Enabling Ice Core Science on Mars and Ocean Worlds

The MSIS instrument can be best described as having two main components: Melter and Sublimator (Figure 1).

On Earth, drilling, extraction, and handling processes result in contamination of an ice core’s outer surface [11]. Accessing pure, uncontaminated regions of the core and rejecting contaminated areas is a requirement for some downstream (elemental, chemical, biological) analyses [12]. Thus, like the melter design from McConnell et al. [6, 8, 7, 9, 10], the MSIS Melter component divides an ice stick into three distinct cross-sectional areas as it melts, providing individual access to the inner, middle, and outer core sections.

MSIS is currently designed to accommodate 3.3 cm x 3.3 cm ice sticks, following [6], but in principle could be adapted to any core cross section. Using the three-ring design adopted from [6], water from the outer 70% of the core is discarded under ideal operating conditions. The inner, most pristine 10% is collected separately from the middle 20% to facilitate downstream analysis applications requiring different levels of purity and volume flow rate. The geometry of these rings can be redesigned to suit any specific set of downstream flow requirements.

The Sublimator component, sitting atop the Melter, enables preconcentration of material for in-situ analyses or Earth return of residues, avoiding the complexities involved in storing and returning large samples to Earth. MSIS will be able to preconcentrate organic samples through near-full sublimation then controllably melt a reduced sample, allowing for sensitive in-situ characterization of non-volatile materials such as organics.

The Melt Head is a stainless steel, 3D-printed structure that sits compressed against a heated copper plate with a layer of thermal paste at the interface. The Melt Head has two  3.5-mm-tall partitioning walls that separate the ice stick as it melts into the inner, middle, and outer Melt Head cavities (Figure 2). Copper was not used for these partitioning walls because it would be chemically reactive. Also, a stainless steel Melt Head does not permit trace metal-clean operation like a silicon carbide one would; however, stainless steel is an inexpensive compromise for our initial testing. Melt Head cavity dimensions were adapted from the dimensions of the McConnell et al. design [6, 8, 7, 9, 10], with the outer cavity having a large enough area to melt through a 3.3 cm x 3.3 cm ice stick. Within each cavity are drainage holes that lead into the fluidic channels and to the outlet ports on the sides of the Melt Head.

The Frame and Bottom Cover are Formlabs Clear resin-printed structures that compress the Melt Head to the Heating Plate via bolts and thumb nuts. The Frame also serves as a mounting point for the Fluidic Inserts. The Melt Head is press-fitted by its four corners into the Frame and held only at those locations. Gaps were designed in the Frame to separate it from the Melt Head and the Heating Plate, minimizing heat transfer from these components to the plastic; This is for safe handling of the MSIS system and to conduct as much heat directly from the Heating Plate to the Melt Head as possible. For similar reasons, the Bottom Cover minimizes its contact with the Heating Plate while maintaining structural integrity.

The Fluidic Inserts interface with the outlet ports of the Melt Head. They were designed to mimic nuts commonly used with fluidic tubing and ferrules. There are four Fluidic Inserts: two host 1 fluid line, one hosts 2 fluid lines, and another hosts 3 fluid lines. Fluidic tubing in this design is $1/8$-in-outer diameter (OD), $1/16$-in-inner diameter (ID), fluorinated ethylene propylene (FEP) tubing from IDEX Health & Science. A custom design for hosting the fluidic tubing was preferred over a commercial product because it can easily be adapted for various design configurations and tubing sizes. Threaded inserts and bolts are used to mount the Fluidic Inserts to the Melter Frame. When tightened down, the Fluidic Inserts compress the tubing with custom ferrules against the Melt Head so that meltwater can be transported from the Melter component to downstream devices via the tubing. O-rings between each custom ferrule and the Melt Head add another measure in preventing meltwater leakage in the system.

Four pumps were added along the fluid lines to provide additional control over the flow rate and to prevent the meltwater from overflowing beyond the Melt Head. Maintaining a higher flow rate at the outer cavity and lower flow rates in the inner two cavities is important for reducing contamination of the inner cavities [12]. Two small, 3 V diaphragm pumps yield higher flow rates and are appropriate for draining the outer cavities, which nominally send contaminated meltwater to waste. Peristaltic pumps are preferred for the middle and inner cavities because the meltwater does not contact the pump internals, preventing sample contamination. One peristaltic pump is used for the middle cavity drain lines, and another for the inner cavity. Each Melter pump can be powered by varying voltages for different flow rates.

The Heating Plate is a $1/2$-in-thick piece of copper manufactured to hold up to five, $1/4$-in-diameter cartridge heating elements. The interfaces between the cartridge heaters and the Heating Plate are covered with thermal paste to lower any contact resistance. Copper was selected as the Heating Plate material because it has high thermal conductivity, $k\approx 385\text{ }\frac{W}{m\cdot K}$, while having a relatively reasonable price compared to more thermally conductive materials like silver or diamond. The MSIS Heating Plate is much thinner than that of other melter designs, meaning the thermal load is lower for the cartridge heaters; This allows for increased thermal efficiency in our instrument. Also, the Heating Plate tends to oxidize with use, so semi-frequent sanding of the surface is done to reduce the thermal load.

We now describe the electrical system that controls the Melter (Figure 3). Inside the Heating Plate sits the $1/4$ in DC cartridge heaters, powered with an external power supply up to 24 V, which is expected on a typical spacecraft bus. In 12 V operation, each of the five cartridge heaters produces 14 W, therefore the theoretical maximum power supplied to the melter system is 70 W; 12 V was selected for operating the cartridge heaters in the lab, as 70 W of power was adequate for MSIS benchtop testing, and working with lower voltages is good safety practice. Per the slim Heating Plate, the cartridge heaters lie horizontally, in contrast to their vertical orientation in other melter designs [12]. A NOYITO 5 V relay board is used to switch the cartridge heaters via an Arduino Uno microcontroller. Each relay can switch loads up to 10 A and 30 VDC. The relay board in this design was selected because it offers low and high-level triggering, accepts serial commands, and has enough channels to switch five cartridge heaters.

Temperature is monitored and controlled by a closed-loop Proportional-Derivative (PD) system that relies on temperature measurements taken by two thermistors mounted to the Melt Head (Figure 5). Each thermistor is first in series with a 100 k$\Omega$ resistor and powered by 2.5 V. As the temperature at the thermistor rises, its internal resistance is lowered, and the Arduino microcontroller can measure a larger analog voltage. The 10-bit analog-to-digital converter in the microcontroller processes this analog signal using a lookup table, comparing a given voltage value to an experimentally determined range of voltages versus temperatures.

A temperature setpoint is specified in the microcontroller. The control system evaluates its current temperature versus the setpoint and begins heating. Then, the PD controller detects when the current rise in temperature is sufficient to turn off the heaters, at which point the control system maintains the temperature setpoint with varying bursts of heat. This PD controller is adaptable to different setpoints and starting temperatures.

The Sublimator component of MSIS (Figure 4) consists of a resin-printed cylindrical sublimation chamber that adapts to the top of the Melter component. The Sublimator is sealed against the Melter by a face-seal O-ring. This allows containment of the ice sample, whether it be fragments or a full ice stick. The current design can house an ice stick measuring 3.3 cm $\times$ 3.3 cm at the base and a height of up to 15 cm. The sublimation chamber is sealed at the top by a cap that screws into the chamber, and three horizontally mounted Lee Company (Westbrook, Connecticut) solenoid valves are attached. The solenoid valves expose the ice core to vacuum when opened. Valve opening is controlled by three solid-state relays connected to the Arduino microcontroller. Through serial command, the operator specifies the number of valves they wish to open, depending on the desired outflow. A CoolCube 50R “hit-and-hold” circuit is placed in series between each relay and each valve. The valves require 12 V to actuate but do not require this voltage to remain actuated. The “hit-and-hold” circuitry reduces the operating voltage to 6 V to prevent overheating while keeping the valve open for extended periods (e.g., during sublimation).

The outlet to the solenoid valves extends into barb connectors to allow the channeling of water vapor into a common moisture trap, which consists of a sealed basin with a cylindrical water and particulates filter. This helps protect the downstream vacuum pump to which the vapor is being vented. Such an adapter may not be necessary in a low-pressure atmosphere, as on Mars, where additional vacuum may be superfluous depending upon the local pressure.

In its operation, the Sublimator would reduce the pressure in the sublimation chamber to below the triple point of water and raise the temperature of the Melt Head to 0^∘C (Figure 5). To simulate Martian conditions more accurately, the Sublimator would reduce the pressure to 1 Torr. Lower pressures can be attained with strong vacuum sources, but pressures significantly lower than 0.75 Torr can negatively impact the sublimation rate by hindering convective heat transfer into the ice [13]. Higher sample temperatures also favor faster sublimation rates; however, we must maintain a lower temperature at the Melt Head throughout the initial exposure to vacuum to prevent premature melting [13]. Since sublimation is endothermic, the temperature of the ice stick will decrease throughout the process. Our control over the Melt Head temperature allows us to set temperatures appropriate for any state of the ice stick during the sublimation process.

Depending on the application, the Sublimator does not require all the functionalities of the Melter. The Sublimator component is designed to both interface with the entire Melter component or only the Heating Plate. This allows for the use of a thermal plate inside a vacuum chamber to heat a sample through the Melter Heating Plate, thereby eliminating the need for a cartridge heater-relay setup where space is limited (Figure 5). For other applications, only a reduced solid sample may be desired, so melting may not be necessary.

The MSIS instrument has thus far demonstrated melting of a lab-made ice core sample in a benchtop setting. Upon the success of 3D printing the Melt Head, a future MSIS design could involve inner fluidic channels with more complex geometry, with outlet ports on a single side of the instrument; this modification would make for a more organized system in the lab and for future space missions. The next Melt Head design could include additional measures for preventing contamination of the inner core sections. For instance, the Melt Head partitioning walls could be made taller to ensure separation of the central core from the outer portion. Also, ultraviolet or infrared LEDs could be placed in a ring pattern at the outer base of the Melt Head to melt the contaminated outer region of the ice core before the inner regions, reducing the risk of meltwater mixing across Melt Head cavities. The next critical step in instrument development would be processing a Greenland or Antarctic ice core sample in the lab with MSIS, integrated with downstream analytical devices for processing validation. Upon such validation and necessary modifications, the MSIS instrument could begin utilization in the lab or in the field for Earth-based ice core studies.

At its current state, MSIS requires human operation and sample loading. A future vision is to make the MSIS system autonomous and compatible with a rover platform exploring ice deposits on other bodies. Automating a sublimator would require robust monitoring and data acquisition capabilities. MSIS currently relies on a user judging when enough sublimation has taken place based on the size of the sublimated ice sample. To process samples back-to-back, a cleaning process would be needed within its fluid lines. A one-size-fits-all approach to cleaning debris of different sizes would be challenging. However, the Sublimator component of MSIS does not limit itself to processing fully intact ice cores.

Considering the deposition of dust and debris on Mars, subsurface Martian ice has the potential to be stratified like permafrost ice deposits on Earth [11]. The drilling process would inevitably accrue regolith in the ice cores which may pose a problem to the MSIS plumbing network. In addition, near-surface Mars ice could be more porous due to lower gravity and there may be a commensurate increase in the depth of pore close-off; this could lead to more fragile ice.

Rougher drilling techniques tend to produce ice “chips,” which are fragments defined to have a 6 mm maximum dimension [11]. These chips lower the resolution and are difficult to decontaminate but are still considered useful for science.

Sublimation is an experimentally inefficient process [14] which does not pose an issue for benchtop experimentation on Earth but complicates its context in a space instrument. The convenient presence of low atmospheric pressure on Mars and other celestial bodies [13] may allow for more efficient sublimation if it reduced or eliminated the need for a vacuum pump, as is needed on Earth. Without a vacuum pump, the sublimator instrument would save upwards of 1 kW of power and consume an estimated 76 W at its maximum power setting. Up to 300 W of power consumption in a space instrument is realistic for the capabilities of some rovers (e.g., Mars Oxygen ISRU Experiment, or MOXIE, instrument on Perseverance [15]) and serves as our metric when characterizing the MSIS power consumption.

The long-term objective is for future versions of MSIS to operate beyond Earth. With the Artemis III crew proposed to target regions near the Moon’s South Pole¹¹1https://www.nasa.gov/press-release/nasa-identifies-candidate-regions-for-landing-next-americans-on-moon, lunar water ice analyses are likely to take place. MSIS could contribute to Artemis III demonstrations of in-situ resource utilization for pre-processing lunar water ice for oxygen and hydrogen extraction to be used in fuel and life support systems. This would serve as a precursor to Mars missions with similar aims in processing ice to serve as an ingredient for propellant production while potentially searching for traces of life [16]. Also, in the context of preparing for the first human mission to Mars, potentially to mid-latitude regions in the 2030s, NASA has identified sampling subsurface ice as a critical science requirement for searching for life and investigating the evolution of Mars’ climate [11]. If the Artemis III and Mars crews have the means to retrieve and manually handle ice core-like samples, MSIS—at its current state—could prove beneficial in preparing lunar water ice samples for in-situ refueling demonstrations, as well as helping meet NASA’s Mars human mission goals.

An important caveat of our proposed analyses is that the formation and preservation mechanisms for different ice deposits on Mars is under study and without ground truth [17]. At present, this means the availability of contiguous ice core records on Mars is unknown. This could complicate our ability to establish a high-resolution contiguous history from Mars ice.

Mars’ obliquity, the angle between its orbital plane and its spin axis, is currently around 25^∘, and it varies on a 120-thousand-year (Ky) timescale, with the amplitude of this cycling varying chaotically on  1.2 My timescales [18, 19]. During periods of high obliquity, polar ice is mobilized and falls as snow in mid-latitude and equatorial regions [20]. This process has resulted in repeated mid-latitude “ice ages” for an estimated total of  680 My out of the last 3.6 billion years (Gy). Thus, it is reasonable to assume that current remnants of these ice ages are mainly locally-contiguous deposits, even if a global contiguous record is not expected.

Based on analysis of nutrient-limited Earth environments, modeling of energy limitations, and other factors, estimates for cell abundances at Ocean Worlds such as Enceladus and Europa may be as low or lower than 1000 to 100 cells/ml [44]. Most instrumentation flown or under development may not have adequate analytical limits of detection without concentration. Concentration approaches include filtration, flow focusing, and sublimation, among others. A prototype sublimator was previously developed and was able to demonstrate the freeze-drying of a 5 ml sample in 5 hours at under 5 W power [13]. This was achieved by accounting for the available cold temperatures and low pressure expected to be available under Europa or Enceladus surface conditions. Even so, sublimating large volumes of ice is likely infeasible due to the high energy cost of sublimation; we can expect to have higher energy budgets on Mars where higher power operation could process larger volumes via sublimation. Cold and low-pressure conditions are also available on Mars, where surface pressure is  600 to 700 Pa; a vacuum pump will be required to maintain pressure below the triple point of water,  610 Pa.