Project RAINBOW, nuclear magnetic resonance and E. Purcell
J - Some time ago, in one of our first meetings, you told me that the so-called 'Philadelphia Experiment' was some sort of a precursor of Purcell's discovery of nuclear magnetic resonance in late '44. I think you said then that Bitter thought that residual magnetism might be a nuclear phenomenon - ?
W - That's right, that's how this whole story really started for Bitter and for the MIT Rad Lab - it's also where Bitter falsified what he actually did for the government. Anyway, he'd been sworn to secrecy - and never received any public recognition for his contribution to the discovery of nuclear magnetic resonance. He suggested that there might be resonant magnetic energy levels, literally unoccupied quantum levels, that had to be reached before the residual could be completely eliminated. He also claimed - following [I.] Rabi's studies - that there was no doubt that nuclei were magnetic dipoles, and so we had to consider the possibility that residual magnetism was a gross external magnetic effect resulting from the magnetic properties of ferromagnetic nuclei. This part of the work refers to the obliteration of the ship's magnetic image - not to what we discussed before, the creation of a false ship image, a ghost.
J - Twin ghosts? really! - one made into a ghost ship, and now the other, a ship ghost...
W - Hah-hah, yes... Bitter's idea was that one needed to combine a permanent magnetic field with an alternating one. Achieving this with permanent magnets was impossible - even today, it still is. So instead, one had to drive permanent electromagnets with homopolar generators, to create the permanent field, and then superpose an alternating electromagnetic field that constantly changed the orientation of the atomic dipoles. The idea was a bit more complex, because the permanent field was also periodic, it was also pulsed, but at a much slower rate, and changed its magnetic polarity 180° with each impulse.
J - Fundamentally, this is the method employed by [E.] Purcell at Harvard to discover proton magnetic resonance in December of '45!
W - ...and don't forget [F.] Bloch at Stanford. Yes, after the studies of Rabi at Columbia in the 1930's. The problem was that, back in '42, Bitter lacked most of the equipment needed to produce or detect these nuclear magnetic resonances, and no one knew where they were, at what energy levels. As you know, nuclear magnetic resonance is --
J - Please explain -
W - Yes, in a permanent magnetic field, atoms and nuclei do not so much oscillate like small permanent magnets would, along an axis parallel to the magnetic field or the force vector, as they rotate like precessionary gyroscopes around the direction of the constant applied magnetic field. From discussions with colleagues at the Rad Lab, Bitter got the idea that it might be possible to superpose over the constant magnetic field, a resonant oscillating radio-frequency field that would selectively flip the direction of some nuclei. He was encouraged by both Purcell and Bloch in this.
J - I don't see the rationale...
W - In a permanent magnetic field, atoms and their nuclei align themselves in predictable ways. In a magnetized piece of iron, all the molecular dipoles (12) will tend to orient close to, or align with, an axis parallel to the magnetic lines of force. Some, as I said, align parallel and others antiparallel. This is a little more complicated than I am making it, because these alignments are precessionary. In iron, the lowest energy state is the parallel alignment. But there are substances whose dipoles lock in parallel and antiparallel orientations within the same magnetic domain - they are called antiferromagnetics, like permanganate. If they are heated above a certain temperature, they become paramagnetic. Atomic hydrogen is like an iron dipole - it preferentially settles in the parallel orientation. The problem then became how to flip sufficient parallel, lower energy states into antiparallel, higher energy states, so that residual magnetism is cancelled - much as it happens in antiferromagnetic substances. From Boltzmann's thermodynamics, Purcell did not expect the distribution of the two orientations to be homogenous - their quantum energy levels or excitation states would not be symmetrically distributed between the two main alignments. The presumption, as I said, was that the lower energy states predominated to generate the permanent magnetic field. Maybe residual magnetism was due to this predominance, no matter how small. With a particular radio frequency signal at the right frequency, it might be possible to shift more atoms or nuclei from the lower to the higher excitation levels, and achieve a balanced distribution. There was no equipment designed to detect these energy absorptions in water, let alone in ferromagnetic materials. But such a balanced distribution could result in the elimination of residual magnetism, and be used to prevent or cancel out induced magnetism.
J - So RAINBOW was a precursor of NMR, is that it?
W - You might say so. In an NMR machine, the 'samples' are placed in a permanent, static magnetic field, and a transverse 'radio-frequency' field is continuously applied at the Larmor frequency to cause zeemanizing, or the splitting of the excitation states. If the permanent field is increased, any nucleus precessing in parallel orientation becomes more resistant to being flipped into the antiparallel orientation, and so higher frequency radiation is required to flip it. When the particular combination of an external magnetic field and the applied RF field causes atomic nuclei to flip, the nucleus is said to be in resonance, in a state of nuclear magnetic resonance. Part of the original RAINBOW protocol was similar to this, except that the permanent field was periodically switched 180 degrees, and the RF field was also pulsed.
J - OK. I think I 'm beginning to understand this at last --
W - A constant magnetic field applied to paramagnetic substances always induces the creation of a molecular magnetic field. Because the induced magnetic field has two orientations for a precessing paramagnetic substance, parallel or anti-parallel, when one forces the parallel into anti-parallel flip at electromagnetic resonance, the induced molecular field ceases to aid the applied magnetic field, and opposes it instead. As the magnetic field strength is increased, the parallel nuclei become more resistant to flip, and higher energy RF must be injected at a higher resonant frequency for the flipping to occur. When there is permanent magnetism, even residual, in a substance, and the applied permanent magnetic field is parallel to it, the two fields add and the lower energy molecular dipoles are said to be shielded, because more RF energy is required to flip them into the antiparallel orientation. If the applied permanent magnetic field opposes the permanent magnetism of the target, then the lower energy dipoles are said to be deshielded, because less RF energy is required to flip the dipoles.
J - I see, that's why Bitter wanted to pulse the permanent magnetic field and reverse its orientation - he would periodically deshield the parallel dipoles, making it easier to have them flip.
W - Uh-huh.
J - And if the permanent applied field was constantly opposing the residual magnetization, would one reach a balanced state like that of antiferromagnetism?
W - That's roughly the idea, but these are quantum processes, and one never even gets close to it. As Purcell found out in late '44, stimulated emission compensated for the absorption --
J - But the idea was to dissipate residual magnetism by balancing the parallel states with more antiparallel states --?
W - Yes, a very tough problem indeed. No one knew where these resonances were, and by [J.] Ewing's theory of molecular magnetization, magnetic alignment of the molecular domains or the molecular dipoles is a step process that takes time and never reaches saturation. There were lags in demagnetization and relaxation - and much heat could be expected from partial gyrations and countergyrations of the magnetic domains.
J - What would happen if one could, say, align all domains, make them all either parallel of antiparallel?
W - That's the problem of superconducting magnets. Making most of the dipoles parallel is the problem of magnetization or induced magnetization of a sample. Tesla had already encountered this problem when he designed his electromagnet rectifiers with iron cores. Fields much greater than 3,000 gauss were needed to bring the cores to saturation. Beyond that limit, Tesla claimed that one had to employ low-frequency disruptive discharges -- which means pulsing the coil.
J - So Tesla preceded Kapitza, the French and Bitter in this technique?
W - Yes, in the case of Tesla, you're trying to get as many domains as possible into a parallel orientation - but as for turning half of all domains into antiparallel orientation, that is well nigh impossible, at least with ferromagnetic or ferrimagnetic cores. By 1938, at Bitter's Magnet Laboratory at MIT, fields of 100,000 gauss were attained. Your question is important though, because Einstein's general theory proposed that gravitational fields have two main components, one static and present in space devoid of matter, and the other dynamic and caused by the gravitational coupling of two or more bodies in relative motion. Two spinning bodies would exert a mutual force of gravitation. If the nuclear gyroscopes of a rotating body could be aligned in a preferred direction - say, parallel or antiparallel to the body's axis of rotation - they would generate a force field normal to that axis. The alignment condition is called spin polarization. If the force field varied periodically, then it might be possible to generate a secondary gravitational field.
J - Would this then superpose another curvature upon the local spacetime continuum?
W - If we disregard magnetic torsion, one can only think of it as either intensifying the existing curvature or relaxing it. This is equivalent to saying that it would either increase the density of the gravitational field flux lines, or decrease it.
J - How could we disregard magnetic torsion? Isn't there a relation, for example, between the earth's rotation and its axis, and the orientation of the magnetic field and its axis? Couldn't the apparent magnetic axis offset be the mean of the precessionary motion? After all you jumped from magnetism to gravity - but you were assuming, implicitly, that the two fields were coincident or nearly so, no??
W - I understand --
J - ...even the idea of varying the line density goes back to your discussion of Faraday's notion of magnetic lenses --
W - Well, yes - hmm, hold those magnetic field notions in your mind, and suppose that it is possible to give a single description of them which could be made identical to a description that would apply to any gravitational field. Now, in the general theory where there is only one metric tensor g to express local geometry, g is locally determined by the gravitational field, by its local flux line density. The metric tensor simply expresses the acceleration of the frame of reference. Therefore, the flux line density of a spin-polarized, rotating body must decrease with respect to the flux line density of surrounding space, if the gravitational attraction of that body towards any other revolving one is to decrease. The spacetime occupied by the spin-polarized rotating body would have to have less of a curvature than the surrounding spacetime.
J - But I don't understand - that would only allow one to decrease the acceleration of the local gravitational field. How would weightlessness or antigravity be possible?
W - It wouldn't - not from any of Einstein's theories, unified or general, by the way. But some interpretations permit one to think in terms of a shield of the spacetime occupied by the spin-polarized rotating body. The weight of two bodies with respect to each other would only exist if the secondary gravitational field within that shield had a line density greater than that of the space surrounding either body. In contrast, weightlessness would be a condition of degravitation, where the density of gravitational flux lines within the shield of either of the spin-polarized rotating bodies would be equal to that of surrounding space. And antigravity would be a negative weight characteristic reached when the line density within a shield was less than that of surrounding space. If this condition occurred, then spin polarized nuclei would align themselves antiparallel to the weight vectors.
J - Something like a curvature that can be straightened or even inverted to form a geometric negative?
W - That's the idea. Yet this interpretation (13) cannot be entirely correct. Geometry results from gravitational fields, from states of acceleration. It is, of course, hardly possible to understand how space devoid of matter is subject to a state of acceleration when it is equally supposed to be devoid of energy, when time itself is taken to be one-dimensional and simultaneity is seen as a relative state. Somehow, one has to imagine space as being subject to an acceleration, without being able to treat that same space as a physical property of energy...
J - If I understand you correctly, Reich wanted Einstein to do just the reverse - to treat that space empty of matter as a property of massfree energy --?
W - Yes. One might think of energy in flux as subject to an acceleration, even if treating the energy as massless presents conceptual difficulties. But to say that space is in motion, or is subject to acceleration, is physically meaningless. It is at least as mysterious as Faraday's lines of force and Maxwell's superposed, counter-rotating vortices. But suppose nuclear spin polarization could permit us to alter or even invert the curvature of spacetime. Then, the main technical difficulty would be to come up with a process that flips most of those gyroscopes from parallel to antiparallel orientations with respect to the mutual weight vectors. That's where magnetism, or some form of it, comes back into focus. After all, outside of a critical distance, two mutually gravitating bodies do not fall towards each other - some force already keeps them apart. And unless one takes recourse to the gravitational shield interpretation, one is only left with the spacetime torsion.
J - ...and neither is satisfactory? What about nuclear magnetic resonance - could it not be used to cause that torsion at very high field intensities?
W - Well, the solution can only come from asking whether a body which had a majority of its magnetic dipoles oriented in antiparallel magnetic direction could also function as a body which was nuclei spin polarized in antiparallel orientation with respect to its main weight vector. Without introducing the question of the magnetic properties of space or of some form of massfree energy, it's difficult to see how the gravitational relationship could be modified or inverted. You may regard the shield as a finite spacetime region. And one can treat the curvature as a matter of flux density - but there is no physical mechanism to create relative differences in this density that would generate opposing curvatures, 'convex' curvatures.
J - So RAINBOW was going to test for the complete removal of stray magnetism with methods analogous to NMR. But it was also going to test for general relativity and the unified field - to see if a high intensity rotating magnetic field generated a secondary gravitational field, and if either its curvature or torsion - or both? - would change...?
W - Yes, engaging magnetic resonance with very intense fields might change the local curvature of spacetime that 'contained' the ship - that was the idea. This could bend all light rays further than they are already bent on the surface of the earth and...
J - ...give the wrong optical or electromagnetic image of the ship's location.
W - Yes, the electromagnetic image of the ship would be red-shifted, and the ship would appear to be further away than it was. The radar bounce would take longer to return, the bounce would be stretched or bent longer, and would not return therefore on the expected frequency for which the receiver was tuned. You see, this is the reason why RAINBOW was such an important project - and its outcome would bear on many different fields. It would address a number of very different questions, all of them of crucial importance.