1 Introduction
Mars, as a primary target for human deep space exploration, poses significant logistical support challenges for crewed missions. Carrying all mission-required supplies from Earth is not only costly but also significantly increases launch mass and mission risks. The concept of
In-Situ Resource Utilization (ISRU) has emerged, aiming to substantially reduce the mission’s reliance on Earth-based resupply by utilizing local Martian resources to produce essential items such as life support supplies and propellants (
Sanders and Larson, 2013). Every unit of mass saved on the Martian surface is equivalent to saving 8−10 units of launch mass in low Earth orbit. This “amplification factor” effect makes ISRU an inevitable choice for crewed Mars exploration (
Sanders and Duke, 2005).
Water is one of the most valuable
in-situ resources on Mars. It is not only a necessity for astronauts’ survival, but can also produce oxygen through electrolysis, or combine with carbon dioxide in the atmosphere through the Sabatier reaction to generate methane/oxygen propellants, providing power for Mars ascent vehicles (
Rapp and Inglezakis, 2024). Although the concept of Mars ISRU dates back to the 1970s, early research progress was relatively slow due to low technological maturity. In recent years, with the accumulation of a large amount of evidence on water ice distribution from detection missions such as MRO, Phoenix, Curiosity, and Perseverance, research on Mars water ice ISRU has regained great attention (
Hoffman et al., 2017;
Yu et al., 2025).
Currently, the core challenge in the research on Martian water ice has shifted from simply “detection and discovery” to the accurate assessment of “extractability”. This requires us not only to grasp the distribution patterns of water ice but also to conduct in-depth research on its occurrence forms and the impact of the characteristics of ice-containing media on engineering extraction. This review focuses on the occurrence characteristics of Martian water ice and its extraction technologies. First, this paper will elaborate on the occurrence forms of Martian water ice in detail according to the latitudinal zoning law, covering mineral-bound water in low latitudes, buried glaciers in mid-latitudes, and polar cap systems in high latitudes, with a focus on evaluating the resource potential of each region. Secondly, this paper will summarize the physical and chemical properties of ice-containing media and clarify their background impact in the extraction process. Finally, this paper will systematically evaluate the different technical routes for extracting the occurrence state, and a comparative analysis of the technical logic and application prospects of schemes such as mechanical excavation, thermal sublimation, and in-situ melting (e.g., the Rodwell system). Through the above multi-dimensional comprehensive evaluation, it aims to provide a reference basis for the design and technical selection of future Mars water resource exploitation schemes.
2 Occurrence characteristics of Martian water ice
2.1 Low-latitudes
2.1.1 Mineral-bound water
Under the atmospheric environment of low pressure and low temperature on the Martian surface, free liquid water cannot exist stably for a long time. The water resources in the low latitudes (within 30° north and south latitudes) of the planet mainly exist in the regolith and bedrock in the form of mineral-bound water (
Bibring et al., 2006;
Murchie et al., 2009;
Ehlmann and Edwards, 2014). The reserves of hydrated minerals on Mars are very abundant. Numerical model estimates show that mineral-bound water alone can form an equivalent water layer (GEL) of 130 m to 260 m on a global scale, and its resource potential is of great significance in the ISRU scenario (
Jakosky and Phillips, 2001;
Sam Kim et al., 2022). Figure 1 shows a global map of Martian hydrated minerals.
Mineral-bound water is embedded in the mineral lattice in the form of hydroxyl groups (–OH) or crystal water, with significantly higher binding energy. It usually can only be released through heating-induced mineral dehydration reactions (
Bibring et al., 2006). The desorption of structural water is often accompanied by changes in the crystal structure and even irreversible phase transitions, which results in higher energy consumption requirements during engineering extraction processes.
Orbital remote sensing and
in-situ detection results together indicate that although the low-latitude regions of Mars lack stable surface ice, they are widely distributed with regolith and ancient strata rich in mineral-bound water (
Murchie et al., 2009). Systematic analysis of OMEGA and CRISM spectral data reveals that phyllosilicate minerals rich in structural water are mainly concentrated in Noachian strata, such as the Arabia Terra, Mawrth Vallis, and Nili Fossae regions (
Poulet et al., 2005). These minerals are believed to have formed in a relatively warm and humid environment with near-neutral pH during the early Martian period through water-rock reactions, with a water mass fraction of approximately 3 wt% − 9 wt% (
Mustard, 2019;
Wernicke and Jakosky, 2021). In contrast, hydrated sulfate minerals (such as gypsum and kieserite) appear more frequently in Hesperian strata, including areas such as Meridiani Planum and Gale Crater, with a water mass fraction of about 1.7 wt% − 5 wt%, and some sulfate-rich areas can reach 5 wt% − 7 wt% (
Rampe et al., 2016;
Wernicke and Jakosky, 2021). When the SAM instrument aboard Curiosity conducted in situ heating analysis of sediments in Gale Crater, it detected a significant water release peak in the temperature range of 100°C–300°C, confirming the presence of crystal water in sulfate minerals (
Leshin et al., 2013).
2.1.2 Subsurface ice in the Medusae Fossae Formation (MFF)
Although low-latitude regions are generally considered difficult to maintain stable ice bodies, the Medusae Fossae Formation (MFF) near the equator is an extremely notable exception. The MFF is a giant sedimentary structure extending approximately 5000 km along the Martian dichotomy boundary, composed of loose and highly weatherable materials.
Early studies based on radar data from Mars Express’ MARSIS indicated that the MFF has extremely low radar reflectivity and a low dielectric constant, which is similar to the characteristics of highly porous volcanic ash or sediments, but also suggests the possibility of large-scale subsurface ice (
Watters et al., 2007). However, only recently, through reanalysis of MARSIS data and gravity field simulations, researchers found that there are large-scale layered sediments up to several kilometers thick inside the MFF, and their radar signal characteristics are highly consistent with those of polar layered sediments (
Watters et al., 2024).
The latest radar inversion results show that the thickness of the underground ice layer in the Medusae Fossae Formation (MFF) can reach 2.5 to 3.7 km, with extremely high purity, and the content of dust and impurities may account for only a very small proportion. It is estimated that if all the ice in the MFF melts, the amount of water would be sufficient to form an equivalent water layer with a depth of 1.5 to 2.7 m globally on Mars (
Dotson et al. 2024). Compared with the mineral-bound water described in Section 2.1.1, the ice in the MFF exists in a form close to pure ice, and its physical state is closer to the underground buried ice in mid-to-high latitudes. Since the surface temperature in the equatorial region is higher than the sublimation point of ice all year round, the ice in the MFF can be preserved to this day mainly due to the thick layer of dry dust or regolith covering it, which is hundreds of meters thick. This “protective shell” effectively isolates the heat exchange between the underground ice and the atmosphere, slowing down the sublimation rate of the ice. Figure 2 shows the thickness map of the suspected ice-rich portions of the Medusae Fossae Formation (MFF) deposits.
2.2 Mid-latitudes
2.2.1 Pore ice and excess ice
Mid-latitude regions of Mars (approximately 30°‒60° north and south latitudes) are considered the main occurrence area of shallow subsurface water ice (
Leighton and Murray, 1966). This region has attracted much attention due to its balance between the accessibility of water ice and relatively moderate landing conditions for crewed missions (
Hoffman et al., 2016). Although the polar regions have more abundant water ice reserves, extreme low temperatures, permanent night, and other environmental factors pose significant challenges to landing and operations; while the low-latitude regions have relatively mild environments, the stability of shallow water ice is poor. Therefore, the mid-latitude regions have become the most strategically significant compromise choice for Mars water ice ISRU. NASA has set the latitude boundary for crewed missions within 50° north and south latitudes. This decision is based on in-depth research on the distribution of water ice on the Martian surface, ensuring that astronauts can obtain key water ice resources, while considering the performance of the Mars Ascent Vehicle and winter conditions.
These water ices are usually found beneath a dry regolith or debris cover ranging from several centimeters to several meters thick (
Mellon et al., 2004). For example, the Phoenix Lander directly detected a subsurface ice layer at a depth of approximately 4.6 cm in the region of 68° north latitude on Mars (
Mellon et al.,2009). Orbital detection data (GRS and FREND) indicate that in regions above approximately 50° north and south latitudes, there may be “excess ice” with a volume percentage of up to 50%‒100% within 1 m of the surface, extending to lower latitudes (approximately 45°) such as Arcadia Planitia and Promethei Terra (
Feldman et al., 2011;
Pathare et al., 2018). The phenomenon that pure water ice exposed by new impact craters subsequently sublimes and disappears also confirms the existence of shallow subsurface ice, whose distribution is consistent with the expected subsurface ice layer (
Byrne et al., 2009). Some studies have even found that water ice exists beneath only 1‒2 m of soil cover in regions of 55°‒58° north and south latitudes (
Webster et al., 2018). Figure 3 shows the possible distribution of Martian subsurface water ice, based on the latest data from SWIM (
Morgan et al., 2025).
The water ice beneath the Martian regolith is mainly divided into two categories according to its occurrence state: pore ice and excess ice. Their distribution structure under the regolith soil is shown in Fig. 4 (
Bramson et al., 2017). There are significant differences in their formation mechanisms and distributions.
Pore ice fills the gaps between soil particles (
Mellon et al., 2009). It forms when atmospheric water vapor seeps underground due to temperature differences and directly undergoes deposition and freezing. This type of ice is tightly bound to soil particles, and its content does not exceed the porosity of the soil. It is the main form of “ice-containing soil”. For example, among the near-surface ice bodies excavated by the Phoenix lander in its landing area, about 90% of the water ice is of the pore ice type.
In contrast, excess ice refers to ice bodies that exceed the soil porosity in volume and exist in the form of relatively pure ice layers or ice lenses. The reasons for its formation are quite complex, and the phenomenon of its high enrichment cannot be fully explained solely by water vapor diffusion. Currently, the mainstream view holds that it may be related to the migration and recrystallization of salt-containing brine under periodic temperature changes, or formed by the freezing and expansion of water in soil microcracks under specific conditions (
Style et al., 2011;
Sizemore et al., 2015).
The evolution processes of the ice bodies on Mars have all been strongly influenced by sublimation (
Dundas and Byrne, 2010;
Patel, 2022). Under the current atmospheric conditions of low pressure and low temperature on Mars, ice bodies are prone to sublimation, a process in which they directly transition from a solid state to a gaseous state, and this is a major pathway for the loss of subsurface ice to the atmosphere. Exposed ice in fresh impact craters often disappears rapidly within months to years, while the widely distributed fan-shaped marginal depressions in regions such as the mid-latitude Utopia Planitia (
Zanetti et al., 2009;
Smith et al., 2021) are considered typical geomorphic evidence of surface collapse caused by the sublimation of shallow subsurface ice.
It is worth noting that some special ice forms can also provide us with clues about Mars’ ancient climate. For example, some of the ice observed by the Phoenix lander exhibits the characteristics of firn (
Sizemore et al., 2015). Firn is an intermediate stage in the transformation of snow into ice, and its discovery on Mars today suggests that the region may have experienced brief snowfall or stronger water vapor condensation events in the past (
Smith et al., 2009), providing valuable evidence for studying the evolution of Mars’ climate.
2.2.2 Underground Thick Layered Ice and Buried Glaciers
Orbital radar on Mars (such as SHARAD and MARSIS) provides important evidence for detecting the distribution of subsurface ice layers on Mars, revealing that thick layers of pure ice deposits and debris-covered glacial structures are widely distributed in the mid-latitude regions of Mars (
Holt et al., 2008; Stuurman et al., 2016).
Thick buried water ice layers are widespread beneath the Martian surface, reaching tens to hundreds of meters in thickness. They originated from large-scale ancient snowfalls and were later sealed under layers of dust. The Phoenix lander once discovered such nearly pure (dustiness < 1%) subsurface ice (
Sizemore et al., 2015), providing strong support for this theory. In addition, the Arcadia Planitia and Utopia Planitia contain subsurface ice layers tens to hundreds of meters thick, with an overlying layer thickness of only 1–20 m, showing certain development potential (
Lefort et al., 2009;
Mank et al., 2021).
The Martian surface is also widely distributed with a type of glacial-related landform known as “viscous flow features”. Figure 5 shows a typical martian Glacier-like form, including Lobate Debris Aprons (LDAs), Lineated Valley Fill (LVFs), and Concentric Crater Fill (CCFs), etc. These features are mostly formed in mid-to-high latitude regions and are usually interpreted as debris-covered glaciers (VFFs), with a thickness of hundreds of meters and showing traces of past flow. These ice bodies are mostly considered to be relics from the period when Mars had a high obliquity. At that time, the polar ice caps sublimated, and water migrated to mid-to-low latitudes, resulting in large-scale snowfall and deposition (
Head et al., 2003). Radar echo analysis indicates that the internal dielectric constant of VFFs is extremely close to that of pure water ice, suggesting that their ice content is usually more than 80%‒90%, and the local thickness can reach 300‒800 m (
Holt et al., 2008,
Plaut et al., 2009). These macroscopic ice bodies show obvious traces of plastic flow driven by gravity. The reason why they have not completely sublimated in the current semi-stable thermal environment is entirely due to covered with a layer of debris retention layer about several decimeters to several meters thick on its surface. This protective shell acts as an excellent thermal insulator, effectively inhibiting the moisture exchange between subsurface ice and the atmosphere (
Schorghofer, 2007), making VFFs the most ideal large-scale
in situ water resource reserve targets in the mid-latitude regions of Mars.
2.3 High-Latitudes
2.3.1 Polar Caps and Layered Deposits
The north and south poles of Mars are covered by a vast polar cap system, which constitutes the largest current water resource reservoir on the planet. Structurally, the polar cap system consists of permanent polar caps that exist year-round and seasonal dry ice layers that grow and shrink with the seasons. The northern permanent polar cap has a diameter of approximately 1000 km and is mainly composed of high-purity water ice; the southern permanent polar cap has a diameter of about 350 km, with a layer of solid carbon dioxide approximately 8 m thick covering its water ice base. The core of these polar caps is composed of polar layered deposits (PLDs) up to several kilometers thick, showing significant layered rhythms, which reflect the global ice and dust distribution patterns caused by periodic fluctuations in Martian orbital parameters (
Plaut et al., 2007; Byrne, 2009).
Based on long-term detections by the MARSIS and SHARAD radars, the total volume of the North Polar Layered Deposits (NPLD) is estimated to be approximately 1.14 × 10
6 km
3, with the purity of water ice usually exceeding 95%, and the remaining components being dust, salts, and rock debris (
Grima et al., 2009). At the bottom of the NPLD, there is a special sedimentary structure called the basal unit, which is composed of ice layers with high sand content. The cross-bedding, with a thickness of about several kilometers, records intense aeolian activities in the early Martian polar regions (
Nerozzi and Holt, 2019). In comparison, the South Polar Layered Deposits (SPLD) are more massive, with a diameter of approximately 1000 km, a maximum thickness of up to 3.7 km, and a total volume of water ice of about 1.6 × 10
6 km
3. Although radar data in 2018 suggested that there might be a liquid brine lake at the base of the SPLD, subsequent full-polarization radar echo analysis indicated that the reflection signal is more likely to come from salt-bearing ice layers or clay minerals rather than liquid water bodies (
Orosei et al.,2018,
Bierson et al., 2021). It is estimated that if all the water ice in the north and south polar caps were melted, the amount of water would be sufficient to form a global equivalent layer (GEL) with a depth of approximately 20 to 30 m on a global scale of Mars (
Lasue et al., 2013). Figure 6 shows the captured images of the Martian north and south polar caps and the North Polar Layered Deposits (
McEwen et al., 2007).
2.3.2 Permafrost and peripheral polygonal landforms
Within the vast area extending from the polar cap boundary toward the equator, water ice exists stably in the form of underground deposits. Unlike the water ice dispersed in pores in mid-latitude regions, the permafrost in high-latitude regions exhibits an extremely high ice content, with local volume percentages often approaching 100%, and its physical properties are more similar to the underground branches of polar ice sheets extending to the surrounding plains (
Morgan et al., 2021). The
in-situ detection results of the Phoenix probe at 68°N show that the underground ice layer in this area is extremely shallow, and only about 4 cm of dry regolith needs to be removed to reach the high-purity hard ice (
Smith et al., 2009), as shown in Fig. 7. These shallow ice bodies are believed to be formed by atmospheric moisture entering the regolith through diffusion and deposition, or by snow accumulation and burial during paleoclimatic cycles, and have maintained thermodynamic stability for millions of years under the current harsh polar low-temperature environment (
Mellon et al., 2009).
3 Physicochemical properties of ice-bearing media on Mars
The physical and chemical properties of Martian hydrated minerals, ice-bearing regolith, and thick layered ice directly determine the selection of water ice extraction technology schemes, the complexity of system design, as well as the final mining efficiency and quality.
3.1 Hydrated Mineral Regolith
Although mineral-bound water has advantages in terms of spatial distribution and resource scale, its engineering extractability is significantly restricted by the physicochemical properties of the regolith. Based on the inversion results of MER wheel disturbance and trench experiments, the internal friction angle of Martian regolith is usually distributed between 30° and 37°, while the cohesion is mostly lower than 10 kPa, showing an overall low-strength loose medium (
Sullivan et al., 2010;
Hanley et al., 2014). However, due to the small particle size and large specific surface area, significant adhesion and agglomeration are likely to occur between regolith particles, which significantly increases the risk of blockage during material transportation and crushing (
Peters et al., 2008).
In addition, perchlorates with a mass fraction of 0.5 wt%–1 wt% are ubiquitous in Martian regolith. They decompose in the heating range of 500°C–600°C, releasing highly corrosive gases such as chlorine and hydrogen chloride (
Hecht et al., 2009). This characteristic not only imposes requirements on the corrosion resistance of materials used in extraction systems but also significantly increases the complexity of post-treatment and purification of produced water. Combined with the influence of the long-term high-energy particle radiation environment, the mineral-bound water extraction system faces multiple challenges in terms of reliability and lifespan design (
Hassler et al., 2014).
3.2 Ice-bearing regolith
Ice-bearing regolith is the most common form of resource in the mid-to-high latitudes of Mars, and its density, porosity, strength, and thermal conductivity are crucial for the excavation and extraction processes.
1) Density
The bulk density of Martian surface soil varies from place to place. For dry soil, its value is usually between 1.15 and 1.6 g/cm
3(
Ash et al., 2016). The presence of water ice is the most critical factor affecting its density; as the ice content increases, the overall bulk density of the surface soil increases significantly (
Mueller et al., 2013). By analogy with lunar data, its density may also increase with depth (
Connolly and Carrier, 2023).
2) Porosity
Porosity is closely related to density. It is usually higher in the surface layer but gradually decreases with increasing depth and compaction. For example, the porosity of lunar soil can decrease from about 65% at the surface to less than 40% at depth (
Connolly and Carrier, 2023), and this trend can provide a reference for Mars. When water ice fills the pores of soil particles in the form of cement, it directly changes their pore structure, which has an important impact on the transmission of heat and substances.
3) Shear strength
The shear strength of soil is mainly determined by cohesion and internal friction angle (
Dotson et al., 2024). For example, the measured internal friction angle of JSC Mars-1 simulant is about 41° (
Ash et al., 2016), which is at a relatively high value compared with the internal friction angle range of 16°‒30° for fine sand in the reference table of foundation soil bulk density and internal friction angle. For dry Martian soil, its cohesion is usually very low, only a few kilopascals. However, once water ice exists as a cementing agent, the cohesion of the soil will increase exponentially (
Dotson et al., 2024), which is one of the key reasons for the significant increase in the strength of frozen soil.
4) Compressive, flexural, and tensile strengths
When describing the mechanical strength of Martian frozen soil, compressive strength, flexural strength, and tensile strength constitute the three basic indicators, among which uniaxial compressive strength is particularly crucial. It is extremely sensitive to changes in ice content and temperature. An increase in ice content or a decrease in temperature will lead to a sharp increase in its strength (
Zhang et al., 2025). For example, the compressive strength of Martian permafrost with a water content of only 4% is approximately 1.5 to 2.0 megapascals; when the water content increases to 10%, this value can soar to 20 to 35 megapascals, with a hardness comparable to that of concrete (
Carrato et al., 2023). The flexural strength of ice-containing Martian simulants is usually less than 10 MPa. The tensile strength of pure Martian water ice at 212 K is estimated to be 1.2 MPa; the tensile strength of Martian permafrost may be close to 4 MPa (
Ash et al., 2016).
5) Thermal conductivity
Thermal conductivity is a decisive parameter for the efficiency of thermal extraction methods. The dry soil on Mars is an excellent insulator. Due to limited particle contact and low atmospheric pressure, its thermal conductivity is very low, approximately 0.02‒0.1 W/m·K (
Putzig and Mellon, 2007). When water ice is formed through vapor deposition and fills the pores, the thermal conductivity of the icy surface soil increases approximately linearly with the increase in ice content. However, the microscopic morphology of ice, such as “ice tendrils” that connect distant particles formed through vapor diffusion under small thermal gradients, may result in the actual thermal conductivity being lower than the predicted value based on the “ice neck” model. This implies that the heat transfer dynamics of Martian icy soil may be significantly different from that of terrestrial icy soil (
Siegler et al., 2012).
3.3 Thick layered ice
The purity of thick layered ice on Mars varies significantly, which directly affects the selection and efficiency of extraction processes. Polar layered deposits typically have high purity with low dust content, for example, the impurity content may be less than 5%‒15%. In contrast, ice in mid-latitude and equatorial regions, such as that in the Medusae Fossae Formation, has a higher content of dust or soil (
Yu et al., 2025). Samples obtained by the Phoenix mission from the Snow White trench contain ~30 ± 20 wt% ice (
Cull et al., 2010).
The impurities commonly found in Martian soil and ice mainly include dust and salts (
Banin, 2005). Martian dust is ubiquitous and mixes with ice, affecting the albedo and thermal properties of ice, and may cause wear or blockage during the extraction process. There are also various salts in Martian soil, such as perchlorates, sulfates, and chlorides (
Chen et al., 2025). Of particular concern are perchlorates (ClO
4−) and [another substance, name missing], which have been found globally (detected by Phoenix, Curiosity, and Perseverance rovers). Their concentration in icy surface soil is approximately 0.4 wt% ‒ 0.6 wt% (
Hecht et al., 2009), and may be higher in brine. These salts are strong oxidants, toxic, and highly corrosive to equipment materials. They can also form brine through deliquescence, significantly lowering the freezing point of water.
Due to the complex composition of Martian native water, traditional purification methods (such as ion exchange resin method) are too bulky and expensive for Mars missions. For example, treating water containing 6 tons of perchlorate requires approximately 26 tons of resin (
Song et al., 2017), which is engineeringly unfeasible. Therefore, it is urgent to develop new and efficient technologies such as biological reduction. In addition, considering the differences in pollutant characteristics in various regions, the purification system needs to have corresponding adaptability, or be tailored to specific extraction technologies and target resources. In summary, Table 1 summarizes the characteristics of Martian water ice.
4 Water ice extraction technology
4.1 Extraction of mineral-bound water
For the extraction of structural water, existing studies generally adopt the technical route of “mechanical pretreatment + high-temperature thermal desorption”. Representative schemes include the Water Extraction from Martian soil (WEXMA) and the regolith water extractor (
Bertolini, 2021). In systems like WEXMA, water-bearing minerals are first crushed to increase their specific surface area, aiming to shorten the internal heat conduction and moisture diffusion paths, and then the desorption of crystal water or hydroxyl groups (–OH) is induced by an external heat source.
In terms of process details, the regolith water extractor system uses solar concentrators or nuclear waste heat to heat the circulating CO
2 scavenging gas, and takes away the desorbed water vapor through convective heat exchange. Experimental and thermogravimetric analysis results show that the crystal water of typical Martian hydrated sulfate minerals (such as gypsum and kieserite) can be effectively released in the temperature range of 150°C–200°C; while the hydroxyl structural water in phyllosilicates (such as montmorillonite and sepiolite) needs to be heated to 500°C–800°C or even higher temperatures to undergo irreversible dehydration reactions (
Bishop, 2005,
Vaniman et al., 2014). Under typical operating conditions and engineering constraints, the recovery rate of structural water by such systems is limited by heating uniformity and condensation efficiency, and the currently evaluated effective recovery rate is approximately between 45% and 58% (
Trunek et al., 2018). Therefore, the structural water extraction scheme in engineering applications tends to target mineral-rich areas with high water content and is deeply coupled with high-temperature heat sources (such as nuclear reactors or high-efficiency concentrating systems).
Current research trends are evolving toward hybrid ISRU systems that combine multiple heating technologies. For example, solar concentrators are used for initial mineral dehydration, supplemented by microwave or radio frequency heating technologies for deep pyrolysis to improve thermal efficiency. In addition, given the abundant sunlight and significant day-night temperature differences in the low-latitude regions of Mars, hybrid systems can directly convert solar energy into heat during the day for the dehydration process, and use the atmospheric environment after cooling at night as a natural condensation sink, thereby significantly reducing the complexity and mass load of the thermal management system (
Zorzano et al., 2024).
4.2 Extraction of pore ice
Regarding the occurrence characteristics and physicochemical properties of pore ice in Martian regolith, various extraction technologies have been proposed and studied, which can be mainly divided into two categories: mechanical excavation technology and thermal extraction technology.
4.2.1 Mechanical excavation technology
Mechanical excavation technology crushes, excavates, and collects icy regolith through physical means, then transports it to a heating reaction vessel to extract water. Mars ISRU considers adopting various mechanical excavation technologies, with the selection based on the loose, compacted, or ice-cemented state of the regolith.
Common mechanical excavation tools include buckets, augers, bucket wheels, and some auxiliary scraping and cutting tools.
1) Bucket + impact vibration
As a basic excavation tool, the bucket is suitable for excavating granular and loose materials, as shown in Fig. 8(a) (
Höber, 2021). For example, the end of the robotic arm of the Phoenix lander is equipped with a bucket with a scraper, which successfully excavated ice-containing soil (
Arvidson et al., 2009). However, for hard ice-cemented soil, the penetration capacity of traditional buckets is limited, and the required excavation force increases sharply (for example, for soil containing 5 wt% ice, the excavation force may exceed 3500 N).
To overcome this challenge and adapt to the low-gravity environment on Mars, impact and vibration-assisted technologies have received widespread attention.
The principle is that by applying periodic impact forces or high-frequency vibrations to cutting tools (such as buckets), the downforce and traction required for excavation are significantly reduced, thereby reducing the overall mass of the excavator and lowering energy consumption. Experiments have proven its effectiveness: for example, the impact-type backhoe bucket (VIPER system) developed by Honeybee Robotics can reduce the excavation force by up to 50% under impact energies ranging from 13.6 to 30.5 J and impact frequencies as high as 700 times per minute (
Mueller et al., 2013). In a study on JSC-1A simulated lunar soil, it was found that impact-assisted technology can significantly reduce the downforce; specifically, the downforce can be reduced to 1/15 of the original (
Craft et al., 2010). Using this technology, researchers have successfully excavated hard frozen FJS-1 simulant and reduced the required force by 50%. Similarly, applying vibration to the cutting edge can also produce a similar effect, with benefits far exceeding the required additional power.
2) Auger
A auger is a spiral tool that can be used for drilling and material conveying in vertical or horizontal directions, as shown in Fig. 8b. For example, the Mobile In Situ Water Extractor (MISWE) concept (
Zacny et al., 2012) and NASA’s auger dryer (
Kleinhenz et al., 2018) both use deep-groove augers for mining and in situ water extraction. If the groove depth and pitch of the auger are insufficient to accommodate the size of the excavated material, clogging is likely to occur. Heated augers can combine excavation with the initial heating process (
Sanders, 2018).
3) Bucket wheel
The bucket wheel is a continuous excavation mechanism. By installing multiple buckets on a rotating wheel, it realizes the simultaneous execution of soil excavation and transportation, effectively reducing energy consumption and improving construction efficiency (
Trunek et al., 2018), as shown in Fig. 8(c) (
Johnson and King, 2010). This design is particularly suitable for low-gravity environments, such as the regolith mining on the Moon and Mars, and has been considered for improving the efficiency of resource extraction on the surfaces of these celestial bodies. After optimizing the design of the bucket wheel excavator through simulation analysis, a high excavation rate can be achieved. For example, it can reach 150 kg/h in the JPL-1 simulant (
Johnson and King, 2010). However, its design may need special consideration to handle large rocks to avoid difficulties in practical applications. The bucket wheel is adopted in NASA’s “Open Air Processor” concept (
Kleinhenz et al., 2018), but its performance may be affected by large stones.
In addition, some auxiliary scraping and cutting tools have also been studied for reducing excavation force. The Phoenix lander, for example, used a scraper to level the excavation surface (
Arvidson et al., 2009). For particularly hard icy soil samples, the Phoenix lander also used cutting tools such as files for sampling. However, the depth capability of the excavation tools deployed to Mars so far is very limited. For instance, the maximum drilling depth of the drill on the Curiosity rover before conducting deeper drilling missions was only 6.5 cm (
Horne, 2017), while the target excavation depth for future missions may reach several meters.
4.2.2 Thermal Extraction Technology
Thermal extraction technology mainly heats the water ice in Martian regolith through different heat transfer mechanisms, such as electromagnetic waves (microwave heating), thermal radiation (radiant heating) or direct thermal conduction (probe heating), causing it to sublimate into water vapor, which is then collected. In addition, for the extraction methods of water from the excavated ice-containing soil, they can be divided into open heating and closed heating according to different reactor designs.
4.2.2.1 Microwave Heating and Radiant Heating
In the technology of thermal extraction of water ice on Mars, microwave heating and radiant heating, as two non-contact heating technologies, have become key research directions in recent years due to their respective potentials in heat transfer efficiency and device simplification. A schematic diagram of the device is shown in Fig. 9 (
Linne et al., 2017).
Microwave heating technology, as a highly promising method for extracting water ice on Mars, its core principle lies in utilizing the direct interaction between electromagnetic waves and water ice to achieve the extraction of water ice. Specifically, microwave energy can penetrate the Martian regolith to a certain depth and be selectively absorbed by water molecules (including bound water and free water) and certain specific minerals within it (
Wiens et al., 2001). After absorbing microwaves, water molecules vibrate violently and heat up, promoting the efficient sublimation or vaporization of water ice into water vapor. This process is called “volumetric heating”. Compared with traditional surface heating or conduction heating methods, its significant advantage is that energy acts directly on water rather than being largely consumed in heating the entire soil matrix. Therefore, the energy utilization efficiency is higher, especially when dealing with soils with low water content.
The efficiency and penetration depth of microwave heating are significantly affected by the dielectric properties (dielectric constant and loss factor) of icy surface soil. These parameters are complex functions of soil composition (such as water content, mineralogical characteristics), microwave frequency, and temperature. Generally speaking, the penetration depth of microwaves can reach several feet. Lower microwave frequencies have deeper penetration capabilities, while higher frequencies can heat substances to higher temperatures at shallower depths (
Ethridge and Kaukler, 2007).
To meet the demand for continuous processing, various microwave-based extraction concepts have been proposed. One approach is in situ heating, which involves using a horn antenna to radiate microwaves to the ground surface or transmit microwave energy downward through a borehole to directly heat the underground ice layer, thereby reducing the amount of excavation work. The generated volatiles will migrate upward to the surface for capture (
Linne et al., 2017).
In addition to the concept of in situ microwave heating, more research has focused on the continuous processing of excavated soil. A large number of experimental studies on lunar soil and Martian simulants have fully verified the feasibility of extracting water using microwaves under low-temperature and vacuum conditions, and have investigated the penetration depth and heating effect at different frequencies (such as 2.45 GHz, 0.9 GHz, and 10 GHz) (
Ethridge and Kaulker, 2012). For example, NASA proposed a resonant cavity-based scheme (
Kleinhenz et al., 2018), where granular materials are fed into a cavity containing a porous container for heating. This design aims to optimize the efficiency of water release and collection. Another common design is the conveyor belt system (
Wiens et al., 2001), such as the one conceived by the JFEET team, which allows the soil to move continuously on the conveyor belt through the microwave field. Its advantage is that the processing flow is smooth and less prone to blockages. These laboratory-scale experiments have achieved significant results, demonstrating that up to 99% of water ice can be removed through microwave sublimation, and 95% of the water vapor can be captured (
Ethridge and Kaukler, 2007). In addition, the effectiveness of microwaves in releasing bound water has also been corroborated. For instance, although the WAVAR reactor is used for atmospheric water extraction (
Hilstad and Adan-Plaza, 1998), it utilizes microwaves to efficiently heat zeolites, thereby desorbing the adsorbed water.
In contrast, radiative heating is a technology that uses devices such as solar concentrators or infrared radiant heaters to transfer thermal radiation energy to the surface of Martian soil in a non-contact manner, with the aim of heating it to extract water (
Wiens et al., 2001). Its core concepts include direct heating and indirect enhancement, such as improving the absorption efficiency of solar energy by changing the optical properties of the soil surface (e.g., reducing albedo), thereby promoting the sublimation of water ice (
Mungas et al., 2006).
The main advantage of the radiative heating method lies in its relatively simple concept, as it can directly utilize solar energy available locally on Mars. This is particularly important in Mars exploration missions because radiative heating has a significant impact on thermal protection design. However, its actual application effect is closely related to the target temperature: although heating the soil surface to 60°C can generate water vapor (
Linne and Kleinhenz, 2016), extracting bound water from minerals may require high temperatures ranging from 200°C to 500°C (
Wiens et al., 2001). The main drawback of this method is its low efficiency. Heat only acts on the surface layer, and due to the poor thermal conductivity of Martian soil, it is difficult for heat to transfer deeply, resulting in high energy consumption for
in-situ heating of large volumes of soil. Therefore, due to the fundamental limitation of poor heat permeability, pure radiative heating is not suitable for large-scale in situ water ice extraction. Instead, it is more appropriate for processing small batches of excavated materials in closed containers, or it must be combined with mechanical methods such as turning and thinning to improve heat transfer efficiency (
Linne and Kleinhenz, 2016).
4.2.2.2 Probe Conduction Heating
The core concept of probe-based conductive heating is to directly insert one or a set of probes integrated with heating elements into the ice-containing surface soil layer by means of drilling or hammering (
Sowers and Dreyer, 2019). Heat conduction transfers the heat from the probe surface to the soil and ice, causing them to melt or sublimate, which is then collected through a preset channel, as shown in Fig. 10 (
Sowers and Dreyer, 2019).
The most significant advantage of this technology lies in its deep extraction capability. Combined with a drilling system, the probe can reach rich ice layers tens of meters below the surface, avoiding the excavation of dry soil layers. As demonstrated in Earth’s glacial drilling (Rodwell Well), its advantages have been verified (
Hoffman et al., 2020a). However, this technology faces severe challenges in Martian applications, the core of which is low heat transfer efficiency (
Grott et al., 2021). Martian soil is inherently a poor conductor of heat (similar to an insulator), making it difficult for the probe to achieve perfect physical contact with the uneven frozen soil. Any tiny gap will greatly hinder heat conduction, resulting in a large amount of energy being wasted or an extremely slow extraction rate.
4.2.2.3 Open Heating
Open heating refers to heating the excavated surface soil in a non-closed environment. Heat is applied through radiation, heating plates, or hot air flow, causing volatile substances to escape into the surrounding Martian atmosphere (or into an open collection hood).
For example, the “open air” heating tray concept proposed by NASA’s Glenn Research Center, as shown in Fig. 11 (
Kleinhenz et al., 2018), operates on the principle that the excavated surface soil is evenly spread on the heating surface, and direct heating methods such as radiant heaters are used to promote the sublimation of water ice. The thin atmosphere of Mars CO
2 can be used as a purge gas to carry the generated water vapor to the condensation collection system. In a specific open reactor concept, a bucket wheel deposits the surface soil onto an inclined, vibrating copper tray, which is heated to approximately 150°C, thereby raising the soil temperature to about 100°C. CO
2 The purge gas transports the water vapor to the condenser at a pressure of approximately 7 torr. The choice of heating temperature depends on the existing form of the target moisture: to recover moisture from ordinary soil, the temperature may need to reach 300−400°C; to extract moisture from hydrated minerals such as gypsum and sulfate, it needs to be controlled between 150°C and 250°C; while for the direct sublimation of subsurface ice, only 100°C−150°C is required.
The advantage of this design is that it avoids complex seals that are difficult to maintain in abrasive environments, and directly utilizes the Martian atmosphere to achieve efficient continuous heating while reducing energy consumption and maintenance requirements. However, its main drawback stems precisely from this openness: the sublimated water vapor is difficult to be completely captured, and some water resources will inevitably escape into the environment, thus limiting the final overall water recovery efficiency (
Linne et al., 2017).
4.2.2.4 Closed Heating
Unlike open systems, closed heating involves heating icy surface soil in a sealed or semi-sealed chamber, which effectively controls internal pressure and improves water vapor capture efficiency (
Kleinhenz et al., 2018).
A typical closed heating system includes a heating auger and a closed reactor/oven. The heating auger includes NASA’s auger dryer (
Kleinhenz et al. 2018) or Honeybee Robotics’ thermal corer (
Lee et al., 2023; Figs. 12(a) and 12(b)). Its principle is that the surface soil is continuously conveyed and heated through a heated auger assembly. The change in the pitch of the auger blades and the plug formed by the surface soil in the hopper form a seal, driving the water vapor to be transported to the condenser under pressure. Honeybee Robotics’ thermal corer integrates small channels for heating fluid inside the auger and has been tested with simulants containing 5% ice (
Lee et al., 2023). Kettle-type, pot-type, or oven-type reactors adopt a batch processing method, where the soil is loaded into a closed container, heated by electric heating or other methods to release water vapor, and then condensed and collected (
Wiens et al., 2001).
In a closed system, the main heat transfer methods are conduction and convection. For a closed system, effective sealing and the smooth discharge of treated tailings are crucial. The auger-based system cleverly uses the surface soil itself to form a seal, while the tailings (treated dry surface soil) must be able to be efficiently discharged from the system.
4.2.2.5 MISWE mobile in-situ water extraction
Given the characteristics of Martian ice-containing regolith, which is widely distributed but has a harsh mining environment, Honeybee Robotics has developed the Mobile
In-Situ Water Extractor (MISWE), as shown in Fig. 13. The core design concept of this system is to achieve “
in-situ drilling and closed-loop extraction”. That is, multiple process links such as drilling, sampling, thermal desorption, and moisture condensation are highly integrated into a single mechanical structure, thereby minimizing the moisture sublimation loss caused by complex material transfer and the complexity of the system mechanism (
Zacny et al., 2012). MISWE is mainly composed of a segmented auger integrated with a high-frequency impact drive module, a sealed deep-well thermal extraction chamber, and an efficient condensation collection device.
MISWE adopts advanced impact-assisted drilling technology, which uses high-frequency vibration or impact loads to effectively break the ice-soil combination, reducing the static downforce required for excavation by more than 50%. This enables it to be deployed on small rovers with mass constraints or fixed landing platforms. In the actual operation process, the spiral drill directly conveys the collected ice-containing cuttings into a sealed extraction chamber; then, the sample in the chamber is heated to above the triple point of ice using built-in resistance heating wires or microwave radiation sources. The generated water vapor is transported to the condenser at the top through the gas guide channel inside the drill pipe under the drive of pressure difference for liquefaction and recovery. Compared with traditional large-area mining schemes, MISWE not only solves the problem that ice sublimes immediately upon exposure under the low-pressure conditions on Mars through a closed environment, but also, through its modular and mobile design, can achieve precise positioning and efficient acquisition of shallow subsurface ice resources in the mid-to-high latitudes of Mars.
4.3 Extraction of excess ice and polar ice caps
4.3.1 In-situ underground melting and extraction (rodwell system)
The underground pure ice on Mars also has huge resource potential. The underground in situ melting and extraction technology is suitable for ice layer mining. Its core lies in placing the heat source directly in the ice-rich layer, so that the solid ice is melted into liquid water in place, and then pumped out.
A typical example of this method is the Rodwell well (Fig. 14(a)), a land-based concept originally developed for drilling and water extraction in the Antarctic ice sheet. It uses circulating hot water or thermal probes to melt and infiltrate the ice layer, forming and maintaining an expanding underground water cavity for water extraction (
Mank et al., 2021). The RedWater system developed by Honeybee Robotics is a specific implementation of this concept for Mars applications. This method is mainly applicable to pure and concentrated massive ice deposits, such as the Martian glaciers studied by NASA (with a burial depth of about 1 m), rather than ice crystals dispersed in the regolith. Its main advantage is efficient and continuous extraction without large-scale excavation and transportation. In the initial stage of operation, the system uses a mechanical drill to penetrate the dry regolith on the surface and enter the deep pure ice layer. Then, it injects thermal fluid (such as heated water or steam) into the bottom of the hole or deploys a downhole heater to melt the surrounding ice to form an inverted pear-shaped “bulb” filled with liquid water. By continuously pumping the cold water at the bottom of the molten pool back to the surface for heating and re-injecting it, the system can maintain and expand the scale of the molten pool. After reaching a stable state, the liquid water can be extracted for life support or propellant production.
However, applying this technology to Mars also faces severe challenges. First, the system itself is relatively complex, and it is necessary to overcome the difficulty of drilling through the overburden of unknown depth. More crucially, the low-pressure and low-temperature environment on Mars poses fundamental problems: the boiling point of water decreases, which may lead to premature boiling or violent sublimation in the water cavity. This not only seriously affects thermal efficiency, but the resulting water vapor may also recondense in the colder upper regions of the borehole, causing blockages (
Hoffman et al., 2020b). In addition, soil or salt impurities in the ice layer may be deposited after melting, affecting the operation of equipment and polluting water quality. Therefore, although the Rodwell well is regarded as the preferred solution for extracting Martian water ice, its successful implementation still requires in-depth simulation experiments and field studies to overcome a series of key environmental adaptability challenges.
4.3.2 MDTR mobile vehicle transportation device
NASA has also developed a Mobile Drilling and Transport Rover (MDTR) (Fig. 14(b);
Hoffman et al. 2016), a specialized exploration vehicle that combines drilling and transportation capabilities. Its mission is to first drive to the area containing subsurface ice, use its built-in drill to penetrate the surface covering layer; after reaching the ice layer, it will deploy a thermal probe called “Cryobot” to go deep into the ice layer for heating operations, prompting the solid ice to melt or directly sublimate into water vapor; this water vapor will be captured by the recovery hood above the vehicle and recondensed into ice in the on-board “ice bucket” through the cold trap system; when the ice bucket is full, the MDTR will return to the main base (such as MAV/fuel factory) for unloading and processing, and can make multiple round trips as needed to continuously extract water resources.
Based on the above extraction methods for Martian water ice, Table 2 summarizes and analyzes them.
5 Conclusions and prospects
As a critical in situ resource for future manned exploration missions, the study of Martian water ice’s occurrence characteristics and extraction technologies has become a cutting-edge interdisciplinary field in planetary science and deep space exploration engineering. This paper systematically reviews the distribution patterns, occurrence states, medium properties, and corresponding extraction techniques of Martian water ice. Overall, the occurrence of Martian water ice exhibits distinct latitudinal zonality: low-latitude regions predominantly contain mineral-bound water, offering immense resource potential but requiring high energy consumption for extraction; mid-latitude regions currently serve as key areas for in situ resource utilization (ISRU) exploration, featuring widespread shallow pore ice, excess ice, and thick-layer buried glaciers, achieving a favorable balance between resource accessibility and engineering implementation conditions; high-latitude and polar regions harbor high-purity permafrost and extensive polar ice cap systems, with extremely abundant reserves but harsh environmental conditions. Simultaneously, the physicochemical properties of ice-bearing media—such as the dehydration temperature of mineral-bound water, mechanical strength and thermal conductivity of ice-containing weathered layers, and impurity content (e.g., perchlorates) in ice bodies—directly determine the selection of extraction processes, the complexity of system design, and the final mining efficiency and water quality.
To address these distinct subsurface conditions, researchers have developed diversified extraction methodologies. For mineral-bound water, the prevailing approach combines mechanical pretreatment with high-temperature thermal desorption, while hybrid systems integrating solar energy and microwave heating are being explored to enhance efficiency. In the case of pore ice prevalent in mid-latitude regions, extraction techniques focus on mechanical excavation (e.g., impact-assisted excavators, spiral drills, and bucket-wheel excavators) and thermal methods (e.g., microwave heating, radiation heating, and probe-conduction heating). Mobile in-situ extraction systems like MISWE demonstrate promising applications due to their high integration and low energy loss. For large-scale excess ice and polar ice sheets, underground in-situ melting technologies such as the Rodwell system are considered viable for continuous extraction, though their successful implementation requires overcoming critical challenges including phase change control of water and impurity deposition under Martian’s low-pressure environment.
Looking ahead, the development and utilization of Martian water ice still face numerous scientific and engineering challenges. Future research should focus on several key directions. First, it is essential to integrate data from next-generation orbital and roving probes to achieve precise evaluation of water ice distribution, burial depth, purity, and medium characteristics, providing reliable basis for mission site selection. Second, efforts should be made to develop lightweight, efficient, and highly reliable extraction systems, promoting intelligent integration and system integration of mechanical, thermal, and chemical processes. Third, breakthroughs must be achieved in in situ purification technologies for impurities such as salts and dust in water ice to ensure safe and available water resources. Additionally, systematic-level verification and reliability testing through ground simulations, near-Earth orbit demonstrations, and future unmanned Mars missions are crucial steps toward mature application. Ultimately, the extraction and utilization of Martian water ice must be integrated into larger closed-loop systems for life support and propellant production, achieving coordinated energy, material, and environmental management to establish a solid resource foundation for long-term sustainable human habitation on Mars. The development and utilization of Martian water ice is not only a technical challenge but also a critical step for humanity’s journey into deep space and interplanetary survival, with its progress profoundly shaping the future landscape of deep space exploration.
PDF
(5913KB)
