Martian regolith is the loose layer of broken rock, sand and dust covering most of Mars. It is mostly crushed basalt, tinted red by a thin film of iron-oxide dust and laced with about 0.5% toxic perchlorate. Under a wheel it behaves like dry sand with a little glue: friction angles of roughly 30 to 40 degrees, a few kilopascals of cohesion or less. At 38% of Earth's gravity, a tire on it gets about 40% of its Earth grip. Mars Rally Championship, a free browser rally game, models its tire grip the same way.
What Martian regolith is made of
Regolith, from the Greek for "blanket" and "rock", is any loose layer lying on bedrock; Martian "soil" is its fine fraction, with almost no organic matter. Gravity and thin air decide most of what driving a car on Mars would be like; the regolith decides the rest.
In October 2012 Curiosity ran the first X-ray diffraction analysis of Martian soil, at a sand-and-dust patch called Rocknest, and found feldspar, pyroxene and olivine, a mix JPL compared to weathered volcanic soils in Hawaii. Roughly half the sample was not crystalline: volcanic glass and its weathering products.
The red is skin-deep, a coat of dust over darker basalt. A 2025 study in Nature Communications attributed it mainly to ferrihydrite, a water-rich iron mineral, suggesting Mars rusted while still cold and wet.
Then come the salts: sulfates, chlorides and perchlorate, the last measured by Phoenix at 0.4 to 0.6% by weight in 2008 and by Curiosity at similar levels in Gale Crater. Heated to 835 °C, Rocknest soil gave up about 2% of its weight as water.
Is Mars soil toxic?
Yes, chiefly because of perchlorate, which blocks the thyroid's uptake of iodide and rides in fine dust a crew would breathe and track indoors. A 2025 review of Martian dust hazards in GeoHealth, which also flagged silica and gypsum, noted that a few milligrams of dust a day would exceed the safe perchlorate dose, and put prevention first: filtered air in suits and habitats, with iodine supplements for what gets through.
Microbes fare worse. In 2017 University of Edinburgh researchers found that perchlorate under Mars-like ultraviolet light killed Bacillus subtilis within minutes at Martian concentrations, and that iron oxides and hydrogen peroxide, also at the surface, lifted cell death after 60 seconds to 10.8 times that of ultraviolet alone.
Can you grow plants in Mars soil?
Probably, after treatment. In 2008 Phoenix dissolved magnesium, sodium, potassium and chloride out of an alkaline soil that its wet-chemistry lead, Samuel Kounaves, likened to a backyard's, where asparagus might do well and strawberries not.
In 2014 a Wageningen University team reported plants growing for 50 days in a Mars regolith simulant with no added nutrients, but also no perchlorate. When a 2021 study added calcium perchlorate at Martian levels to three simulants, none could support plant growth. Real regolith would need its perchlorate washed out or broken down, perhaps by bacteria that feed on it, and organic matter worked in.
Mars sand vs dust, duricrust and polygons
Martian sand and dust differ by grain size. Dust grains average about 3 micrometres, fine enough for wind to loft and slightly electrostatic, so they cling to camera windows and solar panels; they are the stuff of Mars dust storms. Sand starts at about 0.06 mm, some twenty times larger, mostly dark volcanic grains that hop along into ripples and dunes. Some of the Bagnold Dunes, the first active dunes studied up close off Earth, migrate up to a metre a year.
Duricrust is salt-cemented regolith. InSight's self-hammering heat probe, built to reach several metres down, stopped about 40 cm deep: a duricrust some 20 cm thick held a pit open around it, leaving its hull too little friction to hammer against. The regolith there is about 3 m deep, over lava flows. Polygonal ground is split by cracks into cells a few metres across, where ice-rich soil contracts in the cold; Phoenix landed on it.
How strong is Martian soil?
Soil strength comes down to two numbers, measured on Mars since Viking in 1976. Engineers write shear strength as τ = c + σ·tanφ, the Mohr–Coulomb rule: c is cohesion, the hold grains have on each other under no load, and σ·tanφ is friction, growing with pressure σ at a rate set by the internal friction angle φ.
| Mission (landed) | Site | Measured by | Friction angle | Cohesion |
|---|---|---|---|---|
| Viking 1 and 2 (1976) | Chryse and Utopia Planitia | Arm trenches | Mostly 27–39°; drift near 18° | Mostly 0.2–10 kPa |
| Sojourner (1997) | Ares Vallis | One wheel dug in | 34–39° cloddy; 26–28° drift | Small, barely resolved |
| Spirit and Opportunity (2004) | Gusev Crater, Meridiani Planum | 7 wheel trenches, 20 scuffs | 30–37° | 0–2 kPa; up to 11 kPa in scuffs |
| Phoenix (2008) | Northern polygonal plains | Arm trenches, dump piles | 38° ± 5° | 0.2–1.2 kPa |
| InSight (2018) | Elysium Planitia | Heat-probe hammering | Assumed, not measured | About 2–15 kPa, in duricrust |
| Mars Rally Championship, REGOLITH | Sculpted terrain, no real site | Tire model | µ 0.70, the tangent of 35° | ≈2.6 kPa |
Outside loose drift, friction angles run from the high 20s to the high 30s; Spirit's and Opportunity's, Sullivan and colleagues wrote in 2011, fit dry, rigid grains dominated by very fine sand. Cohesion is small but real, from almost nothing to over 10 kPa in cemented crust.
Why a tire gets so little grip on Martian regolith
Multiply shear strength by a tire's footprint for the most push the soil can return: A·c + W·tanφ, area times cohesion plus weight times tanφ. Only the weight carries gravity. At 3.72 m/s², 38% of Earth's, a tire presses down with 38% of its Earth weight while cohesion stays put, so relative to friction, cohesion counts about 2.6 times as much on Mars, rewarding a wide footprint over a heavy one.
Friction still supplies most of the grip. On a slope, weight presses a tire in and pulls it downhill in proportion, so friction alone holds about 35° at µ 0.70 on either planet; driving up Olympus Mons is a matter of cliffs and energy, not grip.
Low gravity also weakens the ground. In 2020 a Concordia University team ran an ExoMars rover wheel prototype through Mars soil simulant on parabolic flights and on the ground at equal wheel load. Traction still fell, by 5 to 10% on average in Mars gravity, which left one cause: gravity's effect on the soil itself. Lightened-rover tests on Earth miss it.
Slip does the rest. A driven wheel in sand turns faster than it travels, and that slip shears the soil into traction, but it also digs: the wheel sinks, pushes a growing heap of soil ahead of it, the bulldozing resistance of terramechanics, and slips more, until it is excavating instead of driving.
Stuck in Martian sand: Opportunity, Spirit and Curiosity
Opportunity at Purgatory, 2005
On 26 April 2005, sol 446, Opportunity drove into a wind-shaped ripple about 30 cm tall on Meridiani Planum and dug all four corner wheels in by more than a wheel radius. Escaping what the team named Purgatory Dune took more than five weeks; it reached firm ground on 4 June and drove on for 13 more years.
Spirit at Troy, 2009
Spirit, already dragging a right-front wheel that had not driven since 2006, broke through a crust in 2009 at a site called Troy, into soft sand with the highest sulfur content yet measured on Mars. In an extrication drive that November it moved less than 1% of the distance its wheel turns would cover on firm ground. On 26 January 2010 NASA made it a stationary research platform; unable to tilt its solar panels toward the winter Sun, it last called home on 22 March 2010. Its churned soil still paid out: layers pointing to thin films of water, perhaps from frost or snow, seeping in fairly recently.
Curiosity at Hidden Valley, 2014
Curiosity's drivers had been steering toward sand to spare wheels torn by sharp rocks, and in August 2014 the rover entered Hidden Valley, a sandy swale about the length of a football field. The wheels slipped more than the team expected from test drives on California dunes, nearly 80% toward the end of one drive, so it backed out: Hidden Valley, project manager Jim Erickson said, was "not a good location for experimenting."
What Mars Rally Championship simulates, and what it leaves out
Mars Rally Championship builds every tire's grip from those two soil terms. On regolith, µ 0.70 and about 2.6 kPa of cohesion over a ≈0.14 m² contact patch give the rover 2.60 m/s² of friction plus 0.48 of cohesion at 3.72 m/s²: ≈3.1 m/s², against ≈7.35 m/s² on Earth gravel.
Each wheel reads its own ground:
- REGOLITH, the default loose soil: grip ×1.00, 0.12 m/s² of cohesion per tire, ≈3.1 m/s² in all; a resting tire settles 5 cm in.
- HARDPACK, ancient lakebed crust, the nearest thing to duricrust: grip ×1.15, 0.16 per tire, ≈3.6 m/s², 30% less rolling resistance, 1.5 cm of sink.
- POLYGON, cracked plates with the crust already sheared: grip ×0.75, 0.05 per tire, ≈2.2 m/s², 30% more rolling resistance, 3.5 cm of sink.
A tire that slips, locks or spins digs up to 3 cm deeper, and one sliding sideways ploughs a wedge of regolith against its sidewall, Bekker's bulldozing: up to 35% more sideways grip on regolith, 20% on polygon, 6% on hardpack. That is why a drift bites on loose ground; the per-wheel tire model write-up shows where cohesion enters the friction circle.
What it leaves out is what stopped Spirit. Sinkage tops out at 8 cm, so no sand trap can swallow the rover; no crust hides soft sand, and gravity does not weaken the ground. The surfaces are laid out by a noise pattern, not geology, and the terrain is sculpted to Martian morphologies, not copied from a real place.
Where to try this
- Free roam has no timer. When the tire noise drops and the dust fades from rust to grey, you are on hardpack; polygon ground rumbles louder. Throw the rover sideways on regolith to feel the bulldozing bite.
- Stage 1, "The Opener": a 1,164 m timed line with medals to chase.
- Mars Rally Championship is free at /play with no download and no account. It runs on a phone in landscape or a school Chromebook; up to 8 people can race a stage live.