Neutron Production and Absorption in Uranium

H. L. A11."'DERSON, E. FERMI AND LEo SziLARD

Reprinted from THE PHYSICAL REVIEW, Vol. 56, No. 3, August 1, 1939 AUGUST 1, 1939 PHYSICAL REVIEW VOLUME 56 Printed in U. S. A. Neutron Production and Absorption in Uranium*

H. L. ANDERSON, E. FER.liH AND LEo SziLARD Columbia University, New York, New York (Received July 3, 1939)

1 3 T has been found - that there is an abundant placed in the center of the tank. The geometry I emission of neutrons from uranium under the was such that practically all neutrons emitted action. of slow neutrons, and it is of interest to as­ by the source and by the uranium oxide were certain whether and to what extent the number of slowed down and absorbed within the tank. Each neutrons emitted exceeds the number absorbed. irradiation extended over several half-life periods This question can be investigated by placing a of radiomanganese and the observed activity of photo-neutron source in the center of a large the solution was about four times the back­ water tank and comparing, with and without ground of the ionization chamber. Alternating uranium in the water, the number of thermal measurements were taken with the cans filled neutrons present in the water. In the previous with uranium oxide, and with empty cans of the 1 experiments of this type • 3 it was attempted to same dimensions. The activity proved to be have as closely as possible a spherically sym­ about ten percent higher with uranium oxide metrical distribution of neutrons. The number of than without it. This result shows that in our thermal neutrons present in the water was de­ arrangement more neutrons are emitted by termined by measuring along one radius the uranium than are absorbed by uranium. neutron density p as a function of the distance r In order to find the average number of fast from the center, and then calculating fr2pdr. neutrons emitted by uranium for each thermal A difference in favor of uranium of about five neutron absorbed by urani percent was reported by von Halban, Joliot and um, we have to de­ Kovarski.4 termine what fraction of the total number of neutrons Since one has to measure a small difference, emitted by the photo-neutron source is, slight deviations from a spherically symmetrical in our experiment, absorbed in the thermal region by uranium. distribution might give misleading results. The The number of photo-neutrons present experiments which are based on the same general principle do not require such symmetry. In order to measure the number of thermal neutrons in the water we filled the tank with a ten-percent solution of MnS04. The activity in­ duced in manganese is proportional to the number 0000 of thermal neutrons present. A physical aver­ 000000 aging was performed by stirring the solution 00000000 before measuring the activity of a sample with an 0 0 O~Q?O 0 0-s ionization chamber. To obtain an effect of suffi­ oooo oooo cient magnitude, about 200 kg of U30s was used. 00000000 The experimental arrangement is shown in . 000000 Fig. 1. A photo-neutron source, consisting of 0000 about 2 g of radium and 250 g of beryllium was

IOcm *Publi cation assisted by the Ernest Kempton Adams Fund for Physical Research of Columbia U niversity. tv. Halban, Joliot and Kovarski, Nature 143, 470 (1939). FIG. 1. Horizontal section through center of cylindrical 2 L. Szilard and W. H. Zinn, Phys. R ev. 55, 799 (1939). tank which is filled with 540 liters of 10-percent MnSO, a Anderson, Fermi and Hanstein, Phys. Rev. 55, 797 solution. A Photo-neutron source composed of 2.3 grams (1939). of radium ~nd 250 grams of beryllium. B, One of 52 cylin­ 4 v. Halban, Joliot and Kovarski, Nature 143, 680 drical cans 5 em in diameter and 60 em in height, which (1939). are either empty or fill ed with uranium oxide. 284 285 NEUTRON PRODUCTION AND ABSORPTIO

emitted by the source is indicated by the activity section for neutron production and neutron ab­ of the solution in the tank when the irradiation sorption in uranium is greater than one, and is carried out with empty cans surrounding the probably about 1.5. \Nhat fraction of the neu­ source. Vve obt<~:ined a measure of this number by trons will reach thermal energies without being taking into accoun t that in our solution about absorbed will, however, depend on the ratio of 20 percent of the neutrons are captured by the average concentrations of hydrogen and manganese and the rest by hydrogen. In order to uranium. Since there is an appreciable absorption obtain , in the same units, a measure of the num­ even far from the center of the resonance band, ber of neutrons absorbed by uranium we pro­ it follows that the fraction of neutrons absorbed ceeded in the following way: A mixture of sand by uranium at resonance will increase with de­ and manganese powder, having the same thermal creasing hydrogen concentration. This has to be neutron absorption as uranium oxide replaced taken into account in discussing the possibility the uranium oxide in t of the cans which were of a nuclear distributed uniformly among the other uranium chain reaction in a system composed essentia oxide-filled cans. After irradiation, all this powder lly of uranium and hydrogen. A chain was mixed together, a ten-percent MnS04 solu­ reaction would require that more neutrons be tion was prepared from a sample, and its activity produced by uranium than absorbed by uranium was measured with our ionization chamber. and hydrogen together. In our experiment the In this way we found that about SO percent of ratio of the average concentration of hydrogen t he neutrons emitted by the source are absorbed to uranium atoms was 17 to 1, and in the experi­ as t hermal neutrons by uranium in our arrange­ ment of von Halban, Joliot and Kovarski this ment. It follows that, if uranium absorbed only ratio was 70 to 1. At such concentrations the thermal neutrons, the observed ten-percent in­ absorption of hydrogen in the t hermal region will crease in activity obtained with uranium present prevent a chain reaction. By reducing the con­ would correspond to an average emission of centration of hydrogen one would obtain the about 1.2 neutrons per thermal neutron absorbed following effect: On the one hand a larger fraction by uranium. This number should be increased, of those neutrons which reach thermal energies to perhaps 1.5, by taking into account the will be absorbed by uranium; on the other hand neutrons which, in our particular arrangement, fewer neutrons reach t he thermal region due to are absorbed at resonance in t he nonthermal an increased absorption by uranium at resonance. region by uranium, without causing neutron Of t hese two counteracting factors the first is emissiOn. more important for high h ydrogen concentra­ From this result we may conclude that a tions and the second is more important for low nuclear chain reaction could be maintained in a hydrogen concentrations. Starting with high system in which neutrons are slowed down with­ hydrogen concentrations, the ratio of neutron out much absorption until t hey reach thermal production to total neutron absorption will thus energies and are then mostly absorbed by ura­ first rise, t he~ pass through a maximum, and, nium rather than by another element. I t remains as the hydrogen concentration is decreased, there­ an open question, however, whether this holds after decrease. We attempted to estimate the for a system in which hydrogen is used for quantities involved from the information avail­ slowing down the neutrons. able about resonance absorption in uranium•-7 In such a system the absorption of neutrons and from the observed net gain of 0.2 in the takes place in three different ways: The neutrons number of neutrons in our experiment. The effect are absorbed at thermal energies, both by hydro­ of the absorption gen and uranium, and they are also absorbed by at resonance turns out to be so uranium at resonance before they are slowed 5 Meitner, Hahn and Strassman, Zeits. f. Physik 106, down to thermal energies. Our result is inde­ 249 (1937). pendent of the ratio of the concentrations of s v. Hal ban, Kovarski and Savitch, Comptes rend us hydrogen and uranium, insofar as it shows that, 208, 1396 (1939). 7 H. L. Anderson and E. Fermi, Phys. Rev. 55, 1106 for thermal neutrons, the ratio of the cross (1939). ANDERSON, FERMI AND SZILARD 286 large that even at the optimum concentration of We wish to thank Dr. D . W. Stewart, of the hydrogen it is at present quite uncertain whether Department of Chemistry, and [r. S. E. Krewer, neutron production will exceed the total neutron for advice and assistance in carrying out some of absorpt-ion. More information concerning the these experiments. We are much indebted to the resonance absorption of uranium as well as more Eldorado Radium Corporation for enabling us to accurate measurement of some of the values work with large quantities of uranium oxide in which enter into our calculation are required our experiments, and to the Association for before we can conclude whether a chain reaction Scientific Collaboration for the use of the photo­ is possible in mi.xtures of uranium and water. neutron source and other facilities. LANCASTER PRESS, INC., LANCASTER, PA.