Formation and Stability of Salty Soil Seals in Mars‐Like Conditions. Implications for Methane Variability on Mars

Methane spikes observed in the Martian atmosphere require the abrupt release of large amounts of methane from the Martian subsurface. The mechanism for such release has not been identified. We tested whether gas traps can form under Mars‐like conditions in the shallow Martian regolith due to salt migration in the icy soil. Experiments were performed on various soil samples in a Mars Simulation Chamber comprising different perchlorate salts and water concentrations mixed with JSC Mars‐1A. Inside the chamber, the samples were exposed to a range of temperatures from −20°C to 10°C and maintained CO2 gaseous pressure between 8 and 10 mbars. As a methane analog, Neon was injected periodically underneath the soil sample. It was found that over a wide range of Mars‐like soil parameters, a gas impermeable soil seal can form over a relatively short period (3–13 days) but requires 5%–10% of perchlorate salt content in the soil. It was determined that such a seal could sustain several mbars of neon above the Martian atmospheric pressure in the soil. Based on our experiments, substantial amounts of gaseous methane may accumulate under the soil seal and get released abruptly into the atmosphere upon seal cracking. An abrupt release of methane from the shallow subsurface may help explain methane variability at the Martian surface, as the Mars Science Laboratory detected.


Introduction
Atmospheric methane on Mars is of great interest because of its possible association with potential microbial activity in the Martian subsurface.Besides microbial life, methane is a key molecular indicator of important abiotic processes that might occur on Mars, namely Fischer-Tropsch synthesis (Mumma et al., 2010) or degradation of meteoritic organic matter (Keppler et al., 2012).All these processes could very well lead to methane reservoirs in the Martian subsurface that are later released into the atmosphere.
All well-established photochemical models state that the photochemical lifetime of methane in the Martian atmosphere is about 300-500 years (Atreya et al., 2007;Krasnopolsky et al., 2004;Summers et al., 2002).The photochemical lifetime is much longer than the time of atmospheric mixing, suggesting that methane in the atmosphere would be evenly distributed over the planet and would not show change for many years (Summers et al., 2002).Yet, several independent groups observed methane on Mars by both ground-based telescopes and Martian orbiters (Geminale et al., 2008;Krasnopolsky et al., 2004;Muma et al., 2009).Mumma's study reported that methane column abundance exhibited both spatial and temporal variability at about ∼10 ppb levels.However, those methane detections were challenged by Zahnle et al. (2011).
In August 2012, the Mars Science Laboratory (MSL) Curiosity rover began roaming on the surface of Mars in the Gale crater and continues to do so until now.The Tunable Laser Spectrometer-Sample Analysis at Mars (TLS-SAM) instrument on the Curiosity rover not only detected methane but also determined that its abundance is variable (Webster et al., 2014), sometimes changing by up to 6 ppbv within a few sols (Martian days).Further TLS-SAM observations established that the average "background" methane abundance of ∼0.4 ppb correlates with seasons.In the warmer spring and summer months, methane abundances increase up to 0.65 ppb and decrease again during the colder winter season to 0.24 ppb (Webster et al., 2018).In June 2019, TLS-SAM detected a spike of 20.5 ppb of methane in the atmosphere, even though only 1 ppb was detected earlier (Webster et al., 2021).
To add to the methane mystery on Mars, recent observations by the ExoMars Trace Gas Orbiter (TGO) of the ExoMars mission did not detect any methane above altitudes of 4-8 km (Knutsen et al., 2021;Korablev et al., 2019), even though TLS-SAM continued to detect methane at the ground level in the Gale crater.There was an attempt to reconcile TLS-SAM and TGO methane data by proposing diurnal variations (Moores et al., 2019).Methane would accumulate in the planetary boundary layer (PBL) at night when TLS-SAM conducts its measurements at the ground level.Then, methane would disperse during the daytime when PBL's thickness increases and TGO makes its methane observations.Potentially, the diurnal methane variation mechanism can "hide" the background (0.4 ppb) of methane from TGO.However, diurnal variations cannot explain the detection of either occasional methane spikes or seasonal methane variability.Recent modeling studies (Luo et al., 2021;Viudez-Moreiras et al., 2021) agree that reconciling TGO and TLS-SAM methane data requires some local, small, unknown sources of methane within the Gale crater.
However, to explain rapid variations of methane abundance in the atmosphere (like the methane spikes or methane seasonality in Webster et al. (2021)), methane has to be released abruptly rather than slowly and gradually seeping from the deep interior (Oehler & Etiope, 2017).Not only does the methane-release process have to be abrupt but it also has to happen locally in the Gale crater.Moreover, the release of methane from the subsurface has to be somewhat dependent on seasons.Such a process of methane release is currently unknown.
The Phoenix mission discovered permafrost just centimeters below the surface (Smith et al., 2009) and perchlorate salts in the Martian regolith (Hecht et al., 2009).Other salts are also present in the Martian regolith (Clark & Van Hart, 1981;Thomas et al., 2019).The Curiosity rover confirmed the ubiquitous presence of perchlorate salts in the Martian subsurface (Kral et al., 2016).
The presence of perchlorates and shallow ice points to the possibility of some amount of mobile water in the shallow Martian subsurface due to the low eutectic point of perchlorates.Under low-pressure conditions, water can migrate to the surface and evaporate once it reaches the top.With the presence of salts in the ground, some salts are inevitably dissolved in the water and transported to the surface via water migration (see Figure 1).Once most of the water evaporates or ice sublimates, hydrated salts are left in the top few centimeters of the soil.We hypothesize that the soil with migrated salts can become cemented similar to the terrestrial duricrust (Lewinger et al., 2018;Miller, Hibbitts, & Mellon, 2020;Miller, Mellon, et al., 2020), and form a solid seal barrier capable of trapping volatiles underneath.Under such soil seals, methane gas, regardless of its origin, can accumulate.If those areas are disturbed (e.g., impacts, seasonal temperature variations, or motion of the rover itself), then methane can be released abruptly into the atmosphere.Our laboratory experiments aimed to test the formation, strength, and stability of salty soil seals formed under Mars-like (pressure, temperature) conditions.

Materials and Methods
Mars-like simulations of the shallow permafrost layer were performed using a Mars Simulation Chamber (Figures 2a-2c).In this section, we describe the general structure of the Chamber (Section 2.1), sample holder details (Section 2.2), sample preparation procedure (Section 2.3), and experimental run procedure (Section 2.4).

Chamber Description
The Martian Simulation Chamber (Figure 2a) is made up of a glass dome that sits on top of a cylindrical steel housing.The chamber was connected to a vacuum pump capable of pumping down the chamber pressure to as low as 0.1 mbar, which created a vacuum seal between the dome and housing.Figure 2b shows a close-up view of the sample in the chamber.Above the sample holder, there was an aluminum foil dome which had a light source and a camera.The pump evacuated air from the chamber and beneath the sample (Figure 2c) and used valves 3, 5, and 6 to control which area was pumped down.Between the pump and the chamber sat a moisture collection system that used the low temperature of liquid nitrogen to condense and trap water before it entered the pump.The moisture collection system was used only during short periods in the run when there was high moisture content coming off the soil sample inside the chamber (usually at the beginning of the run).Inside the chamber sat another smaller cylindrical metal housing (Figure 2c) in which the sample holder was placed.During experimental runs, CO 2 was continuously injected into the chamber through valve 7 (Figure 2c) and pumped down from the chamber at the same time through the valve 6 (Figure 2c).Such a flow-through system allowed the setting of the desired CO 2 pressure in the chamber above the soil sample.The chamber's CO 2 -flowthrough system also ensured that water vapor from the sample was steadily removed from the chamber.However, at present, there is no humidity sensor in the chamber.Before each experimental run, the entire system was pumped down and purged with nitrogen gas through valve 2.
The sample holder was placed on the cold plate and inside the metal housing (Figure 2b) with an open top.A solenoid valve was used to control the flow of liquid nitrogen through the cold line, which went through the cold plate (Figure 2c).Therefore, by adjusting the solenoid valve, it was possible to control the temperature of the sample and to bring sample temperatures as low as 100°C (Figure 2c).The chamber setup had the option to change the surface temperature by using the light source located in the aluminum foil dome (Figure 2b), but the light source was not used in this study.Several thermocouples were used to monitor the temperature of the cold plate, sample holder housing, cold line temperature, and the gas temperature in the chamber.Baratron 1 was used to measure the pressure inside the capillary tube beneath the sample and Baratron 2 was used to measure the pressure inside the chamber above the sample (Figure 2c).The chamber had a separate capillary tube that was connected to a mass spectrometer (not used in this study).The soil sample was placed in a sample holder (Figure 3).

Sample Holder Description
The sample holder was a thin stainless steel cylinder that sat on the cold trap plate and was inside the larger cylindrical metal housing (Figure 2b) inside the chamber.The sample holder was 6.7 cm in internal diameter and 9.5 cm in depth.To have better contact between the sample material and the metal walls, the sample holder was built with ridges and grooves that started about halfway up the holder and continued to the top (see also Figures S3 and S4 in Supporting Information S1).Ridges and groves provided a longer path for volatile molecules to move along the metal wall of the sample holder.Thus, if a microcrack between the steel wall and the frozen soil developed at some depth, there was still sufficient soil column mass along the wall to seal volatiles in the soil.Inside the holder, there were two thermocouples-one that extended 2 cm up from the bottom to measure the temperature of the bottom of the sample and another that extended 6 cm from the bottom to measure the temperature of the upper part of the sample (∼2 cm below the top of a sample's soil level).Thermocouples could measure temperatures with an accuracy of 0.01°C.In this study, the only relevant temperature readings were from the top thermocouple because salty soil seals were formed toward the top of the sample.In the middle of the sample holder, a capillary tube extended 3 cm up from the bottom of the holder.The capillary tube was connected to the neon tank and had a 0.13-mm hole at the top to inject the neon beneath the sample.

Sample Preparation Procedure
Depending on the experiment, samples of different salt and water concentrations were mixed with JSC Mars-1A.JSC Mars-1A is the Martian regolith analog due to its similar porosity and bulk density to soil found at the Viking landing site on Mars (Seiferlin et al., 2008).Before each experiment, JSC Mars-1A was washed with deionized (DI) water for at least five consecutive times to remove any residual salt from previous experiments.The "desalted" JSC Mars-1A sample was then placed in a small pan and dried in the oven at +70°C for at least 20 hr to remove the introduced moisture.Dry and desalted JSC Mars-1A was then mixed with DI water or a solution of DI water and salts.The salt concentration was reported in M salt / (M dry soil + M salt ) % (see Text S1 in Supporting Information S1).Salt was dissolved in DI water at room temperature before mixing with soil.For the consistency of various experimental runs, it was important to ensure a uniform salt distribution in the soil sample before the experiment.We determined that JSC Mars-1A became completely saturated with water at M water /(M dry soil + M water ) ∼35%-37% by weight.Therefore, the initial water content at 40% by weight was kept for all permafrost samples.After mixing, the salty soil- water mixture was poured into a sample holder (Figure 3) and placed in the freezer ( 20°C) overnight for at least 15 hr (see Figure S1 in Supporting Information S1).After the sample was completely frozen, it was placed in the Mars Simulation chamber for an experimental run.
The samples were then placed in the Mars Simulation Chamber.Inside the chamber, samples were exposed to a range of temperatures from 20°C to +5°C and the chamber pressure was kept between 8 and 10 mbars.2-5 mbars of neon were injected periodically beneath the sample.Neon was used as an analog to methane because it had a similar molecular weight to methane.Like methane, neon was not reactive with soil and had a similar diffusivity in water.Therefore, neon was expected to diffuse similarly to methane in the soil, but neon was much safer to work with in the laboratory.

Experimental Run Procedure
Before each experimental run, the sample holder, the cold plate, and the capillary inlet were cleaned with ethanol to remove any residual salt crystals or dust particles from previous runs.The empty chamber was then pumped down to ∼1-5 mbar at room temperature and kept at low pressure for at least 12 hr to remove the residual moisture from previous runs.Each time during placement of the sample holder into the chamber a new copper washer was used to connect the capillary tube (Figure 2b) and its flex line and then tightly screwed the two together.The thermocouples on the sample holder were connected to their corresponding connectors in the chamber.Ethanol was used to carefully clean the points of contact between the rubber on the base of the glass dome and the metal housing on the chamber (Figure 2b).
After the sample holder with a sample was placed in the chamber, each experimental run followed the following steps: 1.The glass dome was placed on the metal housing and the chamber was pumped from above the sample by opening valve 6 (chamber) and beneath the sample by opening valves 3 and 5 (capillary tube) (See Figure 2b).2. The solenoid connected to the liquid nitrogen cold line was set to 90°C to allow the chamber to cool quickly and for a sample to remain ∼ 20°C.3. Once the chamber pressure decreased to ∼10 mbar, valve 3 (Figure 2b), which pumps from the capillary tube, was closed and allowed the volume below the sample to equilibrate with the rest of the chamber.The chamber pressure gradually decreased down to 0.3-1 mbar due to constant pumping from the chamber through open valve 6. 4.Then, using valve 7 (See Figure 2b), CO 2 was injected slowly into the chamber to increase the chamber pressure to the desired CO 2 level (between 8 and 10 mbars) while continuing constant pumping from the chamber through open valve 6 (see Section 2.1). 5.The solenoid's setting was adjusted to get the sample to its desired temperature.6.Once temperature and pressure in the chamber and the capillary tube were equilibrated and stabilized, between 3 and 5 mbars of neon (measured with Baratron 1) were injected directly into the capillary tube using valve 1, and the evolution of pressure in the capillary tube was recorded.If the P capillary remained elevated compared to the P chamber , it was considered that a seal had formed.7. If a seal had formed, then the temperature of the soil was raised to >0°C to determine the stability of the seal at higher temperatures (up to +5°C).Note that once the seal was formed, the capillary tube was isolated from the chamber gases.Thus, the background gas in the "isolated" capillary tube was the water vapor from the soil below the seal as opposed to the chamber's CO 2 /H 2 O gas mixture.Therefore, the background pressure in the capillary tube could be a function of soil temperature and soil water abundance.8.The duration of the experimental run depended on the stability of the seal.If a seal was formed (typically within 1-2 days), the run could continue for several days or weeks.The soil temperature and pressure were recorded every 5 s during neon injections and shortly thereafter.9.If a seal was not formed within 1-2 days from the experiment's start, then the temperature of the sample was increased.Sometimes temperature increases result in seal formation, but prolonged exposure to higher temperatures (>0°C) eventually breaks soil seals.10.Once the seal was broken, then the experiment was stopped, and the sample was removed from the chamber within 24 hr.No long-duration experiments on the potential seal reformation were conducted.

Results
Table 1 provides a summary of the experimental runs in our study.The formation of soil seals was observed in both salt-icy soil mixtures.Samples with pure ice and pure permafrost did not exhibit gas-impermeable seal formation.

Non-Salt Samples
Intuitively, pure ice as a solid body should be gas-impermeable.To confirm that our first experimental run was to test the gas permeability of the pure ice in the Mars Simulation Chamber (Figure 2b). Figure 4a shows the results from a sample of frozen deionized water.The sample was placed in the chamber and maintained at 2°C to 3°C.After temperature and pressure stabilized, 3 mbar of neon was injected beneath the sample.The pressure beneath the sample quickly equilibrated with the rest of the chamber (P capillary = P chamber ), which meant that pure ice was not able to trap neon beneath.Our results indicated that pure deionized ice samples developed microcracks, allowing rapid neon escape into the chamber.
As demonstrated by the Phoenix mission, shallow permafrost is abundant at Mars high and mid-latitudes (Piqueux et al., 2019).Therefore, our next run was focused on permafrost gas permeability.Figure 4b shows the results from a sample consisting of JSC Mars-1A and 40% deionized water.The sample was placed in a freezer overnight (at 20°C) and then placed inside the chamber.In the chamber, the sample was kept at 12°C, and the chamber pressure was maintained at ∼8 mbars.The pressure immediately equilibrated when 3 mbar of neon was injected beneath the sample.We continued the run, periodically injecting neon to see if a seal would form, but repetitive neon injections did not result in any gas accumulation below the permafrost (P capillary = P chamber ).The sample was kept for 52 hr total in the chamber.Based on the observed pressure equilibration, the pure JSC Mars-1A permafrost without the presence of salts also developed microcracks and could not form a nonpermeable gas seal.

Samples With Perchlorate Salt
In the next runs, permafrost samples with Mg-perchlorate salts were studied.

10% Mg(ClO 4 ) 2 , Initially Uniform Permafrost
Figures 5a-5d show the results from Sample 3 (Table 1), which contained 10% Mg(ClO 4 ) 2 and 40% deionized water frozen into permafrost (Figure S4 in Supporting Information S1).During the first day in the chamber (Figure 5a), the seal formation was not observed-capillary tube pressure re-  1).(a) The temperature was kept at 7.5°C.Neons were injected 26 hr after being placed in the chamber."Capillary Tube Pumped Down" means pumping down the capillary tube itself and the sample's soil matrix from below.(b) 3 mbar neon injection showing a stable 25-hr long seal.The temperature was kept between 14.5°C and 12.5°C.Neon was injected 8 days after being placed in the chamber.(c) 66-hr seal also shows pressure decrease until reaching seal limit.The temperature was kept between 14.5°C and 11.5°C.Neon was injected after 9 days in the chamber.(d) Stable 25-hr seal below seal limit.The temperature was kept between 6°C and 3°C.Neon was injected after 13 days in the chamber.(e) Seal break.The temperature was maintained at ∼+2.5°C for about 20 hr before breaking.Neon was injected after 16 days in the chamber.
equilibrated quickly after neon injections (P capillary = P chamber ).It was verified that gas diffused through the soil both ways.During the "Capillary Tube Pumped Down" event (Figure 5a), we deliberately brought the capillary tube pressure below the chamber's pressure.Since there was no seal in the soil at the beginning of the run, CO 2 from the chamber above the sample quickly diffused through the soil into the capillary tube and pressures re-equilibrated (P capillary = P chamber ).
Pressure equilibration between the chamber and capillary tube occurred within 5 s after the neon injection.Over the next few days (not shown), there were signs of a possible incomplete seal formation-the capillary tube pressure still equilibrated with the chamber pressure, but over increasingly longer periods.After the fourth day in the chamber, a seal formation was observed with persistent P capillary > P chamber .It was established that the salty soil seal could hold 3-4 mbars of additional neon pressure.However, there was a limit to how much pressure could be held beneath the seal.
Figure 5b shows a stable seal after ∼8 days in the chamber (hour 187 of the experimental run).It was verified that the formed seal is gas impermeable from both directions.Specifically, during the "Capillary Tube Pumped Down" event (Figure 5b), the gas was pumped down from the capillary tube itself and the sample's soil matrix from below.However, since there was a seal in the soil, CO 2 from the chamber could not diffuse through the soil down to the capillary tube and P capillary remained lower than P chamber after pumping stopped.This was proof that the soil seal was impermeable to both neon (injections from below) and CO 2 (lack of equilibration after capillary tube pumping) in both directions.In subsequent "Capillary Tube Pumped Down" events (Figures 5c and 5d), the pumping of the capillary tube was stopped when the capillary tube pressure was about equal to the chamber's pressure.
The seal was able to withstand 3 mbars of neon gas pressure differential between the capillary tube and the chamber (P capillary -P chamber ∼ 3 mbars).However, when a larger neon injection was made after 9 days of being in the chamber (hour 217, Figure 5c) and 7 mbars of pressure differential was established, the neon slowly leaked until the pressure differential decreased down to ∼3.5 mbars.After ∼13 days in the chamber (hour 306, Figure 5d), the soil seal's stability was tested again under higher temperatures (Figure 5d).The temperatures were kept between 6°C and 3°C.Just as in the pure ice and pure permafrost runs, the vertical temperature gradient between the top and the bottom thermocouples was <1°C. Figure 5d validates that the seal can hold ∼3.5 mbars pressure differential even if the temperature of the soil varies from 6°C to 3°C for over 25 hr.As shown in Figure 5e, once the sample reached the temperature of about +2.5°C and was held at that value for ∼20 hr, the seal broke, and the neon was released.In total, soil seals were observed during a total of 12 days: 16th day (seal break) -4th day (first observed seal) = 12 days.Over those 12 days, seals were tested under multiple temperatures, starting at 14°C and slowly increasing to +5°C and multiple capillary tube pressures.We were not able to reform the seal in Sample 3 by decreasing the temperature to below freezing once the seal was broken.Further comprehensive studies are needed on the possibility of soil seal "reformation."10% Mg(ClO 4 ) 2 permafrost experiment has been repeated 3 times (Table 1).Soil seals formed consistently in all three repetitions within 3-5 days from the start of each run.
Figures 6a and 6b demonstrate Sample 3 after the run.A soil seal was extracted from Sample 3 after 19 days in the chamber (Figure 6b).The soil seal was solid and about 1 cm thick.The seal absorbs atmospheric water over time outside the chamber, and the soil becomes noticeably wet and breakable.Sporadic salt crystals can be seen in the soil seal with the naked eye (Figure 6b).

5% Mg(ClO 4 ) 2 Initially Uniform Permafrost
After the soil seal formation was observed in the sample with 10% Mg perchlorate, we decided to test whether a seal could form at a lower perchlorate soil concentration.Therefore, in the next run (Sample 4, Table 1), a permafrost sample with the perchlorate abundance at 5% by weight (Figure S5a in Supporting Information S1) was prepared.Similar to the 10% perchlorate sample, the 5% perchlorate sample also did not show seal formation at the start of the experiment (Figure 7a).Sample 4 was tested with 3-5 mbar injections of neon at 13°C, 12°C, 10°C, 7°C, 5°C, and 2.5°C.Only after the sample temperature (top thermocouple Figure 3) was raised to 1.7C, a decrease in neon diffusion through the soil sample was observed (Figure 7b).Figures 7a, 7b, and 7c show the results for Sample 4, which contains 5% Mg (ClO 4 ) 2 and 40% deionized water frozen into permafrost.Soil seal formation (P capillary = P chamber ) was not observed for the first 2 days (Figure 7a).However, the equilibration times started at ∼5 s, but after 48 hr of being in the chamber and raising the temperature from 14°C to 1.75°C, the equilibration time increased to ∼10 min, as shown in Figure 7b.An equilibration time increase indicates that a seal has started to form.Once the sample reached ∼ 1°C, a clear sign of seal formation (P capillary > P chamber ) (Figure 7c) was observed.We continued to increase the soil temperature up to ∼+1°C, but the seal lasted for 19 hr before breaking (Figure 7c).No attempt was made to reform the soil seal by decreasing the temperature afterward.5% Mg(ClO 4 ) 2 permafrost experiment has been repeated 3 times (Table 1).Soil seals formed consistently in all three repetitions within 2-3 days from the start of each run when temperatures of the samples were brought up to 2°C, 1°C . We did not explore the possibility of soil seal formation at lower temperatures for 5% Mg(ClO 4 ) 2 permafrost runs.
In a separate 5% Mg(ClO 4 ) 2 permafrost run (not listed in Table 1), the seal formation was observed after 70 hr since the start of the run.The experiment was conducted under 10°C sample temperature and 8-10 mbars chamber pressure.After 92 hr since the start of the run, we accidentally brought (P capillary -P chamber ) pressure differential under the soil seal in a significant excess of 5 mbars.That soil pressure spike resulted in an explosive breakup of the soil seal-a piece of the cemented soil was blown away from the sample (Figure S4 in Supporting Information S1).Unfortunately, the exact soil pressure differential for the explosion was not recorded digitally but it was observed by the chamber operator to be ∼10-15 mbars.The run was stopped shortly after the explosion for chamber cleanup.No intentional attempts to reproduce the explosion in the chamber were made.
In a separate 10% Mg(ClO 4 ) 2 permafrost experiment (one of the repetitions for Sample 3, Table 1), a short-duration test was made on the stability of the soil seal to colder temperatures.Conditions of seal formation described in Figure 5b were maintained (∼ 15°C sample temperature, 8-10 mbars chamber pressure).Shortly after the formation of the soil seal, the sample temperature was rapidly dropped to 30°C and then down to 40°C.The soil seal's gas permeability was tested at both temperatures by injecting 3 mbars of neon.Each pressure test injection was done for 1 hr.At both temperatures, the seal was holding 3 mbars of excess pressure.The run was stopped after the 40°C seal permeability test without breaking the seal.

Limitations, Assumptions, and Future Experimental Studies
In our study, the evidence of the soil seal formation was based on measuring the gas pressure in the chamber above the soil and comparing it to the pressure in the capillary tube below the soil sample (Figure 3).If after injection of neon P capillary remained elevated compared to the P chamber , it was considered that a soil seal had formed.As shown in Figure 6b, it was sometimes possible to extract the salt-cemented soil layer relatively intact after the end of the run.However, our setup did not have a third pressure sensor inserted in the middle of the soil.Therefore, a potential question might be raised whether the neon gas blockage could occur at the point of neon injection into the soil at the tip of the capillary tube (see Figure 3) rather than at the top of the sample in the salt cemented soil layer.Note that the tip of the capillary tube was only 0.13 mm, while the internal diameter of the soil sample holder at the top opening was ∼7 cm (Figure 3).
There are two reasons why the salt-cemented soil layer (Figure 6b) had to be gas-impermeable to explain the observed data (Figures 4, 5, and 7) regardless of whether the tip of the capillary tube was plugged at any point.First, all experimental runs presented in Table 1 started without a detectable soil seal (see Figures 4a, 4b, 5a,  and 7a).Therefore, pure ice, pure uniform icy permafrost, and uniform icy permafrost with perchlorate samples were all gas-permeable.Hence, the 0.13 mm tip of the capillary tube frozen into those icy matrices had to be gas-permeable as well.When the seal formed (e.g., Figure 5b, ∼1 week in the chamber), the temperature of the entire sample temperature was maintained in the range of 14.5 12.5°C.Therefore, the capillary tube was clogged with the same salty icy permafrost as at the beginning of the run.The only difference in the sample structure, compared with the starting conditions, occurred in the top ∼2 cm of the sample, where the ice sublimated to the chamber (due to low chamber pressure) and salt accumulated cementing the topsoil layer.Thus, the pressure differential observed in Figure 5b was consistent only with the gas-impermeable cemented topsoil layer.
The other evidence that the seal must occur in the cemented soil layer came from the "explosion" run shown in Figure S4 in Supporting Information S1.If the seal were to occur only at the 0.13 mm opening of the capillary tube while the soil cemented layer would be gas permeable then there was no feasible way to build up enough pressure in the soil to cause the explosion and lift of a chunk of cemented soil.At 10 mbar neon pressure, a capillary tube with an opening of 0.13 mm would generate an upward force of only ∼10 5 N, which is at least 2000 times smaller than the force necessary to lift 1 cm 3 of cemented soil (see Figure S4 in Supporting Information S1).The only way to lift several grams of soil is to accumulate critical pressure inside the soil under the gas-impermeable soil seal.
Several planetary parameters can affect soil seal formation on Mars, which were not studied in our chamber simulations.For example, as stated in Section 2.1, the CO 2 -flow-through system constantly removed water vapor from the chamber, but relative humidity above the soil was not measured.In theory, if the air above the soil is dry, then the sublimation rate from the sample should increase.Thus, the rate of salt-cemented soil formation on Mars could be more effective.
Lower gravity would result in easier vertical diffusion of the salty water toward the top of the soil.Thus, the rate of soil seal formation might be faster on Mars under the same pressure/temperature conditions.
We discovered the formation of gas-impermeable soil seals in soil samples with Mg(ClO 4 ) 2 perchlorate salt abundances of 5%-10% under Mars-like pressure/temperature conditions (Figure S2 in Supporting Information S1).A regolith salt abundance of 5%-10% by weight is consistent with some locations on Mars reported by Wray et al. (2011).Using thermal emission spectra of minerals (e.g., gypsum, hydrated Mg/Fe-sulfates, and kaolinite), they found that in the Columbus crater in the Terra Sirenum region of Mars, salt abundances are in the tens of percent range.Overall, the Phoenix lander mission and MSL reported 0.5%-1% of perchlorates widespread in Martian soils and regolith.However, some areas in the Martian regolith have concentrated amounts of salt (Fergason et al., 2006;Golombek et al., 2006;Hurowitz & Fischer, 2014;Wang et al., 2008).
Several experiments were conducted with 2% perchlorate permafrost samples (Sample 5, Table 1).In all low-salt experiments, the formation of isolated soil clumps was noticed but not a gas-impermeable layer of soil (Figure S3 in Supporting Information S1).However, laboratory experiments are limited to several weeks and may not capture the effect of seal formation at lower salt abundances.On Mars, such a process can occur naturally over a long period of time in the shallow permafrost regions, and it may be possible for enough salt to accumulate in the top layer to form a seal, even if the average abundance of salt in regolith is lower.Further long-duration experiments are needed to determine the salt-abundance threshold for seal formation.Also, our experiments should be conducted for sulfates, given their high abundance on Mars.
The exact concentration of Mg(ClO 4 ) 2 perchlorate salt in the solid seal was not measured in our experiment.However, upon extraction from the chamber, salt crystals (Figure 6b) can be clearly seen filling the soil pores in the solid soil seal layer, suggesting the upward transfer of salt by evaporating water during the experimental run.Once the impermeable gas seal forms, methane from any biological or abiotic source can accumulate under the seal.
During the 10% Mg(ClO 4 ) 2 permafrost run, we noticed a limit in the amount of pressure the seal could hold.Figure 5c shows that if the excess pressure under the seal is ∼5 mbar, neon starts seeping through the seal, but ∼3 mbar of excess neon pressure is stable under the seal (Figures 5b and 5d).This phenomenon was consistent throughout the entire run under multiple injections at different temperatures.In the 5% Mg(ClO 4 ) 2 permafrost run (Sample 4), soil seal stability was not tested and neon injections were kept at ∼3-3.5 mbar only (Figure 7c).Therefore, future studies need to explore the dependence of the seal-breaking pressure threshold on salt abundance in the soil.
The soil seals for perchlorate soil mixtures were found to be stable over a wide range of temperatures and had a fairly strong stone-like appearance (Figure 6b).Not only were the soil seals able to hold ∼3 mbar of pressure differential (neon excess pressure) beneath, but they were able to do it repeatedly.Most of our neon injections were ∼3 mbar, and the soil seals, once established, would hold such pressure excess indefinitely in all our experiments as long as the sample temperatures remained below 0°C (top thermocouple [see Figure 3]).When the soil temperatures were increased above 0°C, the soil seals did not break immediately.It took hours (Figure 7c) of continuous soil temperatures >0°C to break the seals formed in soils with 5%-10% salt abundance by weight.We hypothesize that the prolonged increase in the soil temperature above 0°C allows the water in the seal layer to evaporate, and eventually, the water content decreases to a point where microcracks appear in the soil, causing the seal to lose its gas impermeability.
All experimental runs that resulted in the formation of soil seals were started from uniform salty-permafrost samples (Figure S1 in Supporting Information S1).As ice rapidly sublimated due to low Martian-like chamber gas pressure above the sample, water with salt migrated toward the top layer of the sample, and an ice-free soil seal formed at the top of the sample (Figures 6a and 6b; Figure S2 in Supporting Information S1).This type of process might occur at present in shallow permafrost regions at high and mid-latitudes on Mars where shallow permafrost and ice are present (Dundas et al., 2021;Piqueux et al., 2019).However, further experiments are needed to determine the timescale of seal formation at lower temperatures.We observed soil seal formation in salty permafrost at temperatures 14.5°C 12.5°C (Figure 5b).Martian surface temperatures at high and midlatitudes are substantially lower.Lower temperatures would most likely result in significantly longer times to develop salt-cemented soil seals due to the lower mobility of salts.
At low latitudes (e.g., Gale crater), the temperatures of the surface and shallow subsurface (Martin-Torres et al., 2015) overlap with the temperature range in our experiments.However, there were no reports of permafrost in the Gale crater, and thus, no active formation of soil seals is expected at present.Salt-cemented soil seals could have formed at low latitudes and equatorial regions (specifically in the Gale crater) at the end of a period of high obliquity (e.g., Emmett et al., 2020;Levrard et al., 2004) when ice and permafrost were abundant at low latitudes.As ice started to sublimate and retreat to the poles, salt-cemented soil seals would form.Once soil seals are formed and buried by even a couple of centimeters of loose soil, the seal becomes sufficiently isolated from surface temperatures and can survive indefinitely.

Relating Experiments to Mars Methane Observations
To relate our gas trap stability experiments to Mars' shallow subsurface, we used the MarsWRF general circulation model output of the subsurface temperature at 2 cm depth, which is also the soil depth at which the temperature readings in our experiments were reported (top thermocouple, Figure 3).MarsWRF is the Marsspecific version of the PlanetWRF general circulation model (GCM) (Richardson et al., 2007;Toigo et al., 2012).We used a 5°horizontal resolution simulation with prescribed dust corresponding to the Mars Climate Database Mars Global Surveyor scenario (Montmessin et al., 2004) and correlated-k radiative transfer as described by Mischna et al. (2012).The simulation was run for four Mars years with hourly output, with Figures 8  and 9 representing average temperatures across that duration.MarsWRF model temperatures on 15 subsurface soil layers, and the surface skin temperature.Soil temperatures were modeled through subsurface heat conduction parameterization and sensible heat exchange with the surface layer (Toigo et al., 2012).MarsWRF assumed a constant heat capacity in the soil of 837.2 J/kg/K, representing a reasonable average value for basalt-dominated regolith.It used thermal inertia from Thermal Emission Spectrometer measurements (e.g., Christensen et al., 2001) to calculate thermal conductivity.The near-surface soil density was set to 1,500 kg/m 3 (Shorthill et al., 1976).Our experiments determined that prolonged exposure (several days) of a salt soil seal to temperatures above 273 K usually results in seal breaking.This experimentally observed quality of soil seals can be used to estimate their global stability.
Figure 8 shows that the temperature at 2 cm depth never exceeds the freezing point of water over most of the Martian surface (gray area).That means that over the vast majority of the Martian surface, the salt soil seals would remain stable, and any methane from the subsurface, produced by either abiotic or biological sources, can accumulate under such seals for a long time.
Surprisingly, Gale Crater is located right on the edge of the area where the salt soil seal can become unstable.Therefore, methane detection within Gale Crater during the warmer months of a Martian year (Webster et al., 2018) is consistent with the breaking of salt seals in Gale and the local release of methane.
The potential link between seasonal variations in methane and shallow subsurface temperatures is demonstrated in Figure 9.It is noteworthy that the lowest background levels of methane were observed during the Northern Spring, which is the season when salt soil seals are the most stable due to lower temperatures and the most effective at trapping methane.
Figure 9 shows that the maximum soil temperatures do not quite reach 273 K at 2 cm depths but can exceed the freezing point of water at 0.5 cm depths during the Northern Hemisphere Autumn in Gale.However, it should be re-emphasized that Figure 9 was generated using the output of MarsWRF with a 5°horizontal resolution simulation.Therefore, the model does not fully resolve the local topography, specifically the low elevation at the Curiosity rover location.Accounting for the model's coarse resolution and the prescribed dust opacities, it seems plausible that subsurface temperatures on the Gale Crater floor would intermittently reach freezing during Northern Hemisphere Summer and Autumn.
It has to be noted that the large sporadic spikes of methane detected with TLS-SAM (Webster et al., 2021) cannot be explained by the seasonal instability of soil seals in the Gale crater.Gale crater is not a unique crater at the low latitudes on Mars; thus, the abrupt large releases of methane should also happen at other craters.Without an extremely efficient unknown methane destruction mechanism, the released methane would have been detected by TGO, which scans all of the Martian atmosphere.The possibility that MSL just happened to land near the only active methane source on Mars is infinitesimal (Luo et al., 2021;Viudez-Moreiras et al., 2021).However, the Gale crater is unique because it is one of only two craters on Mars where a massive rover drives and drills the surface rocks.We propose a hypothesis for future studies that the motion of the rover itself or the drilling can cause the disturbance of the shallow subsurface soil seals and release methane.The heavy rover can crack a 2 cm soil seal, and Curiosity drills down to 5 cm.If our hypothesis is correct, then it might be able to explain the abruptness and seasonality of methane release while the release would remain local to Gale.
To test this hypothesis, it would be beneficial to take methane measurements when the rover just arrives at a location with abundant high-salt content features (like salt veins).Another test would be to try to ingest Martian air while drilling into the salt-rich surface.

Broader Applications of Soil Seals to Astrobiology on Mars
Even though this paper primarily focuses on the potential role of perchlorate salty soil seals in the variability of atmospheric methane on Mars, soil seals can have broader applications in astrobiology.Soil seal traps can be potential habitat locations for Martian microbial life in the shallow subsurface.One of the severe restrictions on active microbial growth is the low atmospheric pressure, as demonstrated by Nicholson and Schuerger (2005).Subsequent studies (Schuerger & Nicholson, 2016) show that some terrestrial microorganisms can grow at pressures as low as 0.7 mbars.However, all experiments  on low-pressure bacterial growth required continuously hydrated nutrient-rich media.On Mars, exposure of hydrated surface soil to low atmospheric pressures would inevitably result in rapid soil desiccation.Any potential microorganisms in the soil would be desiccated as well.Following our experiments, the overall gaseous pressure and the water vapor in the soil below the soil seal can be significantly higher than in the atmosphere.Therefore, below the soil seal, Martian microorganisms would be protected from extreme desiccation and might be actively metabolizing.Another benefit of soil seals is that they can block UV (Cockell & Raven, 2004) and atmospheric oxidants (Bullock et al., 1994) from penetrating below the seal.These factors can potentially contribute to the habitability of the shallow subsurface environment under the soil seals.On Earth, surface mineral crusts exist in different varieties and require water for their formation, entrap organic molecules, and harbor microbial life.
Biosignatures of surface mineral crusts can be extracted and assayed similarly to search for evidence of life on Mars (Brolly et al., 2018).
Salt domes found in the Gulf of Mexico basin show extensive carbonate deposits that are formed as cap rock on top of salt structures (Caesar et al., 2019).These authors have shown that microbial anaerobic oxidation of methane contributes to carbonate formation in the presence of sulfates.In another recent study (Bayles et al., 2020), the authors point out that in the Bonneville Basin in the western U.S., similar to what may have happened on Mars, evaporation of the freshwater lake has left behind huge amounts of evaporite mineral deposits with high salt concentration where the haloarchaea survive.Such microorganisms may be able to survive the harsh extreme environment on Mars, especially if they are preserved under the salty soil seals throughout Mars history.

Conclusion
Gas impermeable seals can form in Mars-like conditions, but they require elevated perchlorate salt content (5%-10% salts by mass) in the soil and a shallow permafrost layer.The soil seals are several centimeters thick and gasimpermeable.Pure ice and salt-free permafrost do not produce gas impermeable barriers for gases such as methane.Substantial amounts of gaseous methane (∼3 mbars of pressure differential) could accumulate under the seal in the shallow subsurface of Mars and be released abruptly into the atmosphere when the seal breaks due to possibly rover movements, drilling, or impacts.Based on modeled subsurface temperature constraints, soil seals should be stable over most of the Martian surface indefinitely in the current climate.Gale crater is at the boundary of soil seal stability, and if present, soil seals can be more susceptible to breaking during late NH Summer and Autumn.Soil seals provide UV protection, impede water loss, and allow the buildup of gas pressure in the soil above the triple point of water.Thus, they can provide a habitat for hypothetical Martian microorganisms.

Figure 1 .
Figure 1.Concept of the soil seal formation and the migration of water and perchlorates.

Figure 2 .
Figure 2. (a) Mars Simulation Chamber, (b) a close-up of the sample in the sample holder in the chamber, (c) block diagram of the entire setup.The Chamber could reach temperatures below 100°C and pressures down to as low as 0.1 mbar.

Figure 3 .
Figure 3. Diagram of the side profile of the sample holder with a typical sample inside.

Figure 6 .
Figure 6.(a) Sample 3 (Table1) directly after the sample run.(b) Salt cemented soil layer carefully removed from the top of the sample.Upon removal from the chamber, the solid soil layer absorbed atmospheric water vapor and softened in about 10 min.

Figure 8 .
Figure8.Total hours per year when the temperature exceeds 273K at 2 cm depth.The gray color represents a permanently frozen area at 2 cm depth.A narrow band between the edge of the gray area and a 100-hr contour line represents an area where soil temperature exceeds 273 K at 2 cm for 0-100 hr.Gale crater (black star) is right on the edge of that band, and therefore soil seals can be less stable.The central band (between 100-hr contours) represents an area on the Martian surface where salt soil seals would likely be unstable yearly.

Figure 9 .
Figure 9. MarsWRF subsurface soil temperatures at 0.5 cm (solid line) and 2 cm (dashed line) depths.The average daily soil temperature (top) is plotted in addition to the maximum daily temperature (bottom) over the year.

Table 1
Summary of the Samples Tested for Gas Permeability Figure 4. (a) Pure ice sample (Sample 1,Table1).Neon injections showed immediate pressure equilibration.Temperature at 2°C to 3°C (top thermocouple [see Figure 3] reading in all Figures).(b)Pure permafrost (Sample 2, Table 1) no salt.Neon injections showed immediate pressure equilibration.The temperature of the sample was kept at ∼ 12°C.Figure 5. Various neon injections into 10% Mg(ClO 4 ) 2 permafrost before, during, and after seal formation (Sample 3, Table A. P. was supported by NASA's Planetary Science Division Research Program through the SEEC ISFM work package at NASA Goddard Space Flight Center.Financial support from the NASA Early Opportunities Program for Underrepresented Minorities in Earth and Space Sciences, MUREP Other Opportunities Program (NASA Award# NNX16AC90A) is gratefully acknowledged.S.G. is supported by the MSL Participating Scientist program.A.P. is supported as a SAM-MSL science team member.