I saw that some missions end with the lunar night time, and the rovers are trashed permanently instead of trying to survive the night.

The concerns have to do with trying to keep the rover at operating temperatures without a power source.

However, I don't really understand why these temperatures have to be maintained in general.

In overclocking, liquid nitrogen is used to cool CPUs to almost -200 °C. So it's not like silicon has trouble functioning at these temperatures.

So why can't the rover just power down completely for the night time, and resume operations in two weeks? What is it about the low temperatures that kills it completely?

In general, circuits should be able to operate at any temperature, from -200 to 100 °C. What kills circuits is temperature change, which enlarges and contracts pieces, which end up breaking off. This is how heat cycles kill graphics cards.

You could use metals with less thermal expansion. You could design the rover with thermal insulation, so the circuits are cold and wouldn't get warm.

Is the problem that the batteries don't survive the freezing? Could just use a maglev flywheel or do solar panels crack from the temperature change? It could just use a high number of individual parallels cells instead of the series.

  • 6
    $\begingroup$ "In general circuits should be able to operate at any temperature, from -200 to 100C" Well, that means dipping ICs alternately in liquid nitrogen and boiling water - most commercial integrated circuits are going to have a tough time operating continuously and reliably over that entire range, so I don't really believe this sentence. Can you show evidence that this is true? Or do you mean they can withstand that range in storage only? One fundamental challenge is the variation of carrier density with temperature, which is different for each kind of doped region within various devices on chip. $\endgroup$
    – uhoh
    Sep 5, 2023 at 20:43
  • 4
    $\begingroup$ "In overclocking liquid nitrogen is used to cool cpus to almost -200 C" I don't believe that either. Certainly LN2 provides a very generous "cold source" (i.e. heat sink) but is the IC really immersed in the liquid? Heat flow is somewhat analogous to electrical current flow and Temperature analogous to voltage. While the LN2 is "ground" the silicon inside the overclocked CPU is the "high end" and at a much higher temperature.Through various transitions through the packaging and heat sinking finite thermal resistance of each material and each interface results in... $\endgroup$
    – uhoh
    Sep 5, 2023 at 20:51
  • 17
    $\begingroup$ While they use LN2 to cool stuff, the silicon itself never gets down to these temperatures. Standard chips often have problems starting at a bit under -40 for various reasons, one is simply the bond wires breaking due to shrinkage, another is that various processes stop working as required ( e.g. bandgap references, certain transistor characteristics ). It generally is possible to create devices that function at those low temperatures (particle accelrator sensor assemblies anyone?) but it gets harder and harder the lower the temperature and the harder the temperature swing. $\endgroup$
    – PlasmaHH
    Sep 5, 2023 at 21:23
  • 21
    $\begingroup$ The purpose of liquid nitrogen in extreme overclocking isn't to cool the CPU to -200C, it's to keep the CPU from heating to 100C. Fast-running CPUs generate a lot of heat. $\endgroup$
    – Mark
    Sep 5, 2023 at 23:58
  • 14
    $\begingroup$ "In general circuits should be able to operate at any temperature, from -200 to 100 °C". That's the mother of all bold assumptions. If you inspect the datasheet of just about any semiconductor part out there, you will find that they're all rated for something like -40 to +85 °C ambient, maybe a few tens of degrees wider if non-operating or a special version for harsh conditions. And checking the temperature dependence of various properties will tell you that they vary wildly across that temperature range, so while the part will be able to operate, it certainly won't operate the same at all Ts. $\endgroup$
    – TooTea
    Sep 6, 2023 at 9:26

2 Answers 2


Yes, generally, the things that kill an unpowered probe are the physical changes associated with extremely low temperatures. Even if the processor itself continued working correctly, at the extremes of cold we're talking about, the differences in contraction between materials can pull apart components. I'm talking about things like the metallic circuit traces breaking off of a circuit board, wires snapping, possibly even solar panels delaminating or battery cases cracking open.

In addition, extreme cold and sitting in a deeply discharged state for a long time can permanently damage the internals of batteries, making them impossible to use later even if charging voltage becomes available. For example, lithium-ion batteries are famous for being difficult to recharge if they drain deeply enough, and any liquid electrolyte solutions are subject to freezing, which could both physically damage the battery and change the qualities of the solution when it starts to defrost under sunlight later. It may not be possible to design a battery that would still produce power at the lunar night-time temperature of -200 °F (-130 °C) without heaters, regardless of what conditions the processor is capable of functioning under.

You mentioned avoiding heat/cool cycles with insulation. But, there's not really any way to go from full sunlight to deep-space shadow without heat/cool cycles. There is no insulation in the world that could maintain temperatures across weeks of constant direct multi-hundred-degree sunlight and weeks of total darkness. Just look at how much work the JWST has to go through to keep its sensors at deep-space temperature, and that's with the entire design revolving around pulling off that trick, with constant power to run the coolers. Maintaining that sort of cold while still having attached components sitting in constant sunlight is a non-trivial problem, and that without actually sitting on a surface that's being baked by the same light.

To avoid getting damaged by freezing, most solar-powered probes use heater circuits to stay warm when they're going to be in a shadow for an extended time. But heaters cost battery power, and that puts a hard limit on how long they can be shaded before they have to shut down and risk taking damage from the cold. Most probes will have a 'safe mode' where they can kill the heaters and just wait for dawn with minimal power usage (it's actually the hibernation mode they were using during the flight out, in most cases), but it's generally expected that something will go wrong and it won't ever wake up. And sometimes the probe needs to actively keep an antenna pointed at Earth to keep talking to us, so even if it does wake up in the morning, it may not be able to reacquire and get stuck yelling into the void. So most lunar landers just assume that sundown means the show is over. (Of course we still hope. Chandrayaan 3 is currently shut down waiting for dawn. We hope it wakes up. We don't count on it.)

Now, the alternative is using radioactive materials for heat and/or power. A few lunar rovers, including the USSR's Lunokhod 1 & 2 and China's Chang'e 4, have used radio-thermal heaters to survive the night and continue their work in the morning, but in all those cases the radioactive decay provided only heat, with electrical power depending entirely on solar panels. (The downside there is the cost and the weight, not to mention the risk; the first attempt at a Lunokhod lander blew up on launch and scattered radioactive polonium across a wide area.) An RTG — a radioisotope thermoelectric generator — uses the heat of decaying radioactive materials to produce electricity. RTGs have been used on most of the probes heading to Jupiter and beyond, where the weak sunlight is not a viable source of power, and on a couple of Mars rovers (as made famous by The Martian). In that case, cold is no worry at all; usually in space, the problem is getting the RTG to radiate away enough energy that it won't melt!

But RTGs and similar technologies are usually not used on the moon. The costs and risks would have to be justified, and the moon just doesn't seem to do that. The moon is a long way off, but it's way, way closer than Mars or Jupiter. Astronomically speaking, it's easy to get there. It's probably more beneficial to land a probe on the moon, gather data for two weeks, and then shut it down and use what we learned to build a new probe that can do new experiments instead of repeating the same ones for months or years on end, even if they were fairly interesting experiments.

  • 16
    $\begingroup$ A couple of additional downsides of RTGs: they are expensive (because the Pu 238 fuel is getting scarce), and, they tend to be fairly heavy (because they need to be sufficiently shielded to stay intact in the event of a rocket explosion) $\endgroup$ Sep 5, 2023 at 21:33
  • 6
    $\begingroup$ @plasticinsect Apart from the price, don't also forget the piles of paperwork and likely even high politics needed before you can get your hands on a chunk of highly radioactive plutonium. Craigslist does have its limits (I hope). $\endgroup$
    – TooTea
    Sep 7, 2023 at 8:21
  • 1
    $\begingroup$ @TooTea Agreed. And this also feeds back into the price (because navigating regulations costs money) and the weight (because quite a bit of the weight is heavy shielding mandated by regulations). $\endgroup$ Sep 7, 2023 at 16:31
  • 2
    $\begingroup$ The USSR's Lunokhod rovers used radioisotope heat generators to stay warm during the lunar night. They lasted several lunar days. The downside was that they used polonium-210 as their radioactive source (lots of heat, easy to shield electronics against the alpha radiation). There was a launch accident that spread one rover's worth of Po-210 across much of the USSR. $\endgroup$ Sep 7, 2023 at 17:00
  • 1
    $\begingroup$ @TooTea The politics don't stop with finding plutonium. Every time we try to launch a probe with an RTG, there are protests about the dangers of launching radioactive material (on the theme of "what if the rocket fails, what then?"). They can't actually stop the launch but there can be public backlash against politicians seen as "risking public safety", so there has to be a will to endure that. $\endgroup$ Sep 7, 2023 at 22:00

First of all, there are possibilities to keep a (uncrewed) lunar mission alive for longer than a lunar day and they have already been used: e.g. Apollo Lunar Surface Experiments Package, which used RTGs to keep up power supply for longer time spans.

Which leads to your actual question: Why do most some end at lunar dawn? -> As you already described, it is mostly the power supply. (I said "some", because e.g. the Chinese rover Yutu survived the lunar night (partially). The Indian rover Pragyan is planned to be woken up after the night.)

So why do the (other) rovers not use a power supply that actually can survive the cold lunar night? -> Simply, it was not required.

Such power sources would take away volume and mass. You get a bigger rover, you need more mass on the moon, you have to get the mass to the moon,..., your rocket is not powerful enough. OR: You push away some science instrument to get that power supply.

On the other side: what is the benefit? Most scientific experiments are already done. And let us be honest: the science onboard is (unfortunately) just incidental. The big goal of such a mission is to generate headlines in the newspaper the next day:

[Generic Country] landed softly on the moon!

As you can guess, it is not necessary to survive longer than a couple of hours to get this. But it is a risk when making your lander/rover more complex than necessary for fulfilling this job!

  • 4
    $\begingroup$ "big goal of such a mission is to generate headlines in newspaper": although it's difficult to disprove, you are ascribing an intention to others that might not be there. Many are scientists, who would like to have a long term exploration project. $\endgroup$ Sep 6, 2023 at 19:30
  • 3
    $\begingroup$ I think you're editorializing to say that the goal of a rover mission is to generate headlines. For example, the stated objectives of the upcoming Canadian rover make no such claim. Although science objectives may be secondary to technology demonstrations, tech demos are still more important than the objective of just headlining. Science goals on tech demo missions are hardly "incidental": science proposals just compete for time money and resources at lower priority than the tech goals. $\endgroup$
    – Wyck
    Sep 7, 2023 at 2:24
  • 2
    $\begingroup$ I fully agree that science and technology demonstration is important but - i'm sorry I did not make this clear enough - that is not the reason why scientist and engineers are getting money for. I and my colleagues all over the word do space exploration because we are idealistic. Governments are spending money for space exploration because of: "economic promotion", "national pride", "military purposes". Space agencies hat to make the balance act and sell science and technology demonstration as one of the mention topics, while letting their employees believe, its about something bigger. $\endgroup$
    – CallMeTom
    Sep 7, 2023 at 5:34
  • $\begingroup$ “Without it, they don’t talk to us, they don’t correct their trajectory, they don’t turn the heat shield around… we gotta turn everything off. Now.” - Solar panels might delaminate, but frozen batteries will not work even if those still do. $\endgroup$
    – Mazura
    Sep 8, 2023 at 0:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.