LuSEE-Night: Lunar Surface Electromagnetics Experiment

• LuSEE-Night is a low‑frequency radio astronomy pathfinder using four 3‑m monopole antennas to study the lunar farside and its plasma environment.
• It employs innovative calibration and measurement techniques—like 16 correlation products and Wiener filter map-making—to mitigate subsurface and beam uncertainties.
• The experiment paves the way for future lunar Dark Ages cosmology and exoplanet radio studies by validating autonomous, RFI‑free nighttime operations.

LuSEE‑Night, the Lunar Surface Electromagnetics Experiment–Night, is a low‑frequency radio astronomy pathfinder for the lunar far side. In current descriptions it is a joint NASA–DOE–ESA instrument that employs four 3‑m monopole antennas arranged as two horizontal cross pseudo‑dipoles on a rotational stage, measures 16 correlation products as a function of frequency, and is designed to operate autonomously through the lunar night in the radio‑quiet farside environment. Its scientific role spans low‑frequency sky characterization, local plasma and exo‑ionospheric studies, and technology and analysis development for Dark Ages and Cosmic Dawn 21‑cm cosmology and later, much larger lunar arrays (Yousuf et al., 23 Apr 2026, Camacho et al., 22 Aug 2025).

1. Scientific scope and frequency regime

LuSEE‑Night is framed as a pathfinder for observing the ultra‑low‑frequency sky from the lunar far side. Its science program includes measuring the low‑frequency Galactic synchrotron spectrum and its polarization, characterizing the local plasma and exo‑ionosphere above the lunar surface, and providing a technology and analysis pathfinder toward Dark Ages / Cosmic Dawn 21‑cm cosmology. In this context, the broader LuSEE‑Night band is described as $0.1\text{–}50\ \text{MHz}$, while several instrument and analysis papers formulate the operational sky band as $1\text{–}50\ \text{MHz}$; the Dark Ages global 21‑cm signal is in turn described as expected at $3\text{–}30\ \text{MHz}$, and the broader band is used to characterize foregrounds and systematics (Saliwanchik et al., 2024, Yousuf et al., 23 Apr 2026).

A central cosmological motivation is the redshifted 21‑cm transition of neutral hydrogen. One instrument paper writes the observed frequency as

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$

for the high‑redshift regime targeted as a precursor to dedicated Dark Ages arrays, while the power and operations study states that $0.1\text{–}50\ \text{MHz}$ covers $27 < z < 1100$ (Bale et al., 2023, Saliwanchik et al., 2024). In this framework, LuSEE‑Night is not presented as a full Dark Ages detection experiment; rather, it is a foreground and instrumentation pathfinder aimed at establishing whether the spectral smoothness, calibration, and environmental control required for mK‑level cosmology are achievable on the Moon (Bale et al., 2023).

The astrophysical scope is broader than cosmology. Below $50\ \text{MHz}$, the sky is dominated by Galactic synchrotron emission, and the Sun and Jupiter are among the brightest discrete sources. LuSEE‑Night is described as capable of time‑variable and polarized emission studies and of occultation measurements as bright sources set behind the lunar limb (Bale et al., 2023). In later work on exoplanet radio astronomy from the Moon, it is also treated as a pathfinder for $1\text{–}10\ \text{MHz}$ exoplanet studies: LuSEE‑Night and ROLSES are expected to study Solar System planets as exoplanet analogs and to place the first meaningful upper limits on exoplanetary radio flux below $10\ \text{MHz}$, with the LuSEE‑Night sensitivity curve averaged over the entire two‑year mission and including post‑processing using a Lomb–Scargle periodogram (Turner et al., 11 Aug 2025).

2. Instrument architecture and measurement products

The instrument concept is a compact crossed system of electrically short monopoles above the lunar regolith. One detailed instrument model describes four 3‑m BeCu stacer monopoles, labeled N, E, S, and W, arranged in a cross‑dipole configuration; the monopoles form two orthogonal dipole pairs and thus provide sensitivity to two linear polarizations. The antenna cluster is mounted on a motor‑driven turntable that can rotate in azimuth in the plane of the lunar surface, enabling beam/sky rotation and polarization characterization (Yousuf et al., 23 Apr 2026).

The receiver chain is described in complementary ways across the LuSEE‑Night literature. A mission overview characterizes the payload as a 4‑channel, $50\ \text{MHz}$ Nyquist baseband receiver with two orthogonal $1\text{–}50\ \text{MHz}$0 tip‑to‑tip electric dipoles, sampled at $1\text{–}50\ \text{MHz}$1, corresponding to a Nyquist band up to $1\text{–}50\ \text{MHz}$2. An analysis paper on map‑making then states that the correlator produces 2048 spectral bins covering $1\text{–}50\ \text{MHz}$3 with $1\text{–}50\ \text{MHz}$4 spacing, and that the four analog channels yield 4 auto‑correlations and 6 complex cross‑correlations, or 16 independent real‑valued observables per time–frequency sample when the real and imaginary parts of the cross products are kept separately (Bale et al., 2023, Camacho et al., 22 Aug 2025).

The basic observable in the electromagnetic forward model is the voltage power spectral density. For a single antenna, the model is

$1\text{–}50\ \text{MHz}$5

with a receiver coupling factor

$1\text{–}50\ \text{MHz}$6

and, in the idealized receiver model used in that study, $1\text{–}50\ \text{MHz}$7 with $1\text{–}50\ \text{MHz}$8 and $1\text{–}50\ \text{MHz}$9 (Yousuf et al., 23 Apr 2026). The map‑making formalism instead works directly with the 16 correlation products $3\text{–}30\ \text{MHz}$0, each represented as a beam‑weighted integral of the sky,

$3\text{–}30\ \text{MHz}$1

where the effective intensity beam $3\text{–}30\ \text{MHz}$2 depends on the monopole pair, the antenna orientation on the lander, the turntable angle, and the lunar orientation relative to the celestial sphere (Camacho et al., 22 Aug 2025).

A recurring design point is that LuSEE‑Night is not treated as a classical large‑baseline interferometer. The antennas are electrically short, the beams are broad, and the system is strongly coupled to the lander and the regolith. This suggests a hybrid measurement philosophy: total‑power and cross‑correlation spectroscopy for global and low‑resolution sky structure, rather than high‑fidelity imaging in the conventional radio‑interferometric sense (Camacho et al., 22 Aug 2025, Yousuf et al., 23 Apr 2026).

3. Farside environment, night operations, and power system

The lunar farside is central to the experiment’s rationale. LuSEE‑Night is designed to observe in drift scan during lunar night while the Moon shields it from radio frequency interference from both the Earth and Sun, and it transmits science and telemetry data back to Earth via an orbital relay during the lunar day. The far side is described as uniquely suited for ultra‑long‑wavelength astronomy because it is shielded from terrestrial RFI, because there is no terrestrial ionosphere, and because the local night removes direct solar radio emission (Saliwanchik et al., 2024, Yousuf et al., 23 Apr 2026).

Electromagnetic cleanliness is treated as a mission requirement rather than a convenient by‑product. One mission description states that the CLPS lander must cease all operations before the first nightfall and remain powered down, so that LuSEE‑Night operates through the lunar night without the electromagnetic interference of an operating lander system. This requirement is presented as a response to earlier experience in which lander‑generated EMI severely affected low‑frequency lunar radio measurements (Bale et al., 2023).

The power and thermal design are correspondingly mission‑defining. LuSEE‑Night is described as a $3\text{–}30\ \text{MHz}$3 instrument with a $3\text{–}30\ \text{MHz}$4 Li‑ion battery of nominal capacity $3\text{–}30\ \text{MHz}$5-hr, required to survive and operate through a $3\text{–}30\ \text{MHz}$6 lunar night while keeping the spectrometer running under hard radiation, with external temperatures reaching $3\text{–}30\ \text{MHz}$7 at night and $3\text{–}30\ \text{MHz}$8 during the day. The photovoltaic system is distributed across the top, east, and west faces of the instrument, and the battery can be fully recharged in $3\text{–}30\ \text{MHz}$9 of effective charging time within the lunar day (Saliwanchik et al., 2024).

Operations are organized into four normal modes: Maintenance Mode, Transmission Mode, Science Mode, and Powersave Mode. Science Mode is the nighttime default, with communications off and the spectrometer active; Powersave Mode turns the spectrometer off and enables a low‑power heater to preserve battery temperature and state of charge. This mode structure follows directly from the power simulations: continuous full‑load operation throughout the lunar night is not feasible with the allowed battery mass, so nighttime observing must be duty‑cycled (Saliwanchik et al., 2024).

A common misconception is that lunar farside radio quiet automatically implies easy ultra‑low‑frequency astronomy. The LuSEE‑Night design literature instead presents a more constrained picture: the farside solves the terrestrial RFI and Earth‑ionosphere problem, but precision observations remain limited by power, thermal design, lander compatibility, and local lunar electromagnetic coupling (Saliwanchik et al., 2024, Bale et al., 2023).

4. Calibration, subsurface coupling, and foreground inference

A major technical issue for LuSEE‑Night is that its antennas do not operate above an electromagnetically passive ground. The most detailed calibration study states that the unknown dielectric properties of the subsurface at the LuSEE‑Night landing site impose the most significant limitation for precision instrument calibration, because reflections from the lunar subsurface can change the primary beam at the $$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$0 level. Simulations show that the subsurface, modeled as a lossy dielectric, can absorb a large amount of the sky power and, more importantly, can complicate the beam pattern and the transfer function between sky and measured voltages (Yousuf et al., 23 Apr 2026).

In that study the subsurface is represented as a two‑layer, lossy dielectric half‑space with

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$1

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$2

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$3

for 216 combinations, all with loss tangent $$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$4. The strongest effect is near the first antenna resonance, at $$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$5, where changing the subsurface primarily alters the resonance amplitude, position, and width (Yousuf et al., 23 Apr 2026).

The observed sky temperature is modeled through a beam convolution,

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$6

and the foreground itself is written as a modified power law in log frequency,

$$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$7

with $$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$8, $$\nu \sim \frac{1420\ \text{MHz}}{1+z} \sim 10\text{–}50\ \text{MHz},$$9 taken from the ULSA map, and free parameters $0.1\text{–}50\ \text{MHz}$0 (Yousuf et al., 23 Apr 2026).

The statistical result is that the spectral signatures of subsurface structure and Galactic foregrounds are sufficiently distinct to permit joint inference. The authors build an RBF emulator over the seven‑dimensional parameter space

$0.1\text{–}50\ \text{MHz}$1

and use a Gaussian likelihood,

$0.1\text{–}50\ \text{MHz}$2

with $0.1\text{–}50\ \text{MHz}$3 taken as $0.1\text{–}50\ \text{MHz}$4 of the mock PSD in each frequency bin. In the idealized test case, all parameters are recovered within $0.1\text{–}50\ \text{MHz}$5 of the inputs, and the authors conclude that parameters of both the galaxy and subsurface properties can be estimated jointly (Yousuf et al., 23 Apr 2026).

The caveat is explicit. The same paper stresses that these results are optimistic because the foreground model is spatially simple, the subsurface model is only two‑layer, the lunar thermal emission is assumed perfectly calibrated, and the RBF emulator degrades when all three subsurface parameters are off‑grid. A plausible implication is that LuSEE‑Night calibration will be limited less by raw sensitivity than by how well the regolith, beam chromaticity, and voltage transfer function can be modeled in a single forward framework (Yousuf et al., 23 Apr 2026).

5. Linear map‑making and low‑resolution lunar farside sky imaging

Although LuSEE‑Night has only four antennas, later work shows that its 16 correlation products contain enough information for low‑resolution sky mapping. The formal data model is written as

$0.1\text{–}50\ \text{MHz}$6

where $0.1\text{–}50\ \text{MHz}$7 is the stacked data vector of all auto‑ and cross‑correlation products, $0.1\text{–}50\ \text{MHz}$8 is the sky map vector in spherical harmonics, $0.1\text{–}50\ \text{MHz}$9 is the time‑ and frequency‑dependent response matrix, and $27 < z < 1100$0 is the noise vector with covariance $27 < z < 1100$1 (Camacho et al., 22 Aug 2025).

Map‑making is performed with the Wiener filter. Assuming a Gaussian signal prior with covariance $27 < z < 1100$2 and Gaussian noise with covariance $27 < z < 1100$3, the maximum a posteriori estimator is

$27 < z < 1100$4

equivalently

$27 < z < 1100$5

In the LuSEE‑Night implementation, the prior $27 < z < 1100$6 is built frequency‑by‑frequency from the ULSA angular power spectrum, and the reconstruction is carried out up to $27 < z < 1100$7 (Camacho et al., 22 Aug 2025).

The thermal component of the noise is described with a radiometer expression for a correlation product $27 < z < 1100$8,

$27 < z < 1100$9

with $50\ \text{MHz}$0 and $50\ \text{MHz}$1 in the fiducial simulation. The same study then enlarges $50\ \text{MHz}$2 to marginalize over calibration systematics, treating gain fluctuations and beam uncertainty as additional noise covariance terms rather than as fixed corrections (Camacho et al., 22 Aug 2025).

For gains, the time‑dependent perturbations $50\ \text{MHz}$3 are modeled as Gaussian processes with covariance

$50\ \text{MHz}$4

For beams, the uncertainty enters through

$50\ \text{MHz}$5

where $50\ \text{MHz}$6 is the mismatch between true and modeled response matrices. This treatment is important because it demonstrates that beam knowledge uncertainty and gain drifts need not be ignored; they can be propagated explicitly through the map posterior (Camacho et al., 22 Aug 2025).

Under reasonable assumptions about instrument performance, the map‑making analysis concludes that LuSEE‑Night should be able to map the sub‑$50\ \text{MHz}$7 sky at a $50\ \text{MHz}$8-degree resolution. A misconception that a four‑monopole system cannot produce sky maps is therefore misleading: the individual beams are broad, but the aggregate information accumulated over lunar rotation and turntable angle is sufficient for low‑resolution deconvolution (Camacho et al., 22 Aug 2025).

6. Programmatic lineage, pathfinder role, and longer‑term significance

LuSEE‑Night occupies an intermediate position in the developing lunar low‑frequency radio program. Earlier CLPS studies described LuSEE and ROLSES as landed pathfinders for a later $50\ \text{MHz}$9-km interferometric array, FARSIDE, composed of 128 pairs of dipole antennas. In that sequence, LuSEE and ROLSES were intended to characterize the plasma environment above the lunar surface and to measure the fidelity of radio spectra on the surface, while FARSIDE would extend the effort to $1\text{–}10\ \text{MHz}$0 interferometric imaging and Dark Ages science (Burns et al., 2021, Burns et al., 2021).

Within this lineage, LuSEE‑Night is the more explicitly cosmology‑oriented farside pathfinder. The FARSIDE study argues that the lunar farside is the only location in the inner solar system where sky‑noise‑limited observations can be carried out at sub‑MHz frequencies, and that a smaller precursor should demonstrate RFI‑free, sky‑noise‑dominated performance, beam and bandpass control, and realistic foreground removal methods before a much larger Dark Ages array is attempted. LuSEE‑Night fits that role: it is smaller than FARSIDE, operates as a single station rather than as a 128‑node interferometer, and is aimed at establishing the low‑frequency lunar farside measurement problem in practice (Burns et al., 2021).

ROLSES provides an operational precursor on the lunar surface. One ROLSES paper states explicitly that the instrument was designed to provide useful information on the development and observational strategy for LuSEE‑Night, and another calls ROLSES‑1 “NASA’s first radio telescope on the Moon” and a trailblazer for lunar radio telescopes, with LuSEE‑Night identified as a later farside successor whose primary science goal is placing the first limits on the cosmological 21‑cm line of the Dark Ages. ROLSES‑1’s detections of terrestrial technosignatures and the low‑frequency Galactic background, and the statistical tools developed for its analysis, are presented as directly relevant to LuSEE‑Night calibration and data reduction (Gopalswamy et al., 23 Mar 2026, Hibbard et al., 12 Mar 2025).

The future scientific reach extends beyond cosmology. In exoplanet radio studies, LuSEE‑Night is treated as the first exoplanet‑oriented lunar pathfinder: it can observe the Sun and the radio‑loud Solar System planets as exoplanet analogs, use two years of data and periodic stacking with Lomb–Scargle periodograms, and place the first meaningful upper limits on exoplanetary radio flux below $1\text{–}10\ \text{MHz}$1. In that framework it is the precursor to larger lunar arrays such as FarView and FARSIDE, which are proposed to carry exoplanet magnetospheric science from upper limits to detections (Turner et al., 11 Aug 2025).

Taken together, these papers define LuSEE‑Night less as an isolated instrument than as a convergence point for several lines of lunar radio development: the environmental characterization begun by LuSEE and ROLSES, the calibration and deployment requirements of large farside arrays, the foreground and systematics control required by Dark Ages cosmology, and the opening of the $1\text{–}10\ \text{MHz}$2 window for exoplanet radio astronomy. This suggests that its lasting significance will depend as much on quantified systematics, validated forward models, and demonstrated night‑side operations as on any single astrophysical result.