The Aircraft Reactor Experiment positive temperature coefficient of reactivity

Originally, ORNL designed the Aircraft Reactor Experiment as a sodium-cooled, beryllium oxide-moderated, solid UO₂ fueled reactor. At the very high power densities and temperatures required to propel aircraft with nuclear heat, however, they realized that there would be a serious positive temperature reactivity effect caused by the Xenon-135 neutron capture cross section. To operate in this condition, they’d need a fast-acting, super-reliable automated control system.

Instead, they decided to switch the design to tubes of stagnant, non-circulating fluoride molten salt fuel. The expansion of the fluid fuel would overcome the effect from Xenon (not to mention Xenon bubbling to the top), so the reactor would be stable at high power. The updated design incorporated these tubes into the already-ordered BeO moderator blocks.

Further analysis suggested that stagnant salt would have a radial temperature gradient of a few hundred degrees: the centerline temperature would be close to the boiling point of the fuel! So they switched it over to the circulating fluid fuel system that we now know as the world’s first molten salt reactor. They did keep the sodium cooling for the reflector.

So it was molten salt and sodium cooled, both of which transferred heat to helium, which then went to water. Talk about advanced! It operated for 4 days and leaked a lot of fission products, but overall we learned a ton and progressed our ability to run interesting reactors.

I had to think a little about the xenon-related positive temperature coefficient of reactivity thing for a sec. After plotting these thermal neutron energy distributions vs. average temperature and the cross section, you can clearly see the issue:

At high power density there is a lot of xenon during operation, and at high temperature, the hot tail of the thermal neutrons starts falling off this cliff-edge cross section. If temperature increases in that condition, more neutrons move to the right, significantly reducing the number of them that get absorbed in Xe, and therefore increasing overall reactivity. Crazy!

This effect can be seen in HTGRs where the temperature coefficient is less negative when Xenon builds up, but it doesn’t generally become net positive and typical HTGR power densities.

See Also

• Bettis, E. S.; Schroeder, R. W.; Cristy, G. A.; Savage, H. W.; Affel, R. G.; Hemphill, L. F. (1957). “The Aircraft Reactor Experiment—Design and construction”. Nuclear Science and Engineering. 2 (6): 804–825.
• ARE Wikipedia page – it’s good, I wrote a lot of it ;)
• Temperature coefficients for the Ft. St. Vrain initial core
• X post on this topic