Cesium¹³⁷ has a half-life of 30 years and emits a gamma ray with an energy of 0.6 million electron-volts. Strontium⁹⁰ has a half-life of 28 years and emits an electron with an average energy of 0.22 million electron-volts. The daughter nucleus in this process is yttrium⁹⁰, which emits another electron with an average energy of one million electron-volts. The half-life of yttrium⁹⁰ is 64 hours. In effect, therefore, strontium⁹⁰ emits two electrons, each with an average energy of 0.6 million electron-volts. For the long-term radioactive hazard, particularly the world-wide fallout associated with atomic explosions, the two isotopes cesium¹³⁷ and strontium⁹⁰ are the most significant. Strontium⁹⁰ is the more dangerous to living organisms because it is deposited in the bones and retained in the body for long periods.
Besides radioactivity there is another feature of the fission process which is so conspicuous that it may seem hard to understand how Fermi failed to notice it—namely the large amount of energy released. The fission of a single nucleus of uranium releases an energy of 200 million electron-volts as contrasted with ordinary radioactive decay energies of 5 to 10 million electron-volts. (The energy released from the burning of one atom of coal is only 4 electron-volts.)
Of the 200 million electron-volts released in fission, about 10 million go into gamma rays and neutrons created in the fission process itself. This energy contributes to the immediate and local radiation danger. Another 24 million electron-volts go into radioactivity of the fission products, and of this, about half go into neutrinos, which are neither dangerous nor useful; the other half is carried by electrons and gives rise to the delayed radioactive hazard. But the bulk of the energy, over 160 million electron-volts, goes into kinetic energy of the two primary fission fragments. Of this amount, 100 million, on the average, go to the lighter fragment.
One hundred million electron-volt fission fragments should certainly have been noticed by Fermi’s radioactive counters—if they had been able to reach the counters. The fragments were not able to reach the counters, however. The reason is that Fermi was a careful worker. He knew that his sample of uranium would emit some radioactive particles even before neutron bombardment. This natural radioactivity he did not want to get mixed up with the radioactivity that would be produced in the experiment. So he put an absorbing foil between the uranium sample and the radioactive counters. The fission fragments could not get through the foil.
It is amusing that shortly afterward another noted physicist repeated Fermi’s experiment, but this time without the foil. He reported that he was unable to get any significant results because his counter, for reasons unknown, started to spark.
Thus fission remained a secret. But in England Leo Szilard obtained patent papers on the nuclear chain reaction. He pointed out that in some nuclear reactions free neutrons might be released. These neutrons might then succeed in producing further reactions which would produce more neutrons. Provided that at least one neutron made in each reaction were able to induce a reaction in another nucleus, a chain reaction would take place.
The main problem, of course, was to avoid excessive neutron losses. There are two ways in which the losses mainly occur. One is by wasteful, nonreproductive capture in the nuclei; the other, by neutron leakage from the material surface. This second loss, Szilard showed, could be minimized by using a sufficiently large amount of chain-reacting material.
The point is that a neutron born in a nuclear reaction must travel on the average a certain distance before it can produce another reaction. If the size of the chain-reacting material is much less than this distance, practically all of the neutrons produced will be able to escape through the material surface, and no chain reaction will be possible. If the size of the material is large compared to this distance, the leakage loss becomes negligible, and the possibility of a chain reaction depends entirely on the magnitude of the first kind of loss, the wasteful captures in nuclei. If this loss is not too great, and a chain reaction is possible, there will be a critical size of the material at which on the average exactly one neutron per reaction will be able to induce another reaction. A just critical chain reaction of this kind is what is needed for an atomic reactor.
If the size of the material is greater than the critical size, on the average more than one neutron per reaction will cause another reaction and the chain reaction will run away. If, for example, two neutrons can cause another reaction, there will be two neutrons after the first generation, four after the second, eight after the third, and so forth. This is the principle of the atomic bomb.
After about 80 generations, an appreciable fraction of all the nuclei in the material will have undergone a nuclear transformation and so much energy will have been released that the material will not stay together even for the short time needed to produce the next generation. The whole material begins to fly apart, the system becomes sub-critical, and the chain reaction stops. The entire process lasts only a fraction of a microsecond.