f)r\0Dc-77/ )?/ i>ti'a»4f MDDC 771 UNITED STATES ATOMIC ENERGY COMMISSION OAK RIDGE TENNESSEE A THERMAL NEUTRON VELOCITY SELECTOR AND ITS APPLICATION TO THE MEASUREMENTS OF THE CROSS SECTION OF BORON by E. Fermi, J. Marshall, and L. W. Marshall Argonne Laboratory University of Chicago Published for use within the Atomic Ene:-gy Commission. Inquir- ies for additional copies and any questic-.ns regarding reproduction by recipients of this document may be : ferred to ti.° Documents Distribution Subsection, Publication Section, Technical Information Branch, Atomic Energy Commission, P. O. Box E, Oak Ridge, Tennessee. Inasmuch as a declassified document may differ materially from the original classified document by reason of deletions necessary to accomplish declassification, U;is copy does not constitute au- thority for declassification of classified copies of a similar docu- ment which may bear the same title and authors. Document Declassified; 4/8/47 This document consists of 7 \ ages. MDDC 771 -1- A THERMAL NEUTRON VELOCITY SELECTOR AITO ITS APPLICATION TO THE MEASUREM E NT OF THE CROSS SECTION OF BORON E. Fermi, J. Marshall, and L. W. Marshall Argonne Laboratory, University of Chicago* * All three authors now at Institute for Nuclear Studies, University of Chicago. Introduction Slow neutrons emerging from various moderators with different geometries usually have average velocities comparable, but by no means equal to the thermal agitation velocity. Large differences, both positive and negative, are observed depending on the nature and the geometry of the moderating substance. This phenomenon has been observed by vari- ous experimenters. (1) (2) (3) (4) (1) J. Rainwater and W. W. Havens, Jr., Phys. Rev.JO^ 136 (1946). (2) W. W. Havens, Jr. and J. Rainwater, Phys. Rev. 70, 154 (1946). (3) J. H. Manley, L. J. Haworth, and E. A. Luebke, Phys. Rev. 69^ 405 0946). (4) R. F. Bacher, C. P. Baker, and B. D. McDaniel, Phys. Rev. 69,443 a946). In this paper we have collected some typical examples of the vari- ations of average velocity of slow neutrons using different moderators as indicated by changes in the apparent cross section of boron. Since boron is often used as a standard substance in slow neutron measurements its cross section has been determined also using monochromatic neutrons obtained with a velocity selector of new design operated in connection with the thermal column of the Argonne graphite pile. The observed temperatures of the neutrons emitted from the vari- ous moderators and arrangements of moderators appear to be in accord- ance with the individual arrangements employed. Within the experimental errors of the method the cross section of boron varies as the 1/v law and the measured cross section is 703 x 10*24 cm^ per atom for neutrons of velocity 2200 meters per second. MDDC 771 Temperatures of Neutrons from Various Sources With the thermal purification column of the graphite pile at the Argonne Laboratory as a primary neutron source, a number of measure- ments were made of the cross section of boron. In all cases the detector was a proportional counter filled with BF3 gas. By the use of cadmium diaphragms a neutron beam was obtained with small angular dispersion. The absorber and detector in these experiments were both boron and consequently both obeyed the I/v law of neutron absorption. It was possible, ther-fore, to use the cori ection method given by Bethel^) to calculate the cross section of boron for monoenergic neutrons of energy (5) H. A. Bethe, Rev. Mod. Phys. 9, No. 2, 134 (1937). kT where T is the absolute temperature of the Maxwellian distribution emitted from the source. Since the cross section of 2200 meters per second neutrons (kT at 2930K) is known, one can then determine the ef- fective temperature of the neutron beam. It must be understood that these effective temperatures are bas d on the assumption that the neutron beam is Maxwellian in velocity distribution. This is certainly not strictly true for most sources employed. The results of these experiments are given in Table I. It is quite clear from an inspection of the table that the effective temperature of the neutron beam depends strongly on the source of neutrons. During these experiments the temperature of the thermal column was in the neighbor- hood of 30OC or 303°K. Table I Source of Neutrons Absorber Cross Section for Effective kT neutrons Temp. "K 1. Beam from, surface Gaseous BF3 '^B = 855xlG-24cm2 198 of thermal column 2. Beam passed through " 598 xlO"^* 408 a 3.7 cm. slab of paraffin 3. Beam passed through " corrected to 20.4OC 288 7.6 cm. of heavy water <7 B = 710xlO"24 gm^ at 33.7°C in a container 18" diam. MDDC 771 -3- Tablel (continued) Source of Neutrons 4. Beam passed through a 22 cm. column of graphite 10 cm. square 5. Beam from hole in thermal column 125 cm. deep, 10 cm. square 6. Beam from a "block hole" in thermal column, a hole 10 cm x 10 cm x 22 cm high connected to sur- face of thermal column by a 42 cm. tube of cadmium of internal diam. 2.5 cm. Absorber Cros s Section for Effective kT neutrons Temp.oK Pyrex plate cal- 2800 X 10-24 18.4 ibrated in velocity selector -/ Gaseous BFg 701 X 10-24cm2 293 755x10-24 255 The source arrangement given opposite 1 produces low temperature neutrons because of the filtering action of the graphite in the pile and thermal column. '"' Very slow neutrons whose De Broglie wave lengths (6) H. L. Anderson, E. Fermi, and L. Marshall, Phys. Rev. 70, 815 (1946). are longer than periodicities encoutered in the graphite crystals are scattered very little and can penetrate to the surface of the column mora?- easily than the faster neutrons. In case 2 the slower neutrons are re- moved preferentially because both the absorption and scattering cross sections of hydrogen are larger and also scattering in the forward direc- tion is preferred at higher energy. Heavy water (case 3) acts somewhat in the same way because also for deuterium compounds the scattering cross section and the coherence of successive free paths vary with the energy in the same direction as for hydrogen compounds. Therefore the effective temperature of the neutrons is raised from the initial 198° K to 288°K. The fact that this last temperature is quite close to the actual temperature of the heavy water probably is coincidental. In case 4 the filtering effect of the graphite is shown very strongly. Most neutrons that are scattered are removed from the beam and the graphite column is so long that almost none of the warm neutrons can travel the whole distance without being scattered. Case 5 gives a rather good approxima- tion of the temperature of the source. The neutrons in the beam from the MDDC 771 deep hole should be a fair sample of the neutrons present at the bottom of the hole. Essentially it is a case of black body radiation from a hole in the wall of a furnace. Case 6 was expected to give a good temperature value, but failed to do so, probably because the hole was not deep enough. Velocity Selector The velocity selector makes use of a rotating shutter to interrupt the beam of neutrons from the thermal column of the pile. The shutter was constructed by inserting a multiple sandwich of .004" to .008" cad- mium foils and 1/32" aluminum sheet tightly into a steel cylinder about 1|" in diameter with walls 1/32" thick. The shutter was mounted in ball bearings on a heavy steel base plate and was belt and pulley driven by a Dumore grinder motor. Maximum rotational speeds of 15000 revolutions per minute were possible. It was constructed in the shops of the Metal- lurgical Laboratory under the direction of Mr. T. J. O'Donnell who is re- sponsible for its mechanical design. A cross section of the shutter is shown in figure 1, From the thick- ness of the aluminum spacers between the cadmium foils, and from the dimensions of the shutter, one would estimate that no neutrons from a parallel beam would be able to get through when the shutter was more than 1.2° from its full open position. On the experimental arrangement used it was impossible to use a strictly parallel beam of neutrons. The collimators actually used allowed a maximum divergence of neutron direc- tion in the beam of approximately 3° Consequently one would expect the shutter to be completely closed during each 180° of rotation except for an interval of 3° + 2 x 1.2° = 5.4°. Actually it was found that the counters indicated background intensity except when the shutter was in a 6o inter- val. Through one end of the shutter was inserted a steel rod with its axis perpendicular to the axis of the shutter and with a minor surface ground and polished perpendicular to its axis at each end. Light from a projection lamp and lense system was reflected from these surfaces onto two photocells so placed that each photocell was illuminated twice during each revolution. One of the photocells was used with an amplifier and scaling circuit as a revolution counter. The other, adjustable and cali- brated as to angular position, was connected to an electronic switch cir- cuit which allowed pulses from the proportional counter to be recorded only when the photocell was illuminated. BF3 filled proportional counters were used as-the neutron detector. A nest of four was connected in parallel and mounted at a distance of 146 cm. from the shutter. A thick shield of wood, iron and paraffin was placed MDDC 771 -5- between the counters and the pile to compensate somewhat for the fact that the top shield of the graphite pile is not so thick as might be desired. A hole in this shield allowed neutrons from the shutter to reach the count- ers. The neutron beam between the shutter and the counters was colli- mated to make sure that no slow neutrons from sources other than the shutter could enter the counters. Slow neutrons reflected from the walls and roof of the building were eliminated by protecting the sides and back of the counters with a i" thick layer of boron carbide. The shutter and an improved velocity selector arrangement are more fully described in the paper of Brill and Lichtenberger which ac- companies this paper for publication. Deter minatioii of Boron Cross Section for Neutrons of Known Velocity The cross section of pure BF3 at several different pressures was measured for neutrons from the thermal velocity selector for velocities ranging from 1700 to 5000 meters per second. Within the experimental accuracy of the method the cross section of boron varied according to the 1/v law. After corrections for scattering were made the average cross section of boron for neutrons of 2200 meters per second velocity ' was 699 X 10"24 cm^ per atom. 2200 meters per second is the velocity of a neutron of energy kT where T is 293°K. In order to verify this value a similar measurement was made with a different boron compound as absorber. Na2B407 was ignited at about 400° and dissolved in heavy water. The solution was enclosed in a thin walled aluminum cell and a second cell of identical wall thickness was prepared containing an amount of heavy water equal to that in the solution. The transmissions of these two absorbers for neutrons from the velocity selector were measured and the value of the boron cross section for 2200 meters/second neutrons was foimd to be 700 x 10~24 cm^ corrected for scattering. In good agreement with these values was the cross section as cal- culated from measurements at the indium resonance energy. ") Trans- mission measurements were made using a collimated beam of neutrons from the interior of the graphite pile of the Argonne Laboratory. The (7) J. Marshall, Phys. Rev. 70, 107 0946). indium foil detectors were protected from thermal neutron activation by thick cadmium covers. Background measurements were made using an MDDC 771 -6- indlum filter. Thus the measurements were limited in more than one way to neutrons absorbed strongly by indium. BF3 gas in a steel cylinder was interposed in the cfillimated beam. The BF3 was highly purified (the same gas as used in the thermal neutron transmission experiments described above). The transmission of the steel container filled with BF3 at 44 and 68 lbs. per in.2 was compared with the transmission of the empty container. The density of gas used was determined by weighing the cylinder. The pressures used and the length of the cylinder (30 cm.) were such that the transmissions were in an accurately determinable range (approximately a 2/3 transmission for the 68 lb. sample). The total cross section of BFg for indium resonance neutrons was measured as 107.1 x 10"24 cm^/atom. Assuming •^ scattering (F) = 3.7 x 10"^"^ cm^ o^ scattering (B) = 2 x 10-24 ^^2 bidium resonance energy = 1.44 ev., the boron absorption cross section for neutrons at velocity 2200 meters/ second is 710 x 10"24 cm2/atom. The results of the three measurements are given in Table n. Table II Measurement ' ^ kT (B) at 293PK Na2B40^^D20 velocity selector 700 x 10-24 cm2 BF3 " 699 x 10-24 cin2 BF3 In resonance 710 x 10-24 cm2 Average 703 x 10-24 ^^ This report is based on work done at the Argonne Laboratory, The University of Chicago under the auspices of the Manhattan District, U. S. Corps of Engineers, War Department. gl 8. RSITY OF FLORIDA