Lunar Gravitational-Wave Antenna (LGWA)

• LGWA is a proposed lunar gravitational‐wave detector that uses the Moon's elastic response to capture decihertz signals.
• It implements a four-station array with cryogenic superconducting inertial sensors positioned in permanently shadowed lunar regions to optimize sensitivity.
• The concept bridges observational gaps between LISA and terrestrial observatories, advancing multiband astronomy and lunar geophysics calibration.

The Lunar Gravitational-Wave Antenna (LGWA) is a proposed lunar gravitational-wave detector that uses the Moon itself as the gravitational-wave-responsive body and an array of next-generation inertial sensors to read out the resulting surface motion. In mission studies, LGWA is positioned in the $\sim 1\,\mathrm{mHz}-1\,\mathrm{Hz}$ interval, with particular emphasis on the decihertz regime, and is presented as the missing link between space-borne detectors such as LISA and future terrestrial observatories such as Einstein Telescope and Cosmic Explorer (Ajith et al., 2024). Historically, it revives the resonant-body idea associated with Weber and the Apollo 17 Lunar Surface Gravimeter, but in a modern form based on cryogenic inertial sensing, multi-station array processing, and explicit modeling of the Moon’s elastic response to passing gravitational waves (Harms et al., 2020).

1. Concept and historical lineage

LGWA belongs to the lineage of resonant-mass gravitational-wave detectors, but it scales that concept from laboratory bars to a planetary body. The motivating observation is that the dimensions of a resonant body set its characteristic response frequencies: Weber bars targeted the kHz range, whereas the Earth or Moon can respond in the mHz regime. The Apollo 17 Lunar Surface Gravimeter was partly motivated by this logic, but a technical failure rendered its data unusable; LGWA is the contemporary re-examination of the same physical idea with vastly more sensitive instrumentation and a defined mission architecture (Harms et al., 2020).

A common misconception is to treat LGWA as a lunar analogue of a long-baseline Michelson interferometer. The mission concept is different. LGWA is not a free-falling laser interferometer like LISA, and it is not a surface-arm interferometer laid across the Moon. Rather, the Moon acts as the detector body, and the payload acts as the readout of the Moon’s GW-driven elastic motion. This viewpoint is explicit in sensor studies that describe an array of four cryogenic superconducting inertial sensors as the readout of a detector constituted by the Moon itself (Badaracco et al., 2022).

This conceptual framing determines the mission’s scientific niche. Because the Moon is the antenna, LGWA’s sensitivity is set jointly by inertial-sensor noise, the lunar seismic background, and the transfer function from incident GW strain to lunar surface displacement. That combination is what makes LGWA a proposed observatory for the $10^{-3}$–$1$ Hz band rather than merely a transplanted terrestrial instrument (Ajith et al., 2024).

2. Measurement principle and lunar response

The basic observable is lunar surface displacement induced by a passing GW. In the normal-mode formulation, the measured displacement channel may be written as

$$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$

where $f_n$, $Q_n$, and $L_n$ are, respectively, the lunar quadrupole-mode resonance frequencies, quality factors, and effective baselines (Harms et al., 2020). At low frequencies, the detector behaves as a resonant-body antenna dominated by quadrupolar elastic modes; the lower end of the LGWA band is set by the Moon’s lowest quadrupole normal mode (Ajith et al., 2024).

For surface-response calculations, later work writes the GW-induced displacement on the lunar surface as

$$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$

where $T_r(f)$ and $T_h(f)$ are radial and horizontal lunar response functions (Yan et al., 27 Apr 2026). This representation makes explicit that LGWA does not infer strain directly; instead, strain sensitivity is obtained by dividing inertial displacement noise by the lunar response. In mission studies, the dominant calibration uncertainty is therefore not the local sensor transfer function but the lunar GW transfer function itself: below $10^{-3}$0 Hz the limiting uncertainty is the deep lunar interior model, whereas above $10^{-3}$1 Hz it is local geology and topography (Ajith et al., 2024).

The choice of horizontal readout is likewise a consequence of the response model. Mission studies state that each station is planned to host two orthogonal horizontal Lunar inertial GW sensors because the instrumental noise is more favorable for horizontal sensing and the lunar GW response is expected to be stronger in the horizontal direction (Ajith et al., 2024). That choice introduces a geophysical penalty—horizontal channels are sensitive to ground tilt—but it is central to the nominal LGWA design.

The response above $10^{-3}$2 Hz is a particularly active area of modeling. An extension of Dyson’s half-space model to horizontally layered geology shows that the decihertz response can depend strongly on shallow structure, low near-surface shear speed, and deeper shear-wave return mechanisms. In that framework, modifications of the geological model required to explain Apollo seismic observations can boost the lunar GW response, with low-$10^{-3}$3 shallow layers and strong S-wave reflection amplifying the horizontal response by factors $10^{-3}$4 and in some cases up to $10^{-3}$5 relative to a homogeneous half-space (Bi et al., 2024). This implies that LGWA sensitivity forecasts are intrinsically geology-dependent and, plausibly, site-dependent.

3. Array architecture and sensor technology

The current mission concept is a four-station array deployed in a permanently shadowed region near a lunar pole, with each station carrying two orthogonal horizontal channels, for a total of eight horizontal seismic channels (Ajith et al., 2024). In the default implementation used in GWFish studies, the detector is modeled as an array of four stations in permanently shadowed regions, each station equipped with two orthogonal horizontal Lunar inertial GW sensors and operated over a 10-year mission duration (Song et al., 5 Feb 2025). Permanently shadowed regions are central to the concept because they offer low and stable temperatures; mission and sensor studies repeatedly cite values below $10^{-3}$6 K as a key environmental advantage for cryogenic operation (Badaracco et al., 2022).

The enabling payload technology is a cryogenic low-frequency inertial sensor. A dedicated sensor study presents a Cryogenic Superconducting Inertial Sensor based on a $10^{-3}$7 proof mass suspended in a Watt’s linkage configuration, implemented in a monolithic niobium mechanical structure fabricated by electrical discharge machining. The same study specifies niobium’s superconducting transition temperature $10^{-3}$8, a modeled operating temperature $10^{-3}$9, and an expected mechanical quality factor $1$0. The quoted displacement sensitivity is $1$1 at $1$2, with a few $1$3 from $1$4 onward, approximately a factor $1$5 better than the most sensitive existing seismic sensors (Badaracco et al., 2022).

That sensor concept combines cryogenic operation, superconducting Meissner-effect actuation, and dual readout. The interferometric channel uses a simple Michelson configuration with an additional beamsplitter to form a differential signal, while the Rasnik channel provides larger dynamic range and damping of the proof-mass resonance. The paper reports a demonstrated Rasnik sensitivity of $1$6, with sub-pm performance expected after improvements (Badaracco et al., 2022). Mission studies also discuss an alternative silicon Watt’s-linkage concept expected to improve the inertial displacement sensitivity by about a factor 10 relative to the niobium baseline, with a corresponding factor $1$7 improvement in GW strain sensitivity (Ajith et al., 2024).

The architecture is explicitly array-based because LGWA must distinguish the coherent lunar GW response from the lunar seismic background. A pathfinder mission, Soundcheck, is proposed to measure the seismic environment, temperature stability, magnetic fluctuations, and deployment systematics of a permanently shadowed site before committing to the full observatory (Ajith et al., 2024).

4. Sensitivity band and source populations

The defining observational band is $1$8, with the scientific emphasis in the decihertz interval where neither LISA nor terrestrial detectors are expected to operate optimally (Ajith et al., 2024). In earlier mission studies, representative detection ranges were quoted as $1$9 for a $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$0 binary neutron star and $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$1 for a $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$2 binary black hole, and simulations suggested that 28 of the 80 GWTC-1/2/3 sources would have been observed with LGWA (Ajith et al., 2024).

More recent BBH forecasts sharpen the multiband case. One study of catalogued and simulated BBHs finds that LGWA alone would have observed $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$3 of the 176 public LVK events up to GWTC-4.0 at $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$4, and $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$5 at $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$6. For simulated populations consistent with the latest LVK reconstruction, the same work forecasts $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$7 detections per year at $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$8 out to $$\xi(f) = \frac{1}{2}h(f)\sum\limits_{n=0}^\infty L_n\frac{-f^2}{f_n^2-f^2+ i f_n^2/Q_n},$$9, and $f_n$0 at $f_n$1 out to $f_n$2. It further argues that third-generation detectors would observe most of the BBHs detected by LGWA in the simulated mass range $f_n$3, making systematic joint analyses feasible (Iacovelli et al., 10 Dec 2025).

Intermediate-mass black holes are a flagship LGWA science case because their inspiral predominantly occupies the decihertz band. Population studies comparing LGWA and ET find that ET is more suited to binaries with primary masses below $f_n$4, whereas LGWA is especially effective over roughly $f_n$5–$f_n$6 and can reach beyond $f_n$7. At $f_n$8, the LGWA detection rate for IMBH binaries with primary masses above $f_n$9 is reported to be largely insensitive to orbital inclination and mass ratio, whereas the joint LGWA+ET network achieves nearly complete coverage across the IMBH parameter space considered (Dong et al., 14 Jul 2025).

Double white dwarfs constitute a different and more source-specific motivation. In the decihertz band, short-period DWDs approach merger, including systems relevant to Type Ia supernova progenitors. A population study based on SeBa forecasts that LGWA could detect approximately $Q_n$0 monochromatic Galactic DWDs and $Q_n$1 extragalactic mergers over a 10-year mission in the contact scenario, with host-galaxy identification usually possible for detected extragalactic events because the inferred source volumes are generally below galaxy-confusion limits (Benetti et al., 9 Sep 2025). Mission studies also single out WD tidal disruptions by $Q_n$2–$Q_n$3 black holes, light IMRIs, and lower-mass MBH binaries as especially well matched to the decihertz window (Ajith et al., 2024).

The quantitative reach, however, is model-dependent. A dedicated IMBH study notes that under the original LGWA PSD used in its main analysis, favorable IMBH binaries are detectable out to $Q_n$4, but with an updated, less optimistic PSD the horizon can shrink to $Q_n$5 (Song et al., 5 Feb 2025).

5. Multiband, multimessenger, and cosmological roles

LGWA is repeatedly framed not only as a detector of new sources, but as an observatory that changes the temporal logic of multimessenger astronomy. A 2040s white paper argues that access to the $Q_n$6 band enables continuous observation of compact binaries over months to years prior to coalescence, with days-to-months warning times for neutron-star and black-hole systems. In that picture, LGWA shifts electromagnetic follow-up from reactive to predictive mode, enabling scheduled pre-merger observations, deep contemporaneous reference imaging, minute-to-hour cadence campaigns starting at merger time, targeted spectroscopy, and pre-selection of likely host galaxies. The same paper emphasizes that fully exploiting such warnings would require new optical/NIR infrastructure capable of handling hundreds to thousands of transient candidates per event (Patat et al., 18 Dec 2025).

The multiband role extends beyond early warning. For IMBH binaries, LGWA and ET are described as complementary because LGWA captures the earlier inspiral SNR in the decihertz band while ET contributes the high-frequency late inspiral and merger for lower-mass systems (Dong et al., 14 Jul 2025). In dark-siren cosmology, a three-band network composed of Taiji, LGWA, and ET is forecast to outperform all two-detector configurations considered. Under the study’s baseline assumptions, the LGWA-inclusive three-band network constrains $Q_n$7 and $Q_n$8 to $Q_n$9 and $L_n$0, respectively, in flat $L_n$1CDM from a 4-year dark-siren sample, while in flat $L_n$2CDM it constrains $L_n$3 to $L_n$4 from GW data alone and $L_n$5 after adding BAO and SNe Ia (Song et al., 13 Mar 2026). Those results are explicitly sensitive to IMBHB population assumptions and galaxy-catalog completeness.

LGWA has also been proposed as the decisive decihertz element in tests of frequency-dependent GW propagation. In an effective-field-theory modified-gravity model with fiducial transition frequency $L_n$6, the decihertz band is the location of the turnover in $L_n$7. Multi-band observations involving LGWA, CE or ET, and optionally LISA are then forecast to measure the effective-theory scale $L_n$8 with $L_n$9 precision for CE-ET+LGWA and $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$0 precision for CE-ET+LGWA+LISA using only $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$1 high-SNR events (Praveen et al., 30 Apr 2025). In this sense, the scientific role of LGWA is not only astrophysical coverage, but access to a propagation regime that neither mHz nor ground-based detectors probe directly.

6. Uncertainties, lunar geophysics, and mission development

The principal technical uncertainty in LGWA is that the detector response is inseparable from the Moon’s geophysics. Mission studies state explicitly that the main calibration error is expected to come from the lunar response model rather than from the local sensor calibration (Ajith et al., 2024). Above $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$2 Hz, the layered-response analysis indicates that sensitivity may depend strongly on local stratigraphy and megaregolith structure; below that, normal-mode calculations depend on the deep lunar interior and mode $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$3-values (Bi et al., 2024).

The second major uncertainty is the seismic background and its subtraction. LGWA was conceived as an array because data processing should distinguish the GW signal from the lunar seismic background, but until recently that claim lacked a quantitative benchmark. A two-station analytic study in an isotropic, random, Gaussian seismic field finds that the equivalent seismic ASD can be improved by approximately a factor of $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$4 at $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$5 relative to a single station, with the gain strongly controlled by station spacing relative to the seismic correlation length. The same work argues that suppression of an isotropic seismic background is generally constrained to within a factor 3 for the two-station case, and that the oscillatory residual structure is a direct consequence of the Bessel-function correlation of diffuse Rayleigh noise (Yan et al., 27 Apr 2026). This does not invalidate the array concept, but it constrains expectations for how much raw seismic noise can be mitigated by geometry alone.

Forecast sensitivity is therefore model-sensitive on both the signal and noise sides. The IMBH study already noted that replacing the optimistic baseline PSD with an updated lunar-response-motivated curve can reduce the horizon from $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$6 to $$\vec{\xi}(\hat r) = 2 T_h \mathbf{h} \cdot \hat r + (T_r - 2T_h) (\hat r \cdot \mathbf{h} \cdot \hat r) \hat r,$$7 (Song et al., 5 Feb 2025). Conversely, the layered-response study argues that Apollo-motivated modifications of the geological model can boost the decihertz response substantially (Bi et al., 2024). A plausible implication is that the present spread in LGWA science forecasts reflects real uncertainty in the coupled problem of lunar response, site geology, and environmental noise, rather than merely differences in data-analysis choices.

Because of this coupling, lunar science is not ancillary to LGWA but constitutive of it. Mission studies emphasize that LGWA data could detect very weak moonquakes, meteoritic hum, and lunar normal modes, constrain crust, mantle, core, and megaregolith structure, and characterize the seismicity of permanently shadowed regions, all of which feed back into GW calibration and site optimization (Ajith et al., 2024). Data analysis itself has also been shown to be unusually geometry-driven: for long-duration decihertz signals, using an SSB-comoving frame with an optimally shifted origin can reduce sampling time by roughly an order of magnitude, making parameter estimation for LGWA a problem in detector motion, timing conventions, and signal evolution as much as in raw SNR (Tissino et al., 3 Jun 2026).

In that sense, LGWA is best understood as a coupled observatory concept at the boundary of gravitational-wave astronomy, cryogenic inertial sensing, and lunar geophysics. Its promise lies in opening a band that is otherwise largely inaccessible; its difficulty lies in the fact that the detector is the Moon.