LuSEE-Night: Lunar Radio Pathfinder

• LuSEE-Night is a low-frequency radio astronomy payload on the lunar farside that conducts full-Stokes spectral-density measurements below 50 MHz.
• It utilizes four 3 m monopole antennas on a motor-driven carousel with a 4-channel, 50 MHz receiver system to capture detailed sky and calibration data.
• Its mission design leverages lunar night operations and duty-cycled power management to minimize terrestrial and spacecraft interference for Dark Ages studies.

LuSEE-Night—described in the literature both as the Lunar Surface Electromagnetics Explorer “Night” and the Lunar Surface Electromagnetics Experiment at Night—is a low-frequency radio astronomy payload for the lunar farside, designed to operate through the lunar night and to make full-Stokes, spectral-density measurements of the radio sky below 50 MHz in an environment shielded from terrestrial radio-frequency interference and, during night operations, from lander-generated electromagnetic interference. Across the mission papers, it is consistently framed as a pathfinder for lunar low-frequency radio astronomy and for future efforts to measure the global 21 cm signal from the Dark Ages, with strong emphasis on sky characterization, foreground control, and end-to-end calibration (Bale et al., 2023).

1. Programmatic origin and mission context

LuSEE was selected by NASA in 2019 under the Lunar Surface Instrument and Technology Payloads (LSITP) program. The project is described as having heritage from the FIELDS instrument on Parker Solar Probe, and in early 2020 NASA partnered with the DOE to develop LuSEE-Night specifically as a low-frequency radio pathfinder. Later publications describe it as a joint NASA–DOE–ESA low-frequency radio telescope, reflecting an expanded institutional framing (Bale et al., 2023, Yousuf et al., 23 Apr 2026).

Published descriptions also document an evolving mission architecture and schedule. The 2023 overview states that LuSEE-Night would be delivered to the Moon by NASA’s Commercial Lunar Payload Services (CLPS) program in late 2025 or early 2026, as part of the CS-3 mission carrying ESA’s Lunar Pathfinder relay satellite. The 2024 power study describes the payload as mounted on a Blue Ghost Lunar Lander and delivered by an Elytra Transfer Vehicle, both from Firefly Aerospace. The 2026 subsurface-calibration paper instead states that the telescope will land on the lunar farside in 2027 (Bale et al., 2023, Saliwanchik et al., 2024, Yousuf et al., 23 Apr 2026).

This programmatic trajectory is significant because LuSEE-Night is not presented merely as a single stand-alone instrument. It is positioned as an enabling experiment for a broader lunar-radio-observatory concept: a compact farside system intended to demonstrate site quality, instrument stability, calibration methodology, and scientifically useful data products before more ambitious Dark Ages missions are attempted.

2. Scientific scope and low-frequency rationale

The mission literature treats the lunar farside as the best available site for radio astronomy at very low frequencies because Earth-based observing is limited by three factors: the terrestrial ionosphere becomes opaque below roughly its plasma frequency, around $\mathcal{O}(10\,\mathrm{MHz})$; human-made radio interference is severe; and the low-frequency sky is dominated by bright Galactic synchrotron emission (Bale et al., 2023). During the lunar night, the Moon also shields the instrument from radio-frequency interference from both the Earth and Sun, further improving observing conditions (Saliwanchik et al., 2024).

The scientific program has two closely related layers. The immediate objectives are empirical characterization of the poorly explored sky below 20 MHz, including the global low-frequency sky spectrum, spatial structure in the Galactic synchrotron background, and bright variable sources such as the Sun and Jupiter. The longer-term objective is to support cosmological measurements of the highly redshifted 21 cm signal from the early Universe, especially the Dark Ages (Bale et al., 2023).

For the cosmological use case, the mission papers are explicit about the dynamic-range problem. The Dark Ages 21 cm feature is expected to be buried more than 5 orders of magnitude below the Galactic foreground, and the 2024 power study states a science requirement of constraining any non-smooth monopole signal at the $10^{-3}$ level relative to foregrounds. The same study describes the target observing range as 0.1–50 MHz, corresponding to $27 < z < 1100$ for the redshifted 1.42 GHz hyperfine transition of neutral hydrogen (Bale et al., 2023, Saliwanchik et al., 2024).

For that reason, LuSEE-Night is repeatedly characterized not as a definitive Dark Ages detection mission, but as a foreground and instrument characterization pathfinder. Its scientific value lies as much in measuring the radio environment and calibration-relevant systematics as in the sky measurements themselves.

3. Instrument configuration and measured observables

LuSEE-Night is a compact radio observatory with four 3 m monopole antennas arranged as two horizontal cross pseudo-dipoles or, equivalently, as two orthogonal $\sim 6$ m tip-to-tip electric dipole antennas formed from 3 m BeCu stacer elements on each side of the dipole. The antenna system is mounted on a motor-driven carousel or turntable that allows rotation in the lunar surface plane during lunar daytime, when solar power is available. The rotation is intended to help disentangle intrinsic antenna response, coupling to the lander structure, and coupling to the local regolith dielectric environment (Bale et al., 2023, Camacho et al., 22 Aug 2025).

At the receiver level, the instrument is described as a 4-channel, 50 MHz Nyquist baseband receiver system. Each of the four single-ended antenna voltages is measured with a high-impedance, low-noise JFET front end, and the spectrometer samples the voltages at 102.4 Msamples/s. In the map-making study, the digital back-end is described as a 4-channel spectrometer/correlator that channelizes the signals into 2048 frequency bins spanning 0–51.2 MHz with 25 kHz spacing (Bale et al., 2023, Camacho et al., 22 Aug 2025).

Subsystem	Published specification	Measurement role
Antennas	Four 3 m monopoles / BeCu elements forming two orthogonal $\sim 6$ m pseudo-dipoles on a rotational stage	Wide zenith-pointing beams; two orthogonal linear polarizations
Front end and sampling	4-channel, 50 MHz Nyquist baseband receiver; high-impedance, low-noise JFET front end; 102.4 Msamples/s	Voltage acquisition for low-frequency spectrometry
Correlator products	2048 bins over 0–51.2 MHz with 25 kHz spacing; four $V_{ii}$ and six unique complex $V_{ij}$	16 independent real-valued correlation products per time sample

The correlator forms the four auto-correlations $V_{ii}$ and the six unique complex cross-correlations $V_{ij}$ for $i\neq j$. Since each auto-correlation is real and each complex cross-correlation contributes real and imaginary parts, the instrument yields $10^{-3}$0 independent real-valued correlation products at each time sample. This measurement basis is sufficient for polarization-sensitive radio astronomy, and the 2023 overview notes that Stokes parameters can be computed using only cross-correlation products if desired, thereby avoiding the antenna shot noise present in autocorrelations (Bale et al., 2023, Camacho et al., 22 Aug 2025).

The physical payload is described as an enclosure about 1 m $10^{-3}$1 1 m $10^{-3}$2 0.7 m, including a spectrometer, power system, S-band communications hardware, and an azimuthal rotation platform. The electronics are housed in the Main Electronics Crate (MEC) inside the Inner Equipment Assembly (IEA), which combines thermal insulation with a daytime heat-rejection path (Saliwanchik et al., 2024).

4. Farside site, lunar-night operations, and power-constrained observing

The landing site is given as a 100 m ellipse centered at $10^{-3}$3, or equivalently $10^{-3}$4 latitude and $10^{-3}$5 longitude in the power study. The site was chosen to minimize terrestrial radio interference, provide a relatively flat horizon, reduce spectral chromaticity from sky occultation, support favorable thermal and communications geometry, and place the antennas over well-mixed regolith so as to reduce asymmetric dielectric effects beneath them (Bale et al., 2023, Saliwanchik et al., 2024).

The operational concept is built around the lunar night, about 14 Earth days or 328 hours. This interval is central to the mission identity. The lander must cease all operations before nightfall and remain powered off, allowing LuSEE-Night to operate standalone without lander-generated electromagnetic interference. The mission literature highlights this as a major differentiator, noting that spacecraft EMI affected earlier lunar radio experiments, including Chang’e-4 (Bale et al., 2023, Saliwanchik et al., 2024).

The enabling engineering problem is energy storage and thermal survival. The 2023 overview describes a large battery of about 40 kg; the dedicated 2024 power paper gives a more specific design point: a lithium-ion battery with nominal capacity 7160 Wh—approximately 7 kWh—and battery mass 50 kg. The battery must power the instrument throughout the 328-hour lunar night, while the battery operating temperature is maintained near 268–303 K in an external environment that can reach 100 K at night and 390 K by day (Bale et al., 2023, Saliwanchik et al., 2024).

The power architecture couples photovoltaic generation during the lunar day, battery storage, a Peak Power Tracker (PPT), a Power Distribution Unit (PDU), and a Picket Fence Power Supply (PFPS). PV panels are mounted on the top, east, and west faces. The top panel uses an azimuthally symmetric ring-like layout of four strings of 13 Coverglass Interconnected Cells (CICs), while the side panels are motivated by low-sun-angle charging near dawn and dusk. Simulations indicate that a side-panel fraction near 50% of total PV area, with 0.4–0.6 also acceptable, provides a good balance between daily energy harvest and dawn/dusk availability (Saliwanchik et al., 2024).

A central conclusion of the power study is that continuous observations through the whole lunar night are not possible with the available battery mass. The spectrometer therefore must be duty-cycled, and the autonomous Concept of Operations (ConOps) uses four normal modes—Maintenance Mode, Maintenance [Transmit] Mode, Science Mode, and Powersave Mode—plus two safe modes. During nighttime science, the transmitter is kept off to suppress self-generated RFI; during powersave intervals, the spectrometer is off and a low-power heater keeps the battery warm enough for survival and later recharge. The battery is expected to be fully recharged in about 100 hours of charging time, leaving daytime margin for communications and calibration (Saliwanchik et al., 2024).

5. Calibration strategy and the lunar subsurface as a dominant systematic

Accurate calibration is treated as a first-order scientific requirement because low-frequency global-spectrum measurements are highly sensitive to foreground spectral structure and to instrumental chromaticity. One planned calibration element is the Far-Field Calibration Source (FFCS), designated CS-4, envisioned as a transmitter on another CLPS payload, possibly on an orbiter or CubeSat, emitting a known pseudo-random waveform. The signal is specified to have flux density corresponding to $10^{-3}$6 to $10^{-3}$7, known to about 1%, over a 10 s interval at the LuSEE-Night site, and to produce a comb-like response across the LuSEE-Night band up to 51.2 MHz. The source is required to make at least 30 passes and operate for 50 Earth days. As it moves horizon to horizon, LuSEE-Night correlates against it to measure the antenna pattern, system voltage response, and frequency-dependent chromaticity (Bale et al., 2023).

The most serious calibration uncertainty identified in later work is the unknown dielectric properties of the lunar subsurface at the landing site. The 2026 study states that reflections from the lunar subsurface can change the primary beam at the 10–20% level, making the subsurface among the dominant uncertainties for precision calibration. The subsurface is modeled as a two-layer lossy dielectric with top-layer thickness $10^{-3}$8, relative permittivities $10^{-3}$9 and $27 < z < 1100$0, and loss tangent 0.01, sampled on a $27 < z < 1100$1-point parameter grid for HFSS beam simulations over 1–50 MHz (Yousuf et al., 23 Apr 2026).

A useful diagnostic introduced in that study is the fraction of beam power coupled to the ground,

$27 < z < 1100$2

The simulations show that a significant portion of the beam points into the subsurface at all frequencies, so the telescope is strongly sensitive to subsurface reflections and absorption. These effects are especially consequential near the antenna resonance, where changing the subsurface alters the resonance amplitude, position, and width, and for frequencies above about 25 MHz the resonance can cause the beam pattern to bifurcate (Yousuf et al., 23 Apr 2026).

The same study emphasizes that the subsurface changes not only the beam but also the antenna impedance, and therefore the conversion from sky temperature to measured voltage power. For this reason, antenna temperature is not the actual observable for LuSEE-Night forecasts; the relevant quantity is the observed voltage PSD. The Galactic foreground is modeled with a smooth frequency-dependent law referenced to the ULSA sky model at 25 MHz, with free parameters $27 < z < 1100$3, $27 < z < 1100$4, $27 < z < 1100$5, and $27 < z < 1100$6. Using an RBF-based emulator and pocoMC with Preconditioned Monte Carlo (PMC), the paper finds that foreground and subsurface parameters can be jointly recovered in an idealized one-night mock analysis because their spectral signatures are distinct: foreground variations are smooth and broadband, whereas subsurface effects are concentrated around the resonance. The authors explicitly describe this as a proof of concept, noting that the emulator grid is not dense enough for faithful interpolation everywhere in parameter space (Yousuf et al., 23 Apr 2026).

6. Linear map-making, deconvolution, and expected sky products

Although LuSEE-Night is not a conventional interferometer, the 2025 map-making study argues that its combination of 16 independent correlation products, modulation by the Moon’s rotation, and optional stepping of the turntable provides enough diversity to reconstruct a low-resolution sky map. Each correlation product is modeled as its own beam-weighted integral over the sky,

$27 < z < 1100$7

with $27 < z < 1100$8 the real-valued intensity beam for that product and $27 < z < 1100$9 the sky intensity (Camacho et al., 22 Aug 2025).

The map-making problem is then written as a linear inverse system,

$\sim 6$0

with a Gaussian prior on the sky map and the corresponding Wiener estimator

$\sim 6$1

This formalism treats the time-ordered correlation data as measurements coupled to the sky through the beam model and observing geometry. The paper also develops a systematic-marginalization strategy in which gain fluctuations and beam-model uncertainty are absorbed into the effective noise covariance rather than imposed as exact calibration corrections. In particular, the total covariance can be augmented by a gain term, $\sim 6$2, and beam uncertainty can be represented by $\sim 6$3 (Camacho et al., 22 Aug 2025).

The fiducial simulations use the ULSA sky model and HFSS-derived antenna beams, assume that the sky is unpolarized, and reconstruct each frequency independently up to $\sim 6$4. For one full lunar sidereal rotation sampled every 2 hours with a fixed turntable angle, the data are already sufficiently overconstrained to recover the main large-scale Galactic structure. Performance is quantified with a harmonic-space cross-correlation coefficient $\sim 6$5 and an effective signal-to-noise ratio $\sim 6$6; the results show strong correlation to about $\sim 6$7–35, depending on frequency and assumptions, corresponding to recovered features on angular scales of roughly 5–10 degrees (Camacho et al., 22 Aug 2025).

The headline result is that, under reasonable assumptions about instrument performance and calibration, LuSEE-Night should be able to map the sub-50 MHz sky at about $\sim 6$8 resolution. Longer integration and varied rotation improve intermediate angular scales, while reduced observing campaigns still detect the Galaxy but with degraded fidelity and an effective resolution closer to $\sim 6$9 (Camacho et al., 22 Aug 2025).

In that sense, LuSEE-Night combines two roles that are sometimes separated in low-frequency cosmology: it is simultaneously a radio pathfinder for the Dark Ages global-signal problem and a compact, rotating, correlation-based instrument capable of producing scientifically useful maps of the lunar-farside low-frequency sky.