PRIMA: Probe-Class Far-Infrared Observatory

• PRIMA is a far-infrared observatory concept featuring a 1.8 m telescope cooled to 4.5 K that bridges the 30–300 μm gap between JWST/MIRI and ALMA.
• The mission employs dual complementary instruments—PRIMAger for imaging and polarimetry, and FIRESS for low- and high-resolution spectroscopy—covering 24–264 μm.
• It is optimized for rapid, wide-field surveys and community science, enabling detailed studies of galaxy evolution, star formation, and planetary system formation.

PRIMA (PRobe far-Infrared Mission for Astrophysics) is a Probe-class far-infrared observatory concept built around a 1.8 m telescope actively cooled to 4.5 K and a two-instrument payload, PRIMAger and FIRESS, spanning the mid- to far-infrared window from 24 to 264 μm. In the mission and instrument papers, PRIMA is framed as a next-decade facility for reopening the 30–300 μm regime largely left without a space observatory since Herschel and SOFIA, bridging JWST/MIRI and ALMA, and functioning as a community observatory for problems ranging from planet-forming disks to obscured galaxy growth, magnetic fields, and the buildup of dust and metals (Ciesla et al., 1 Sep 2025, Moullet et al., 2023).

1. Mission concept and observatory architecture

PRIMA is described as a cryogenically cooled far-infrared observatory concept selected into NASA’s APEX/Probe line following Astro2020 and currently in Phase A (Ciesla et al., 1 Sep 2025). Its central architectural feature is a modest-aperture but very cold telescope: a 1.8 m primary actively cooled to 4.5 K, with passive V-groove shielding and a JWST/MIRI-like J-T cooler bringing intermediate stages to about 18 K, so that observations are set primarily by astrophysical backgrounds rather than by warm optics (Ciesla et al., 1 Sep 2025).

The mission is consistently positioned as a far-infrared complement to existing facilities. JWST/MIRI reaches to about 28 μm, while ALMA operates from about 300 μm upward; PRIMA occupies the intervening band where dust-reprocessed radiation, PAH emission, many fine-structure lines, water lines, and FIR polarization diagnostics are concentrated (Ciesla et al., 1 Sep 2025, Pontoppidan et al., 1 Sep 2025). Several PRIMA papers emphasize that this spectral domain contains roughly half of the cosmic luminous energy emitted in the far-infrared, and that the observatory is intended to provide orders-of-magnitude improvement in mapping speed over previous far-infrared missions (Ciesla et al., 1 Sep 2025, Kane et al., 30 Apr 2026).

The mission is also explicitly organized as a General Observer facility. The PRIMA General Observer Science Book states that 75% of observing time over a nominal 5-year mission is intended for GO programs, corresponding to more than 26,000 hr of community science time, and documents 76 contributed science cases spanning galaxy evolution, star and planet formation, Solar System science, transients, and ISM magnetism (Moullet et al., 2023). Later mission papers retain this emphasis, and the FIRESS science-driver paper notes that about two-thirds of current science cases use FIRESS, while the PRIMAger instrument paper states that more than a third of the proposed GO cases require PRIMAger alone and another third require both instruments (Pontoppidan et al., 1 Sep 2025, Ciesla et al., 1 Sep 2025).

2. Payload and instrument suite

PRIMA carries two instruments with deliberately complementary roles. PRIMAger is the imaging camera, optimized for dust continuum, PAHs/macromolecules, and polarized dust emission; FIRESS is the spectrometer, optimized for gas diagnostics, line surveys, and high-resolution spectroscopy (Ciesla et al., 1 Sep 2025).

Component	Coverage	Principal capability
PRIMAger PHI	24–84 μm	LVF-based hyperspectral imaging, $R = 8$
PRIMAger PPI	80–264 μm	Four-band polarimetric imaging at 92, 126, 183, and 235 μm, $R \approx 4$
FIRESS low-resolution mode	24–235 μm	Four slit-fed grating modules, $R \approx 85$–130
FIRESS high-resolution mode	24–235 μm	Fourier-transform-assisted spectroscopy, $R \sim 2000$–20000

PRIMAger consists of two distinct focal planes, both based on kinetic inductance detectors. The PRIMAger Hyperspectral Imager (PHI) covers 24–84 μm, split into PHI1 at 24–45 μm and PHI2 at 45–84 μm, each at $R=8$. The PRIMAger Polarimetric Imager (PPI) covers 80–264 μm with four broad bands centered at 92, 126, 183, and 235 μm, at $R \approx 4$ (Ciesla et al., 1 Sep 2025). The telescope supplies a 50 mm collimated beam through a Beam Steering Mirror at the telescope exit pupil, which is also the PRIMAger entrance pupil. Inside PRIMAger, the common and channel-specific optics reimage the field onto detector planes at $f/21$ for PHI and $f/12$ for PPI; the optical structure is carried by a 1 K Mechanical and Optical Thermal Assembly, while the two focal plane assemblies operate at 125 mK and are cooled by a continuous ADR chain with heritage from Hitomi/XRISM (Ciesla et al., 1 Sep 2025).

PHI and PPI share the telescope field of view but are offset on the sky by 4.7′ edge-to-edge. The total telescope field is about $37' \times 27'$, and the individual PRIMAger sub-fields are about 4′ on a side, relocatable anywhere within the telescope field by the Beam Steering Mirror (Ciesla et al., 1 Sep 2025). Across both focal planes the instantaneous sampling is $F\lambda$, coarser than Nyquist, so scanning is intrinsic to the observing concept.

FIRESS is the Far-Infrared Enhanced Survey Spectrometer, covering 24–235 μm in all modes (Pontoppidan et al., 1 Sep 2025). In low-resolution operation it uses four slit-fed grating modules with resolving power of about 85–130 and broad simultaneous wavelength coverage; in high-resolution operation it inserts a Fourier Transform Module upstream of the gratings, giving a tunable resolving power described in the instrument science paper as $R \approx 4$0, corresponding to about $R \approx 4$1 at 25 μm and $R \approx 4$2 at 112 μm (Pontoppidan et al., 1 Sep 2025). FIRESS is therefore the payload element that turns PRIMA from an imaging mission into a wide-band spectroscopic observatory.

3. Measurement strategies and performance

PRIMAger’s hyperspectral mode is based on linearly variable filters rather than a dispersive spectrograph or an FTS. Each PHI pixel sees a narrow band with $R \approx 4$3, and the central wavelength varies monotonically along the long axis of the array. A source spectrum is therefore acquired by scanning the source along that axis so that it crosses pixels with progressively changing central wavelength (Ciesla et al., 1 Sep 2025). The instrument paper is explicit that PHI never operates in a stare mode: all observations are scans. For large surveys the spacecraft executes long quasi-rectangular scan legs, while the Beam Steering Mirror can impose quasi-periodic patterns such as Lissajous modulation to move the astronomical signal above the detector $R \approx 4$4 knee and to provide sub-pixel sampling independent of scan angle (Ciesla et al., 1 Sep 2025).

PPI uses a different polarimetric architecture. Each pixel is a polarization-sensitive KID with an antenna at one of three orientations separated by 120°, and there is no rotating half-wave plate. Polarization modulation is therefore geometric, produced by analyzer-angle diversity together with the scan pattern (Ciesla et al., 1 Sep 2025, Dowell et al., 2024). In the notation used in the PRIMAger paper, the signal measured by a detector at orientation $R \approx 4$5 is

$R \approx 4$6

and map-making solves the coupled linear system for $R \approx 4$7, $R \approx 4$8, and $R \approx 4$9 after destriping and relative-baseline fitting (Ciesla et al., 1 Sep 2025). The end-to-end simulator for the PRIMA Polarimetric Imager, using deliberately pessimistic $R \approx 85$0 assumptions, finds near-fundamental-limit recovery of simulated $R \approx 85$1, $R \approx 85$2, and $R \approx 85$3 maps in nearby-galaxy observations, validating the basic no-HWP architecture (Dowell et al., 2024).

The PRIMAger sensitivity requirements are stated in a survey language tied to specific programs. For a 1 deg$R \approx 85$4 region mapped in 10 h in large-map mode, the 5σ background-subtracted point-source total-flux-density requirements are 1.18–2.2 mJy in PHI1, 2.2–4.1 mJy in PHI2, and 1.77, 2.56, 3.39, and 4.59 mJy in PPI1–4; the corresponding 5σ polarized flux-density requirements in PPI are 2.50, 3.62, 4.65, and 6.49 mJy (Ciesla et al., 1 Sep 2025). The same paper gives a polarized surface-brightness requirement for galaxy magnetic-field studies of

$R \approx 85$5

at 235 μm in an effective 600 arcsec$R \approx 85$6 diffraction beam for a 10 arcmin$R \approx 85$7 field mapped in 2 h (Ciesla et al., 1 Sep 2025). Relative to Herschel/PACS, the PRIMAger requirements already imply 10–25× greater depth in the same time, or up to about 200× faster mapping for a fixed depth at 100 μm (Ciesla et al., 1 Sep 2025).

FIRESS performance is expressed in line-survey terms. The GO book and the FIRESS science-driver paper describe low-resolution 1 hr 5σ line sensitivities of about $R \approx 85$8–$R \approx 85$9 across 24–235 μm, together with orders-of-magnitude faster spectral mapping than Herschel or SOFIA (Moullet et al., 2023, Pontoppidan et al., 1 Sep 2025). In the FIRESS cosmic-noon survey study, a blind 200 arcmin$R \sim 2000$0 spectroscopic survey with 640 hr on source reaches a 5σ limit of $R \sim 2000$1 at 24 μm, sufficient to detect about 1000 galaxies via the 11.3 μm PAH feature and/or [O III] 52 μm (Fernández-Ontiveros et al., 8 Sep 2025).

4. Scientific program across astrophysics

The science case for PRIMA is organized around a small set of top-level themes but expands into a broad set of application domains. In the PRIMAger instrument paper, the three primary drivers are black hole and star-formation co-evolution in galaxies, the evolution of small dust grains and PAHs over cosmic time, and interstellar magnetic fields in the ISM, star-forming regions, and galaxies (Ciesla et al., 1 Sep 2025). In the FIRESS science-driver paper, the three mission-level science objectives are the origins of planetary atmospheres, the co-evolution of galaxies and supermassive black holes, and the buildup of heavy elements in the Universe (Pontoppidan et al., 1 Sep 2025).

In extragalactic astrophysics, PRIMA is designed to access the obscured side of galaxy growth. The FIRESS cosmic-noon study states that around 90% of UV/optical photons from young stars and AGN are absorbed by dust and reradiated in the mid- to far-infrared at the epoch of peak activity, and develops a blind FIRESS survey strategy for measuring star-formation rates and black-hole accretion rates for hundreds of galaxies at $R \sim 2000$2–4 (Fernández-Ontiveros et al., 8 Sep 2025). PRIMAger contributes the continuum and low-resolution spectral information required to identify PAH complexes and dust SED shapes, while FIRESS provides line diagnostics such as [Ne II] 12.8 μm, [O IV] 25.9 μm, [O III] 52 μm, [N III] 57 μm, and OH absorption/emission features (Ciesla et al., 1 Sep 2025, Pontoppidan et al., 1 Sep 2025).

A distinct extragalactic application is the census of deeply obscured nuclei. The obscured-AGN study argues that PRIMAger can identify heavily buried nuclei through the deep 9.8 μm silicate absorption feature between $R \sim 2000$3 and $R \sim 2000$4, while FIRESS can obtain $R \sim 2000$5 spectra of such systems out to $R \sim 2000$6, detecting PAHs, ices, ionized gas, and molecular gas (Donnan et al., 14 Mar 2025). In the same study, a 1500 hr deg$R \sim 2000$7 survey model yields about 9000 detected galaxies at $R \sim 2000$8–2.5 under a six-band criterion and still 53 galaxies at $R \sim 2000$9–7.5, illustrating the mission’s reach into the obscured early Universe (Donnan et al., 14 Mar 2025).

For magnetism and dust physics, PRIMA’s primary tool is PPI. The PRIMA Polarimetric Imager simulation paper demonstrates near-fundamental-limit recovery of galaxy polarization maps in four bands from 91 to 232 μm (Dowell et al., 2024). The extragalactic MHD study based on synthetic PRIMA observations finds that PRIMA will better sample magnetic turbulence than SOFIA, will recover unresolved intrinsic magnetic-field orientations to about $R=8$0 precision, and will resolve observables such as polarized fraction or magnetic alignment down to scales comparable to the simulation resolution of about 10 pc for galaxies up to 0.5 Mpc away (Maglione et al., 2 Sep 2025). The mission science case therefore extends from Milky Way molecular clouds to nearby galactic disks and their dense star-forming environments (Ciesla et al., 1 Sep 2025, Maglione et al., 2 Sep 2025).

For planetary-system formation, FIRESS is designed around HD, water, HDO, CO, and solid-state features. The FIRESS science-driver paper emphasizes the HD 112 μm line as the only practical far-infrared tracer of total disk gas mass, since its abundance relative to H$R=8$1 is well constrained, and shows a simulated disk spectrum in which an HD line flux of $R=8$2 is detected at S/N $R=8$3 with 10 h per setting in high-resolution mode (Pontoppidan et al., 1 Sep 2025). The same configuration gives access to numerous H$R=8$4O and HDO lines, [O I], [C II], and ice features, enabling direct links between disk volatile budgets and exoplanet-atmosphere composition (Pontoppidan et al., 1 Sep 2025).

The GO literature also extends PRIMA into more specialized applications. One paper proposes FIR cooling-break measurements in AGN jet hotspots, arguing that the synchrotron cooling break frequency

$R=8$5

is expected to fall in or near PRIMA’s band for standard hotspot parameters, with typical hotspot flux densities of order 0.2–2.5 mJy across the PRIMAger bands and exposure times of order 0.2–2 h per channel in $R=8$6 maps (Isobe et al., 2 Sep 2025). Another study uses PRIMAger and FIRESS to address dust stripping in cluster galaxies, concluding that PRIMA can detect diffuse dust emission in ram-pressure-stripped tails, measure FIR cooling lines such as [C II] 158 μm in diffuse gas, and use FIR polarimetry to constrain turbulent magnetic fields in stripped material (Boselli et al., 2 Sep 2025). A further AGN study proposes dust reverberation mapping at 25–30 μm, the peak of the intrinsic AGN infrared SED, using repeated PRIMA visits to measure mid-infrared echoes of optical accretion-disk variability (Gorjian et al., 19 Oct 2025).

5. Surveys, community use, and deconfused far-infrared cartography

PRIMA is explicitly conceived as a survey observatory. The GO Science Book assembles 76 initial science cases, and the observatory concept assumes that most of the mission time will be competed for by the community (Moullet et al., 2023). The large-area survey paper for PRIMAger develops this idea into the $R=8$7-IR survey, a proposed wide-field GO program covering 25% of the sky, or about $R=8$8 sr $R=8$9, in 2059 hr at a scan speed of $R \approx 4$0 (Burgarella et al., 22 Sep 2025). That program uses simultaneous PHI and PPI observations to deliver hyperspectral imaging from 25 to 80 μm and four-band polarimetric imaging from 80 to 260 μm, with 5σ point-source limits of 2.54 mJy at 34.3 μm, 3.44 mJy at 64.5 μm, 0.765 mJy at 92 μm, 1.09 mJy at 126 μm, 1.49 mJy at 172 μm, and 2.58 mJy at 235 μm (Burgarella et al., 22 Sep 2025).

The same $R \approx 4$1-IR study projects a galaxy sample of about $R \approx 4$2 objects to $R \approx 4$3, with deep ecliptic-pole components extending to still higher redshift and wide-area coverage suited to environments ranging from voids to clusters and proto-clusters (Burgarella et al., 22 Sep 2025). The survey is designed to leverage PHI’s $R \approx 4$4 spectral sampling for PAH measurements and PPI’s multi-band polarimetry for large statistical studies of dust polarization in galaxies (Burgarella et al., 22 Sep 2025). In the GO-book framework, $R \approx 4$5-IR sits alongside other wide programs, including all-sky or degree-scale surveys, deep “wedding cake” extragalactic tiers, Galactic plane mapping, molecular-cloud polarimetry, Kuiper Belt studies, and large protoplanetary-disk programs (Moullet et al., 2023).

A central question for any far-infrared survey is confusion. The deblending study “XID+PRIMA, II” uses realistic PRIMAger simulations and a Bayesian method, XID+stepwise, to exploit the mission’s hyperspectral imaging for confusion mitigation (Donnellan et al., 15 Dec 2025). 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 factors of 1.3–3.4 fainter than the classical confusion limit, and down to 0.9, 2.5, 7.6, and 14.8 mJy at 92, 126, 183, and 235 μm, corresponding to factors of 3–5 below confusion (Donnellan et al., 15 Dec 2025). With deeper Euclid-based priors and weak 25 μm flux priors, the study reaches up to a factor $R \approx 4$6 fainter than the confusion limit at about 96 μm, and even demonstrates that PRIMAger data alone can bootstrap positional priors through blind short-wavelength detection followed by deblending (Donnellan et al., 15 Dec 2025).

This survey architecture has direct consequences for core extragalactic science. In the same deblending study, IR-luminous galaxies at $R \approx 4$7 are robustly detected in more than 98% of cases in 12 of 16 representative PRIMAger channels, yielding densely sampled FIR SEDs even several factors below the classical confusion floor (Donnellan et al., 15 Dec 2025). That result is important because it converts PRIMA’s formal instrumental sensitivity into an effective depth that is limited more by survey design and prior information than by the traditional far-infrared confusion barrier.

6. Enabling technologies, calibration, and development status

PRIMA’s instrument strategy depends on large-format KID arrays operating near the photon background limit. The PRIMAger instrument paper reports background-limited performance for prototype KIDs at both 1.5 THz (200 μm) and 12 THz (24 μm), with white noise across the science band, $R \approx 4$8 roll-up only below about 0.1 Hz, and dark NEP of about $R \approx 4$9, well below the mission requirements; it places the detector modules at TRL 5 and identifies vibration, radiation, cosmic-ray, and thermal-cycling tests as the path to TRL 6 (Ciesla et al., 1 Sep 2025).

Array-level detector demonstrations support those requirements. A 210 μm KID prototype optimized for PRIMA achieved an NEP of $f/21$0 at 10 Hz under low loading, and extrapolation indicated that it could remain photon-noise limited up to 20 fW, providing high dynamic range for bright-source observations (Hailey-Dunsheath et al., 2023). A later flight-like 210 μm array, read out with a Radio Frequency System on a Chip in multitone mode, found that 92% of the measured KIDs had NEP below $f/21$1 at 10 Hz, meeting the PRIMA threshold for subdominant detector noise (Kane et al., 2024).

Optical coupling is another enabling technology. The FIRESS lenslet-array paper develops monolithic silicon lenslet arrays with 1008 pixels on a 900 μm pitch, fabricated by grayscale lithography and deep reactive ion etching, and anti-reflection coated with Parylene-C (Dahal et al., 13 Nov 2025). In the updated Band 4 design, extending the lens profile into the hexagonal corners increases the collecting area and directs about 14% more optical power to the detectors than the earlier circular-lens profile, while destructive sectioning confirms epoxy bond layers of $f/21$2 for Band 1 and 1–4 μm for Band 4, within the optical-loss budget (Dahal et al., 13 Nov 2025).

Radiation tolerance at Sun–Earth L2 has also been tested directly. A fully cryogenic alpha-irradiation experiment exposed FIRESS KID arrays to approximately 62% of the expected 5-year L2 displacement-damage dose, then remeasured quasiparticle lifetimes, resonant frequencies, and quality factors (Kane et al., 30 Apr 2026). The reported changes were small: mean quasiparticle lifetime shifted from 0.37 ms to 0.36 ms, the mean fractional resonance shift was about $f/21$3, and no significant degradation of internal quality factor was found, implying that displacement damage is unlikely to be a mission-limiting issue for the KID architecture (Kane et al., 30 Apr 2026).

Calibration strategy is correspondingly multi-layered. For PPI, the polarimetry simulation paper develops a sky-based relative-gain calibration using repeated observations at multiple roll angles, an internal calibration source mounted on the Beam Steering Mirror structure, and beam-matching strategies to reduce $f/21$4 leakage from detector-to-detector beam differences (Dowell et al., 2024). For PRIMAger and FIRESS more broadly, the instrument papers emphasize hardware uniformity, cold stops and filter stacks, shared readout heritage from balloon and ground-based KID instruments, and map-making techniques derived from Herschel/SPIRE and Planck (Ciesla et al., 1 Sep 2025, Pontoppidan et al., 1 Sep 2025).

Taken together, these developments define PRIMA not as a single instrument but as an observatory system: a 4.5 K telescope, dual payload, hyperspectral and polarimetric scanning, wide-band low- and high-resolution spectroscopy, deconfusion-aware survey design, and a detector/readout chain already benchmarked against the dominant instrumental risks.