MOLTEN-SALT FAST REACTORS
L. G. Alexander
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Thorium, plutonium, and uranium chlorides and fluorides are soluble in mixtures of the halides of Li, Be, Na, K, Mg, and other metals. Their fluoride solutions, at least, are compatible with INOR (an alloy consisting primarily of nickel). They are also compatible with graphite, a structural material as well as a moderator having many desirable properties for high-temperature application. Thus, many embodiments are possible for molten-salt reactors, ranging from simple one-fluid, one-region systems externally cooled to complex internally cooled, two-region, two- fluid systems. The capabilities of only a few of the more obvious systems have been studied so far.
The nuclear and economic potentials of several thermal reactors have been evaluated heretofore, (1-3) and recently a limited program for the preliminary evaluation of fast molten-salt reactor concepts was instituted. Appropriate background studies were performed: (1) a survey was made of data available about the thermal and physical properties of molten fluoride and chloride salt mixtures, (2) the compatibilities of selected reactor materials with various reactor coolants and with molten-salt fuels were studied, (3) potential processing methods for irradiated molten fluoride and chloride reactor fuels were reviewed, and (4) data available for the nuclear properties of chlorine, nickel, and other nuclides of interest were reviewed.
The compositions and physical properties of typical molten-salt reactor fuels in various reactor concepts are compared in Table 1.
Molten-Salt Reactor Technology
A molten-salt reactor program was initiated at the Oak Ridge National Laboratory in 1957 to exploit, for purposes of economic civilian power, the technology of molten-salt fuels developed at ORNL in connection with the Aircraft Nuclear Propulsion (ANP) project. Molten- salt fuels were conceived originally as a means of satisfying the requirements for very high temperature and extremely high power density necessary for aircraft propulsion, and a large amount of work on the physical, chemical, and engineering characteristics of uranium- and thorium-bearing molten fluorides was carried out as part of the ANP program.
The technology of molten-salt reactors was first introduced into the open literature in 1957 by Briant and Weinberg. (4) Bettis et al.(5,6) and Ergen et al. (7) reported on the Aircraft Reactor Experiment, a beryllium-moderated reactor fueled with UF4 dissolved in a mixture of the fluorides of sodium and zirconium, and contained in Inconel. The reactor was successfully operated in 1954 for more than 90,000 kW-hr without incident at thermal powers up to 2.5 MW and temperatures as high as 1650°F.
For thermal power reactor fuels, fluorides of 7Li and beryllium are preferred because of their low parasitic neutron capture rate. Mixtures of these with fluorides of thorium and uranium (in the range from 5 to 15 m/o) have liquidus temperatures below 1000°F, have good heat transfer properties, and are nearly inert with respect to INOR, the preferred container material.
In fast reactors, sodium and potassium may be used in place of 7Li and beryllium, which are costly. Fluorine may be a suitable anion for use with 233U in a fissile stream, but for plutonium fuels, chloride mixtures are required to obtain sufficiently high concentrations of plutonium at temperatures below 1000°F. Another important advantage of chlorine is that it has considerably less moderating power than fluorine. On the other hand, 35Cl exhibits a disadvan– tageous (n,p) reaction, and it is necessary to use the separated isotope 37Cl contaminated with less than 5% 35Cl. Since it is not so important to obtain a hard spectrum in the blanket and plutonium concentrations are never high there, fluoride mixtures are satisfactory for the fertile stream.
The potential usefulness of molten-salt fuels for civilian power was recognized from the start. The features that attracted attention were the high temperature at which the fuel could be used (permitting use of modern stream technology and attainment of high thermal efficiency), combined with a low vapor pressure, the unsurpassed stability of halide salts under radiation, and the usual advantages that a fluid fuel provides. These include a negative temperature coefficient of reactivity, absence of the need for initial excess reactivity and its wastage in control elements, no limitation to burnup by radiation damage or loss of reactivity, the absence of a complicated structure in the reactor core, removal of the heat-transfer operation from the core to an external heat exchanger, and the potential for a low-cost fuel cycle.
The ORNL molten-salt reactor program(8) has comprised a reactor evaluation program for selecting the most promising concepts for civilian power and for pinpointing specific develop- ment problems; an extensive materials development program for fuels, containers, and moderators; an equally extensive program for the development of components, especially pumps, valves, and flanges suitable for extended use with molten salts at 1300°F; a modest program for the discovery of supplementary chemical processes for recovering valuable components (other than uranium) from spent fuel; and a program for the development and definitive demonstration of the feasibility of maintenance of molten-salt reactor systems.
These programs have substantially reached their initial goals,(l) and the work is now mainly devoted to the design and construction of a molten-salt reactor experiment, (9,10) which will demonstrate the utility of the developments achieved and resolve remaining areas of uncertainty. The MSRE will produce up to 10 MW of heat in a fuel consisting of a solution of highly enriched 235UF4 dissolved in a mixture of the fluorides of lithium (99.990 % 7Li), beryllium, and zirconium, having a liquidus temperature of 842°F. Construction and installation of the entire system are scheduled for completion in mid-1964, and criticality is planned for early in 1965.