Case Study: Designing a Mars Ascent Vehicle — the ISRU Payoff

"The cheapest kilogram of Mars propellant is the one you never launched from Earth."

Executive Summary

The hardest single object in a crewed Mars mission is not the ship that gets there — it is the small rocket that has to get the crew back off the surface: the Mars Ascent Vehicle (MAV). In the first case study we took two finished vehicles apart; here we design one on paper, and in doing so we quantify the most important architectural decision in all of crewed Mars planning: make the return propellant on Mars, or bring it from Earth? We will size the MAV with the rocket equation, then trace what each kilogram of its propellant costs depending on where it comes from — and watch the tyranny of the rocket equation (Chapter 3) turn a modest propellant load on Mars into a mountain of mass in low Earth orbit. The result is why in-situ resource utilization is not a nicety but the linchpin of the whole enterprise.

Skills applied

• Inverting the rocket equation to size a stage's propellant (Ch. 3, §3.5).
• Tracing a "gear ratio" — how mass compounds backward through mission legs (§39.2).
• Applying ISRU chemistry to a mass budget (§39.2; Ch. 34).
• Judging a design by what it forces onto the rest of the architecture, and tiering every number.

Background

The requirement

• Job: lift a crew from the Martian surface to a low Mars orbit, where the Earth-return vehicle waits.
• Ascent delta-v: $\Delta v \approx 4.3\ \text{km/s}$ (Tier 2 — Mars's gravity is gentler than Earth's, but you still fight it and a thin atmosphere; sources cite roughly $3.8$–$4.3\ \text{km/s}$).
• Engine: methane/oxygen, $I_{sp} \approx 370\ \text{s}$, chosen precisely because both propellants can be made from Martian resources (the Sabatier reaction of §39.2). So $v_e = I_{sp}\,g_0 = 370 \times 9.81 \approx 3{,}630\ \text{m/s}$.
• Burnout mass: the empty ascent stage plus the crew capsule, $m_f = 5\ \text{t}$ (illustrative, Tier 3).

The one honest complication

Unlike an Earth launch, every option here is shadowed by a second cost: whatever the MAV needs must first be delivered to the Martian surface, and landing mass on Mars is itself brutally expensive (the entry-descent- landing problem of Chapter 34). That second cost is what makes the ISRU-versus-import decision so lopsided, as we will see.

Phase 1: Size the MAV propellant

Invert the rocket equation. The mass ratio for the ascent is $$\frac{m_0}{m_f} = e^{\Delta v/v_e} = e^{4300/3630} = e^{1.185} \approx 3.27.$$ So the fuelled MAV is about $3.27$ times its burnout mass: $$m_0 = 3.27 \times 5\ \text{t} = 16.35\ \text{t}, \qquad m_p = m_0 - m_f = 16.35 - 5 = 11.35\ \text{t}.$$

The MAV needs about $11.4\ \text{t}$ of methane/oxygen propellant. (Sanity check: a mass ratio of ~3.3 for a ~4 km/s stage is in the right range — compare the Falcon 9 stages of Chapter 3, whose ratios ran ~4–6 for ~4–6 km/s.) That $11.4\ \text{t}$ is the number the entire architecture now has to source. Everything hinges on the next question: where does it come from?

Phase 2: Option A — bring the propellant from Earth

Suppose we ship all $11.4\ \text{t}$ of ascent propellant from Earth, pre-landed on Mars before the crew arrives. What does that cost upstream? We trace the mass backward through the mission legs, each of which multiplies the requirement through the rocket equation. We will use deliberately rough "gear ratios" and label them Tier 3 (illustrative) — the precise values depend on the architecture, but the direction and order of magnitude are robust.

• Landing on Mars. Delivering payload to the Martian surface takes several kilograms in Mars-approach orbit per kilogram landed (heat shield, descent propellant, structure). Take a gear ratio of ~2 (illustrative). $11.4\ \text{t}$ landed $\rightarrow \sim 23\ \text{t}$ delivered to Mars.
• Getting to Mars. Pushing mass from LEO through trans-Mars injection and cruise takes roughly ~2.5 kg in LEO per kg arriving at Mars (Tier 3). $23\ \text{t} \rightarrow \sim 57\ \text{t}$ in LEO.
• Reaching LEO. From Chapter 3, a launcher delivers only ~2–3% of its lift-off mass to LEO as payload — call it a factor of ~35. $57\ \text{t}$ in LEO $\rightarrow \sim 2{,}000\ \text{t}$ on the launch pad.

So, roughly and illustratively, ~11 t of ascent propellant on Mars implies on the order of ~2,000 t on Earth's pad — the mass of a large orbital rocket or more, spent only to pre-position the fuel for the ride home. The gear ratio from Mars surface back to the Earth pad is enormous because every leg pays the rocket equation, and the legs multiply.

💡 Intuition: This is the rocket equation's exponential wearing a mission-architecture costume. In Chapter 3's second case study we saw a chain of mass ratios multiply to turn a half-tonne satellite into a 21-tonne rocket. Here the same compounding runs across planets: land, cruise, launch — three mass ratios in series — and a modest tank of Martian fuel balloons into a launch campaign. Bringing propellant from Earth means paying that whole chain.

Phase 3: Option B — make the propellant on Mars (ISRU)

Now the alternative. Instead of $11.4\ \text{t}$ of propellant, we land a compact propellant plant plus a power source, and let it fill the MAV's tanks from Martian air and water over the ~500 days before the crew even arrives.

The chemistry (§39.2): the Sabatier reaction, $\mathrm{CO_2} + 4\,\mathrm{H_2} \rightarrow \mathrm{CH_4} + 2\,\mathrm{H_2O}$, makes methane from atmospheric $\mathrm{CO_2}$; electrolyzing the product water (and mined water ice) makes the oxygen. Methane/oxygen burns at a mixture ratio of roughly $3.6$ kg of oxygen per kg of methane, so of the $11.4\ \text{t}$ of propellant, about $2.5\ \text{t}$ is methane and about $8.9\ \text{t}$ is oxygen — and the oxygen, the heavier share, comes essentially for free from Martian $\mathrm{CO_2}$ and water.

What must we land? A plant, a power supply, and — in the leanest version, if we do not mine water ice — a seed supply of hydrogen. From the stoichiometry, making $2.5\ \text{t}$ of methane needs about $0.5 \times 2.5 \approx 1.25\ \text{t}$ of hydrogen (§39.2 worked the $0.5\ \text{kg H}_2$ per $\text{kg CH}_4$ ratio). Add a plant and power source of, say, $2$–$4\ \text{t}$ (illustrative, Tier 3). So ISRU lands on the order of ~3–5 t instead of 11.4 t — and if the base mines water ice for its hydrogen, even the $1.25\ \text{t}$ of seed hydrogen disappears.

Phase 4: The comparison, and the compounding payoff

Put the two options side by side. The headline is the landed mass; the shock is what it becomes upstream.

ISRU roughly halves-to-thirds the landed mass — but because that saving sits at the bottom of the gear chain, it compounds backward into hundreds to thousands of tonnes saved on Earth's pad. And the plant is an investment: it does not vanish after one flight but keeps making propellant for the next crew, amortizing its landed mass across the whole program. Option A pays the full gear ratio every mission; Option B pays it once.

🔧 Engineering Reality: These gear ratios are deliberately rough and clearly labeled Tier 3 — a real architecture (NASA's reference designs, SpaceX's Starship plan) would compute them in detail and get different specific numbers. But the conclusion is not sensitive to the details: because propellant sits at the bottom of a multiplicative chain of mass ratios, manufacturing it in place instead of hauling it down the chain is the highest-leverage decision in the mission. This is why every serious crewed-Mars study, however much else they disagree on, converges on ISRU for the return propellant. The rocket equation forces the agreement.

Phase 5: Iterate — what if the ascent is harder, or the plant is heavier?

Design is iteration, and it is worth testing how fragile the case is.

• Harder ascent. Suppose $\Delta v = 4.8\ \text{km/s}$ instead of $4.3$. Then $m_0/m_f = e^{4800/3630} = e^{1.32} \approx 3.75$, and $m_p = 2.75 \times 5 = 13.75\ \text{t}$ — a 20% heavier propellant load, which makes the import option worse still and the ISRU case stronger.
• Heavier plant. Even if the ISRU plant plus power masses $6\ \text{t}$ — more than half the propellant it replaces — it still wins, because it is landed once and reused, while the imported propellant must be re-landed (paying the full gear chain) for every mission. The break-even is not close.

The design conclusion is robust: for a repeated crewed Mars campaign, ISRU is not a trade to be studied — it is the foundation the rest of the architecture is built on. The only serious question is engineering the plant to run reliably for years, unattended, in dust and cold (theme 2), on the power a Mars base can generate (Chapter 25).

Discussion Questions

1. Trace, in your own words, why a saving of ~7 t in landed mass becomes a saving of ~1,000 t on Earth's pad. Which chapter-3 idea is doing the work?
2. Methane/oxygen was chosen partly because both can be made on Mars. Why is oxygen (the heavier share of the propellant) the easier win for ISRU than the methane?
3. Option B lands a plant "once" and reuses it. Explain how that reusability changes the comparison for a program of five crewed missions versus a single flight.
4. The gear ratios here are labeled Tier 3. Why is the conclusion nonetheless robust to their exact values?

Your Turn: Extensions

• Option A (design). Resize the MAV for a burnout mass of $8\ \text{t}$ (a larger crew capsule). Recompute the propellant and both gear-chain totals. Does ISRU's advantage grow or shrink in absolute tonnes?
• Option B (computation). Write mav_propellant(dv, isp, m_dry) returning the ascent propellant mass, and gear_chain(mass, ratios) that multiplies a landed mass up a list of gear ratios. Hand-trace them to reproduce the $11.4\ \text{t}$ and ~$2{,}000\ \text{t}$ figures, with an # Expected output: comment. Do not run the code.
• Option C (your mission). If you are on Track B (lunar) or Track C (Mars), add an "ISRU sensitivity" note to your MDR: identify one consumable (propellant, oxygen, water) your mission could in principle make in place, and estimate — even roughly — the landed mass it would save.

Key Takeaways

1. The MAV propellant is the pivot of Mars mission design. Sizing it is a one-line rocket-equation inversion (~$11\ \text{t}$ for a 5-t vehicle at ~4.3 km/s); sourcing it is the whole architecture.
2. Propellant sits at the bottom of a multiplicative gear chain (land → cruise → launch), so importing it pays a compounding penalty that turns ~11 t on Mars into ~2,000 t on Earth's pad (illustrative).
3. ISRU halves-to-thirds the landed mass and is reusable, so its saving compounds backward into hundreds to thousands of tonnes and amortizes across a program.
4. The conclusion is robust and tiered. The exact gear ratios are illustrative, but because the rocket equation makes the chain multiplicative, making propellant on Mars is the highest-leverage decision in the mission — which is why every serious study adopts it.