I talk all the time about the weird nerdy epiphanies I had as a kid. One of those epiphanies involved driving a car around on the outside of a space station. I realized that the car would have to bring along its own air supply, because an internal combustion engine can’t run on vacuum. I know that sounds obvious, but when you consider I was like nine years old at the time, it’s almost impressive that I figured it out. Almost.

Now that I’m older, I realized “Hey! I can actually figure out how much air I’d need to bring with me!” Conveniently, the worldwide craze for automobiles (some say they’ll replace the horse and buggy. I think that’s a pretty audacious claim, sir.) means that all sorts of vital statistics about gasoline engines are known. For instance: the air-fuel ratio. It’s as simple as it sounds. It’s the mass of air you need to burn 1 mass unit of fuel. The “ideal” ratio is 15:1: combustion requires 15 grams of air for every gram of fuel burned. Of course, if you’ve watched Mythbusters, you’ll know that stoichiometric (ideal) mixtures of air and fuel detonate, often violently. You don’t actually want that happening in a cylinder. You want subsonic combustion: deflagration, which is rapid burning, not an actual explosion. Supersonic combustion (detonation) produces much higher temperatures and pressures. At best, it’s really rough on the. At worst, it makes the engine stop being an engine and start being shrapnel. So, in practice, mixtures like 14:1 and 13:1 are more common. I’ll go with 14:1, although I freely admit I don’t know much about engines, and might be talking out my butt. No change there.

Either way, we now know how many mass units of air the engine will consume. Now, we need to know how many mass units of fuel the engine will consume. There are lots of numbers that tell you this, but for reasons of precision, I’m using one commonly used in airplanes: specific fuel consumption (technically, brake specific fuel consumption). The Cessna 172 is probably the most common airplane in the world. It has a four-cylinder engine, just like my car, though it produces 80% more horsepower. Its specific fuel consumption, according to this document, is 0.435 pounds per horsepower per hour. The Cessna engine produces 180 horsepower, and my car produces 100, so, conveniently, I can just multiply 0.435 by 100/180 to get 0.242 pounds per horsepower per hour. Assuming I’m using 50% power the whole way (I’m probably not, but that’s a good upper limit), that’s 50 horsepower * 0.242, or 12.1 pounds of gasoline per hour.

So, we know we need 12.1 pounds of gasoline per hour, and from the air-fuel ratio, we know we need 169.4 pounds of air per hour. That’s all fine and dandy, but I’m not sure how much room 169.4 pounds of air takes up. Welders to the rescue! According to the product catalog from welding-gas supplier Airgas, a large (size 300) cylinder of semiconductor-grade air has a volume of 49 liters, and the air is stored in that bottle at about 2,500 PSI. (I don’t know what you actually do with semiconductor-grade air, but it’s got the same ratio of gases as ordinary air, so it’ll do.) At room temperature, the bottled air is actually a supercritical fluid with a density 1/5th that of water. Therefore, each cylinder contains about 10 kilograms (22 pounds) of air. Much to my surprise, even when it’s connected to an air-hungry device like an internal combustion engine, a single size-300 cylinder could power my car for over seven and a half hours.

But you guys know me by now. You know much I like to over-think. And I’m gonna do it again, because there are a lot of things you have to consider when driving a car in a vacuum that don’t come up when you’re driving around in air.

Thing 1: Waste heat. This is a major issue for spacecraft, which live in a vacuum (unless you’ve really screwed up). The problem is that there’s only one good way to expel waste heat in a vacuum: radiation. Luckily, the majority of automobile engines are already radiator-cooled. Normally, they depend on heat flowing from the engine to the cooling water, into the metal fins of a radiator, and into the atmosphere. In vacuum, the cooling will run engine-water-radiator-vacuum. The engine produces 100 horsepower at maximum, which is about 75 kilowatts. A radiator operating at the boiling temperature of water radiates about 1,100 watts per square meter, for  a total area of 68 square meters, which means a square 27 feet (8.2 meters) on a side. You could play tennis on that. Luckily, the radiator is two-sided, which cuts the radiator down to a square 19 feet (5.8 meters) on a side. It’s still going to be larger than my car, but if I divide it into ten fins, it would only be absolutely ridiculous, rather than impractically ridiculous. That’s already my comfort zone anyway.

Thing 2: Materials behave differently in a vacuum. Everything behaves differently under vacuum. Water boils away at room temperature. Some of the compounds in oil evaporate, and the oil stops acting like oil. Humans suffocate and die. To prevent that last one, I’m going to have to beef up my car’s cabin into a pressure vessel. And since I’m doing that, I’ll go ahead and do the same to the engine bay, so that I don’t have to re-design the whole engine to work in hard vaucuum. I’ll make the two pressure vessels separate compartments, because carbon monoxide in a closed environment is bad and sometimes engines leak.

I’ll also have to put a one-way valve on the exhaust pipe, because my engine is designed to work against an atmospheric pressure of 1 atmosphere, and I feel like working against no pressure at all would cause trouble. I’m also going to have to change the end of my exhaust pipe. I’ll seal it off at the end and drill lots of small holes down the sides, to keep the exhaust from acting like a thruster and making my car spin all over the place.

Thing 3: Lubrication. A car’s drivetrain and suspension contain a lot of bearings. There are bearings for the wheels, the wheel axles, the steering linkages, the universal joints in the axles, the front and rear A-arms… it just goes on and on. Those bearings need lubrication, or they’ll seize up and pieces will break off, which you very rarely want in engineering. Worse, in vacuum, metal parts can vacuum-weld together if they’re not properly protected. We can’t enclose and pressurize every bearing and joint. That would make my car too bulky, for one. For two, there would still have to be bearings where the axles came out of the pressurized section, so I’ve gotta deal with the problem sooner or later. Luckily, high-vacuum grease is already a thing. It maintains its lubricating properties under very high vacuum and a wide range of pressures, without breaking down or gumming up or evaporating. We’ll need built-in heaters to keep the grease warm enough to stay greasy, but that’s not too big a hurdle.

Thing 4: Tires. My car’s owner’s manual specifies that I should inflate my tires to 35 psi (gauge). I’ll have to inflate them to a higher gauge pressure in vacuum, since they’ll have almost no pressure working against them. If I don’t, they’ll be under-inflated, and that’ll make them heat up, and in vacuum, that goes from a minor problem to a potentially fatal tire-melting and tire-bursting disaster. Actually, I think I’ll eliminate that risk altogether. I’ll do what most rovers do: I’m getting rid of pneumatic tires altogether. Because my car’s going to be fast, heavy and have a human passenger, I can’t do what most rovers have done and just make my wheels metal shells. I need some cushioning to stop from rattling myself and my car to pieces.

That’s Gene Cernan driving the Lunar Roving Vehicle (the moon buggy). It’s about five times lighter than my car, but it proves that airless tires can work at moderate speed. Michelin is also trying to design airless rubber tires for military Humvees, and while they don’t absorb shocks quite as well as pneumatic tires, they can’t puncture and explode like pneumatic tires. So I’m going with some sort of springy metal tire, possibly just composed of spring-steel hoops or something like that.

Thing 4: Fuel. If I was sensible, I’d have chucked the whole idea of powering a vacuum-roving Toyota with a gasoline engine. (Actually, I’d have chucked the whole idea of a vacuum-roving Toyota and started from scratch…) We know I’m not sensible, so I’m going to demand that my Lunar Toyota run on gasoline. 10,000 liters of gasoline (I like to mix units, like an idiot) will let me drive 42,500 kilometers. Enough to go around the Moon’s equator three (almost four) times. You might think that carrying a small tanker’s worth of gasoline to the Moon is an impossible feat, but when you consider that the mass of my car (about 1,000 kilograms) plus the mass of all that gasoline (7,300 kilograms) plus tankage is less than the weight of the Apollo Command-Service Module and the Lunar Module, not only does the Apollo program seem that much more audacious and impressive, but it becomes possible to talk sensibly (sort of…) about putting my car, my air tanks, and a lifetime supply of gasoline on the Moon. That also takes care of…

…Thing 5: Getting my car on the Moon. We can just use a Saturn V, or wait for the engineers to finish building the Falcon Heavy or Space Launch System. Lucky for me, the rocket scientists have already solved the problem of landing a heavy vehicle, too: the ballsy sky-crane landing used during the Curiosity rover’s descent would almost certainly work just fine for my car, since it’s only 200 kilograms heavier than Curiosity. The fuel and air can just be landed under rocket power, or by expendable airbags.

So it wasn’t all that insane for my nine-year-old self to imagine driving an ordinary street car around on the Moon. That is, from the point of view of fueling and aspirating (ventilating? aerating? Providing air to, is what I mean…) the engine and the passenger. But the physics of driving around in vacuum and/or under low gravity pose another challenge, and that challenge is interesting enough to get a post of its own. Watch this space!