Smoothed Particle Hydrodynamics model for volatile emissions from Mars’ subsurface triggered by the drill onboard the ESA Rosalind Franklin rover

Introduction

The ESA Rosalind Franklin rover, whose launch is scheduled for 2028 as part of the ExoMars mission, is equipped with a drilling system capable of analyzing down to 2 meters in the subsurface of Oxia Planum [1,2]. The primary aim of the drill is to collect subsurface samples with significant astrobiological potential for detailed analysis. Although subsurface ice was not initially expected near the Martian equator, the Fine-Resolution Epithermal Neutron Detector instrument on board Trace Gas Orbiter has revealed traces of its presence [3]. Consequently, understanding how drilling activities could affect the presence and stability of subsurface volatile materials is crucial. 

For this aim, we are developing a theoretical model using the Lagrangian Smoothed Particle Hydrodynamics (SPH) approach [4]. In the SPH framework, the fluid is represented by particles, and each particle’s properties are determined by neighboring particles within the range of an interpolation function. The model presented here integrates hydrodynamic equations while accounting for key microscopic processes, including phase transitions, viscous coupling between solid and gas phases, dynamic and thermal interactions with solid boundaries, and also atmospheric effects.

Model 

We model a 3D cylindrical fracture with a radius 1 mm larger than the 12.5 mm radius of the drill rod [2] and a variable depth. The drill tip is represented as a smaller cylinder, 5 mm in radius and height. A layer of volatile material - a mixture of water vapor, ice, and dust - with variable thickness is initialized inside the fracture. Fig.1 (left panel) shows a setup with a 50 cm deep borehole and a 5 cm thick volatile layer. While the detailed geometry of the drill has minimal impact on the outcome, the temperature profiles of both the drill and the borehole walls are critical. Fig.1 (right panel) illustrates a plausible temperature profile: surface temperature is set at 230 K, representing the average diurnal peak during early Martian spring [5]. This value decreases linearly within the skin depth and then stabilizes at a subsurface temperature of 217 K [5]. Near the drill tip, the temperature is assumed to rise exponentially, reaching a peak that remains constant along the height of the tip before decreasing back to the subsurface temperature. In this study, we explore two scenarios based on experiments conducted in a thermovacuum chamber (CISAS, Padua) designed to replicate Martian drilling conditions: a cold case with a 250 K peak and a hot case with a 290 K peak. These values are consistent with other simulations reported in [6]. Fig. 1 corresponds to the hot scenario, however the cold case follows the same profile with a lower peak. Given the drill's slow penetration rate, the thermal disturbance is expected to remain confined to a narrow region. Water vapor is initialized at the drill tip temperature, whereas ice and dust are assumed to be in thermal equilibrium with the surrounding subsurface—consistent with conditions where ice may be present. Table I summarizes these initial temperatures.

Table I: Overview of initial temperature values.

At this stage, beyond solving the hydrodynamic Euler equations of the multi-component fluid, we also model key microphysical processes occurring on short timescales [7]. Ice-vapor phase transitions are treated statistically, while particles interacting with the subsurface exchange heat based on their dynamic timescales, potentially leading to the formation of deposited ice. Gas-dust interactions are modeled following [8], treating icy grains as dust particles that can undergo phase changes.

Fig.1: 2D initial conditions (left) and borehole walls temperature profile (right).

Results

In the hot scenario, the vapor saturation pressure exceeds the Martian atmospheric pressure, making its effect negligible in first approximation. The initial conditions here assume vapor, ice, and dust fractions of 0.2, 0.4, and 0.4, respectively. As shown in Fig.2, the vapor fraction rapidly decreases as it condenses into ice. Interactions between the ice, the hot tip, and the walls, cause further evaporation, gradually reducing the in-flight ice fraction. The newly formed vapor rises, condenses as it expands, and deposits ice on the borehole walls and drill rod, as illustrated in Fig.3. Additionally, viscosity coupling between vapor and dust carries some dust out of the hole. This entire process occurs rapidly, within approximately 10^-2 s, due to the small fracture size. The height at which the ice deposits form is consistent with the location of the Ma_MISS instrument, suggesting the spectrometer could detect water ice on the borehole walls, if present. 

In the cold scenario, by contrast, the vapor saturation pressure is too low to overcome the ambient Martian atmospheric pressure. To model this, we are extending our SPH framework to include a carbon dioxide component that fills the borehole [9,10]. Preliminary results indicate that vapor is pushed downward by atmospheric pressure and recondenses near the borehole bottom.

Fig.2:  Evolution of volatile material fractions inside and outside the borehole.

Fig.3:  Deposited ice distribution on the subsurface and drill rod at time t = 0.03 s.

Conclusions

This approach allows us to study not only the conditions under which volatiles are preserved but also their dynamics within the borehole. The results underscore the significant impact of drilling activities and highlight the necessity of exploring a wider range of scenarios to account for the many unknown variables. The next step is to continue refining the model by accurately incorporating the Martian atmosphere. While the current focus is on Mars, this model is also applicable to other planetary environments, such as the Moon.

Acknowledgments 

Work supported by the ASI-INAF grant "Attività scientifica di preparazione all'esplorazione marziana 2023-3-HH.0".

References

[1] Coradini et al. (2001), Adv. Space Res., 28(8), 1203-1208.

[2] Vago et al. (2017), Astrobiology, 17(7), 471-510.

[3] Malakhov et al. (2020),  EPSC 2020.

[4] Monaghan (2005), Rep. Prog. Phys. 68, 1703.

[5] Spohn et al. (2024), Geophys. Res. Lett., 51.

[6] Formisano, M. (2021), Adv. Astron., 2021(1), 9924571.

[7] Teodori et al. (2025), under review. 

[8] Laibe, G. & Price, D. J. 2012a, MNRAS, 420, 234.

[9] Chapman, S. & Cowling, T. G. (1970).

[10] Maggioni et al., in preparation.