Rob Goldston: A Scientist on the Cutting Edge of Fusion and Arms Control Research


Professor Rob Goldston teaches in the areas of nuclear energy and non-proliferation at Princeton University. Rob is a leading researcher in plasma physics and fusion energy. He was director of the DOE Princeton Plasma Physics Laboratory (PPPL), 1997 – 2009. Since then he has published on the tradeoff between climate change mitigation by nuclear energy, fission and fusion, and nuclear proliferation risks. Recently he has collaborated with Professors Alexander Glaser of Princeton and Boaz Barak of Harvard on a Zero-Knowledge Protocol for warhead verification, for which the three were named “Leading Global Thinkers of 2014” by Foreign Policy magazine. He was acting director of the Princeton University Woodrow Wilson School Program on Science and Global Security during the Spring semester of 2015.


What inspired you to become a scientist? Was there a particular person or an event that put you on this path?

As far back as I can remember I was interested in physics, but one incident does stand out. The father of a fellow high school student was a laser physicist – back when those were rare – and he was invited to teach our 8:30 am physics class. We were so enthralled with what he could tell us about modern physics that we made him keep answering our questions until lunchtime. We cut all of the intervening classes.

What are the potential benefits of fusion energy?

Fusion has a number of potential benefits. Its fuel is abundant. It cannot melt down or run away. And its proliferation risks are small – if it is safeguarded.

What are the proliferation risks, if any, from fusion energy? What can and should be done to minimize those risks?

Proliferation risks are conventionally divided into use of clandestine facilities to produce fissile material, covert misuse of declared facilities for this purpose, and breakout. A DT [deuterium-tritium] fusion device capable of producing enough neutrons to transmute uranium or thorium to make 1 SQ [significant quantity[1]] worth of weapons material per year, while much smaller than a fusion power plant, is still quite large and would have very clear environmental signatures. So the risk of a clandestine facility making bomb material is small.

One could imagine, however, placing uranium or thorium targets in the vicinity of a neutron and power producing fusion plasma (the cloud of hot, ionized gas that is the fusion fuel). You would need to have safeguards to assure that no such material was present, but these would be relatively easy to implement, because the baseline amount is zero – so detection of any uranium, thorium or fission products would be a clear signature of misuse of the plant.

Finally, you could worry about breakout. The advantage of fusion is that even an unannounced breakout would be easily detected (again due to the presence of improper materials) and at the time of breakout there would be no fissile material yet produced. It would be relatively easy to disable a fusion power plant without risk of spreading radiation. I have written about these issues[2] and others associated specifically with inertial confinement fusion[3], and worked in an IAEA Consultative Group to suggest ways in which safeguards could best be deployed for fusion systems.

What more needs to be done to deploy the first commercially viable fusion energy power plant? How far away approximately is the world from achieving that breakthrough?

I think we know how to make commercial amounts of power from fusion, and this will be demonstrated by the ITER [Latin for “the way”] experiment now being built in France. ITER is slated to produce up to 500 MW of fusion power in pulses lasting between 400 seconds and an hour. The next challenge – which is my current area of research – is learning how the heat of the plasma escapes from the edge of the plasma[4],[5] and how to capture it most effectively. A parallel challenge is developing materials that can withstand the flux of 14 MeV neutrons from the DT reaction. In a sense our next challenges are set by our successes so far in making fusion power.

It is hard to say when all of this will come together. ITER should demonstrate major power production in the 2030s. We should bring along in parallel the other science and technology so that the device after ITER can put electricity on the grid. This is the structure of the plan that has been articulated in China, Europe, Japan and South Korea. In the U.S. we have been more reticent about articulating such a plan.

In particular, what advice would you give (or have you given) to the U.S. government (both the executive and legislative branches) to further advance the prospects for a commercial breakthrough in fusion energy?

I think that the U.S. needs to commit to being a commercial competitor in fusion energy, which means that we need a focused program with a set of specific goals and milestones. In particular, I think the winner in fusion will be the country that addresses the heat and neutron flux issues most effectively. We should be doing that in the U.S. while supporting the international ITER project, so that we can build a competitive pilot fusion power plant as soon as ITER succeeds.

Please describe in layperson’s terms what the “Zero Knowledge Protocol” is and how it can help address verification problems in nuclear arms control. Please describe the Consortium for Verification Technology.

A key issue for future arms control agreements will be for multi-national inspection teams to be able to verify that a nuclear warhead slated for dismantlement is truly a warhead, and one of the type specified. The problem is that this must be done without revealing anything about the design or composition of the warhead. (In other words, nuclear arms control should not facilitate nuclear proliferation!) Alex Glaser, Boaz Barak, and I have proposed a new interactive “Zero-Knowledge” technique[6] to get around this apparent paradox. We propose that the inspectors would first select one or more warheads, randomly, from actively deployed missiles. At least one of these warheads, we assume, is a live one. If the inspectors are uncomfortable about this, they can select more. Then, say, 50 warheads are pulled out from storage. Now if the inspectors can prove to their satisfaction that these 50+ objects are identical, without learning anything about them, the problem is solved.

Our approach to this next step is a form of differential neutron radiography, and we are just now starting experiments on this at PPPL – using unclassified test objects. If we just were to take a neutron radiograph of a warhead, the resulting image would be highly classified. So our concept is that the owner of the warhead preloads the complement of this image onto an array of neutron detectors of a special type that record neutron fluence by producing small bubbles. Of course, the inspectors do not get to see these preloads either. However, when they irradiate a true warhead with neutrons that ultimately fall onto the preloaded array, the total signal at each detector should add up to a pre-agreed number of bubbles – the number that would have been produced with nothing there. So if we get an image of nothing – we have a real warhead! And we convey no information. The nice trick that made me fall in love with this idea is that if the preload is given a random Poisson distribution, there isn’t even information in the noisy speckle pattern on the image, since Poisson(n) + Poisson(m) = Poisson(n+m).

The astute reader, however, may have noticed a problem. Why can’t the owner of the warheads pull out a bag of rocks from storage, and give the inspectors a detector array preloaded with the complement of the bag of rocks? The answer – and this is where the interactive Zero-Knowledge feature comes in – is that the inspectors get to choose which preloaded array of detectors goes with which putative warhead. So the preload that is complementary to the bag of rocks could well end up behind a real warhead pulled off a missile. If we do this a few times, the odds that the warhead owner can get away with cheating are infinitesimal.

The Consortium on Verification Technology is a multi-institutional activity funded by the National Nuclear Security Administration of the Department of Energy. It provides funding for universities to collaborate with National Labs to work on a number of kinds of verification technology, not just associated with warheads. Princeton University and PPPL are members of this Consortium, working together on Zero-Knowledge warhead verification.

In early August, you joined 29 leading scientists in a letter[7] to President Obama in support of the nuclear deal with Iran. Why did you sign the letter?

When I read the JCPOA, I was amazed at how strong it was. Viewed in the frame of other non-proliferation agreements, it is extremely innovative, very restrictive, and very well verified. I thought it was important for non-technical people to understand that this is indeed a very good deal – indeed the best that has ever been negotiated –and in absolute terms, able to get the job done. After 15 – 25 years of “good behavior” Iran will be constrained only as any other member in good standing of the NPT and signatory to its Additional Protocol is constrained, but this was inevitable after some period of time. As we said in the letter, and I have written separately[8], we need to strengthen the non-proliferation regime for the long run – and this deal gives us the time to do that.

What scientific opportunities do you think American scientists can pursue in collaboration with Iranian scientists?

JCPOA indicates that Iran is interested in fusion, and in particular in ITER. I would be very glad to welcome Iranian scientists to work on these.

What advice would you give fellow scientists who are considering applying their knowledge and skills to societal issues?

First of all, science is great fun. There is no thrill greater than understanding something deeply for the first time. If you are the first person in the world to understand it – that makes it a hundred times better. And if you can be solving societally important problems at the same time, what could be better?


[1] According to the International Atomic Energy Agency, a significant quantity (SQ) represents the amount of fissile material to make a first-generation nuclear explosive device, including manufacturing losses.

[2] Proliferation risks of magnetic fusion energy: clandestine production, covert production and breakout, A Glaser, RJ Goldston, Nuclear Fusion 52 (4), 043004

[3] Inertial confinement fusion energy R&D and nuclear proliferation: The need for direct and transparent review, RJ Goldston, A Glaser, Bulletin of the Atomic Scientists, 05/30/2013

[4] Heuristic drift-based model of the power scrape-off width in low-gas-puff H-mode tokamaks, RJ Goldston, Nuclear Fusion 52 (1), 013009

[5] Theoretical aspects and practical implications of the heuristic drift SOL model, RJ Goldston, Journal of Nuclear Materials 2014.10.080

[6] A zero-knowledge protocol for nuclear warhead verification, A Glaser, B Barak, RJ Goldston, Nature 510 (7506), 497-502


[8] Negotiating with Iran: Breakout and Sneakout, RJ Goldston, Bulletin of the Atomic Scientists 02/10/2015