Over the last few months Popular Sciencei09Space.comWired and the Huffington Post have run stories on NASA experiments on "warp drive" technologies, which are baby steps towards developing starship engines that could send astronauts out of the solar system and across the galaxy. And today it made the New York Times.

How cool is that?  Quick answer: not quite as cool as it sounds.  

Despite everything you have heard about the speed of light, you really can design a warp-drive without repealing general relativity. In the 1990s, Miguel Alcubierre found a solution to Einstein's field equations that describes a "pocket" of space surrounded by a ring of material with negative mass. From the perspective of an outside observer, this pocket moves faster than the speed of light -- just what you expect from a warp drive.

That's the good news, but now for the bad news. First up is negative mass -- there is no known stuff that "weighs less than zero", and there are good reasons to think that negative mass is impossible in principle, rather than something we could find if we search hard enough. So while you can design a warp drive, the universe may not let you build it. It is true that the Casimir effect, from some perspectives, can reduce the energy (and thus the mass) in a cavity to negative values. Even then it's complicated, and with general relativity the devil is always in the details. For our purposes, though, the key word is "cavity" -- the walls of a Casimir cavity are made from regular stuff and to power a warp drive the whole assembly would need a negative mass, and it is hard to see how that would happen. So I'm skeptical, at best, but happy to wait and see what happens. 

I am keen to know more. However, the (publicly available) technical descriptions NASA's efforts are oddly sketchy and incomplete. The news stories focus on Harold "Sonny" White, who is using a White-Juday warp-field interferometer to demonstrate the warping of spacetime in the lab. He compares this to the Chicago pile which demonstrated the possibility of controlled fission, but was useless as an actual source of nuclear energy. Problem is, I can't find a detailed description of a White-Juday interferometer. Wikipedia has a helpful picture but it is short on details.

On the other hand, let me see if I can work it out for myself. Here's the question: Assume you build a little tabletop spacewarp -- how would you check that it is really warping space?

Well, general relativity tells us that matter warps space, which is felt by us as gravity (roughly speaking). We know how much space is warped by a given mass, since the warping creates a "well" around all massive bodies. Matter with negative mass would give us a hill, instead of a well.  Clocks inside a gravitational well tick more slowly than those in flat space. A photon's frequency is  a "ticking clock" -- so a photon in a gravitational well ticks more slowly than one in flat space, leading to gravitational redshift, a well-known and well-observed phenomenon. An interferometer is a device that measures the difference between two beams of light. We put the prototype spacewarp in one beam -- and if it changes the "shape of space" we can look for a change in the output of the interferometer. Without knowing exactly how this interferometer is set up, I can't do a precise calculation, but it is a reasonable guess that the effect of any warp field would be about the same size as the gravitational redshift generated by a similar mass. [For the cognoscenti: a photon in an interferometer also gains energy as it falls into a well, increasing its frequency, but we are just doing an order-of-magnitude estimate.]  

Interferometers are remarkably sensitive machines, so it is a reasonable approach to try. Now reach for the back of the proverbial envelope. A few grams of matter (ok, be generous, say a kilogram) in a radius of a few centimeters produces a TINY gravitational redshift -- a change in frequency of roughly a part in a hundred-trillion-trillionth. For a negative mass, the sign changes but the size of the effect is the same, but this seems like a reasonable estimate of any "warping" you might produce in the lab. To get the overall effect on the interferometer, multiply this by the number of wavelengths of light in the warp region which gives the phase shift -- the extent to which the two beams move out of step with one another. For a smallish spacewarp, its effect on photons in one beam of the interferometer is perhaps a few parts in a quintillion. (And remember that a negative mass of a few grams matches the energy generated in a small nuclear explosion, and a kilogram would be getting very big indeed.) 

I have tried to be optimistic; I could easily have been over-generous by a factor of a million. However, in Warp Field Mechanics 101 (the origin of the picture with the WIkipedia article) White talks about a sensitivity to "warping" of parts in 10,000,000 with an interferometer sensitivity of around 1/4 of a wavelength, which misses my estimates by really massive factors.  And at this point I am really quite puzzled. 

We can apply the sniff-test in other ways. Warp Field Mechanics 101 begins with a fairly straightforward treatment of the Alcubierre warp drive, but veers off into a discussion of a "brane world" model invented by Dan Chung and Katie Freese.  Brane worlds are fascinating. They propose that our (spatial) universe is a effectively a giant three-dimensional ribbon embedded in a larger spacetime. However, there is not a single, solitary shred of evidence that our universe is a brane world -- these are simply one idea among the many possible scenarios theoretical physicists play with. So it is jarring to see them in a write-up of an experiment that has nothing to do with brane worlds themselves. Likewise, many descriptions of the warp drive compare it to inflation -- which is the period of rapid expansion in the early universe. Drawing an analogy between a warp drive and inflation implies that you don't understand warp drives or inflation: inflation involves negative pressure  (which is no more radical than a stretched rubber band trying to contract) not negative mass.  

Last but not least, one of White's collaborators is a chap called Davies, who is affiliated with the "Institute for Advanced Studies at Austin" -- which is most definitely not the more famous Institute for Advanced Study at Princeton. And it does some pretty advanced stuff. You know, free energy machines, cold fusion - that sort of advanced. Their publications page includes a paper in the notorious "journal" Chaos, Solitons and Fractals along with several manifestoes by people who are, scientifically speaking, running for Mayor of Crazytown (to paraphrase Ethan Hawke). Which is not dumping on "alt-science" just for the sport of it -- but a spot of peer review would be more than usually welcome, given that NASA is spending both public money and its own reputational credit on this enterprise. 

So, my take is that even if you can generate a spacewarp in the lab, a quick estimate says its gravitational effects would be billions or trillions of times below the threshold of detectability. The problem is not that warp drives are impossible, but that this description of them does not seem to be self-consistent. The available technical data goes no distance at all towards providing the level of detail one would expect for a potentially foundational experiment, and some of the people affiliated with this effort are working on what can charitably be described as fringe science. Meanwhile, a basic premise of the warp drive proposal -- that we can generate large and controllable quantities of "negative mass" -- is at best an untested assumption, and at worst contradicts our current understanding of fundamental physics.

So, is this the start of our journey to the stars? Probably not. 

Postscript:  I met White at SciFoo this year and he was a remarkably pleasant man, and clearly an expert when it comes to regular rocket science. I did not want to write a "someone is wrong on the internet" blog about this, but between the Times and hearing questions about it from my students it seemed that it was time to weigh in. I would love to be wrong about this. Make it so. 

Note: I have edited to the comparison between my sensitivity estimate and White's.  If anything, it makes things worse, but does not materially change my conclusions.

Postscript 2: This story received a second run in the media in 2014, and I wrote a second blog about it.