A new experimental capability for the study of regolith surface physical properties to support science, space exploration, and in situ resource utilization (ISRU)

The surfaces of many solar system bodies, including the Earth’s moon, small bodies, and Mars, are covered in regolith. The physical properties of regolith are influenced by the mineral composition, particle shape, particle size, and volatile content among other factors. The measurement of the physical properties of regolith is important for advancing knowledge of planetary processes, for enabling robotic and human exploration of the solar system, as well as for the development of in situ resource utilization (ISRU) systems, which aims to improve space mission capabilities and depends on resources found in regolith. Acquisition of resources and their conversion to useful products are dependent on engineering processes (e.g., excavation, trafficability, and vapor transport) governed by regolith physical properties. The study of regolith physical processes in the laboratory environment will inform scientists and engineers of conditions that may be encountered on the Moon, asteroids, comets, small bodies, and Mars, advancing knowledge and improving the ability to work on these bodies. In this paper, the development of a laboratory facility for the study of the physical properties of planetary regolith called the ISRU Experimental Probe (IEP) is presented.

The regolith formation process on many bodies, especially airless bodies, renders extraterrestrial regolith unlike terrestrially formed regolith. The physical properties of the lunar surface were summarized in Carrier et al. (1991), in which the findings of measurements from in situ human and robotic missions, laboratory measurements of returned samples, and remote sensing were reviewed. The surfaces of airless bodies are processed by meteorite impacts from the micrometer to kilometer scale, rendering these surfaces a fine-grained power of highly angular particle shape while also imparting chemical alterations. Due to the orientation of the lunar axis relative to the plane of the ecliptic, craters and other depressions at the lunar poles are permanently shadowed from sunlight. These permanently shadowed regions (PSR) are therefore exceedingly cold (as low as 25 K) and act as traps where volatiles deposit by vapor deposition (Paige et al., 2010). Vapor deposition of ice in regolith is thought to form pore-filling structures that differ from terrestrial analogs (e.g., terrestrial permafrost). The thermal conductivity of icy regolith formed by vapor deposition differs substantially from icy regolith formed under terrestrial conditions (Siegler et al., 2012; 2015). In addition, absorbed gasses on regolith grains affect the measured physical properties of regolith, as shown by Perko (2001), which indicates the need to consider differences between the regolith formation process in space and the terrestrial analogs used for experimentation.

Based on remote observations, fly-by missions, orbiters, surface contact missions, laboratory testing, and theoretical considerations, a number of small bodies are now known to consist of a wide range of unconsolidated materials. Observations indicate that the mean particle size varies greatly across the surface of many asteroids. For example, asteroid 25143 Itokawa was found by the JAXA Hayabusa mission (Fujiwara et al., 2006) to have a varied surface consisting of boulder fields and smooth areas. Asteroid regolith is likely to be less compacted than regolith on planetary surfaces due to reduced gravity (Britt et al., 2002). As the in situ measurement of asteroid surfaces is lacking, the physical properties of these surfaces remain uncertain. The laboratory measurement of meteorites provides some relevant direct measurements, but the meteorite size and selection bias hinder extrapolation to asteroid regolith surfaces and boulder-sized objects. Additionally, carbonaceous chondrite asteroids that contain hydrated minerals, and are therefore of particular interest for ISRU, are not well represented in meteorite collections. Numerical modeling indicates that cohesive forces between grains dominate the physical strength of asteroid regolith (Kulchitsky et al., 2016). Discrete element method simulations show that this initially occurs in a quasi-elastic phase followed by a plastic phase when the bonds between individual grains are broken (Sánchez and Scheeres, 2014 and Sánchez et al., 2017). In addition, there is evidence that some main belt small bodies contain ice under a partially protective surface layer (Snodgrass et al., 2017). These discoveries suggest the need for additional experimental capabilities to aid in the understanding of these bodies and their resource utilization (Dreyer and Abbud-Madrid, 2015 and Dreyer et al., 2014).

Relative to regolith on bodies that lack an atmosphere, the regolith on Mars varies greatly (Certini et al., 2009). Martian regolith is basaltic in composition, pulverized by impacts, and altered by aqueous processes that were likely present on Mars in the past. In addition to impact features, sedimentary formations and wind-blown features appear to be present. Liquid water is not stable on the surface of Mars today, but there is evidence of ancient periods when Mars could have been wet. The polar glaciers contain water and carbon dioxide ice, and water ice is stable in the shallow subsurface at mid-latitudes (Holt et al., 2008 and Dundas et al., 2018). There is evidence that vapor-deposited water ice forms structures within the regolith pore space distinct from those created through freezing of liquid water (Hudson, 2008 and Siegler et al., 2012).

In addition to scientific test facilities, several groups have built facilities to support flight hardware development at the subsystem and system level in relevant environments, which is critical to successful flight hardware development, particularly early in development. One such facility has been developed at NASA Glenn Research Center for technology development in the lunar environment (Kleinhenz, 2014 and Kleinhenz and Wilkinson (2014)). Several groups have studied forces during excavation and other regolith hardware interaction processes, such as regolith-tool interactions using regolith simulants at atmospheric pressures (King et al., 2011 and Johnson and King, 2010), and excavation and drilling in lunar simulants (Green et al., 2013; Green and Zacny, 2014; Zacny et al., 2015; and Kleinhenz et al., 2015).

Research has been limited to specific ranges of temperature and pressure, and certain regolith simulants, including methods of icy regolith formation. Technology developers select test conditions and processes based on perceived need and facility availability. Conditions are typically selected to isolate critical physical processes of the technology, given what is known about the operational environment, which may itself be incomplete. Improved knowledge of physical properties of extraterrestrial surfaces is needed to make informed decisions about technology development.

The new ISRU Experimental Probe (IEP) developed in the Center for Space Resources (CSR) at the Colorado School of Mines is capable of measuring the forces of physical probes interacting with regolith surfaces. The IEP was developed as part of the Institute for Modeling Plasma, Atmospheres, and Cosmic Dust (IMPACT), a node in NASA’s Solar System Exploration Research Virtual Institute (SSERVI). Regolith mixtures can be formed dry or as icy regolith mixtures by vapor deposition in vacuum or as liquid mixtures initially formed in air. Samples can be held at temperatures ranging from −196 to 150 °C and at pressures from 10⁻⁷ Torr to ambient laboratory pressure. In this article, the technical details of the IEP, experimental capabilities of the system, and proof-of-concept results of regolith penetration resistance are presented.

Sections II A–II E describe the IEP design and capabilities. The IEP, shown in Fig. 1, enables the interaction of a physical probe with a bed of regolith in vacuum and at cryogenic conditions. IEP motion is fully computer controlled, although human intervention and control are also available. Data acquired during a test include the position, six-axis force and torque, temperature at user selected locations, and vacuum chamber pressure. The probe shown in Fig. 1(b) is conically tipped, but in principle, it can be of any geometry and additional instrumentation can be added. The measurement of interaction forces of the IEP probe with the regolith surface allows mechanical properties of regolith surfaces to be derived and geotechnical properties of surface strength to be inferred (such as compressive, shear, and cohesive strengths to name a few).

The IEP consists of a three-axis translation stage with a 50 mm travel range in each axis and is capable of complex coordinated motions. A fourth motor can be integrated to provide the rotary action of the probe. The six-axis force-torque sensor mounted on the translation stage aligns the sensor axes with the translation stage axes with high accuracy (±1.5°). The fixture can carry a variety of probes for testing probe penetration, scraping, and anchoring operations on a regolith surface. The IEP support structure is designed for low deflection at maximum loads. The stage position, three orthogonal force, three orthogonal torque, video, pressure, and five temperature readings are recorded simultaneously. The IEP structure fits within a 15 cm × 33 cm × 16 cm volume, and mass is approximately 18 kg. The IEP probe can be translated within a cube 50 mm on a side; thus, interaction with a regolith surface is limited to 50 mm.

To facilitate the description of this instrument, we have defined several components of the IEP. The IEP structure contains the translation systems, force-torque sensor, probe, and structural parts. The IEP sample platform creates a stable base for the sample containers, which hold regolith samples. Finally, supporting equipment includes vacuum systems, sensors, controls, and tools for preparing regolith samples. Sections II A–II E describe the IEP in detail.

The IEP is designed to operate within an 18″ diameter vacuum chamber in the CSR, as shown in Fig. 2. The CSR 18″ vacuum chamber has 45 cm inner diameter and 163 cm inner length and is capable of providing ultra-high vacuum (UHV) (<10⁻⁷ Torr). The IEP is mounted on the fixed end of the chamber and is enclosed in the vacuum volume by the moveable end of the chamber. When UHV is desired, all ports can be sealed using oxygen-free copper seals. When pressure above 10⁻⁶ Torr is adequate, the main chamber seal can be converted to a fluoroelastomer seal for quick and easy access. Vacuum pressures are achieved using various means; for a rough vacuum (>10⁻³ Torr), a rotary-vane or scroll pump is sufficient. For higher levels of vacuum, a turbo-molecular pump can be employed. To achieve UHV, cryogenic and/or ion pumping methods are also available to be used in combination with a turbo-molecular pump. With a sample in an IEP sample container, the pumpdown time to 50 mTorr is less than 20 min, while the time to UHV is approximately an additional hour using a turbo-molecular pump. The vacuum level can be maintained during testing using a manually adjustable valve between the pump and vacuum chamber. The time to switch out samples can be a few minutes, given adequate preparation of samples and containers. Heaters are integrated on the exterior of the chamber, to be used in combination with a full chamber insulation shroud (not shown), to facilitate high temperature bake-out operations.

The IEP is constructed of 6061-T6 aluminum and has been designed to support the loads imparted on the sample by the IEP probe. The structure is supported by two cantilevered fixtures, which are fixed into the stationary end of the chamber. The IEP structure includes three Newmark NLS4 Series linear translation stages. The horizontal axes are in a “Newmark XY” configuration mounted via custom brackets on the vertical NLS4 stage, which in turn is mounted on the IEP support structure. Misalignment in the support structure is mitigated by tight tolerances and by incorporating a rigorous quality control process throughout the manufacturing process. The structure is designed to be vacuum compatible; blind holes and dead volumes were eliminated to minimize “virtual” leaks to speed vacuum pump down.

The IEP support structure allows probe-to-surface angles from surface-normal to 45° from surface-normal in increments of 5°. When using the cryogenic sample container (pictured in Fig. 5), the interface angle is reduced to 30°. In addition, the angle of interaction of the probe with the surface can be increased by making samples with sloped surfaces. The IEP has been designed to give the user flexibility in a relative position and orientation of the sample container and IEP structure.

The three linear translation stages provide 50 mm of travel and are driven by NEMA size 17 vacuum compatible stepper motors with lead screws that facilitates sub-µm resolution at up to 10 mm/s (unloaded). The stall load of the motors was empirically determined to be 250 ± 10 N. A fourth motor can be mounted on the translation subsystem for testing drilling and anchoring operations.

The IEP is capable of testing a range of samples from uncompacted (loose) dry regolith to saturated icy regolith. The first scenario requires low backlash due to the small forces that a probe experiences when interacting with loose materials, and the latter requires high system stiffness due to high forces that a probe experiences when interacting with compacted or icy materials.

The Newmark NLS4 translation stages have zero backlash due to a special drive nut design; a plastic preloaded nut takes up backlash and is designed to wear such that it continues to take up backlash. The material leads to compliance in the nut, giving a bidirectional repeatability of 10 ± 2 μm. Stress analysis using nut dimensions, compliance, and modulus of elasticity of 1.4 GPa (plastic falls between 1.4 and 4 GPa; thus, 1.4 GPa will produce a worst-case) reveals that a compensating force of less than 0.1 N is present when the stage is under load. As the stiffness in the rest of the system becomes dominant after the nut is initially compressed, the compensating force can only affect load conditions under 0.1 N, and the translation stages may be considered infinitely stiff above 0.1 N.

Several test configurations are possible with the IEP. Configurations with high loading will lead to higher stresses and displacements. The worst-case scenario is off-axis loading, which occurs at the maximum angle between the structure and sample platform. Figure 3 shows a finite element model of the IEP structure and sample platform with a color map to visually show displacement with a normal loading condition of 200 N. The maximum displacement of 32 μm is shown in red, but the majority of elements in the stressed region range from orange-red (29 μm) to yellow (24 μm). General loading conditions of other IEP configurations have total displacements of less than 10 μm, indicating that the IEP has sufficient system stiffness for testing under high loading conditions.

The IEP uses a six-axis force-torque sensor (ATI-AI Mini40) mounted on the translation stage using a custom-made fixture, which aligns its axes with those of the IEP. The force-torque sensor uses an arrangement of strain gauges to determine three orthogonal forces and three torques about each axis. The force-torque sensor calibration allows for a wide range of loading conditions. ATI force-torque sensors are calibrated with three defined calibration ranges: (1) the complex loading range, (2) the extended calibration range, and (3) the out-of-calibration range. In the complex loading range, the sensor can simultaneously measure loads and torques up to specified limits with the highest accuracy. In the extended calibration range, simultaneous load and torque measurements are more limited (e.g., high Z-axis loads can be measured only with low torques in the X and Y-axes) and accuracy is decreased. In the out-of-calibration range, the sensor will not return accurate readings. The sensor can store several calibrations, which are used as needed depending on the anticipated experimental loads, torques, and required accuracy. Calibrations range from peak calibration Z-axis loads of 240 N to 30 N with corresponding peak X-axis and Y-axis torques of 4 N m to 0.5 N m, respectively, in the complex loading range. In the X-axis and Y-axis, the sensor has calibration peak loads of 80 N–10 N and peak Z-axis torques of 4 N m–0.5 N m, respectively, in the complex loading range. Sensor resolution in the Z-axis is 0.08 N for the high load calibration and 0.01 N for the low load calibration within the complex loading range. With the highest load calibration in the extended range, the sensor can operate up to 560 N in the Z-axis. Mini40 was selected to provide adequate load measurements in regolith simulants with relatively modest sized probes. Alternate ATI-AI sensors with either a higher load limit or a lower load limit with higher accuracy can be installed in the IEP if different sensor performance is required.

The IEP is run by a custom LabVIEW control and data acquisition virtual instrument (VI) that records the temperature (at five user-selected locations), chamber pressure, and ATI-AI sensor data and communicates with the Newmark NCS-G4 GalilTools four-axis motor controller. Thermocouples can be placed by the user at many locations, including the sample container wall, the IEP structure, the regolith sample, and the sample probe. A combination of custom LabVIEW control code and GalilTools at the motor controller level can create complex motion control. The IEP is located close to a viewport to allow video recordings from a camera placed outside the chamber at a distance of 20 cm. Figure 1(b) shows such an image using a wide angle lens, which provides 0.5 mm resolution at the probe tip. In principle, other cameras can be used to increase resolution.

The IEP can accommodate samples up to 12 cm in diameter including granular medium mixtures such as regolith, as well as breccia and rock. Samples can be mounted directly on the sample platform shown in Fig. 1 or placed in a sample container mounted on the sample platform. Alternatively, the sample platform can be replaced with a custom-designed sample containment system. Two sample containers have been built to date: the Basic Sample Container (BSC) and the Cryogenic Sample Container (CSC).

The BSC, shown in Fig. 1, is a cylinder measuring 5 cm in diameter and 5 cm high designed for regolith mixtures that is mounted on the sample platform. In addition to venting out the open top of the sample, as shown in Fig. 4, the container vents through the bottom via a fine mesh screen (e.g., 304 stainless steel mesh No. 200 with 74 μm openings) and via vent holes in the baseplate to minimize the disturbance of the regolith during pumpdown. The bottom of the baseplate is machined with a relief to create a small gap (0.5 mm) when the BSC is on a sample platform. Resistive heaters and lamps can be used to increase the sample temperature above ambient.

The CSC, shown in Fig. 5, is designed for studies of icy regolith mixtures. The regolith bed within this container can be chilled from the base via a cryogenic cooling loop and heated from the top with resistive heaters or lamps. Icy regolith can be formed either via vapor deposition or by freezing moistened regolith under atmospheric conditions.

Water vapor added to the vacuum chamber freezes within the regolith where the temperature passes through the solid-vapor phase boundary. The control of the thermal gradient within the regolith sample (using the chilling base and heated top plate) aids in directing and controlling vapor deposition within the sample. Vapor deposition of water is slow and requires careful control of sample temperature and chamber pressure conditions. The process of deposition in the laboratory can take days to weeks depending on the quantity of ice needed, the volume of regolith pore space, and the temperature gradient imposed on the regolith sample (Hudson et al., 2009). In principle, vapor-deposited ice can be formed from volatiles other than H₂O under the correct thermodynamic conditions for the volatile.

Figure 5(b) shows CSC thermal model results for the cryogenic puck at −85 °C and the sample cup support plate heated by embedded resistive heaters operating at 5 W. The regolith was modeled with the thermal conductivity of dry JSC-1A in vacuum, a lunar mare simulant (McKay et al., 1994). The model results show that the temperature of the regolith gradually varies from 0 °C at the surface to the puck temperature at the bottom of the sample cup and that temperature profiles are uniform in the horizontal plane near the centerline of the sample cup. The regolith temperature can be varied to simulate a variety of conditions in which ice may be present (such as shadowed lunar polar craters, asteroids, or the Martian surface), from pure solid ice to ice-regolith mixtures at various moisture contents. Preliminary thermal modeling results suggest that with the current material selection and employed thermal isolators, a controllable temperature gradient can be generated within the sample. The ability to establish a thermal gradient in the cryogenic sample container has been experimentally verified.

To control cooling, cryogenic fluid can be circulated through the cryogenic puck while in the vacuum chamber through vacuum fluid feedthroughs. Liquid nitrogen or other cryogenic fluid can be used to achieve low temperatures. When cryogenic temperatures are not required, a pumped closed-loop refrigerant system can be employed using methanol or another refrigerant. The refrigerant flows through a heat exchanger external to the vacuum, in a cryogenic fluid bath such as dry ice or liquid nitrogen. The temperature of the refrigerant can be controlled through the manipulation of the pump throughput and by varying the cryogenic fluid contact with the heat exchanger.

Sample preparation methods depend upon the desired moisture content (i.e., dry or icy) and compaction state (compacted or loose) of the regolith simulant. Regardless of the final sample state, all regolith simulants are subjected to an initial drying phase following ASTM D2216-10 (2010) in a vacuum oven to ensure a minimum of residual moisture. The dried regolith simulant is then kept in sealed containers (mason jars) until selected for sample preparation. The jars are rotated axially and end-over-end at least ten times prior to use to avoid selection bias due to particle settling.

Dry samples are made from dried simulants taken straight from the sealed containers and therefore contain minimum water content (considered negligible). Uncompacted samples are prepared by free-pouring loose simulants into the BSC in a randomized manner, using a small scoop. This reduces the tendency for stratification of the sample due to particle settling during pouring. Simulants within the BSC are carefully leveled to the sample container lip without compacting the upper portion, allowing accurate volume determination. Compacted samples are prepared in a layered fashion using a modified Proctor method (ASTM Standard D698-07, 2012) that relies on a surcharge weight to impart 13.8 kJ of energy to the sample surface. Layers are compacted within the CSC to a desired (constant) density using the surcharge height (above the container lip) as a guide. The number of layers in each sample is kept constant (typically three layers of ∼200 g simulant each), a method empirically determined to provide the highest sample repeatability.

Icy samples are prepared using the vapor deposition method described earlier in this section or using a liquid water method. The choice of method is dictated by considerations of sample fidelity and experimental practicalities. The vapor deposition of ice more closely resembles the process by which ice is deposited in regolith on many solar system bodies and therefore is thought to have higher simulant fidelity. This is an active area of research, and the fidelity of vapor-deposited ice relative to liquid-formed ice is not well known at present. Liquid-formed icy regolith is being used in many research labs due to the ease of quickly forming large volumes of uniform material (Kleinhenz, 2014 and Zacny et al., 2015).

The liquid water method of icy regolith sample preparation begins by adding de-ionized water in a desired amount (e.g., 3%, 5%, or 12% by mass) using a syringe to a known mass of the dried simulant in a plastic bag. The bag is sealed, and the moist simulant is mixed by hand to ensure even distribution of the water and to avoid clumping. The mixture is then left in a sealed container for a minimum of 12 h. Simulant, water, and mixture masses are recorded before and after to ensure accurate moisture content. Compaction of the moist simulant to a specified density is achieved using a vibratory method (ASTM Standard D4253-16, 2016), again relying on a surcharge weight to provide the necessary compaction force. A specialized compaction setup was designed to achieve accurate compaction and enable each sample to be sized to the required cylindrical dimensions (Fig. 6). Following compaction and sample extraction from the compaction assembly, an excess compacted material above the sample container lip is carefully removed using a straight-edge. The sample is then placed in a freezer to cool to −20 °C for at least 12 h.

Sample volumes are controlled and determined in two ways: using the surcharge height above the sample container to indicate the sample volume during compaction and, after removal of excess material, using the sample container volumes after compaction. Sample densities are then calculated using the known masses of moist simulants (measured during compaction) or the measured mass of the sample after compaction.

The IEP can accommodate a wide range of probes that interact with sample surfaces in different ways. Several probe configuration concepts are shown in Fig. 7. Many more probe configurations and motion profiles can be devised than the motion profiles shown. Probes can be mounted directly on the force-torque sensor or integrated into a custom mounting bracket. In addition to the instrumentation described here, the probe can be instrumented with custom sensors.

The behavior of the IEP in penetrating the regolith simulant was characterized using JSC-1A, a lunar mare simulant. Initial test penetrations were made into loose and compacted simulants at ambient pressure and under vacuum. Additional penetrations under vacuum into the simulant at varying degrees of ice saturation served to characterize the IEP under significantly higher loads. Forces and torques in all three principal directions were recorded during insertion as well as for a brief time after probe motion ceased.

The performance of the IEP in penetrating a loose dry regolith simulant is presented in Fig. 8. The 6 mm diameter, 30° probe tip was inserted into a dry bed of JSC-1A in the CSC at a constant rate of 0.25 mm/s to the maximum extension of the probe, around 30 mm depth depending on the initial surface height of the sample. At the maximum extension at 100–120 s, the probe was stopped, and the stress relaxation of the simulant was recorded for at least 50 s. The stress relaxation period appears as a step change in the force curve.

Six penetrations into loose regolith are presented in Fig. 8(a): three at the ambient pressure in Golden CO (617 Torr) and three under a vacuum pressure of 50 mTorr, all at an ambient temperature of ∼295 K. Ambient pressure tests were conducted in the open air. Penetrations at ambient pressure show an initial linear increase followed by a decrease in slope, reaching a maximum penetration resistance of 0.28 N–0.38 N at ∼30 mm depth (corresponding to ∼120 s of penetration time). Under vacuum, the maximum penetration resistance increased by ∼0.1 N (0.38 N–0.48 N). The shape of the curves is similar to those at ambient pressure but appears to lack the aforementioned change in slope after 50 s. Loose samples had initial densities ranging from 1.50 g/cm³ to 1.55 g/cm³, and no significant correlation was seen between density and maximum resistance. Each test run was conducted with a different JSC-1A sample.

The increase in penetration resistance of samples tested under vacuum conditions relative to ambient pressure in air is likely due to a small increase in density of the vacuum samples. The pumpdown procedure can take upwards of 20 min to reach 50 mTorr, during which time vibrations due to the vacuum scroll pump, as well as environmental vibrations and movement of gases out of the pore space of the loose simulant, may contribute to a small increase in density. Such an increase in density would serve to increase resistance and reduce variation in the samples, as well as eliminating or reducing the decrease in slope observed relative to the sample tested at ambient pressure. The decrease in slope seen with some penetrations may be due to minor surface compaction that occurred while scraping away the excess material from the sample surface during preparation.

Figure 8(b) presents six penetrations into compacted JSC-1A under ambient and vacuum pressures. While more variability is seen in the samples tested under ambient pressure (penetration resistances of ∼4.5 N–6.3 N) compared to the vacuum-tested samples (5.5 N–6.8 N), the penetration curves are quite similar and show a non-linearly increasing resistance. Two exceptions are prominent: one ambient sample that resulted in a significantly lower maximum resistance and one sample under vacuum that displayed characteristics of layering in the sample, as seen by distinct steps in the curve. The vacuum-tested samples resulted in a modestly higher average maximum penetration resistance. Compacted samples had densities ranging from 1.78 g/cm³ to 1.84 g/cm³, and no significant correlation was observed between density and maximum resistance.

While stress relaxation profiles of those samples presented in Figs. 8(a) and 8(b) have not been analyzed in detail, they appear similar in shape and maximum stress drop. Additional information can be gleaned from the curves, and more analysis is required in order to determine the effect of density and testing pressure on the relaxation profile.

Figure 9 presents the results of three penetration tests into compacted regolith: one with an ice content of 3 wt. % and three with 12 wt. %, under a vacuum pressure of ∼50 mTorr and at ∼143 K. The penetration occurred at 0.25 mm/s, and the penetration was programmed to cease when 200 N of axial force was reached.

The single penetration into the 3 wt. % ice sample is presented in two parts: a portion that lies within the complex loading range of the force-torque sensor and a portion that lies outside the complex loading range. What can be seen is a complicated curve that shows significant variation, though it follows a general trend of increasing resistance to 200 N at a depth of ∼7.5 mm at 30 s. The range of resistances at the same conditions is due to variations in density and porosity, and other heterogeneities intrinsic to natural materials. The large drop at 25 s is due to the load and torque exceeding the calibration extended range; however, settling of the support plate standoffs into the IEP interface rails under load cannot be ruled out. The bulk density of the 3% ice sample was 1.53 g/cm³.

Three penetrations into a 12 wt. % ice sample occurred in sequence while the sample was at the cryogenic temperature. Penetrations were sufficiently far from each other (∼15 mm) to minimize stress effects from the previous tests. Figure 9 shows that the strength of the 12 wt. % samples is considerably higher than the 3 wt. % ice sample due to the increased ice content and the cryogenic strengthening of the ice. The maximum penetration resistance of 200 N was reached at less than 1.25 mm depth in each test. Two of the three insertions saw slight variations in the linear increase of resistance between 150 N and 175 N, potentially due to the settling of the standoffs under load or layering within the sample. All penetrations into the 12 wt. % sample fell within the complex loading range of the force-torque sensor due to the shallow depth of the penetration which did not allow significant torque to build.

Stress relaxation characteristics of the icy sample have not been analyzed and should not be considered as presenting information on the behavior of the material at cryogenic conditions, as the probe tip was not cooled prior to insertion. Therefore, the relaxation curves have a contribution of thermal conduction (and likely melting of ice around the probe tip).

Current work with the IEP is focused on geotechnical properties of dry and icy regolith lunar simulants. Experimental conditions range from mTorr pressure to ambient pressure in a dry nitrogen atmosphere, temperature ranges from ambient to cryogenic (∼100 K), and regolith compaction ranges from low (∼1.5 g/cm³) to high (∼1.8 g/cm³). This work is relevant to understanding the strength of ice cemented regolith in permanently shadowed regions of the Moon. Future work will include cohesive force measurements using stress relaxation and comparison of mechanical properties of icy regolith formed by vapor deposition versus the liquid water method.

Future work includes studies of small body surfaces. As a ground facility, the IEP operates in a 1-g environment; therefore, the experimental design for small bodies that exist in a microgravity environment must be carefully considered. A gravity-induced effect that must be mitigated is the gradual compaction of the granular material in gravity. This can be addressed by preparing “loose” regolith simulants and conducting experiments soon after preparation, as was shown in Fig. 8(a). Gravity-induced compaction can be further mitigated by reducing the weight of the simulant. One approach is to incorporate low-density materials into the small body simulant mixture, such as 3M glass microspheres that have densities on the order of 0.1 g/cm³. The reduction of gravity-induced compaction would enhance IEP experiments focused on cohesive forces, which are thought to be governed by van der Waals and electrostatic forces (Sánchez and Scheeres, 2014). In addition, the vacuum environment may have an effect on cohesive forces. It has been shown that absorbed gasses lower surface energy and thus alter grain adhesion (Israelachvili, 1991). Frictional forces may also be affected by the vacuum level (Karafiath and Mohr, 1969). Thus, future work with the IEP will include the preparation of small body simulants that enable experiments to elucidate the cohesive forces of fine grained regolith on small bodies.

Long term work includes additional capabilities that can be added to the IEP, including instrumented probes, new probe geometries, high torque motors (for greater loads), and higher load force sensors. Finally, to advance the understanding of small body regolith mechanical strength, the IEP or a later iteration of the instrument could be developed for microgravity platforms, such as micro-gravity aircraft and the International Space Station. The authors welcome suggestions and collaboration with the community for other topics of future work with the IEP.

A new experimental capability has been developed at the Colorado School of Mines’ (CSM) Center for Space Resources (CSR) as part of the Institute for Modeling Plasma, Atmospheres, and Cosmic Dust (IMPACT) of NASA’s Solar System Exploration Research Virtual Institute (SSERVI), called the ISRU Experimental Probe (IEP). The IEP is a versatile instrument for the study of physical properties of extraterrestrial surfaces and a facility for the development of ISRU systems that interact with regolith surfaces. Regolith mixtures can be studied from low to high compacted states and pressures from ambient to as low as 10⁻⁷ Torr. With sample containers designed for vapor-grown ice formed in regolith or icy regolith mixtures formed outside the vacuum chamber, the regolith temperature can be varied between −196 °C and 150 °C. Probes can be designed to conduct penetration, scraping, gripping, pull-off force, load bearing tests, drilling, and anchoring tests. Initial penetration resistance tests of dry JSC-1A regolith at loose and compacted states in air and vacuum, in addition to icy regolith at 3 wt. % and 12 wt. % water in compacted JSC-1A at ∼143 K and ∼50 mTorr, were shown, which demonstrate the capabilities of the IEP.

This work was supported by the IMPACT node of NASA’s SSERVI. The authors gratefully acknowledge the contributions of Colorado School of Mines students Steven Hackenberg, Brandon Tortorelli, Ted Rand, Jacob Drozdowicz, and Colby Moxham.