Science and Global Security, 16:87–114, 2008 Copyright C Taylor & Francis Group, LLC ISSN: 0892-9882 print / 1547-7800 online DOI: 10.1080/08929880802565115 Compromising Safety: Design Choices and Severe Accident Possibilities in India’s Prototype Fast Breeder Reactor Ashwin Kumar 1 and M. V. Ramana 2 1 Department of Engineering and Public Policy, Carnegie Mellon University, Pittsburgh, PA, USA 2 Centre for Interdisciplinary Studies in Environment and Development, Bangalore, India This article explores the safety capabilities of the 500 MWe Prototype Fast Breeder Re- actor that is under construction in India, and which is to be the first of several similar reactors that are proposed to be built over the next few decades, to withstand severe accidents. Such accidents could potentially breach the reactor containment and disperse radioactivity to the environment. The potential for such accidents results from the reac- tor core not being in its most reactive configuration; further, when there is a loss of the coolant, the reactivity increases rather then decreasing as in the case of water-cooled reactors. The analysis demonstrates that the official safety assessments are based on assumptions about the course of accidents that are not justifiable empirically and the safety features incorporated in the current design are not adequate to deal with the range of accidents that are possible. INTRODUCTION India plans a major expansion of nuclear energy based on fast breeder reactors (FBR). 1 The Indian Department of Atomic Energy (DAE) has been committed to this program for a long time and continues to pursue it even though many other countries have suspended their fast breeder programs. India is constructing an Received 26 December 2007; accepted 14 October 2008. This article has benefited from guidance and comments from many people, in particular, Frank von Hippel, Tom Cochran, Hal Feiveson, and one anonymous reviewer. All errors are the authors’ responsibility. Alex Glaser and Ahnde Lin provided much appreciated assistance and useful references. The authors thank the Ploughshares Fund for finan- cial support. Address correspondence to Ashwin Kumar, Department of Engineering and Public Policy, Carnegie Mellon University, Baker Hall 129, Pittsburgh, PA 15213, USA. E-mail: ashwink@cmu.edu 87