E-Beam Propagation


EB Discharge


James Benford

November 2013

Recently, Alan Bond claimed that this photo shows multiple intense relativistic electron beams (REB) propagating to a target.

 I remember this photograph well because I worked on E-Beam propagation for most of the 1970s. The light you see is NOT REBs.  It’s exploding wires. The photo was made to illustrate the idea.  Nice photo, but gives a false impression.  (For one thing, e-beams are not that visible.)

To verify this I looked up the original paper.  In that paper [1] the photograph caption says “Photograph of exploding wires in air. This view is normally blocked by the vacuum chamber hardware.” Therefore this is clearly a photograph of exploding wires at atmospheric pressure and not about the propagation of electron beams. Therefore Alan Bond’s distant memories are in error.

However, his basic conclusion is correct. A series of experiments in the 70’s did demonstrate efficient propagation of electron beams along current-carrying channels. But it was my work that did it first. The efficiency in the Sandia experiments in the photo was >90%, the explicit figures being 91–94%.  My efficiency was >99%.

These efficiencies are for the propagation of a beam over a distance of 30 cm in a channel caused by current sufficient to explode a thin wire. The electron beams were constrained by the magnetic field of that discharge from a capacitor bank. However, in the cool plasma formed, the electrons of the beam scatter in a short distance. That’s why the efficiency was ~90% over even a short distance.


This subject has come up because of my paper “E-Beam ICF for Daedalus Reconsidered, which you can see on my website here:


I showed that the original scheme, in which e-beams are guided along magnetic field lines in the shape of a cusp, could not work. The reason they wouldn’t work was twofold: first, the cusp does not focus the electrons, but produces a large cloud of electrons on the axis of the system, a cloud far larger than any pellet. Second, we now know that electron beams would not be propagated through any gas to survive a transit of 50 m, as required by the Daedalus design.

In the narrative below I show that the concept being developed by others and myself in the US in the 1970s, using plasma channels to bring the beam at high intensity onto targets, could at least get the beam to the target. Daedalus did not use that scheme.

In fact it’s hard to see how that scheme could be realized in practice. You need a very high temperature plasma to produce a good enough conducting medium that beam losses would be small in transiting a distance of 50 m as Daedalus required. That would have to be done repetitively at 100 Hz. The only proposal that could be successful would be for lasers providing conducting channels and capacitor banks driving current through those channels in advance of each pallet arriving on axis. Frankly, I think that’s really unlikely to be even remotely practical.

I think the Daedalus e-beam scheme could never be made to work by any means whatsoever, based on what we knew then and, even more so, on what we know now. I consider the question of Daedalus e-beam ICF to be a closed subject.

That’s not to say that other forms of inertial fusion couldn’t be made to work, but the propagation of perhaps laser beams over such distances to very small targets in a complex medium provided by the explosion of the previous target, in hundredths of a second seems at the very least problematical, certainly unlikely, if not impossible.

That said, I’ll describe the development of E-beam propagation work in the 1970s in what follows.


To understand this work in time context, I’ve gone back to the literature, which brings back memories of the events of that busy time.

The idea of propagating electron beams along plasma channels containing current was first demonstrated by Roberts in 1968. When I joined Physics International (PI) in 1969, having just finished my PhD, I took up extension of that idea. Over the next 6 years I conducted a long series of experiments, asking and answering the basic questions.  My group demonstrated all the essential features of propagation of beams over several meters, combination of two beams and concentration of beams by tapering the plasma channel. You can see a list of the particle beam papers the I wrote with my colleagues at my website:



The reason these questions were important at that time was that there was a national security need to understand the effects of nuclear weapons all materials, electronics, and other components of defense systems. For this we used the new pulsed power machines to produce extremely high electrical powers to generate intense particle beams. These beams were used directly on targets to simulate the effects of x-ray high fluences. This was a high priority, carried out by the Department of Defense through the Defense Nuclear Agency.

By the middle 1970s this was a well-developed technology and was used extensively in the number of facilities to find out what the effects were.

E-beams were also thought to have possible use as directed energy weapons. And, in the early 1970s the Lawrence Livermore Laboratory had come up with the idea of inertial confinement fusion using laser beams.  Consequently, the e-beams were being developed widely and intensively.

At that time Gerry Yonas ran the electron beam development division at PI, located 20 miles west of Livermore. Several people realized that electron beams could be used for ICF, in competition with lasers. Yonas knew ICF wasn’t a DOD objective. Therefore he sought to start an e-beam ICF program at the only DOE laboratory with any real experience in pulsed power, Sandia Laboratory in Albuquerque.

Yonas sold to DOE his idea of setting up an independent ICF program at Sandia to compete with the lasers at Livermore, evoking the *Russian*threat* of the USSR e-beam ICF program. He gathered up all the information he could at PI and moved to Sandia to begin this program.


In 1976, I wrote a paper advocating that the multiple beam approach to electron beam ICF could be most effectively carried out by propagating multiple electron beams for a few meters onto an ICF target. I considered both propagation in current-carrying plasma channels and the use of longitudinal magnetic guide fields. My figure shows the geometry:

e-beam in plasma channel


You can see this paper at my website here:



 It showed my idea was very clearly advantageous; substantial current densities could be obtained at distance with losses of less than 1%.  At that time I was basing this on my experiments on propagating e-beams in Z-pinch plasma, dense and very hot, typically greater than 10 eV. In those conditions the electrons lose little energy in traversing a plasma channel.

Having figured all that out, I communicated with several colleagues at Sandia, telling how this should work. I also submitted it to the Journal of Applied PhysiMiller 1 experimentcs, where it was published in 1977 [2].

Yonas’s crew realized that this was the right way to solve the electron beam ICF problem and quickly demonstrated propagation over 0.4 m, an experiment documented in Physical Review Letters (PRL) [3].

PRL sent it to me for review and I pointed out that they had not referenced the previous 9 years of work by Roberts and my group. They then corrected this by referencing an extensive set of papers demonstrating the key features experimentally and theoretically, which they had known of all along. After I agreed that the paper should be accepted, the authors then removed all reference to previous work and published it with those references missing.

Subsequently, Paul Miller did an experiment using up to 8 electron beams, 4 of which were channeled to impact a diagnostic target, the rest stopped during propagation to diagnose the beams [1]. In this and subsequent papers Miller claimed that these were the first such experiments done, which was false. We did all that at least 5 years previously.

Sandia later claimed [1] that they were the first to propagate, combine, or compress electron beams. In fact, these were done much earlier and with far greater efficiency. My experiments, done with hotter plasmas, showed no observable loss whatsoever over distances of the meter. Miller’s experiments showed 10% losses over 40 cm.

Sandia discovered that electrons preheat the ICF targets so that poor compressions resulted. Subsequently, this entire approach was abandoned. This was the end of the electron beam fusion program.

Ion beams replaced it, but they too were found to have fundamental flaws. The Sandia group then moved on and, once again, undertook a method that had been previously developed by Physics International: the implosion of wire arrays. Rich Schneider, recently deceased, and Charles Stallings invented that technology. Sandia adopted it and extrapolated it by many orders of magnitude. DOE did not allow lasers to be defeated by this more practical approach. Therefore Livermore continues to receive the greatest PR support from DOE, even after the failure of the vastly expensive NIF ICF experiments.

In the end the e-beam propagation work by Sandia did them no good. But I think it’s important to give keep the historical record accurate: Roberts invented it as a grad student in 1968. I followed up with extensive set of experiments demonstrating electron beam propagation over substantial distances, which were then used for DOD purposes for many years. Years later, Sandia adopted the idea. They later propagated several e-beams to a diagnostic target.  But they also found it didn’t work on ICF targets and dropped it.

I’ve set the record straight; let history record it.

[1] Electron And Ion Beam Transport To Fusion Targets”, J. R. Freeman, L. Baker, P. A. Miller, L. P. Mix, J. N. Olsen, J. W. Poukey, and T. P. Wright, Proc. Beams 1979, Novosibirsk, USSR, pp. 617-626, (1979)

[2] “Electron-Beam Transport for Inertial Fusion”, James Benford, J. Appl. Physics, 48, 2320 (1977).

[3] “Propagation of Pinched Electron Beams for Pellet Fusion”, ”Miller, P.A., Butler, R. I., Cowan, M., Freeman, J. R., Poukey, J.W., Wright, T. P., Yonas, G., Phys. Rev. Letters 92, pp. 93-95, (1977).

Leave a Reply

Your email address will not be published. Required fields are marked *