The Energy Amplifier: Carlo Rubbia's solution to world energy demand
CERN Courier, April/May 1995
Even under the heavy burden of responsibility as CERN's Director General from 1989-3 the fertile mind of Carlo Rubbia the scientist was never still. A long-time Rubbia 'hobby' has been the search for new sources of nuclear energy, exploiting knowledge and skills from high energy physics.
An initial objective was to adopt heavy ion techniques to induce controlled thermonuclear fusion, but in 1994 this quest changed direction. Putting the problems of thermonuclear fusion aside, Rubbia began to explore an alternative route to energy production through controlled nuclear fission.
The idea is to use a particle accelerator producing neutrons by spallation (interaction of particles with a target) to feed a fuel/moderator assembly where the neutrons multiply by fission chain reactions. If the energy liberated becomes substantially greater than that needed to drive the accelerator, the process has a net gain and becomes self supporting. Hence the name "Energy Amplifier" (EA).
Similar systems for energy production or for nuclear waste incineration have been proposed at Los Alamos and in Japan and Russia, but appear to require the prior development of innovative linear accelerators. For Rubbia's Amplifier, the requisite accelerator is a reasonable extrapolation of an existing cyclotron such that at the Swiss Paul Scherrer Institute.
Moreover, the EA would require fuel rods very similar to those of conventional reactors, rather than demanding new technology using liquid fuel loops (molten salts) with on-line separation of radioactive products.
Unlike a reactor, the EA's fission reaction is not self-sustaining: it is sub-critical and needs a continuous supply of neutrons from the accelerator. This makes Chernobyl-type meltdowns unlikely: if the accelerator stops, the-reaction stops too. Another major advantage is that the old dream of using thorium as a fuel is now made possible. Thorium is not itself fissile, but under neutron bombardment can be transformed into highly fissile uranium 233. This fission yields neutrons which, in addition to maintaining the fission chain, in turn regenerate uranium 233 from thorium. This cannot be achieved practically in a normal thermal reactor since the number of neutrons is too small.
Calculations show that in the EA, uranium 233/thorium equilibrium is soon established. In this thorium cycle very little plutonium is produced - 1000 to 10,000 times less than in conventional reactors. Thorium, more abundant in the Earth's crust than uranium, is fully used in the cycle (unlike natural uranium where only the 0.7% sliver of isotope 235 is fissionable). Thorium energy reserves appear to be practically ' inexhaustible.
To illustrate the case for a beam energy of 7 Megawatts (7 mA protons produced by a state-of-the-art 1 GeV cyclotron) the EA would produce 280 MW of thermal energy, corresponding to about 100 MW of electrical power. As the power needed to operate the accelerator does not exceed about 20 MW, there would thus be a net production of over 80 MW. (A normal nuclear power station produces about a Gigawatt.) During 1994 an irradiated sub-critical assembly has been tested in a beam from CERN's PS proton synchrotron at a very low power (of the order of one watt).
Rubbia has also simulated using lead as moderator. This involves working with neutrons of higher energy than in normal fast neutron reactors where the less agreeable liquid sodium is the moderator. Calculations show that reactivity using lead remains very constant since the effect of fission products (poisons) is considerably reduced. This allows more complete combustion before the fuel rods have to be reprocessed (they could stay in place for four years instead of a year with the light-water variant). Moreover, the constant reactivity opens up the prospect of energy gain increasing from about 60 to 100 or 120, with the assembly remaining sub-critical.
Fast neutrons open up interesting opportunities for the reprocessing of the irradiated rods. This reprocessing will be limited to the separation of the fission products since all actinides become combustible, sidestepping the plutonium problem. The only EA actinide waste (0.5%) comes from imperfections in the separation process.
Fission products can be recycled in the rods, the neutron flux transforming them into non-radioactive elements. However, unlike the fission of actinides, this transformation consumes neutrons and reduces energy production. A compromise solution would be to reserve this costly processing to sensitive elements such as cesium 135, iodine 129, etc., which are long-lived (several million years) and potentially polluting.
A study carried out on the thermal version (the fast neutron version, probably even cheaper, has not yet been costed) corroborated by experts from the Laboratoire d'Economie de l'Energie in Grenoble indicates a unit price slightly higher than that of French nuclear power stations, but much lower than other sources (German nuclear power stations, coal, French natural gas).
This may seem surprising ("you're adding a cyclotron, how can you be cheaper?"), but the relative additional investment is offset by the lower fuel cost (no isotope separation) and ease of operation, with less frequent fuel rod manipulation.
To drive the assembly would require a 1 GeV cyclotron with an intensity of 10 mA or more - an entirely feasible prospect. The experiment at CERN, using a beam from the PS, shows that an energy of 800 MeV to 1 GeV is optimal.
The main objective of the test was to confirm computer simulations. The Universidad Politecnica of Madrid provided a subcritical assembly used for teaching purposes with 3.6 tonnes of natural uranium in a tank of demineralized water, which could be used with only minor modifications to take a spallation target. Operation used low intensity (1/100,000 of the PS proton production rate) to minimize radioactivity.
In this test rig, energy production was of the order of 1 watt, giving a temperature rise of the order of 1/100 of a degree, requiring careful thermometry. To prove that the heating was due to fission, plastic sheets (Lexan) were exposed in the thermometric probe sites. The characteristic etching produced by fission fragments on such plastic can be developed and observed under the microscope. The correlation was perfect.
From these local measurements, carried out at four different distances, the rise in temperature could be established from the fission density distribution. This relied on the computer simulation and a detailed mapping from 200 electronic fissionproduct detectors. These silicon solar diodes or ionization counters in pressurized argon count the fragments emerging from thin calibrated deposits of uranium and show the time dependence of the fission rate following a short pulse of protons.
The measured energy gain is of the order of 30, consistent with the simulation predictions.
The test rig was far from optimum. Simulation shows that a larger device would give a gain of about 60, sufficient for 600 MW of thermal energy, i.e. about 200 MW of electrical power, from a 10 mA 1 GeV cyclotron.
The dependence of gain on incident proton energy has been studied between 600 MeV and 2.75 GeV. Above 1 GeV the gain hardly increased while below 800 MeV it decreased markedly. By the end of the year, the CERN group expects to have a detailed feasibility report for a pilot energy production facility.
Elsewhere, several groups - Los Alamos, Brookhaven, Japan (JAERI) and Russia (at least seven institutes around ITEP, Moscow) - are planning accelerator-driven fission for a range of applications (waste and plutonium destruction, tritium production, energy production from thorium, uranium or plutonium). A joint report will be drawn up this year under the aegis of the International Atomic Energy Agency.
With the heavy ecological implications of present nuclear and conventional energy sources, it is surprising how little R&D work is being invested anywhere in this potentially rewarding alternative energy solution.