Case Study: Designing a Lunar Sample-Return Mission

"Getting there is the tyranny. Getting there and back is the tyranny, squared."

Executive Summary

The one-way lunar cargo lander of §40.3 was hard enough: a mass ratio of 6.72, a vehicle 85% propellant, just to set payload down on the Moon. Now we make it one tier harder by demanding a return: land, collect a sample, and bring it home to Earth. This case study is a clean-sheet design — the harder sibling of the audit in Case Study 1 — and its central lesson is a skill the four simple tracks did not require: iterated, staged sizing. A return leg cannot be sized with a single rocket-equation call, because each stage's propellant becomes the next stage's payload, compounding the tyranny. We will design the mission as a stack of stages sized from the top down, discover why a single-stage version is hopeless, and show that even done well, returning a small sample from the Moon demands a vehicle in low Earth orbit some forty-six times the mass of what comes home. That number is why sample return is a flagship undertaking, and why every real one — Luna, Apollo, Chang'e-5 — stages and rendezvouses.

Skills applied

• Iterated ("onion-layer") vehicle sizing across multiple stages (Ch. 3, Ch. 22).
• Building a round-trip delta-v budget across the delta-v map (Ch. 3, Appendix G).
• Recognizing where a leg is aerodynamic, not propulsive — Earth re-entry (Ch. 7).
• Making and defending a staging decision, and checking closure on a launch vehicle (Ch. 30).

Background

A one-way trip spends delta-v getting down the gravity well. A round trip spends it getting down and back up, and the return propellant is the cruelest kind of payload: you must carry it, unburned, all the way to the surface before you can use it. That is what makes sample return the "tyranny squared" of the epigraph. Our objective: return a ~1 kg sample of lunar regolith to Earth's surface. The returned hardware is a small capsule with a heat shield, because — the one mercy in the budget — the final descent to Earth is aerodynamic, paid with a heat shield (Chapter 7), not with propellant.

Phase 1: The round-trip delta-v budget

Trace the route across the delta-v map (Appendix G), out and back, splitting it by which stage pays each leg. The launch vehicle separately delivers the whole stack to LEO (~$9.4\ \text{km/s}$, not the spacecraft's concern here).

The descent stage's job totals $5{,}500\ \text{m/s}$ ideal (the Track-B budget of §40.3); the ascent stage's job totals $2{,}600\ \text{m/s}$. Applying margins as in §40.3 (5% on TLI, 10% on capture/TEI, 15% on the powered descent, 10% on ascent) gives margined budgets of about $\mathbf{5{,}980\ \text{m/s}}$ (descent) and $\mathbf{2{,}900\ \text{m/s}}$ (ascent).

Phase 2: Size from the top down (the onion)

Here is the skill the simple tracks did not need. You cannot add the two budgets and size once, because the ascent stage's fueled mass is payload that the descent stage must land. You size the innermost thing first (what comes home) and work outward, each stage's wet mass becoming the next stage's payload. Use a storable/methalox engine at $I_{sp} = 320\ \text{s}$ ($v_e = 3{,}138\ \text{m/s}$) throughout.

import math
G0 = 9.80665

def size_stage(dv, isp, payload, stage_dry):
    """Wet mass (kg) of a stage that pushes (payload + its own dry mass) through dv (m/s).
    Returns (m_wet, m_prop). The stage's final mass is payload + stage_dry."""
    ve = isp * G0
    m_f = payload + stage_dry              # what remains when this stage's tanks are empty
    m_wet = math.exp(dv / ve) * m_f
    return m_wet, m_wet - m_f

# Layer 0: what comes home -- an Earth-return capsule (heat shield + ~1 kg sample)
capsule = 200.0                            # kg dry, re-enters aerodynamically (no dv)

# Layer 1: ascent stage lifts the capsule off the Moon and injects it toward Earth
asc_wet, asc_prop = size_stage(2900, 320, payload=capsule, stage_dry=150)
print(f"ascent stage: wet {asc_wet:.0f} kg (prop {asc_prop:.0f} kg)")

# Layer 2: descent stage lands the ENTIRE fueled ascent stack on the Moon
des_wet, des_prop = size_stage(5980, 320, payload=asc_wet, stage_dry=500)
print(f"descent stage / mass to LEO: {des_wet:.0f} kg (prop {des_prop:.0f} kg)")
print(f"LEO mass per kg returned: {des_wet/capsule:.0f} : 1")
# Expected output:
# ascent stage: wet 882 kg (prop 532 kg)
# descent stage / mass to LEO: 9291 kg (prop 7909 kg)
# LEO mass per kg returned: 46 : 1

Read the onion from the inside out. A $200\ \text{kg}$ capsule needs an $882\ \text{kg}$ ascent stage to get it off the Moon and toward Earth. That $882\ \text{kg}$ is what the descent stage must land, so the descent stage — hauling the whole fueled ascent stack down the gravity well — balloons to $\mathbf{9{,}291\ \text{kg}}$ in low Earth orbit. To bring home $200\ \text{kg}$ (of which ~1 kg is the actual sample), you must assemble nearly ten tonnes in LEO: a 46-to-1 ratio of departure mass to returned mass. The tyranny of the rocket equation, squared by the return leg.

Phase 3: Why you cannot do this single-stage

The reader who resists staging asks: why two stages and all that separation hardware — why not one vehicle that does the whole round trip? Size it and see. A single vehicle must carry all its dry structure (descent engines, tanks, ascent engine, capsule — call it $850\ \text{kg}$) through every leg, so it faces the entire round-trip budget, $5{,}500 + 2{,}600 = 8{,}100\ \text{m/s}$ ideal, ~$8{,}900\ \text{m/s}$ margined:

$$ \frac{m_0}{m_f} = e^{8900/3138} = e^{2.836} = 17.0, \qquad m_0 = 17.0 \times 850 = 14{,}500\ \text{kg}. $$

The single-stage vehicle needs ~14,500 kg in LEO against the staged design's ~9,300 kg — over 50% heavier — and it is worse than that number suggests, because it drags empty descent tanks all the way back to Earth. Staging saves ~$5{,}200\ \text{kg}$ by setting down the mass it no longer needs: the spent descent stage stays on the Moon; only the small ascent stage flies home. This is the identical lesson as Falcon 9 in Chapter 3 and the lander of §40.3, now compounded across a round trip.

🚪 Threshold Concept. A return mission is not "twice a one-way mission" — it is exponentially more than that, because the return propellant is payload that must itself be transported, propelled, and (until burned) landed. Every kilogram you want to bring back forces you to land the propellant to launch it, which forces you to carry the propellant to land that, and so on down the onion. This compounding is why sample return from anywhere with a gravity well is a national-flagship undertaking, why Mars sample return is one of the hardest missions ever attempted, and why the rocket equation, once learned, reframes every ambition in terms of mass ratios. Getting there and back is the tyranny, squared.

Phase 4: Choose the architecture and check closure

At ~$9{,}300\ \text{kg}$ to LEO, the staged stack closes comfortably on a Falcon 9 (17–22 t to LEO) or any medium-heavy launcher (Chapter 30, Appendix H) — a real, affordable vehicle. The design closes. Two architectural refinements a review board would raise, both real:

• Lunar-orbit rendezvous (LOR). Instead of landing the whole return capability, leave an Earth-return vehicle in lunar orbit and send only a small ascent capsule up to dock with it. This is what Apollo did (with a crew) and what Chang'e-5 did (robotically). It shrinks the landed mass further, at the cost of an autonomous rendezvous and docking in lunar orbit — a capability, not a given.
• Engine choice per leg. The descent and ascent need high thrust (chemical), but a trans-lunar tug could be electric, spiraling the stack out from LEO over months to save launch mass — the SMART-1 trick (Chapter 20), trading time for mass exactly as the other tracks do.

🔧 Engineering Reality: The onion sizing above assumed the same $I_{sp}$ and tidy dry masses for every stage; a real design iterates it dozens of times as subsystem masses firm up, because a heavier capsule (Layer 0) is amplified by both stages above it — the mass-growth amplifier of Chapter 29, now applied twice in series. This is why sample-return programs guard the returned-capsule mass ferociously: a kilogram added to the capsule can add ten or more to the vehicle in LEO, through two nested mass ratios. Mass is the enemy; on a return mission it is the enemy compounded.

Phase 5: Risk and the driving requirement

The driving requirement is the returned sample mass — every kilogram of it is amplified through two nested mass ratios into ~46 kilograms in LEO, so it is the input the whole coupled design is most sensitive to. The dominant risks are the two un-abortable, autonomous events far from home: the powered descent (as in §40.3) and the ascent — a small engine that has sat cold on the lunar surface, perhaps through a two-week night, and must light on the first try with no backup, or the sample never comes home. A review board will spend most of its time on that ascent engine, because it is the single point of failure with the least test heritage and the highest consequence. Naming it first, with its mitigation — qualification testing "like you fly," thermal keep-alive through the lunar night, redundant ignition — is how you defend this design (Chapter 32, §40.6).

Discussion Questions

1. Explain, in terms of the onion sizing of Phase 2, why the ascent stage's wet mass (not its dry mass) is what the descent stage must treat as payload.
2. The single-stage vehicle (Phase 3) came out over 50% heavier than the staged one. State the physical reason in one sentence, connecting it to the Falcon 9 result of Chapter 3.
3. Earth re-entry contributed $0\ \text{m/s}$ to the budget. Why, and what does the capsule pay instead of delta-v? (See Chapter 7.)
4. A proposal doubles the returned capsule mass from 200 kg to 400 kg "to bring back more science." Using the amplifier idea, estimate the new LEO mass and argue whether the science is worth it.

Your Turn: Extensions

• Option A (analysis). Re-run the Phase 2 onion with a lighter capsule (150 kg) and a higher-$I_{sp}$ ascent engine ($I_{sp} = 360\ \text{s}$, a hydrolox or high-performance stage). How much does the LEO mass drop? Which layer's change helped more, and why? (Hand-trace; add # Expected output:.)
• Option B (computation). Implement lunar-orbit rendezvous: split the mission so an Earth-return stage (sized for TEI + reserve) waits in low lunar orbit while only a minimal ascent capsule launches to meet it. Size both and compare the total LEO mass to the direct-return 9,291 kg. Does LOR win, and by how much?
• Option C (your mission). Whatever your track, add a return leg to it (a comsat that de-orbits to Earth for servicing, a Mars sample return, an asteroid sample return) and size it with onion staging. Report the departure-to-returned mass ratio and reflect on what the return did to your vehicle.

Key Takeaways

1. Return missions require iterated, staged sizing. You cannot add the budgets and size once; each stage's wet mass is the next stage's payload, so you size from the returned capsule outward.
2. The return leg compounds the tyranny. Bringing home 200 kg from the Moon demanded ~9,300 kg in LEO — a 46:1 ratio — because return propellant is payload you must transport and land before you burn it.
3. Staging is not optional here. The single-stage version was over 50% heavier; you must set down the mass you no longer need, which is why every real sample return stages and most rendezvous.
4. The returned mass is the driving requirement, amplified through two nested mass ratios; guard it, and interrogate the un-abortable ascent burn hardest of all in review.