No spacecraft has delivered a deliberately collected sample of Mars to a laboratory on Earth. China’s planned Tianwen-3 mission is intended to change that by joining several operations no mission has completed as a chain at Mars: collecting material, sealing it, launching it from the surface, meeting another spacecraft in orbit and carrying it back to Earth.

A 2025 perspective in Nature Astronomy, led by Tianwen-3 chief scientist Hou Zengqian, describes a target of at least 500 grams returned around 2031. The China National Space Administration repeated the plan to launch around 2028 and return samples around 2031 in an April 2026 agency announcement.

That is a mission schedule, not a record of completed hardware. The paper describes the intended architecture and science program; it does not establish that every rocket, spacecraft and containment system is ready to fly.

Mars rocks are already on Earth, but not with their context intact

Martian meteorites have reached Earth without help. Their composition identifies them as pieces of Mars, but an impact tore them from places that were not documented before they crossed interplanetary space and fell through Earth’s atmosphere. Researchers can examine the rocks closely, yet they usually cannot connect a meteorite to a particular outcrop, surrounding layer or sampling decision on Mars.

Robotic laboratories have examined Martian soil and rock on Mars itself. Curiosity has heated powdered samples in its onboard instruments. Perseverance has drilled selected cores and sealed them in tubes for a possible future return. None of those tubes has left Mars.

Returning a documented sample changes the range of possible work. The European Space Agency’s sample-return overview notes that Earth laboratories can use instruments too massive or complicated to put on a rover, compare results across facilities and preserve material for techniques that do not yet exist. The sample’s field context is what lets those measurements become a geological history rather than a list of chemical abundances.

Two launches divide Tianwen-3 into a surface mission and a return mission

The published plan uses two spacecraft stacks. The Orbiter-Returner Combination would travel to Mars and wait in an approximately circular orbit about 350 kilometers above the planet. The Lander-Ascender Combination would carry the landing platform, a Mars ascent vehicle and a small helicopter.

On the surface, a robotic arm would collect material within roughly 1.5 meters of the stationary lander. A drill is designed to reach two meters down, while the helicopter would gather rocks beyond the arm’s reach. The mission team proposes those three methods to obtain surface material, subsurface material and samples from more than one point near the landing site.

The container would then be loaded into the ascent vehicle. That rocket must leave Mars and enter an orbit the waiting spacecraft can reach. The orbiter-returner must locate the ascending vehicle or sample package, approach it, capture the sealed material and transfer it into the system that will travel to Earth.

None of those steps can substitute for another. A successful landing without a successful ascent leaves the sample on Mars. A clean launch into the wrong orbit can leave two working spacecraft unable to meet.

The rendezvous has to be autonomous

Mars is too far away for real-time piloting. Radio signals take minutes to cross the gap, and the delay changes as the planets move. By the time a ground controller saw an error during close approach and sent a correction, both vehicles would already be somewhere else.

The rendezvous therefore depends on onboard navigation, sensing and guidance. The vehicles must establish their relative position, reduce the distance between them safely and complete the capture without continuous steering from Earth. NASA’s description of its own proposed architecture calls the same kind of Mars-orbit capture an autonomous operation and shows why Mars Sample Return has been treated as a campaign rather than a single ordinary spacecraft.

China has relevant experience from the Chang’e lunar sample-return missions, which used ascent vehicles and automatic rendezvous in lunar orbit. Mars still changes the problem: it is farther away, the surface launch follows atmospheric entry and landing, and the hardware must remain healthy through a mission lasting more than three years.

A sealed container protects the science and Earth

The returned material has to be protected from Earth contamination. A stray terrestrial organic compound or microbe could confuse the search for Martian chemistry, especially when the question is whether a faint signal might have a biological origin.

Containment also works in the other direction. Material from Mars must not be released into Earth’s environment before it has been assessed. The Tianwen-3 team says returned samples would go to a dedicated facility with ultraclean and biocontainment zones. A 2026 mission presentation identifies Tianwen-3 as a Category V restricted Earth-return mission under the Committee on Space Research’s planetary-protection framework.

This is a demanding balance. The container has to exclude terrestrial material, prevent any unassessed Martian material from escaping and preserve the chemical and physical evidence that made the trip worthwhile.

The value of 500 grams depends on what the mission collects

At least 500 grams could support several analytical methods while leaving material for later study. Mass is only part of the scientific return, however. A carefully documented grain from the right layer can be more informative than a larger amount of mixed dust with little context.

The landing site must satisfy both science and engineering. The mission paper restricts candidates to 17 to 30 degrees north latitude, no more than three kilometers above the reference level, slopes of eight degrees or less and rock abundance no greater than 10 percent. Such terrain is safer for landing, but those limits also decide which parts of Mars are available to sample.

The two-meter drill would add material that a surface scoop cannot reach, and the helicopter could add some local variety. Neither turns one landing region into a survey of the whole planet. Claims about ancient life, climate or planetary evolution would still have to match the specific rocks, their setting and the limits of each test.

That caution matters because Mars measurements do not always line up neatly. Science Blog has covered the unresolved difference between Curiosity’s local methane detections and orbital searches that found none. A returned rock would not by itself solve that atmospheric question, but Earth laboratories would make it possible to test preserved material with several independent methods instead of relying on one compact instrument package.

The 2031 date depends on every link in the chain

CNSA’s 2026 announcement shows that Tianwen-3 remains an active agency program, including selected international payloads. It does not make the return date certain. Mars launch windows, spacecraft development, landing-site selection and qualification testing can all move a schedule before launch.

If the two stacks depart around 2028, they must still reach Mars, establish the intended orbits, land safely, collect useful material and seal it without compromising the sample. The ascent vehicle must launch successfully, and the orbital vehicles must rendezvous and transfer the container. The returner then has to leave Mars, cross interplanetary space and deliver the sample into controlled handling on Earth.

The next useful evidence will be the final landing-site decision and qualification results for the ascent, rendezvous and containment hardware.