Memorandum On Test Of Implosion Gadget
February 16, 1944
1. The implosion gadget must be tested in a range where the energy release is comparable with that contemplated for final use. . . . This test is required because of the incompleteness of our knowledge. Thus the reaction will proceed at a temperature unobtainable in the laboratory, which corresponds to energies at which nuclear properties are, and will probably remain, rather imperfectly known. Further, pressures under which the gadget will operate are likewise unobtainable in the laboratory and the information which we may obtain on the spacio-temporal distribution of the pressures will in all probability be not only imperfectly known to us, but somewhat erratic from case to case.
Various attempts have been made to propose an experimental situation which would enable a test of the kind mentioned above to be carried out under conditions so controlled that the energy release was small. . . . All present proposals seem to me unsatisfactory, at least in the sense that they cannot replace more realistic tests. The proposals which have been made are the following:
a. That the amount of active material used be so limited that the nuclear reaction proceeds over a matter of some 30 ± 15 [neutron] generations to give a readily detectable radio-activity or neutron burst, but no appreciable energy liberation.
b. That the reaction be limited by the thermal stability and increased time scale of excess hydrogenation.
c. That the reaction be limited with normal or excess hydrogenation by the addition of appropriate resonance absorbers which will quench the reaction at temperatures of the order of tens of volts.
As for the first of these proposals, . . . we do not now have, and probably will never have, information precise enough to predict an appropriate mass with any degree of probability. . . . This would involve, among other things, knowing the radius of the compressed core to within 5 per cent. Furthermore, it is doubtful whether one could approach this limited explosion by gradual stages with any certainty and without very numerous subcritical trials since there is no a priori assurance, and some a prior doubt, that the implosions will be reproducible to the extent required.
As for the second and third proposals, which have been advocated with eloquence by Dr. Teller, it appears at the present time extremely doubtful whether a sufficiently complete knowledge of the hydrodynamics and nuclear physics involved will be available to make these tests either completely safe or essentially significant. We should like to leave open at the present time the possibility that either these experiments or others not yet proposed may, some months from now, be capable of essentially unambiguous interpretation. . . .
4 . . . . It is my decision that we should plan . . . an implosion . . . so designed that the energy release be comparable with that of the final gadget, but possibly smaller by as much as a factor of 10; . . . that no definite decision against more controlled experiments be made at the present time. . .; and that in the light of the above considerations, all methods which hold promise of giving reliable information about the hydrodynamics and nuclear physics of the implosion be pursued with greatest urgency. . . . It would appear to be very much less difficult to predict and interpret the dimensions and construction of a gadget releasing some thousands of tons of TNT equivalent in nuclear energy than to make the corresponding predictions for nuclear explosions whose energy release, though finite, is negligible.