51–210 µm, which reached line RMS values of $R\approx 20{,}500$7 in 1 h, FIRESS is projected to cover the same range to RMS $R\approx 20{,}500$8 in 1 h, corresponding to a sensitivity improvement of roughly 30–200 and a survey-speed gain of approximately $R\approx 20{,}500$9 (Pontoppidan et al., 1 Sep 2025). In the cosmic-noon galaxy-evolution study, the instrument is likewise characterized as offering about 1.5 dex improvement in sensitivity and factors of $R\gtrsim 2000$0 in mapping speed over Herschel and SOFIA (Fernández-Ontiveros et al., 8 Sep 2025).

Its spectral position is also strategically intermediate between existing flagship facilities. The 24 µm short-wavelength edge provides overlap with JWST-MIRI out to about 28 µm, while the long-wavelength end covers the classical FIR fine-structure and solid-state diagnostics inaccessible to JWST and complementary to ALMA and the VLA. In extragalactic use, this means that FIRESS can carry out blind spectroscopic surveys in a regime where JWST lacks wavelength reach and radio facilities are less efficient for the same obscured systems (Pontoppidan et al., 1 Sep 2025).

The design, however, includes explicit trade-offs. Because of the slit pairing, only two of the four bands observe a point source in a single pointing, so full 24–235 µm point-source coverage requires two pointings. The maximum high-resolution resolving power is set by the available optical path difference, and the present slit widths are intentionally conservative, which makes current sensitivity estimates pessimistic rather than aggressive. The development papers also identify full-scale array integration, cosmic-ray response, and final stray-light and slit-width optimization as active engineering topics, although current detector yield, cryogenic performance, lenslet coupling, and radiation testing all support the baseline design (M. et al., 2 Sep 2025).

Taken together, these characteristics define FIRESS as a broad-band, highly multiplexed, background-limited far-infrared spectrometer that couples survey-scale mapping to targeted high-resolution follow-up. Its scientific identity rests on that combination: the same instrument architecture is used to measure HD and water in protoplanetary disks, PAHs and metal lines in galaxies at cosmic noon and beyond, OH and high-ionization feedback signatures in AGN hosts, and even weak solid-state features such as the 44 and 62 µm water-ice bands in the interstellar medium (Pontoppidan et al., 1 Sep 2025).