A high-precision lunar soil water content estimation method designed for the Chang’E-7 Lunar Soil Water Molecule Analyzer (LSWMA)

Abstract

We present a high-precision in-situ inversion method for lunar water ice content, addressing the critical challenge of water vapor loss during dynamic exhaust operations in the Lunar Soil Water Molecule Analyzer (LSWMA) on Chang’E-7 mission. To overcome the partial water vapor loss caused by the instrument’s pressure-triggered exhaust system (1000 Pa to 400 Pa), we established a quantitative relationship between chamber pressure dynamics, cumulative exhausted gas volume, and initial water content through ground-based simulation experiments. This relationship enabled derivation of an inversion equation suitable for in-situ processing. Validation with simulated lunar regolith (1 wt%−5 wt% water content) demonstrated an inversion error within ±20%, a significant improvement over the original ±50% specification. This method ensures accurate water content assessment for lunar In-Situ Resource Utilization (ISRU) applications.

Graphical abstract

Keywords

Lunar water ice / Lunar Soil Water Molecule Analyzer (LSWMA) / Chang’E-7 mission / In-Situ Resource Utilization (ISRU)

1 Introduction

1.1 The significance of lunar water exploration

Water plays a unique and critical role in planetary evolution. The origin and abundance of lunar water provide key constraints on the Moon’s evolution history (Arnold, 1979; Siegler et al., 2015, 2016), with profound implications for understanding the evolution of the Earth and the entire solar system. Of particular importance is that lunar water resources represent a core strategic asset for future deep-space exploration and sustainable lunar operations. As China advances its plans for crewed lunar landing and the establishment of a lunar research station, In-Situ Resource Utilization (ISRU) has become a pivotal technological pathway to enable long-term habitation, reduce mission costs, and enhance operational autonomy (Hurley et al., 2023; Sanders et al., 2024).

1.2 The history of lunar water exploration

In recent years, multi-source remote sensing and laboratory analyses have provided robust evidence for the presence of water (H₂O) and hydroxyl (OH) on the lunar surface. Since the Clementine mission in 1996 first detected signals suggestive of water ice in the lunar polar regions (Nozette et al., 1996), the poles, particularly the Permanently Shadowed Regions (PSRs)—have been regarded as ideal reservoirs for water ice due to their extremely low temperatures (< 110 K) (Watson et al., 1961; Schorghofer and Taylor, 2007). Neutron spectrometry data from the Lunar Exploration Neutron Detector (LEND) further confirmed the presence of abundant hydrogen in these PSRs (Sanin et al., 2012). Additionally, spectral data from the Moon Mineralogy Mapper (M³) aboard Chandrayaan-1, along with observations at other wavelengths, have revealed multiple signatures consistent with surface-bound water within the permanently shadowed regions (Pieters et al., 2009; Gladstone et al., 2012; Li and Milliken, 2016).

In 2008, Saal et al. (2008) first discovered indigenous lunar water within glass beads from Apollo samples. The most direct evidence of polar volatiles came from the Lunar Crater Observation and Sensing Satellite (LCROSS) mission, which detected water ice molecules in the ejecta plume generated by its impactor (Colaprete et al., 2010). Furthermore, both Chang’E-5 and Chang’E-6 missions have identified the presence of hydroxyl in samples collected from the Moon’s nearside and farside, respectively (Zhou et al., 2022; Liu et al., 2025).

Despite these advances, in-situ quantitative measurements of water ice abundance on the lunar surface remain lacking a critical bottleneck hindering the practical implementation of lunar ISRU engineering.

1.3 Chang’E-7 Misson

The Chang’E-7 mission, scheduled for launch in 2026, will conduct in-situ measurements of water-ice content in the shallow subsurface at the landing site and within PSRs for the first time (Wang et al., 2024).

According to neutron data from LEND, the average water ice abundance in the polar regions is approximately 0.5 wt% (Sanin et al., 2017). Meanwhile, Thomson et al. used Miniature Radio-Frequency instrument (Mini-RF) radar data and scattering models to estimate that the upper limit of water ice concentration near Shackleton Crater, the candidate landing site for Chang’E-7 (Mo et al., 2025), could reach 5 wt% to 10 wt% (Thomson et al., 2012).

These measurements will provide new evidence to deepen our understanding of the Moon’s evolutionary history and establish the foundation for lunar base development and in-situ utilization of lunar water resources.

1.4 Instrument overview and technical challenges

1.4.1 Lunar Soil Water Molecule Analyzer (LSWMA)

To unravel the mystery of lunar surface water ice distribution and concentration, the Chang’E-7 mission will carry a LSWMA onboard a mini-flying probe, which will for the first time enter the PSRs of the lunar south pole to conduct in-situ detection of water ice in the upper meter of regolith and provide quantitative assessments of its abundance (Wang et al., 2024). LSWMA integrates four components: a Differential Absorption Spectrometer (DAS), a Lunar Soil Heating Module (LSHM), a Tunable Diode Laser absorption Spectrometer (TDLAS), and a Time-of-Flight Mass Spectrometer (ToF-MS) (Li et al., 2025).

1.4.2 Operating and testing principles

The water content measurement task is entirely performed by TDLAS. The TDLAS quantifies H₂O using a tunable distributed feedback laser at 2.73 μm to avoid interference from H₂S, CO₂, and CH₄.

Firstly, the DAS illuminate PSRs from 30−80 cm distance and measuring reflectance at 1.272, 1.512, 1.854, and 2.004 μm to determine whether water ice is present. Subsequently, the drill will excavate downward (up to 1 m) to collect samples. These samples will then be heated in the LSHM to extract water vapor, which will be analyzed by the TDLAS to determine water content through integrated absorbance calculations.

1.4.3 Limitations and challenges

The instrument’s design detection limit is set at 0.1 wt% (Li et al., 2025), and ground calibration experiments have demonstrated that its actual sensitivity exceeds this design specification (Chen et al., 2024).

However, covering a wide range of water contents presents a significant challenge: during measurements of high-water-content samples, rapid buildup of water vapor pressure inside the spectroscopic chamber can cause spectral signal saturation. To extend the upper measurement limit, the LSWMA incorporates an innovative dynamic exhausting system: when the chamber pressure reaches 1000 Pa, the control system triggers a miniature solenoid valve to open, rapidly reducing the pressure to below 400 Pa. This “exhausting - measurement” cycle can be repeated multiple times, effectively preventing sensor saturation.

It should be noted that the amount of gas exhaust during each exhaust event directly impacts the accuracy of the final water content inversion. To address this critical source of error, this study constructed a ground-based simulation platform replicating the actual instrument geometry and conducted a series of controlled experiments. These experiments established a quantitative relationship among chamber pressure dynamics, cumulative exhausted gas volume, and initial water content, enabling the derivation of an inversion equation tailored for in-situ data processing. This work not only provides the core data processing methodology for the Chang’E-7 mission but also lays essential experimental and theoretical foundations for future in-situ exploration, assessment, and utilization of lunar water ice resources, directly supporting humanity’s deep-space exploration ambitions and lunar base construction strategies.

2 Methods

2.1 Dynamic exhaust unit

The primary components involved in the dynamic exhausting unit are the spectrometer chamber, a solenoid valve, and the dynamic exhausting unit (Fig. 1).The other section contains the LSHM introduced in Section 1.4.1. In exhausting experiment, we’d like to skip the LSHM extraction process and directly inject water vapor into the spectrum using an autosampler. The dynamic exhausting unit is connected downstream of a custom-designed solenoid valve and features a bowl-shaped flow restrictor with a 1 mm diameter orifice. The solenoid valve is controlled by a pressure sensor and, upon receiving a trigger signal, rapidly reduces the pressure inside the sealed spectrometer chamber from 1000 Pa to below 400 Pa. However, during actual measurements, gas continues to be generated from the sample upstream, making both the venting duration and the total vented gas volume uncertain. Therefore, it is urgently necessary to establish an accurate mapping relationship between the dynamic pressure evolution within the chamber and the cumulative vented gas volume to enable precise inversion for high-water-content samples.

2.2 Exhausting experiment

To determine the relationship between pressure and exhaust gas volume, we isolated the spectroscopic chamber and directly connected it to the exhaust system via a tube (Fig. 2(a)). Notably, the tube’s internal volume matches that of the actual flight component; only non-essential peripheral components were removed (compared with Fig. 1). To precisely control the amount of water injected, we employed a PAL LSI Series autosampler (Fig. 2(b)), calibrated to deliver an injection corresponding to approximately 1000 Pa of water vapor pressure. Both the tubing and the spectroscopic chamber were wrapped with heating tapes for temperature stabilization, achieving a control precision of ± 0.5°C.

The exhaust outlet was connected to a vacuum tank maintained at a base pressure below 10⁻⁵ Pa to simulate the high-vacuum environment of the lunar surface. Owing to the vacuum tank’s volume being much larger than that of the spectroscopic chamber, each experiment caused negligible disturbance to the tank’s vacuum level, enabling repeated consecutive tests.

Chamber pressure was monitored using a Pfeiffer CCR362 vacuum gauge, which offers a reading accuracy of 0.2% at room temperature and a sampling frequency of 10 Hz. The gauge has been cross-calibrated by our institution and other payload development teams, resulting in an injection-related pressure uncertainty of less than 10 Pa. Within the target measurement range (< 1000 Pa), the gauge is capable of measuring water vapor pressure; after calibration, its measurement error is less than 4 Pa.

Considering that, in actual operations, the spectroscopic chamber must be fully exhausted between successive sample analyses, our data acquisition covered the full pressure decay from 1000 Pa down to 50 Pa. The segment from 400 to 1000 Pa will be used for high-water-content inversion, while data below 400 Pa will support determination of inter-experiment intervals and studies of surface adsorption/desorption processes.

The instrument is designed to measure at a temperature of 35°C. Therefore, we selected three cavity temperature values 25°C, 35°C, and 45°C for testing also to evaluate the impact of different operating temperatures on the instrument.

Before the experiment began, the entire system was leak-checked to ensure good sealing. The gas inlet valve and the solenoid valve were opened, and the spectroscopic chamber along with the surrounding tubing was maintained at 25°C. The vacuum tank, spectroscopic chamber, and associated tubing were pumped down until the overall pressure fell below 10⁻⁵ Pa. During pumping, the autosampler, vacuum gauge, and spectrometer were powered on and allowed to warm up.

Afterward, all valves were closed, and the status of the measurement instruments was verified. Once all conditions met the experimental requirements, the autosampler was triggered, and the gas inlet valve was simultaneously opened. Upon completion of sample injection, the inlet valve was closed. After the pressure inside the spectroscopic chamber stabilized (allowing time for adsorption effects to subside), the solenoid valve was activated, and the exhaust duration together with the corresponding pressure data were recorded.

The above procedure was repeated at chamber temperatures of 35°C and 45°C, respectively.

Experimental conditions are shown in the Table 1.

The experimental workflow is illustrated in Fig. 3.

2.3 Full-process experimental validation

Through the experiments described in Section 2.2, the relationship between chamber pressure and exhaust gas volume per unit time can be established, along with an understanding of the influence of temperature on this relationship. However, the empirical equation derived from these results requires further validation. Therefore, we used data from full-process experiments to invert the initial water content. During the full process experiment, the instrument was maintained at the designed temperature of 35°C.

Using the engineering prototype of the payload, we tested its capability to detect and measure high-water-content samples with simulated lunar soil containing 1 wt% to 5 wt% water.

The preparation of the simulated lunar soil with water involves three main stages. First, the simulant is placed in a vacuum oven and dried at 205°C for over 48 hours until it reaches a constant weight, after which it is cooled to room temperature. Next, 50 g dried simulant is weighed, and deionized water is added incrementally in small batches; The mixture is thoroughly stirred until the surface color appears uniform, and the total weight of this high-humidity mixture is recorded (corresponding to the total mass of water and simulant). This mixture is then transferred into a glove box, where excess moisture is evaporated by controlling the heating duration to achieve the target water content. Finally, the moisture content is verified using a Karl Fischer Moisture Meter (Model MA-30) via the multi-point sampling method, with the average value taken from three repeated measurements. Seal the prepared samples and store them in a constant-temperature freezer (−80°C). Upon activation, simultaneously take a portion of the sample and measure its moisture content three times for verification. The actual water contents of the prepared samples were measured as 1.3 wt%, 1.68 wt%, 3.795 wt%, 3.942 wt%, and 4.006 wt%, respectively.

The simulated soil underwent the same sequence as in-situ lunar surface operations: sample injection and transfer, thermal extraction, and dynamic exhaust. Each water content level was tested twice. The empirical inversion equation developed in this study was then applied to estimate the initial water content, and the resulting errors were calculated and compared against the instrument’s design specification.

Full-process experimental conditions are shown in the Table 2.

Other potential error sources—such as sublimation loss during sample handling and incomplete efficiency of thermal extraction—were previously characterized through extensive calibration experiments, which collectively indicated an average water loss of approximately 20%. This loss was corrected for in the inversion calculations to realistically evaluate the applicability of the empirical equation under operational conditions.

3 Results

First, the experimental results in Fig. 4 show that the overall gas exhaust rates at the three temperatures are nearly identical. This is because the relative change in absolute (Kelvin) temperature is actually very small. In the higher-pressure range (400–1000 Pa), pressure dominates the exhaust dynamics, and the data across different temperatures exhibit minimal differences, indicating that temperature fluctuations within ± 10°C under real operational conditions will not significantly affect the final water content inversion.

Below 400 Pa, however, the influence of temperature becomes increasingly evident: the higher the temperature, the faster the pressure drops in the chamber.

Given that the tubing in the full-process experiment was temperature-controlled at 35°C, we selected the data acquired at 35°C for further analysis. We performed a curve fit on the pressure and pressure-drop data, and the results are shown in Fig. 5.

At a sampling frequency of f = 10 Hz, the relationship between exhaust rate (represented by pressure drop on the vertical axis) and chamber pressure (on the horizontal axis) follows a quadratic function trend. The adjusted coefficient of determination (R²) is approximately 0.97, indicating a good fit. This quadratic function can be used to accurately estimate the pressure drop at any given pressure during the measurement process:

(1)dp=-8.5{\times 10}^{-6}{p}^2-4.8\times {10}^{-3}p+0.19.

By performing numerical integration of the above equation, the total pressure drop in actual testing can be obtained.

Using the data from the full-process experiments and taking the average of the two experiments for samples with the same water content, we generated a comparison chart (Fig. 6) showing the inverted water content versus the true water content.

From the chart, it can be seen that the trend of inverted water content matches perfectly with the true water content. The mean absolute error is less than 0.5 units, and the relative errors are as shown in Table 3.

The experimental results demonstrate that the proposed inversion algorithm achieves a mean absolute error (MAE) of 12.8% across all test cases (Table. 3), consistently outperforming the instrument’s design specification of ±50%. Notably, the error is confined to within 10% for water contents greater than 3.5%. This improved accuracy directly supports in-situ payload operations by enabling real-time quantification of lunar water ice resources with minimal calibration requirements.

4 Discussions

Because the relative change in absolute (Kelvin) temperature is actually very small. The curves of exhaust pressure drop between 25°C and 45°C show little to no variation. In below 400 Pa pressure regime, gas flow transitions toward molecular flow, where motion is strongly influenced by molecular thermal velocity (i.e., Maxwell–Boltzmann speed distribution). This observation confirms that the exhaust process exhibits a transitional behavior between continuum flow and molecular flow.

The dynamic exhaust–based inversion method proposed in this study effectively extends the applicability of spectroscopic and mass spectrometric payloads in high-volatiles environments. Traditional closed-chamber designs are prone to signal saturation when analyzing samples with water content exceeding 1 wt%. In contrast, our approach employs a closed-loop control strategy, “pressure-threshold triggering to rapid exhaust to repeated measurement”, which significantly raises the upper measurement limit without compromising sensitivity. This concept is readily adaptable to in situ detection of other volatiles (e.g., CO₂, CH₄), laying a technical foundation for future multifunctional platforms dedicated to in-situ resource analysis and utilization in deep space exploration.

Nevertheless, certain limitations of the current inversion model should be acknowledged. First, experiments were conducted using pure water vapor injection, whereas real lunar regolith may release a mixture of coexisting volatiles during thermal extraction; the fluid dynamic behavior of such gas mixtures differs from that of pure water vapor and could introduce systematic biases. Second, the model assumes instantaneous adsorption–desorption equilibrium during exhaust, while actual measurements exhibit minor drifts under non-equilibrium conditions. Third, the underlying physical mechanisms of the exhaust process require further theoretical and experimental investigation to ensure robustness across diverse payload architectures.

Moreover, data from the analyzer should ultimately be integrated into a multi-scale, multi-method validation framework, combining orbital remote sensing (e.g., Mini-RF, LEND), impact probing (e.g., LCROSS), and returned-sample analyses (e.g., Chang’E-5/6), to cross-validate water abundance estimates.

Looking ahead, as lunar research station development progresses, the requirement for in-situ water analyzers will shift from “detection capability” to “mining guidance”. Next-generation instruments could incorporate features such as programmable temperature-gradient heating and enhanced thermal extraction efficiency to directly assess the feasibility of water ice mining and utilization. The dynamic exhaust and quantitative inversion framework established in this work provides a scalable technical pathway toward these advanced functionalities.

In summary, this study not only resolves a critical technical bottleneck for the Chang’E-7 mission but also introduces a novel methodological paradigm for in-situ volatile detection.

5 Conclusions

This paper addresses the signal saturation challenge faced by the Chang’E-7 LSWMA when measuring high-water-content samples, proposing and validating a water content inversion method based on dynamic exhaust conductance experiments. By establishing a quantitative relationship among chamber pressure, exhaust gas volume, and initial water content, we have successfully achieved high-precision inversion of water ice concentration in simulated lunar regolith with water contents ranging from 1 wt% to 5 wt%, demonstrating well-controlled errors that meet mission design requirements.

This achievement not only ensures the scientific return of the analyzer payload but also provides a reliable technical tool for future in-situ exploration and utilization of water ice resources in the lunar south pole. Accurate in-situ water content measurements are essential for evaluating the mine ability of lunar regolith and optimizing key ISRU process parameters, such as heating temperature, extraction duration, and energy budget, thereby offering direct engineering value in supporting lunar research base construction and advancing sustainable extraterrestrial habitation.

References

[1] Arnold J R (1979). Ice in the lunar polar regions.J Geophys Res, 84(B10): 5659–5668

[2] Chen R, Lu W, Li X, Wang X, Chen K, Zhang C, Cao N, Xu Z, Xu Y, Kan R (2024). Chang’E-7 lunar soil water molecule analyzer (LSWMA) prototype for high-precision measurement of water content and hydrogen isotope ratio.J Earth Sci, 35(6): 2180–2182

[3] Colaprete A, Schultz P, Heldmann J, Wooden D, Shirley M, Ennico K, Hermalyn B, Marshall W, Ricco A, Elphic R C, Goldstein D, Summy D, Bart G D, Asphaug E, Korycansky D, Landis D, Sollitt L (2010). Detection of water in the LCROSS ejecta plume.Science, 330(6003): 463–468

[4] Gladstone G R, Kurt D R, Anthony F E, David E K, Paul F M, Joel W P, David H, Paul M R, Maarten H V, Michael W D (2012). Far-ultraviolet reflectance properties of the Moon’s permanently shadowed regions.J Geophys Res Planets, 117(E12): E00H04

[5] Hurley D M, Matthew A S, Joshua TS C, Anthony C, Emily C, Ariel N D, Richard C E, Fa W Z, Cesare G, Paul O H (2023). Surface volatiles on the Moon, Rev Mineral Geochem, 89: 787–827

[6] Li S, Milliken R E (2016). An empirical thermal correction model for Moon Mineralogy Mapper data constrained by laboratory spectra and Diviner temperatures.J Geophys Res Planets, 121(10): 2081–2107

[7] Li , Xiang , Cao N L, Kan R F, Xu Z Y, Zhang X P, Wang Q, Lu W Z, Chen R, Wang X P, Deng H (2025). Miniaturized Tunable Laser Spectrometer for the Simultaneous Detection of Water Ice and Hydrogen–Oxygen Isotopes for the Chang’E-7 Lunar Soil Water Molecule Analyzer.ACS sensors, 10(6): 4679–4686

[8] Liu B, Zeng X, Xu R, Li S, Liu J, Guo L, Ren X, Liu D, Yan W, Chen W, Li J, He Z, Li C (2025). Lunar surface and subsurface water revealed by Chang’E-6.Nat Astron, 9(12): 1–9

[9] Mo , L J, Xu J W, Dong Z H, Li J W, Su Y, Ding C Y (2025). Water ice and the shallow regolith structure of the shackleton crater, the Moon: implications for future Chang’E-7 in situ radar observation.Space: Sci Technol, 5: 0225

[10] Nozette S, Lichtenberg C L, Spudis P, Bonner R, Ort W, Malaret E, Robinson M, Shoemaker E M (1996). The Clementine bistatic radar experiment.Science, 274(5292): 1495–1498

[11] Pieters C M, Goswami J N, Clark R N, Annadurai M, Boardman J, Buratti B, Combe J P, Dyar M D, Green R, Head J W, Hibbitts C, Hicks M, Isaacson P, Klima R, Kramer G, Kumar S, Livo E, Lundeen S, Malaret E, McCord T, Mustard J, Nettles J, Petro N, Runyon C, Staid M, Sunshine J, Taylor L A, Tompkins S, Varanasi P (2009). Character and spatial distribution of OH/H2O on the surface of the Moon seen by M3 on Chandrayaan-1.Science, 326(5952): 568–572

[12] Saal A E, Hauri E H, Cascio M L, Van Orman J A, Rutherford M C, Cooper R F (2008). Volatile content of lunar volcanic glasses and the presence of water in the Moon’s interior.Nature, 454(7201): 192–195

[13] Sanders G B Julie E K, Mark W H (20242024. Using ISRU and Surface Construction to Define Long-Term Lunar Infrastructure Needs. In AIAA Aviation Forum and Ascend 2024: 4840

[14] Sanin A B, Mitrofanov I G, Litvak M L, Bakhtin B N, Bodnarik J G, William V B, Chin G, Evans L G, Harshman K, Fedosov F (2017). Hydrogen distribution in the lunar polar regions.Icarus, 283: 20–30

[15] Sanin A B, Mitrofanov I G, Litvak M L, Malakhov A, Boynton W V, Chin G, Droege G, Evans L G, Garvin J, Golovin D V (2012). Testing lunar permanently shadowed regions for water ice: LEND results from LRO.J Geophys Res Planets, 117(E12): E00H26

[16] Schorghofer N, Taylor J G (2007). Subsurface migration of H₂O at lunar cold traps. J Geophys Res Planets, 112(E20): 117(E12): E02010

[17] Siegler M A, Miller R S, Keane J T, Laneuville M, Paige D A, Matsuyama I, Lawrence D J, Crotts A, Poston M J (2016). Lunar true polar wander inferred from polar hydrogen.Nature, 531(7595): 480–484

[18] Siegler M, Paige D, Williams J P, Bills B (2015). Evolution of lunar polar ice stability.Icarus, 255: 78–87

[19] Thomson B J, Bussey D B J, Neish C D, Cahill J T S, Heggy E, Kirk R L, Patterson G W, Raney R K, Spudis P D, Thompson T W, Ustinov E A (2012). An upper limit for ice in Shackleton crater as revealed by LRO Mini-RF orbital radar.Geophys Res Lett, 39(14): 39

[20] Wang C, Jia Y, Xue C, Lin Y, Liu J, Fu X, Xu L, Huang Y, Zhao Y, Xu Y, Gao R, Wei Y, Tang Y, Yu D, Zou Y (2024). Scientific objectives and payload configuration of the Chang’E-7 mission.Natl Sci Rev, 11(2): nwad329

[21] Watson K, Murray B C, Brown H (1961). The behavior of volatiles on the lunar surface.J Geophys Res, 66(9): 3033–3045

[22] Zhou C, Tang H, Li X, Zeng X, Mo B, Yu W, Wu Y, Zeng X, Liu J, Wen Y (2022). Chang’E-5 samples reveal high water content in lunar minerals.Nat Commun, 13(1): 5336

Rights & permissions

The Author(s). Published by Higher Education Press. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0)