LIFE: Large Interferometer for Exoplanets

Updated 27 November 2025

LIFE is a mission concept that employs formation-flying nulling interferometry to directly detect thermal emission from temperate terrestrial exoplanets in the mid-infrared.
It uses a multi-telescope X-array configuration with advanced kernel nulling to suppress starlight and enable precise extraction of planetary signals.
The mission aims to retrieve key atmospheric molecules and biosignatures, providing accurate measurements of planetary sizes, temperatures, and habitability indicators.

The Large Interferometer for Exoplanets (LIFE) is a mission concept for a formation-flying, space-based nulling interferometer designed to directly detect and characterize the thermal emission from temperate terrestrial exoplanets. Focused on the mid-infrared (MIR) domain (4–18.5 $\ \mu$ m), LIFE targets Earth-size planets in the habitable zones of nearby stars, aiming to establish the prevalence of biosignature gases and to constrain crucial planetary and atmospheric parameters unreachable by transit or ground-based methods. LIFE's architecture and scientific strategy synthesize advances in nulling interferometry, astrobiological retrievals, multiplexed array design, stability engineering, and exoplanet statistic-driven yield optimization (Cacaj et al., 9 Apr 2025, Menti et al., 31 Oct 2024).

1. Science Objectives and Rationale

The science goals of LIFE are to detect and obtain MIR spectra of dozens of terrestrial exoplanets in their habitable zones, measure their radii and temperatures, and retrieve atmospheric abundances of key molecules: $\mathrm{H_2O}$ , $\mathrm{CO_2}$ , $\mathrm{O_3}$ , $\mathrm{CH_4}$ , $\mathrm{N_2O}$ , as well as secondary trace gases and surface pressure proxies (Defrère et al., 2018, Alei et al., 2022, Alei et al., 18 Jun 2024, Angerhausen et al., 16 Jan 2024). The MIR is selected because:

It contains the strongest vibrational transitions for biosignature gases, notably $\mathrm{CO_2}$ (15 $\mu$ m), $\mathrm{O_3}$ (9.6 $\mu$ m), $\mathrm{H_2O}$ (6–8 $\mu$ m), $\mathrm{CH_4}$ (7.7 $\mu$ m), $\mathrm{N_2O}$ (16–18 $\mu$ m), and $\mathrm{PH_3}$ (8–12 $\mu$ m).
Thermal emission directly constrains planetary size and equilibrium temperature, even for non-transiting planets.
Planet/star contrast in the MIR is $10^{-7}$ for an Earth–Sun twin, relaxing suppression requirements compared to visible-light missions [$2504.06776$, $2401.08492$, $2410.23892$].
Key diagnostics for atmospheric habitability and reconstruction of evolutionary epochs can be accessed through Bayesian retrievals on high-S/N and $R\sim50$ spectra [$2204.10041$, $1801.04150$].

2. Instrument Architecture and Array Configurations

The baseline configuration is a formation-flying kernel-nulling interferometer:

Four to five collector spacecraft (each $D=2$ m mirrors) form an $X$ -array (rectangular, double-Bracewell) or a regular pentagonal kernel-nulling array (Hansen et al., 2022, Hansen et al., 2022).
Baseline lengths are adjustable ( $b_\mathrm{null} =$ 10–100 m for nulling, $b_\mathrm{img} =$ 60–600 m for imaging modulation), targeting an angular inner working angle $\mathrm{IWA} = \lambda/(2B)$ comparable to coronagraphic single-aperture telescopes of $D\sim600$ m at $\lambda=4\,\mu\mathrm{m}$ [$2504.06776$].
A central beam-combiner spacecraft implements the nullers, phase-shifters, and chopping required to suppress starlight to $N\lesssim10^{-5}$ – $10^{-6}$ ( $N$ = null depth) over the broad bandpass, with redundancy to handle failed collectors via internal shutters (Hansen et al., 2022).

Kernel nulling, as studied in detail for 3-, 4-, and 5-aperture arrays, achieves second- and fourth-order nulls at the optical axis, delivers robust insensitivity to piston errors ( $\lesssim3$ nm RMS) and enables straightforward extraction of differential planet signals even under beam combiner imperfections (Hansen et al., 2022). The five-telescope kernel design provides up to 23% yield gain for Earth-twin detection compared to the classic 4-telescope X-array, and up to 1.3 $\times$ improvement in SNR for atmospheric characterization (Hansen et al., 2022, Hansen et al., 2022).

3. Nulling and Signal Extraction Methodologies

Nulling Interferometry Principle:

A $\pi$ phase shift is introduced in each short baseline to destructively interfere on-axis starlight, minimizing stellar photon leakage, with modulation via chopping and slow array rotation unveiling off-axis planetary signals through their unique MIR thermal emission time-series (Dannert et al., 2022).
The transmission map $T(\delta_r,\delta_\theta;\lambda)$ is highly structured, leading to nontrivial confusion properties but affording imaging resolution equivalent to a filled-aperture $D=600\,\mathrm{m}$ telescope at $\lambda=4\,\mu\mathrm{m}$ (Cacaj et al., 9 Apr 2025).

Signal Extraction Algorithms:

The LIFEsim pipeline simulates time series in each spectral channel, including photon noise from star, local zodiacal, exozodiacal, and planetary sources; maximum likelihood extraction enables retrievals of angular separation, temperature, radius, and spectrum to $<2\%$ , $<10\%$ , and $<20\%$ accuracy, respectively (Dannert et al., 2022).
Phase-space synthesis decomposition (PSSD) uses spectral modulation rather than time modulation, enables robust inversion even with sparse rotation, and tolerates OPD instabilities up to $\sim10$ nm over minutes, substantially relaxing engineering constraints versus classical baseline rotation schemes (Matsuo et al., 2023).

4. Spatial Resolution, Target Confusion, and Survey Yields

Spatial Resolution and Confusion Criteria:

The classical Rayleigh criterion is fundamentally inadequate for LIFE's rotating multipoint nullers. Model-based criteria, incorporating the Kullback–Leibler divergence $J(P,Q;\Sigma)$ and the principle of parsimony (Occam penalties, e.g., AIC), define regions of possible source blending ("photobombing") and complete destructive interference ("cancellation") [$2504.06776$].
The effective resolution $\delta_1$ (contamination) and $\delta_0$ (cancellation) are derived as radii where the statistical confusion becomes nonzero. For the LIFE X-array, $\delta_1$ at $4\,\mu$ m matches that of a $600$ m monolithic telescope.

Yield and Survey Results:

Comprehensive Monte Carlo surveys using LIFEsim with realistic noise and synthetic nearby stellar populations anticipate detection of $\sim73.4$ habitable planets, with only $\sim2.1$ potentially contaminated—i.e., a $2.8\%$ upper-limit photobombing rate [$2504.06776$].
In a typical 2.5-year search phase with four $2$ m collectors and $5\%$ throughput, yields are $\sim550$ detected planets (radii $0.5$– $6\,R_\oplus$ ), including $25$–$45$ rocky ($0.5$– $1.5\,R_\oplus$ ) habitable-zone objects (Quanz et al., 2021), and $\sim15$ –$22$ rocky HZ planets around Sun-like/FGK stars (Kammerer et al., 2022).

5. Biosignature Retrieval and Atmospheric Characterization

Retrieval Framework:

Atmospheric retrievals employ forward models (petitRADTRANS, PSG/SMART), Bayesian inference (pyMultiNest), and sensitivity studies in $R$ and S/N (Alei et al., 2022, Longstaff et al., 2022).
Baseline mission requirements to distinguish major molecules are $R=50$ , S/N = 10 (per channel at $11.2\,\mu$ m), and $\lambda=4$ – $18.5\,\mu$ m (Konrad et al., 2021).
S/N and $R$ gains improve abundance posteriors for O $_3$ and CH $_4$ by up to 50\%, but diminishing returns arise for $R>100$ .

Performance on Representative Targets:

For an Earth twin at 10 pc: $\log_{10}\mathrm{CO_2} = -3.39\pm0.14$ , $\log_{10}\mathrm{H_2O} = -2.99\pm0.10$ , $\log_{10}\mathrm{O_3} = -6.52\pm0.45$ , and $T_\textrm{surf}=284\pm10$ K (Alei et al., 18 Jun 2024).
Key biosignature gases such as N $_2$ O, CH $_3$ Cl, CH $_3$ Br, and PH $_3$ are also accessible. For golden targets ( $<$ 2 pc), detection of N $_2$ O at 10 $\times$ modern terrestrial flux is achievable at SNR $\sim$ 7 in 1 day; for standard M-dwarf HZ planets ( $\sim$ 5 pc) 10–100 days suffice for SNR $\geq$ 7 (Angerhausen et al., 16 Jan 2024, Angerhausen et al., 2022).

Joint MIR and Reflected-Light Retrievals:

Synergies between LIFE (thermal MIR) and visible/NIR reflected light missions (e.g., HWO) produce multiplicatively stronger constraints: O $_2$ is robustly detected only with reflected light, but CO $_2$ , H $_2$ O, O $_3$ posteriors shrink by up to 50\% and temperature uncertainties by 40\% (Alei et al., 18 Jun 2024).

6. Target Database, Yield Optimization, and Survey Design

A dynamic, virtual-observatory-compliant database provides stellar/planetary/ disk properties for %%%%95 $\mathrm{H_2O}$ 96%%%% systems within 30 pc (LIFE-StarCat, final target lists, and "golden systems"), supporting (Menti et al., 31 Oct 2024):

Yield simulations via LIFEsim, linking telescope/instrument parameters, baseline choices, and exozodiacal dust constraints to optimize planet detection returns.
Figures of merit per star, supporting prioritization under mission time constraints and variable baseline configurations.
Gaps exist in exozodi measurements ( $<30$ \% stars have $<1$ zodi limits), mid-IR photometry, and close companion screening, driving parallel JWST/MIRI and LBTI high-precision campaigns.

Yield trade-offs follow $Y \propto D^{1.5}\,\eta^{0.5}\,z^{-0.15}\,W^{0.25}$ (where $D$ is mirror diameter, $\eta$ throughput, $z$ exozodi, $W$ wavelength span), and optimal survey design is strongly dependent on mirror aperture and total mission time. For 2 m mirrors and $5\%$ throughput, $\gtrsim30$ HZ rocky planet detections require including FGK stars or improvements in throughput (Kammerer et al., 2022).

7. Technological Implementation and Instrumental Stability

Key enabling technologies and error budgets are summarized as follows (Hansen et al., 2022, Matsuo et al., 2023):

Fringe tracking and piston stability: RMS $<3$ nm over $>10$ minutes for photon-limited operation, $<0.5$ nm for suppressing fourth-order leakage in kernel-nulling up to $10^{-7}$ planet/star contrasts (4–19 $\mu$ m).
Beam combiner tolerances: Reflectance uncertainty $|\Delta R| < 5\%$ , phase error $\Delta\phi<3^\circ$ , with redundancy built-in for partial array operation.
System throughput: End-to-end efficiency (optics $\times$ detector QE) of $>5\%$ is critical; increased throughput linearly translates to decreased required integration times, or to enable detection of fainter biosignature gases.
PSSD technique: Adoption of phase-space synthesis decomposition relaxes the OPD and amplitude control requirement from the $\sim1.5$ nm scale to $\sim10$ nm over minutes, a plausible step-change in system engineering effort (Matsuo et al., 2023).
Array configuration: Five-telescope kernel nulling provides the highest yield and SNR, with multi-kernel redundancy and robust operation under component failure.