EM Drive paper passed Peer Review

Or, why passing peer review doesn’t suddenly mean that a technology is either validated or useful.

I just saw an article in Universe today claiming that a paper on the EM Drive is forthcoming. As you may remember from my previous post, the EM Drive is a piece of crank technology that is The One To Bring Them In and In Darkness Bind Them of the crank technology world. As they all know, it is about to change everything! (Or so they say.)

The device is an assymetrical microwave cavity which will apparently generate thrust when microwaves are injected into it without producing an apparent exhaust stream. The creator, Robert Shawyer, repeatedly invokes a crazy wrong interpretation of Special Relativity in order justify why his doodad works and makes grandiose claims about the capabilities of the device. Guido Fetta, a chemical engineer turned speculative technology wonk, has also jumped out into the public about his grand claims to test the device on a cubesat in orbit soon… Fetta’s description of why his “Cannae Drive” works is somewhat more reasonable than Shawyer’s is, but still a bit iffy…

The Cannae Drive also features an asymmetrical cavity, but is flatter than the EmDrive. According to Fetta, it works by deriving force from a reduced reflection coefficient at one of the device’s end plates, due to imbalances in the Lorentz force (a combination of electric and magnetic force on a point charge due to electromagnetic fields). Nasa Eagleworks, on the other hand, suggests that the Cannae Drive works by the cavity pushing against a “quantum virtual plasma” of particles that shift in and out of existence.

This description is actually not terribly aphysical because it’s essentially describing exactly what happens in a laser. Believe it or not, the NASA description is the crankier version since it seems to be invoking something along the lines of Casimir force. I’m not a huge fan of Eagle Labs because they skirt the ragged edge of being cranky themselves sometimes. (If it all works, I will gladly eat my words.)

I think that the one word that may be useful in this mess is the word “propellantless”… I mention this here because there could actually be a big difference in utility between claiming that the drive is “reactionless” (which is impossible) and “propellantless,” but this comes back to one’s definition of the substance of “propellant.”In the end, if the justification for the drive is simply that you don’t need to take along a huge quantity of reaction mass to make it work and can instead use a nuclear plant to power it, that is not necessarily a bad thing.

Still, house needs to be cleaned.

First, the device must be described to work in a way that matches physics. No insane invocations of Special Relativity. This paper coming out is actually a nice first step toward doing just that. Passing Peer Review is a way of saying “Yes, science is being done! We have made measurements by accepted methodology and here are our results!” Which is actually much more impressive than anything that has come out of either Shawyer or Fetta for the last decade.

Making and reporting measurement is really all there is to experimental science: we may not have the interpretation right just yet, but we have numbers that can be compared to everything else in the field. How does the efficiency actually compare to a chemical rocket? Spin the numbers! It is all to show that the methodology is sound and the numbers are honest. And, those numbers will have to ultimately say that momentum and energy are conserved. The device is not… I repeat NOT… a reactionless drive. If it has a propellant, the substance is probably in photons, not gas or plasma like in conventional chemical rockets or ion drives.

The second thing that must happen is that the device should be engineered. The core of engineering is tweaking the physical parameters of the system to optimize the functioning of the device, which requires a model of the behavior… whether you understand the physical rationale behind it or not. Again, this Peer Reviewed paper is a terrific first step because it starts to characterize the actual observed behaviors of the system. If the rumored thrust is actually 1.2 mN/kW, great! A millinewton is a higher thrust than I was estimating in my previous writing, but how big of a powerplant does that require? A nuclear submarine can carry a 500 MW reactor, which would theoretically give hundreds of Newtons of thrust, which is not insignificant at all if the rumored numbers reported by Eagle Labs are true. Now, explain why and begin to tweak the envelop. If it is just a big microwave flashlight, fine, start plugging the physics into that and tell me what the actual performance limits are.

I will admit that my previous post may have been somewhat in error: it may turn out that this research is not a waste of time, but we’ve got to get away from the cranky hopefulness and start figuring out what we’ve actually got so that we can make it better.

Now, I have made something of a shift of stance in my writing of this post. Previously, I flat out called the EM Drive a waste of time. For a very long time it looked like a vanity obsession of a garage crank with delusions of popular fame. As long as it has that air, I won’t have much nice to say. Mutilating physics to build a miracle machine is crankery and there’s way too much of that happening in our world right now. What has changed now is simple: if there is a real, explicable physical phenomenon to measure, steps forward can be taken to find a real thing. It would be nice if there’s a world-altering discovery lurking in here, but that isn’t what we have yet. It really ultimately doesn’t matter to me where the idea came from, whether it came out of somebody’s garage or some rocketry lab… millions of ideas come from everywhere all the time: the point of the science is to sort through and find which observations are actually useful so that we can discard the ones that aren’t.

We’ll at least see if there’s something useful here and hopefully have a real guess about why it works. If the numbers are not reproducible or if there is some huge other way to interpret what has been seen, then it becomes time to discard the EM Drive. I guess that’s kind of the weird thing about frontier science: it always may not survive the meat grinder, no matter the source.

The Waste of Time That is the EmDrive

I stumbled over an article yesterday that I found again today in my newsfeed that finally causes me to be willing to spend time writing about something. The article is here.

This article is about a sketchy physics topic that the popular media loves itself something fierce. Namely, the EmDrive.

This device is supposed to be a form of engine that can drive spacecraft faster and more efficiently than current technology otherwise allows. The creator of the EmDrive loves to claim that the device can solve all the world’s ills, from the energy crisis and global warming to the drip under your sink. Never mind that the excessive claims should set everybody’s danger sense a-tingling, it is a device that has persisted past it’s creator’s obvious lack of background in the basic science of physics.

Here’s a description of how the EmDrive is supposed to work as quoted from the article:

How the EmDrive works

The EmDrive is the invention of British scientist Roger Shawyer, who proposed in 1999 that based on the theory of special relativity, electricity converted into microwaves and fired within a closed cone-shaped cavity causes the microwave particles to exert more force on the flat surface at the large end of the cone (i.e. there is less combined particle momentum at the narrow end due to a reduction in group particle velocity), thereby generating thrust.

The general idea here is that you’re injecting microwaves into a hollow cone and allowing them to bounce around. Because the cavity is asymmetrical, the argument goes, they end up breaking symmetry on the pressure they’re exerting and push the cone only in one direction.

In a way, the set-up is almost an exact duplicate of the thought experiment that Einstein used to come up with the equation of E=mc^2, but beyond that, this is actually a flagrant violation of conservation of momentum. You can think about it this way: a guy standing inside a train car pushes against the wall of the train car… no matter what the shape of the inside of the car, the guy walking around never moves the center of mass of himself with the car –even if the car will actually move slightly as he walks back and forth, if the axles of the car and the rails are frictionless. The only way the car can continue to move is if the guy goes running along the car and jumps out the end, thus enabling his center of mass to be decoupled from the car… if he were to keep running, the center of mass of his system and the car would remain at rest, while he and the car must both be moving in order to conserve the net zero momentum they started with as a system. The analogy breaks down because the guy standing on the ground would be able to exert force to stop running. As Christopher Nolan wrote in Interstellar: the only way to go somewhere in space is to leave something behind. As a physicist’s aside, one has to put in a train car analogy at least once in this discussion because Einstein loved trains during his explanations of special relativity (I’m convinced that this is part of why Sheldon Cooper loves trains).

Breaking conservation of momentum is a pathological, ‘do not pass go’ fault that should immediately consign this whole EmDrive concept to the dumpster the same way Avagadro’s number kills Homeopathy. Despite that, the creator of the EmDrive has a ready response:

However, Shawyer claims that following fundamental physics involving the theory of special relativity, the EmDrive does in fact preserve the law of conservation of momentum and energy.

The author’s recourse is “Don’t worry about it, it’s hidden in special relativity!” Having dealt with special relativity and being aware that Einstein used this very thought experiment to prove E=mc^2, I can assure you that violating conservation of momentum is still completely fatal to an argument. Special Relativity isn’t exactly an impassable mountain that breaks conservation rules the way General Relativity does.

Despite all of that, various labs around the world have built EmDrives to test the idea. In the end, this is sort of like continuing to test whether or not autism is caused by vaccination, but okay, fine.

To everyone’s surprise, some of these labs, including Eagle Labs at NASA, have reported tiny tiny thrust. Something smaller than micronewtons IIRC, but still thrust.

And, of course, the cranks go wild! Here it is, the reactionless Cannae drive that will take us to Alpha Centuari by 2035 and Vulcan by 2150.

Now, the fact is that while these labs have reported thrust, we don’t know exactly why it did. Sure, it did, but we need a theory that sits within physics that explains why it did. Rest assured, the reasons given by the drive’s creator are completely bogus, so new explanations are needed. Is it Casimir vacuum pressure? Is it warped spacetime? Is it Calvin’s Universal Transmogrifier? We need to figure it out.

The Finnish physicist in the IBTimes is remarkably conciliatory even if his tacitly favorably worded response is actually just another huge nail into the EmDrive’s already well-built coffin. Here is this physicist’s explanation:

“The EmDrive is an engine like any other engine. It takes in fuel and produces exhaust. The fuel side is easy for everyone to grasp – microwaves are being fed in. The trouble is, we don’t see anything coming out, which is why people think it doesn’t work,” Annila told IBTimes UK.

“So how could something come out that you can’t detect? Well, the photons bounce back and forth inside the metal cavity, and some of them end up going together in the same direction with the same speed, but they are 180 degrees out of phase. Invariably, when travelling together in this out-of-phase configuration, they cancel each other’s electromagnetic field out completely.

“That’s the same as water waves travelling together so that the crest of one wave is exactly at the trough of the other and cancelling each other out. The water does not go away, it’s still there, in the same way the pairs of photons are still there and carrying momentum even though you can’t see them as light.

“If you don’t have electromagnetic properties on the waves as they have cancelled each other out, then they don’t reflect from the cavity walls anymore. Instead they leak out of the cavity. So we have an exhaust – the photons are leaking out pair-wise.”

Whatever else I might think about everything here, this is actually not a bad explanation. Photons have the quality where they can be superpositioned: if you pick two photons headed in the same direction, of the same polarization, with their E-fields 180 degrees out of phase, the Poynting’s vector still exists, allowing them to still carry momentum, but their field oscillations will cancel out. If they are introduced pair-wise in this manner, there’s not a reason to think that they can interact with matter any longer and they could simply slip straight through the confinement of the drive and off into empty space. So, the thrust from the drive would then be generated by the physical asymmetry of the cone allowing photons to pair up and escape easily in one direction, but not in another.

As a slight aside, I think I disagree with the Finnish physicist’s usage of water waves in the example above. The reason is that macroscopic waves in water are not discrete the way photons are. By effecting the continuum of material in water, the displacement of the wave crest from its resting state is what contains the energy of the wave: for small displacements, the momentum is perpendicular to the direction of travel. By adding a second oscillation 180 degrees out of phase, you completely cancel out the energy of the wave… and no wave remains after the fact. I’ve been thinking and I continue to think about whether or not the same is true with photons. I don’t think that it is mainly because photons are quantum mechanical particles and they have a quality of being discrete objects in the sense of their particle-wave duality. Photons contain linear momentum parallel to their direction of travel, while a water wave does not (the momentum is perpendicular to its direction of travel), and two photons caught traveling in the same direction must conserve momentum, regardless of their phase.

Now, I am granting here that there’s a physical explanation for why thrust is being generated, but we’ve slipped into explicable physics. If you stop and think about what we’re talking about, all we’re talking about is a very specialized form of microwave antenna. If you want thrust from momentum carried away by emitted microwaves, this process of pairing up photons so that they become invisible to the walls of the device (and sensors behind the device) is sort of beside the point. Granted, it would not torch anything behind it, this device is not the most efficient way to produce photons in the form of thrust. A flashlight or a laser would be much more efficient at converting power into thrust by doing essentially the same thing.

You could presumbly do an experiment like this one with a lasing cavity using optical light. I would partially-silver the surface of one mirror with a coating that is about a quarter wavelength thick before you hit the actual mirror. It’d be technically challenging since you’re talking about 1/4 micron thicknesses that are the tolerance of the lasing mode, but that’s something that can be attempted. Provided you stay at a condition of optical gain in the cavity modes (not all the photons are canceling) you should be able to test whether the recoil of the laser body due to the emitted radiation is the same as the recoil at the laser light spot. You could probably just set it up as an interferometer with adjustable arms and forget the coating. Again, this would depend on polarizing the emitted light.

Just thinking about it, I can imagine several more ways to test this in an optical setting. Some of them could be quite cheap to do.

Point is that there’s nothing magic about this.

Again, the problem with the EmDrive is that it’s exploiting physics to not do the most efficient thing it could do at its supposed task. If you start tabulating the amounts of power needed to generate thrust that is appreciable by these methods, you’re going to start tripping over conservation of energy somewhere. This not being magical, the amounts of energy needed to do anything are also not magical and will turn into eyepopping numbers when you start demanding that the thrusts the engine can produce are big enough to move masses humans might want to move with it.

I’ve been quite generous here. This is supposing that the explanation the Finnish physicist has supplied is useful over the other potential sources of noise in the experiment –the micronewtons or nanonewtons Eagle labs reported is so tiny that somebody’s breath could have been hitting the side of the experiment.

Scaling this thing up in force is crazily hopeful and would require you to jettison basically the whole design and go with something that does better what this device is actually doing. At some point, it will be time to forget about this EmDrive and relegate it to the wide-eyed, hopeful crankery that it is.

“Breakthrough Starshot” or Are You Serious, Mr. Hawking?

His heart’s in the right place, I will admit that much.

The “Breakthrough Starshot” project was bound to be proposed by someone eventually. Postage stamp-sized space probes riding a beam of light to Alpha Centuri, then beaming messages back to us of what they find there, all with Stephen Hawking’s stamp of approval. Explore another star! Nifty idea, truly, but can we really do it?

Hawking’s intentions are good. From a long-term perspective, there’s a 100% chance that the Earth will endure another extinction-level event. It has happened repeatedly in the history of our world. All the species of life living here and now are dead. Eventually. If we’re here when it happens, we’re dead too. There’s no chance it won’t happen. It may happen tomorrow, next week, next year, maybe next century or even a million years from now. If anything is living here a billion years from now, our sun will have changed its radiant output enough to render this place uninhabitable anyway. Our little oasis in the void won’t last forever. It will happen, even if the probability of it happening in the life span of anybody living now is very very small.

Admittedly, we’re biasing the statistics a little at the moment. There are enough people spread across the face of this planet that everything bad that ever happens will have a human audience caught up in it. And, worse, news of every bad event can now circle the whole planet in minutes by Twitter. It’s also an active discussion that there are enough humans on this planet that our activities are fouling the environment to the point where we may not be able to live here any longer. The notion is scary especially since we really don’t know enough one way or another to tell what can or should be done, if anything. It’s like being diagnosed with cancer, but not having been told yet how serious the disease is or what form the treatment will take.

By truth or distortion, it’s no mistake that we increase the odds of our long term survival by not leaving all the eggs in one basket. This is Stephen Hawking’s thought: we increase our chances of surviving a thousand years from now if we have people living elsewhere… and not merely on Mars, but as far elsewhere as we can get. Hawking is in the unenviable position of being aware that extinction events may originate not here on Earth, or even anywhere in our solar system, but elsewhere nearby in the stars. From his perspective, the further we spread ourselves, the greater the chances we survive that big event that no one can see coming.

Someone eventually has to propose a trip elsewhere beyond the edge of our solar system and it has to happen in such a way that information can return to Earth quickly enough for us to do something about it. Hawking and a couple billionaires proposing it are not surprising to me.

Where I got tripped up is that the proposal was essentially a big piece of Alistair Reynolds Sci-fi. God love him, it’s an interesting read, but not a very plausible one.

The Yahoo article calls it an “Audacious” plan. Postage stamp-sized, laser-beam riding autonomous mini-robots that will get there in 20 years. Seems like they hit all the sci-fi buzz words: nano-beamriding-laser-AIs. Moore’s Law gone crazy. I’m sure it will sell well to the public and they can wave off the ‘audacity’ by pushing the technology requirements into the future. We will develop what’s needed.

Or can we?

If you stop to think about what they’re proposing, they’re talking about essentially being able to detect emissions from a cellphone at a range of 4 light years. That’s essentially all the more power this device can carry with it… that or less depending on how small you build it. I do like lasers, but they aren’t all-powerful. There are no sensors available to detect such a thing. We are not capable of detecting the albedo light reflected off the surface of a planet at a distance of 4 light years and they’re talking about detecting radiant output from an object not able to carry as much battery power as the typical cellphone…

I had no details on the solar sail they’re planning to propel this probe with, but I did some back-of-the-envelop calculations on the power requirements for propelling a ‘stamp-sized’ object with a laser. The concept itself is not really a fictional one: light contains momentum and can be said to have a pressure –I would recommend studying Poynting’s vector for anyone curious. Knowledge that light carries momentum was actually a prerequisite for Einstein’s E=mc^2 postulate (what, you’ve never heard of E=pc?) You shine light on a surface and that surface rebounds from the light it absorbs or reflects. The whole concept underlies the idea of solar sailing. Using a laser as your light source is not a big leap: lasers are more columnated than sunlight and can deliver a great deal of intensity to a tiny spot. Some calculations about the system are easy to make.

I estimate that a postage stamp sized probe needs to be hit by something like a 600 megawatt/m^2 intensity laser for the duration of a year in order to achieve the speeds they’re talking about. That’s the average amount of light power a 1 gram mass space probe must absorb to be accelerated to 20% of light speed in the course of one year. Laser lights are extraordinarily intense, but this rating becomes even steeper when you start to realize that even laser light must fall off as 1/distance^2 for the spherical radius of the laser wave fronts, however well columnated that light is. For a high finesse laser cavity, where the mirrors are curved, that radius actually ends up being shorter –corrective lenses can help, but this is still imperfect. For all existent lasers, the spot size disperses to meters across at a distance of thousands of kilometers, at maximum. How far does a little laser pointer reach? For radiation delivered for a year, coping with the inverse square fall-off of light intensity as the object moves away from us ever more rapidly, we’re talking about a laser that is probably as intense as terawatts/m^2 (in order to continue to deliver the average 600 megawatts/m^2 to our postage stamp sized craft when it is a distance of 1/10th of a Light year away from us). The power capacity required to drive that is terawatt*hrs, which is the energy necessary to drive a small (or big) country. The US generates something like 25 terawatt*hrs of energy. It would certainly take a billionaire to buy up 1/20th of the power generation capacity of the US for a  year.

This is not even considering that precarious problem of how to know where to aim the laser. If it gets some other force input (space isn’t completely empty) it might veer out of the laser path –it doesn’t have to move far, just meters, or even centimeters where you aren’t expecting it to. And, it won’t even travel straight with respect to gravitational interactions it encounters from the planets in our solar system as it leaves. We can’t detect objects that are a centimeter square on planetary scales, let alone extra-solar scales. You could possibly track it by back reflection from the laser propulsion beam, but if it falls out of the beam, then what? They talk about Stealth aircraft having radar cross sections the size of a pigeon… this has a cross section the size of a postage stamp. Maybe the best way to go would be lidar using the propulsion laser, just scanning the beam around until you find back scatter again, but that would depend on the probe maintaining its orientation. The whole problem gets harder and harder as a function of a square of the distance with the rate at which light intensity falls off… the solar system is light hours across where the propulsion system would need to be workable out into light months of distance, or about a fifth of a light year given how far and fast they’re talking about. Feedback of how to move the steering laser beam if the probe disappears when it’s light minutes away means that you can’t find it again for at least minutes and then your knowledge is always minutes old. What do you do when it’s light days or even light weeks away?

Another pesky problem is what to do about dust. Like I said before, space isn’t completely empty. The fragility of the probe is somewhat inversely proportional to its size. The smaller the probe is, the less likely it is to hit dust, but the more likely a dust strike can do critical damage. A small, thin, unarmored thing with highly reduced electronics and electromechanics getting hit by a relatively larger piece of dust is serious business. How many of these probes do they plan to fly in order to hedge their bets and how many lasers are they using? The Pluto New Horizon’s mission carried a dust counting experiment and it is known that dust will gradually wear a space probe down given enough time. For a probe so small traveling at tremendous speeds, that time is less.

All of this is still just dealing with departure and says nothing about the complication of how to return data once the probe reaches its destination –which will be a rapid flyby since there won’t be any way to ever slow back down again. Like I said, do they really think they can detect a broadcast, even a laser broadcast, from a cellphone at a distance of 4 light years? The stars themselves are just points of light and many not even visible!

I shook my head, “Surely you must be joking, Professor Hawking!”

I think that the problem with the proposal is that they’re trying to have their cake and eat it too. They want to explore another star and do it in a time horizon that is significant to our civilization. 25 years is a long time, but not really that long. Exploring another star is not a technically insurmountable task, I don’t think, but it is demanding of both technical expertise and patience. Realistically, 25 years may be a bit impatient on the scale of our universe. It’s true that we have to perform this task someday and this proposal itself may well simply be trying to force that ‘someday’ to come sooner, but is this a form in which that task can work? Given our current capabilities, no: it’s betting on what we don’t know yet. Of course the proposal sounds like science fiction… it is science fiction still.

The great problem may be that we don’t know how close that existential threat is. Perhaps the achievable exploration mission demands that our civilization remain stable for a longer duration than perhaps it can. Are we better than the Roman Empire? With 25 years, there’s a chance that all the difficulties we’re currently facing haven’t exacted their existential toll yet… with two or three hundred years, a slightly more reasonable time goal for the exploration mission at hand, maybe our civilization will already have disappeared and won’t remain to learn what information the probe brings back. We live in days where pre-Renaissance barbarians are actually knocking on the door… will we withstand that?

I’m skeptical, but… go for it, guys, I wish you the best of luck.