FIRESS: Far-IR Survey Spectrometer

• FIRESS is a cryogenic far-infrared spectrometer designed for PRIMA that operates from 24 to 235 μm using four grating channels with an insertable Fourier-transform mode.
• It employs kinetic inductance detectors coupled with innovative hexagonally packed lenslet arrays to achieve background-limited sensitivity and high mapping speeds.
• The instrument integrates dual observing modes with resolving powers ranging from ~100 in low-res to over 20,000 in high-res, critical for studies of planetary atmospheres, galaxy evolution, and interstellar dust.

The Far-IR Enhanced Survey Spectrometer (FIRESS) is one of two focal-plane instruments on the PRobe far-Infrared Mission for Astrophysics (PRIMA), and in current PRIMA studies it is defined as a cryogenic far-infrared spectrometer that combines broad simultaneous wavelength coverage, highly multiplexed low-resolution grating spectroscopy, and an insertable high-resolution Fourier-transform capability. In the PRIMA literature, FIRESS is associated with a cold 1.8-m telescope, detector operation near the $100$ mK regime, and a science program centered on the origins of planetary atmospheres, the co-evolution of galaxies and supermassive black holes, and the buildup of dust and heavy elements, while also supporting a large General Observer component (Pontoppidan et al., 1 Sep 2025, M. et al., 2 Sep 2025, Dahal et al., 13 Nov 2025).

1. Mission placement and terminological scope

Within the PRIMA mission architecture, FIRESS is the far-infrared survey spectrometer paired with a cryogenically cooled telescope designed for astrophysical-background-limited operation. The mission is described as using a cold, 1.8-m, actively cooled telescope with 4 K optics, large-format microwave-kinetic-inductance-detector arrays operated at $\simeq 100$ mK, broad simultaneous wavelength coverage, and highly multiplexed spectroscopy (Pontoppidan et al., 1 Sep 2025). A lenslet-coupling study describes PRIMA more broadly as a cryogenically cooled 1.8-m space telescope designed to address galactic ecosystems, planetary atmospheres, and the buildup of dust and metals, with unprecedented sensitivity in the $24$–$261\,\mu\mathrm{m}$ range enabled by background-limited kinetic inductance detectors cooled to $120$ mK (Dahal et al., 13 Nov 2025). Taken together, these descriptions place FIRESS within a PRIMA hardware baseline that is consistently cryogenic, survey-oriented, and optimized for low-background far-infrared spectroscopy.

The acronym has also appeared in other contexts. A 2019 notional description, motivated by wide-field far-infrared spectral-imaging science, used “FIRESS” for a probe-class concept that combined a multi-beam heterodyne module and a grating spectrometer over $50$–$350\,\mu\mathrm{m}$ on a $2.5$ m, $6$ K telescope (Goicoechea et al., 2019). A later extragalactic survey study used the same name for a SPIRIT-based interferometric facility with two cold 1 m-class collectors, an effective aperture $D_{\rm eff}\simeq 80$ m, and $\simeq 100$0–$\simeq 100$1 coverage (Bonato et al., 2024). In current PRIMA publications, however, FIRESS refers specifically to the PRIMA far-infrared spectrometer rather than to those broader concept studies (Pontoppidan et al., 1 Sep 2025).

2. Optical architecture and observing modes

In the PRIMA implementation, FIRESS is a post-dispersed far-infrared spectrometer. Light from the 1.8 m cold telescope is brought through fore-optics to a cold pupil stop and slit, split by dichroics into four wavelength channels, and dispersed by reflective diffraction gratings onto detector arrays (Pontoppidan et al., 1 Sep 2025, M. et al., 2 Sep 2025). The approach paper describes four $\simeq 100$2 slit-fed grating modules, each coupled to a $\simeq 100$3 (spatial) by $\simeq 100$4 (spectral) pixel array, with all four arrays read out simultaneously; a point source can be coupled to two of the four bands at a time because Bands 1 and 3 share one pointing while Bands 2 and 4 share a second (M. et al., 2 Sep 2025). The science-driver paper adds that two grating settings per band yield full-band low-resolution spectra in only two settings (Pontoppidan et al., 1 Sep 2025).

Published band definitions are not fully uniform:

Source	Published band definition	Context
(M. et al., 2 Sep 2025, Yoon et al., 2 Sep 2025)	24–43, 42–76, 74–134, 130–235 $\simeq 100$5	Instrument and PAH studies
(Pontoppidan et al., 1 Sep 2025)	24–52, 52–95, 95–160, 160–235 $\simeq 100$6	Science-driver summary

This discrepancy indicates that the documentation contains both hardware-centric and science-planning parameterizations of the same overall $\simeq 100$7–$\simeq 100$8 spectroscopic domain. The low-resolution mode is consistently described as $\simeq 100$9–$24$0, or approximately $24$1, across the full band (Pontoppidan et al., 1 Sep 2025, M. et al., 2 Sep 2025).

The grating mode follows the standard reflection-grating relation

$24$2

with the resolving power set by slit width and camera optics (M. et al., 2 Sep 2025). FIRESS also includes an insertable Fourier-Transform Module (FTM). In the science-driver description, an insertable Fourier-Transform Module in each dispersed beam produces a high-resolution interferogram on the same grating pixel, with maximum optical path difference $24$3 setting

$24$4

The same paper summarizes the resulting resolution as

$24$5

so that $24$6 at $24$7, $24$8 at $24$9, and $261\,\mu\mathrm{m}$0 at $261\,\mu\mathrm{m}$1 (Pontoppidan et al., 1 Sep 2025). The approach paper presents the high-resolution module as a Martin–Puplett interferometer reinserted into the same grating train and gives $261\,\mu\mathrm{m}$2 (M. et al., 2 Sep 2025). A plausible implication is that different documents are quoting closely related implementations or margins around the same interferometric architecture.

3. Detector system and lenslet-coupled focal planes

The instrument-specific PRIMA papers describe FIRESS as using kinetic-inductance-detector technology. Each grating module focal plane consists of two $261\,\mu\mathrm{m}$3 subarrays, making a total $261\,\mu\mathrm{m}$4 pixels per module, with pixels on a $261\,\mu\mathrm{m}$5 pitch (M. et al., 2 Sep 2025). KID resonators are placed at $261\,\mu\mathrm{m}$6 MHz–$261\,\mu\mathrm{m}$7 GHz, logarithmically spaced so that $261\,\mu\mathrm{m}$8, which is larger than the maximum optical-loading-induced shift $261\,\mu\mathrm{m}$9; eight identical readout circuits generate and digitize all tones simultaneously, and a native frame rate of $120$0 kHz permits cosmic-ray pulse identification before down-sampling to $120$1–$120$2 Hz science rates (M. et al., 2 Sep 2025).

A dedicated hardware development effort has focused on the optical coupling between these detectors and the telescope beam. For FIRESS, monolithic kilopixel silicon lenslet arrays were developed to couple incident radiation from the fore-optics onto the KID absorber elements (Dahal et al., 13 Nov 2025). Each lenslet array is hexagonally close-packed in a $120$3 format, for $120$4 pixels per die, with $120$5 center-to-center pitch. The lens apertures are inscribed in $120$6 hexagons to minimize inactive flat regions, and the lenslets are fabricated as plano-convex aspheric approximations (Dahal et al., 13 Nov 2025).

For a plano-convex lens, the focal length is given by

$120$7

with $120$8 in the FIRESS bands, and the effective $120$9-number is

$50$0

where $50$1 (Dahal et al., 13 Nov 2025). The reported designs yield $50$2 in Band 1 and $50$3 in Band 4, with diffraction-limited spot diameters of $50$4 and $50$5, respectively, both smaller than the KID absorber diameters of $50$6 and $50$7 (Dahal et al., 13 Nov 2025).

Fabrication uses grayscale lithography followed by deep reactive ion etching, with a thick positive resist spin-coated to $50$8, lateral resolution of $50$9, and vertical resolution of $350\,\mu\mathrm{m}$0 in the resist profile (Dahal et al., 13 Nov 2025). The silicon transfer process uses alternating cycles of $350\,\mu\mathrm{m}$1 silicon etch and $350\,\mu\mathrm{m}$2 resist etch, with an empirical etch model containing a $350\,\mu\mathrm{m}$3–$350\,\mu\mathrm{m}$4 isotropic component and a remainder that is highly anisotropic. Sidewall angle control is held within $350\,\mu\mathrm{m}$5 of vertical, and lateral-etch compensation is incorporated into the initial resist design (Dahal et al., 13 Nov 2025). Stylus profilometry across five regions gives RMS error $350\,\mu\mathrm{m}$6 for Band 1 and $350\,\mu\mathrm{m}$7 for Band 4 except at edges, while SEM and 3D scans confirm accurate overlap at the hexagonal corners (Dahal et al., 13 Nov 2025).

Anti-reflection control is based on a quarter-wavelength-thick deposition of Parylene-C. The measured refractive index is $350\,\mu\mathrm{m}$8–$350\,\mu\mathrm{m}$9 from 5 K FTS of a $2.5$0 sample, with nominal thickness

$2.5$1

Representative values are $2.5$2 at the Band 1 center and $2.5$3 at the Band 4 center (Dahal et al., 13 Nov 2025). Because each focal plane spans $2.5$4 in wavelength, a single quarter-wave layer cannot cover the full band with $2.5$5 reflectance; the developed solution uses stepped-thickness coatings produced by $2.5$6 plasma ashing through a shadow mask, creating four discrete thicknesses along the 77 mm array length (Dahal et al., 13 Nov 2025).

Lenslet arrays are aligned and bonded to the KID arrays through a thin epoxy layer using a flip-chip bonder. The Smart Equipment FC300 achieves $2.5$7 post-bond alignment, with $2.5$8 typical pre-bond alignment. The epoxy is Epo-Tek 301, with measured $2.5$9 at 5 K, and the bonding force is controlled so that $6$0 in Band 1 and $6$1–$6$2 in Band 4 (Dahal et al., 13 Nov 2025). Electromagnetic simulation predicts $6$3 coupling efficiency per pixel, while FTS measurements of a bonded prototype show peak coupling of $6$4 in Band 1 and $6$5 in Band 4, within $6$6 of model. End-to-end throughput from telescope to detector exceeds $6$7 in Bands 1 and 4, array-averaged beam uniformity is $6$8 in encircled-energy diameter across $6$9 pixels, and the projected NEP improvement is a $D_{\rm eff}\simeq 80$0–$D_{\rm eff}\simeq 80$1 gain in sensitivity over feedhorn-coupled architectures (Dahal et al., 13 Nov 2025).

4. Sensitivity, calibration, and survey performance

The detector-noise formalism used in FIRESS performance papers follows the standard quadrature combination

$D_{\rm eff}\simeq 80$2

The science-driver paper describes FIRESS as background-limited, with $D_{\rm eff}\simeq 80$3 and a typical $D_{\rm eff}\simeq 80$4 (Pontoppidan et al., 1 Sep 2025). The instrument-approach paper sets a detector target $D_{\rm eff}\simeq 80$5 and quotes a current best-estimate detector NEP of $D_{\rm eff}\simeq 80$6, described as meeting requirements with margin (M. et al., 2 Sep 2025).

Published sensitivities depend on observing mode and the quantity being quoted. The science-driver paper gives a low-resolution point-source line-flux sensitivity of $D_{\rm eff}\simeq 80$7 in one spectral setting, a high-resolution FTM broadband line RMS of $D_{\rm eff}\simeq 80$8 in 1 hr, and a mapping-mode $D_{\rm eff}\simeq 80$9 line sensitivity $\simeq 100$00 over $\simeq 100$01 in $\simeq 100$02 hr (Pontoppidan et al., 1 Sep 2025). The general-observer guidelines in the approach paper instead give

$\simeq 100$03

for low-resolution pointed mode and

$\simeq 100$04

for faint high-resolution pointed mode (M. et al., 2 Sep 2025). This suggests that the numerical values depend on spectral setting, unresolved-line assumptions, and whether full-band interferometric operation is being summarized.

The PAH study provides another performance view tuned to low-resolution spectroscopy. It quotes $\simeq 100$05 per spectral channel, a $\simeq 100$06 unresolved-line sensitivity of $\simeq 100$07 in 1 hr on source, and continuum sensitivity $\simeq 100$08 in 1 hr with $\simeq 100$09 binning ranging from $\simeq 100$10 in Band 1 to $\simeq 100$11 in Bands 2–4 (Yoon et al., 2 Sep 2025). For mapping, it states that reaching a $\simeq 100$12 line sensitivity of $\simeq 100$13 over $\simeq 100$14 takes $\simeq 100$15 hr in Bands 1–2 or $\simeq 100$16 in Bands 3–4 (Yoon et al., 2 Sep 2025). The same paper characterizes FIRESS mapping as $\simeq 100$17 faster than JWST/MIRI or Spitzer/IRS at similar wavelengths (Yoon et al., 2 Sep 2025), while the science-driver paper gives overall survey speeds $\simeq 100$18–$\simeq 100$19 faster than Herschel-PACS and ISO-LWS (Pontoppidan et al., 1 Sep 2025).

Some science cases impose unusually stringent calibration requirements. In the search for far-infrared crystalline-water-ice signatures, the relevant $\simeq 100$20 and $\simeq 100$21 bands are broad enough that $\simeq 100$22 is sufficient, but the expected band-to-continuum contrast is only $\simeq 100$23–$\simeq 100$24; this drives a requirement for relative calibration better than $\simeq 100$25 across the $\simeq 100$26–$\simeq 100$27 region and continuum $\simeq 100$28 per resolution element (Onaka et al., 2 Sep 2025). That application illustrates a broader point: for FIRESS, calibration stability and flat-field control are as important as raw line sensitivity in continuum-dominated far-infrared spectroscopy.

5. Scientific program and observational use cases

The PRIMA science-driver paper identifies three principal science domains. The first is the origins of planetary atmospheres, centered on measurement of the ground-state HD $\simeq 100$29–$\simeq 100$30 line at $\simeq 100$31 to derive total disk gas mass using $\simeq 100$32 and $\simeq 100$33, together with broad-band, high-resolution spectroscopy of $\simeq 100$34–$\simeq 100$35 water lines spanning upper-level energies up to several hundred K (Pontoppidan et al., 1 Sep 2025). The second is the co-evolution of galaxies and supermassive black holes, using rest-frame mid- to far-infrared fine-structure lines such as [Ne II] $\simeq 100$36 as a tracer of star-formation rate, [O IV] $\simeq 100$37 as an AGN accretion tracer, and the OH $\simeq 100$38 doublet for molecular outflows; for typical $\simeq 100$39 ultraluminous galaxies with $\simeq 100$40, the stated line-to-continuum requirement is $\simeq 100$41 at $\simeq 100$42 (Pontoppidan et al., 1 Sep 2025). The third is the buildup of heavy elements and dust, using PAH bands such as the $\simeq 100$43 feature redshifted into the FIRESS bandpass at $\simeq 100$44, together with [O III] $\simeq 100$45 and [N III] $\simeq 100$46 as abundance tracers (Pontoppidan et al., 1 Sep 2025).

A large General Observer program is integral to the instrument’s identity. PRIMA allocates $\simeq 100$47 of its nominal 5-year mission to GO, and of $\simeq 100$48 approved GO cases, $\simeq 100$49 invoke FIRESS modes, spanning cometary D/H measurements, protostellar accretion variability, ice-band surveys, nearby-galaxy line mapping, and high-redshift dust and gas studies (Pontoppidan et al., 1 Sep 2025). The same paper sketches an example PI plan of $\simeq 100$50 hr for origins of planetary atmospheres, $\simeq 100$51 hr for galaxy/SMBH co-evolution, and $\simeq 100$52 hr for heavy-element buildup, with the remaining $\simeq 100$53 hr filled by the GO program (Pontoppidan et al., 1 Sep 2025).

The high-redshift PAH program is one of the most fully developed FIRESS use cases. Simulations suggest that PRIMA observations of PAH emission are $\simeq 100$54 times more efficient at detecting galaxies than VLA observations of CO(1–0) for galaxies with the same infrared luminosity, that PRIMA/FIRESS can detect PAH emission from galaxies with $\simeq 100$55 up to the end of reionization and possibly beyond if $\simeq 100$56, and that FIRESS can function as a PAH mapping instrument for star-formation and redshift measurements in high-redshift protoclusters (Yoon et al., 2 Sep 2025). The same study emphasizes that full spectral fitting and simple flux “clipping” yield different PAH band ratios, with deviations $\simeq 100$57 for strong radiation fields, especially in the $\simeq 100$58 and $\simeq 100$59 bands (Yoon et al., 2 Sep 2025).

FIRESS is also explicitly tied to solid-state and ISM-composition problems. In the “missing interstellar oxygen” case, sensitive observations of crystalline water-ice features at $\simeq 100$60 and $\simeq 100$61 are presented as the decisive method for constraining the amount of crystalline water-ice grains up to $\simeq 100$62 or even larger sizes. If no $\simeq 100$63 band is detected at $\simeq 100$64 contrast and $\simeq 100$65, the inferred limit is $\simeq 100$66–$\simeq 100$67 ppm O/H in large crystalline-ice grains, corresponding to $\simeq 100$68 of the missing $\simeq 100$69 ppm oxygen budget (Onaka et al., 2 Sep 2025).

Galaxy-evolution survey papers extend the program toward blind spectroscopy. A $\simeq 100$70 blind low-resolution survey in $\simeq 100$71 h is projected to detect $\simeq 100$72 galaxies at $\simeq 100$73 via PAH $\simeq 100$74 and/or [O III] $\simeq 100$75, including $\simeq 100$76–$\simeq 100$77 galaxies with $\simeq 100$78 at $\simeq 100$79, with $\simeq 100$80–$\simeq 100$81 of sources hosting AGN detectable via O IV. In high-resolution mode, the same work emphasizes P-Cygni profiles in OH $\simeq 100$82, blueshifted absorption wings with $\simeq 100$83, and extended [O IV] and [Ne V] line wings as feedback diagnostics (Fernández-Ontiveros et al., 8 Sep 2025).

6. Development status, documentation differences, and interpretive cautions

The most advanced hardware-specific FIRESS publication is the lenslet-array study. It reports optimized fabrication, lens design, anti-reflection coating, and bonding processes; brassboard arrays meeting the specifications of the FIRESS low and high spectral bands; stepped-thickness AR coatings for broad wavelength ranges; and spectral transmission measurements of the AR coating and epoxy bonding layers (Dahal et al., 13 Nov 2025). The same work states that vibration and cryogenic environmental testing of bonded arrays is ongoing in order to validate readiness for flight integration on PRIMA (Dahal et al., 13 Nov 2025). In that sense, FIRESS is no longer only a science case or mission-architecture construct; it is accompanied by detector-coupling hardware that has already been fabricated and cryogenically characterized.

At the same time, the FIRESS literature is not perfectly uniform. The dedicated instrument papers consistently describe slit-fed grating modules feeding KID or MKID arrays, together with a Martin–Puplett-style Fourier-transform insertion (M. et al., 2 Sep 2025). By contrast, one science-oriented galaxy-evolution survey paper describes “four independent integral-field spectroscopic channels,” quotes instantaneous fields of view of $\simeq 100$84 per channel with cross-pattern offsets, and refers to transition-edge-sensor bolometer arrays rather than KIDs (Fernández-Ontiveros et al., 8 Sep 2025). This suggests that not every publication is using an identical hardware baseline; some papers appear to emphasize science-performance parameterizations rather than the same implementation details.

A similar caution applies to the wider use of the FIRESS name. The notional dual-module probe-class concept of the 2019 white-paper derivative and the SPIRIT-based interferometric survey instrument of the 2024 black-hole/galaxy co-evolution study are scientifically adjacent but technically distinct from the current PRIMA FIRESS baseline (Goicoechea et al., 2019, Bonato et al., 2024). For encyclopedia purposes, the most defensible definition is therefore the PRIMA instrument: a cryogenic far-infrared survey spectrometer with four low-resolution grating channels, an insertable high-resolution Fourier-transform capability, and a focal-plane implementation centered on lenslet-coupled superconducting detectors. The broader literature shows how that definition emerged within a larger ecosystem of far-infrared survey-spectrometer concepts, but the PRIMA version is the one supported by the most detailed architectural and component-level documentation (Pontoppidan et al., 1 Sep 2025, M. et al., 2 Sep 2025, Dahal et al., 13 Nov 2025).