Ever since Jules Verne wrote "From the Earth to the Moon" space guns have long been proposed as a method of launching objects into space. And for non-human, non-fragile payloads they make sense as the payload doesn't have to carry much of it's fuel with it. But apart from a few experiments no successful space gun has been developed.
So I'm wondering, what technological/engineering hurdles that need to be overcome before we can see a working and efficient space gun?
Scaling is probably the biggest issue.
First of all, you always need an upper stage, as a gun cannot get you into a reasonable orbit. It can get you high enough, and maybe fast enough, but not into an orbit.
So now your gun has to be large enough to launch a payload, including at least some kind of upper stage.
Once you scale up to those kind of sizes to deliver meaninful payloads, these guns get VERY large and much more complex than would seem.
There are many issues that come up with scaling. As things get larger, their frontal area must get longer, air resistance gets higher, and now more energy is needed to push all the way through the atmosphere, with sufficient velocity at the end, to be close enough to orbital velocity so that a small upper stage can circularize the orbit.
But as you trade off upper stage size/payload/initial speed, the initial G load gets higher and higher in order to deliver a useful payload.
The biggest problems with a space gun are inherent the simple nature of ballistics.
We can use railguns to accelerate projectiles very fast; much faster than explosive propellants can do over the same distance. The current record is 33 MJ; that's a one-kilo projectile propelled at approximately escape velocity (8127 m/s). So, theoretically, the most powerful railgun we have (currently in the hands of the Navy Labs, who are developing the technology for a new class of warship) could put a 1-kilo object into space.
However, there are those problems I mentioned. First, in the Navy railgun, this amount of energy is currently imparted to the projectile over a distance of approximately 12 meters, in a timespan of 10 milliseconds. An 8000m/s velocity change in a 1kg projectile over .01 seconds equals an acceleration force of 800,000 Newtons, which is about 81,632.65 G-forces. Humans can survive about 14 Gs when lying flat on their back or stomach, and 4 Gs is the maximum we like submitting our astronauts to in a seated position, since they're not all fighter pilots in prime condition able to handle 6-8 Gs for long maneuvers.
To get the same delta-V with acceleration limited to 40m/s2 (just over 4G) would require 200 seconds, during which time you would have traveled 800,000 meters (the required length of the barrel of the railgun). 800,000 meters is just shy of 500 miles; you're talking about engineering a track from Dallas to El Paso, TX, with millimeter tolerances for deviation per kilometer of track from being perfectly straight. It would be the biggest and most expensive single civil engineering project the human race has ever undertaken; bigger by far than the current bullet train systems, bigger than the CERN supercollider, bigger than the U.S. interstate project.
Second, at the instant you leave the barrel, you're travelling as fast as you ever will. Drag from air resistance will slow you down, and the air is thickest right at the surface of the Earth, where it's cheapest to build this 500-mile track. So as soon as you leave the barrel, you're blasted with Mach-25 winds that will instantly begin slowing you down below escape velocity. And if it happens to be raining in Dallas (remember you want to travel west to east, in the direction of the Earth's rotation, to take advantage of the extra 465 m/s of angular velocity), the raindrops will shatter the windshield. We mitigate this with modern rockets by limiting our acceleration and velocity until we exit the troposphere, at which point we throttle up the rocket to maximum power as the air thins. Our railgun, on the other hand, would have to be built on about a 7* incline, with the muzzle about 100,000 feet in the air, in order for the projectile to avoid the worst of the air drag as it exits the barrel. So now, you're taking what's already the most expensive project in history and adding to it the task of building a ramp 36 times taller than the Burj Khalifa, and that's the shorter dimension, by far. We're now also working against gravity as we accelerate, requiring us to add an additional 1.225m/s2 to our railgun's acceleration just to account for gravity (we've already been assuming that air drag and other friction within the barrel is negligible).
Lastly, 33MJ is our current world record for imparting kinetic energy into a projectile via magrail without completely destroying it, and that's only enough to get one kilo up to escape velocity. The Space Shuttle orbiter, empty, weighs 130,000 kg. The lightest vehicle we've ever put into orbit that is capable of supporting human life was the Mercury capsule at 1,400kg; a railgun capable of putting this capsule in orbit would have to be able to impart, in the ideal, not one millijoule less than 46.2GJ of kinetic energy to the capsule. In the grand scheme of things that's not much; it's about the yearly electricity consumption of the average refrigerator. However, that's after all inefficiencies and losses, which when you're dealing with electromagnetic inductors can be several orders of magnitude more than the energy delivered.
Since we've been fielding a lot of questions on this subject lately, I just want to chime in with some of the rules-of-thumb that seem very natural to me. Unlike rockets, a gun sees the maximum density of the atmosphere at its fastest point. So provided you understand the necessity to circularize an orbit, and you understand that high accelerations need hardened equipment, the viability of many proposals can be easily ruled out envelope calculations for the drag.
Here's a helpful metric to look at the atmosphere with: the mass-thickness of the atmosphere is the mass per unit area, looking straight up into the sky at sea-level.
$$ \frac{ \text{Mass of Atmosphere} }{ \text{Area} } := \mu_{atm} \approx 10 \frac{ \text{tons} }{ m^2 } $$
In a naive sense, look at a bullet straight-on. Divide its mass by the area (same area it presents to the atmosphere). A more accurate approach would be to do some manipulation of the drag equation. We're not interested strictly on the force on the projectile. Alternatively, I'll consider the velocity lost due to its trip through the atmosphere, $\Delta v$ here. I obtained the following from the drag equation, under the assumption that the total velocity is notably greater than the loss in the atmosphere (if not, it's nonviable anyway).
$$ \frac{ \Delta v }{ v } = \frac{1}{2} \frac{C_d \epsilon \mu_{atm}}{ \rho D f \sin{( \theta)} } $$
Due to practical considerations, if this ratio is about 1, you don't have a chance. It will burn up in the atmosphere, and even if not, you can't produce those speeds. Speeds near orbital have never been demonstrated. So if you can't push this ratio well below 1, the idea is out the window. I'll go over all the terms below, dividing them into more-or-less into things that are completely impossible to push beyond a certain window, and things that have not inherent limitation.
Hard limits:
Soft limits:
So we have to use our (only) soft limit in order to engineer around the other hard limits. I'll take a velocity loss ratio of 0.2 for now. You could stretch this, but not by much. You actually come out with a set of parameters like:
$$ D \approx 4.5 cm \\ M \approx 11.7 kg $$
This doesn't sound shockingly huge. But if you change it to a sea-level launch, the required mass goes up by a factor of 8, due to the fact that mass scales with D^3. I was also extremely generous with the diameter to length factor. A bullet shape that has a factor of 5 would be, again, 8 times the mass. So if we're looking t something like a practical sea launch to ultimately reach orbit, the minimum mass is closer to the scale of 750 kg.
You can see here how the minimum size depends strongly on the assumptions of the type of system you're using. Nonetheless, 10-750 kg is still a starting range for the minimum bullet mass needed for the system to ever work.
But the bullet size and speed also set a limit on the minimum size of the gun itself. If you can get pressures on the order of 50,000 psi, then the 750 kg case (Quicklaunch) at 8 km/s would need a volume of about 110 m^3. That's big.
Compare, a nuclear reactor pressure vessel is probably about 2-3 times the volume of that. It's also about 1/20th the pressure. And it costs upwards of $100 million.
Straightforward material requirements for a pressure vessel are proportional to the (pressure)x(volume) product. So things are not looking good for our space gun, although the economics of heavy forgings are far from simple. On the other hand, a minimalist space gun mounted on the slope of a tall mountain might have a hard cost minimum below the $10 million scale.
Or you could just buy a Falcon 9 flight, for somewhere around $50 million. Of course you could get better economics from the space gun if you used it enough times to amortize the cost sufficiently. However, the total yearly demand for payload into LEO is only about 240 tons. The Falcon 9 can carry 10 tons in one shot. Only a small fraction of the payloads could be substituted by the space gun (hardened equipment). So maybe there would be enough demand for the space gun to replace 1 or 2 rocket launch equivalents. On top of that, the risks are huge. Global flight frequency would have to be much larger for people to seriously put the necessary capital into this.
Several posts have mentioned atmospheric drag. Drag slows the spacecraft. Traveling 8 km/s through the troposphere also induces heat and stress. The spacecraft would need a sturdy structure and thermal protection system or it'd burn up and/or crumple like beer can.
An important factor would be the flight path angle as the space craft exits the muzzle. If the spacecraft departs at a 0 degree flight path angle (in other words, horizontal), it must travel a long ways through the atmosphere. A huge column of air would be displaced:
That part of the path traveling through the atmosphere is colored red.
Over a trip of around 8000 kilometers the ship endures slowing resistance, heat and pressure.
If the flight path angle is close to 90 degrees (in other words, near vertical), the trip through the atmosphere is a lot shorter. The ship endures extreme atmospheric friction for only about 100 kilometers. But then the orbit would be an elongated ellipse that would quickly return and crash into earth's surface:
A typical rocket starts with a nearly vertical ascent. As the air grows thinner, it leans more towards the horizontal. It doesn't do the major horizontal burn until the ship is well above most the atmosphere. This isn't an option for space guns.
The notion is much more viable for airless worlds like Luna or Ceres.
I think there is another, more fundamental problem which hasn't been mentioned by the other answers. The drag on a body due to friction with the air increases quadratically with velocity. The drag force is given by (taken from this Wikipedia article) $$ F_D=\frac12\rho C_D A v^2, $$ where
This is a problem because a gun would need to impart all of the energy necessary to get the object into orbit right at the beginning, meaning that all of the energy would need to be stored in the objects velocity rather than the chemical energy of its boosters. This would significantly increase the total amount of energy necessary due to the quadratic dependence of the drag force.
Fundamentally, the issues are that a functional spacecraft is only engineered for between 3 G's and 30 G's; typical rockets only to about 100 Gs. (3.5 for the shuttle, 30-60 for ICBMs.) A reasonable launch tube would require an exit velocity in excess of 7.8km/s (the LEO speed). Adding 1 km/s for atmospheric losses, that gives a need for about 8.8km/s; for simplicity, let's round up to 9km/s. Note that a 180 km high suborbital flight was attained by the US Navy HARP project, with a 3.6 km/s launch velocity and near vertical launch.
9 km/s, at a reasonably robust 30 G's, is a hair over 30.6 seconds. That requires (using d=0.5AT^2) 137644 m... almost 138 km. (The HARP space gun was 41m long, and launched at 3600m/s, near vertical, and around 18500 G.)
Peak electronics safe launch is around 15500 G's (used on artillery shells), but designing spacecraft for that is highly unlikely. At that acceleration, and using the same 9m/s target, it's 274 meters long, and in the barrel for 0.06 seconds.
A more reasonable benchmark is the 100 G's that is routinely survived in momentary impact events - it's a good benchmark for frame survival of relatively thin-skinned craft. (After all, most cars survive this level relatively intact. It's also the acceleration of the Sprint Interceptor Missile.) At 100 G's, that's 9.2 seconds in the tube, and about 41.5 km long.
A human-safe launch is 12 G's... for a target of 9 km/s, it needs to be 345 km long, and is 70.6 seconds in the tube.
As can be seen, the length of the tube is a huge limit. Essentially, one has to launch a small missile - it doesn't need a whole lot of delta-V, just enuough to raise the perigee - and thus the reasonable level is a high acceleration missile. At that it's impractical, but doable. (Designs for this have been worked up by various engineering schools as sample problems.) The problem being that the gas can only accelerate so fast, and to maintain a consistent thrust requires adding additional combustion along the tube.
It's possible now, but impractical. It would be monumental architecture, unusable for humans, but a practical method of delivering certain categories of payload to orbit - bulk fluids, such as fuel, water, and air, and raw materials for construction, such as girders. It would require on-orbit tugs. Any failure has no abort mode, as well, since the payload is unpowered in flight.
I just saw Kinetic Launch system. It managed to perform a sub-orbital launch but plan is for a 2024 orbital one. It works by rotating an object and shooting it out with an extremely high speed.
It would still require some second stage burn though.
