Tying Knots With Light 125
thedreadedwiccan points out a summary of a recently released physics paper about tying knots with light. A pair of researchers showed that a relatively new solution to Maxwell's equations allows light to be twisted into stable loops. They are designing experiments to test the theory now, and it could have a big impact on fusion technology. The paper's abstract is available at Nature, though a subscription is required to see the rest. Quoting:
"In special situations, however, the loops might be stable, such as if light travels through plasma instead of through free space. One of the problems that has plagued experimental nuclear fusion reactors is that the plasma at the heart of them moves faster and faster and tends to escape. That motion can be controlled with magnetic fields, but current methods to generate those fields still don't do the job. If Irvine and Bouwmeester's discovery could be used to generate fields that would send the plasma in closed, non-expanding loops and help contain it, 'that would be extremely spectacular,' Bouwmeester says."
Re:The summary misses the key point (Score:5, Informative)
It's just cool (though maybe unrealistic)! (Score:5, Informative)
The (slashdot) summary really does miss some of the key points, and emphasize the "fusion containment" aspect, which I doubt anyone takes seriously as a use of this. One of the points that I think is key is the whole subject of homotopy groups (which I've really just learned about).
Maxwell's equations (and the wave equation, the Helmholtz equation in momentum space, etc.) have a family of solutions characterized by various parameter values. When you first start learning physics, you typically only allow real-valued wavevectors, which leads to only propagating waves and so on. Later on, you start to realize (as did George Green around 150 years ago, and Newton realized experimentally) that allowing for complex wavenumbers is more appealing mathematically (because it allows for more complete solutions), and actually leads to physically realizable solutions that propagating waves just don't give you. The effect of passing from real to complex wavenumbers is, on the face of it, crazy, but easily understandable once the analysis is carried out, and simple to visualize on an Argand diagram.
However, homotopy groups (if I understand it correctly) say that there may be other solutions to such equations (in nonlinear/dispersive media) which one can't get to from just simple replacements of real with complex numbers, and so forth---these divisions are the "families" of solutions. There just isn't a simple projection from one family of solutions to another, and the solutions of from one may bear no resemblance to the solutions from other famililes. This means that there may, in sufficiently complicated systems, be physically realizable behaviors which a system may fall in to, which aren't describable by the "usual" solutions of the equations. Of course, Maxwell's equations work wonderfully in all situations I've ever heard of (no concession to the "Electric Universe" wackos!), so perhaps nature, for some reason, won't allow other families of solutions to make themselves known on any scale I know of.
Re:Ok, questions (Score:5, Informative)
1. How do you bend light without passing it through matter or using a grav field that will crush the experiment?
Magnetic fields will bend light, which I believe is what this paper was based on.
2. If they can bend light, why are we using electron beams for crt's?
Because it's easier to bend a stream of electrons than a stream of photons.
3. If you could build loops of light can they be modulated to store information and read it back again?
I suppose, in theory, but it wouldn't be the most efficient means of data storage.
The reason, I think (IANAP), that this could be important to fusion reactions is that a photon loop within a plasma could heat the plasma to fusion-levels without the plasma trying to burn it's way through the outer walls of the reaction chamber. Current torus designs, I think (IANA nuclear scientist), run the plasma around the inside of a magnetic field, like cars on a racetrack, to get the energies necessary for fusion. This causes that super-hot plasma to push against the outer part of the magnetic field, which has to be extremely strong to contain it.
Re:The summary misses the key point (Score:3, Informative)
Because for the last fifty years, fusion power has been constantly just twenty years in the future, that's why.
No.
The ITER guys [iter.org] state that it will take until the 2050s until the first production fusion powerplant comes online.
Re:Ok, questions (Score:3, Informative)
The reason, I think (IANAP), that this could be important to fusion reactions is that a photon loop within a plasma could heat the plasma to fusion-levels without the plasma trying to burn it's way through the outer walls of the reaction chamber. Current torus designs, I think (IANA nuclear scientist), run the plasma around the inside of a magnetic field, like cars on a racetrack, to get the energies necessary for fusion. This causes that super-hot plasma to push against the outer part of the magnetic field, which has to be extremely strong to contain it.
Not quite. In a tokamak, the plasma isn't accelerated around the torus to heat it. The basic method is ohmic, or resistive heating, where a current is induced in the plasma with magnetic fields. The current across the plasma resistance generates heat. This is kinda like your concept, but not exactly.
Ohmic heating is typically insufficient for reaching fusion energies. The other methods of heating rely on direct energy injection, either through RF or neutral ion beams.
Regarding containment, the magnetic field in a torus is not like a hard wall; it only presents a permeable barrier that particles are still able to diffuse across. If you turn up the magnetic field, you slow down the diffusion, but turn it up too high and you risk plasma instabilities. The key is to control the energy leakage to a point where enough energy stays in the plasma long enough to sustain the reaction.
I haven't read this new paper yet, so I can't comment on its applications to fusion.