LX^f ]C '/,yA\PX)^ -^9^ ^Dc-f7f MDDC - 879 UNITE E' STATES ATOMIC ENERGY COMMISSION THE ABSOLUTE MEASUREMENT OF THERMAL NEUTRON DENSITY by H. L. Anderson P. G. Koontz J. H. Roberts This document consists of 4 pages. Date of Manuscript: August 7, 1942 Date Declassified: January 2, 1947 This document is for official use. Its issuance does not constitute authority for declaGsificaticn of classified copies 3aiiic GJL~ oliiilxar i^Ohlci'il a. and by the same author(s). Technical Information Division, Oak Ridge Directed Operations Oak Ridge, Tennessee Digitized by tiie Internet Arciiive in 2011 witii funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation http://www.archive.org/details/absolutemeasuremOOusat THE ABSOLUTE MEASUREMENT OF THERMAL NEUTRON DENSITY By H. L. Anderson, P. G. Koontz and J. H. Roberts ABSTRACT A method for detcrir.ining the absolute value of the thermal neutron density is described. For measurements in a bijam of thermal neutrons, a boron trifluoride proportional counter filled with a known amount of BF3 gas is used. For measurements inside a medium in which thermal neutrons are diffusing, MnG2 detectors are used. These detectors were standardized against this BF counter in a slow neutror beam. For the Mn02 detector which we used (1.05 grams on an area 5x6 cm^ enclosed in scotch tapa and measi-red by wrapping around a 0.011 cm wall dural counter), the number of neu- trons per second per cm^ traversing the detector is 0.76 x the saturation activity of the detector in counts per mir.ute. STANDARD BORON TRIFLUORIDE COUNTER The capture of slow neutrons by boron leads to the process 5B" + on' ~ ^LC + jHe^ By using a proportional counter filled with boron trifluoride gas and suitable cadmium shielding, it is possible to record all the disintegrations by thermal neutrons which take place in a known volume of the gas in the counter. The cross section of boron for the capture of thermal neutrons has been determined by Anderson and Fermi in Report C-74 by measuring the absorption by BF^ gas of the thermal neutrons emerging from a graphite column. The same gas was taken to fill the proportional counter as was used in the absorption measurements. In the present use of a BF3 proportional counter to make an absolute measurement of thermal neutron density, it is assumed that all the thermal neutron captures by boron lead to disintegrations which will be recorded by the counter.* If this is the case, then the number of disintegration observed per second when the counter is placed in a beam of slow neutrons will be given by ■}' = nvMffg (v) (1) where nv is the number of neutrons which pas.s through 1 cm^ per second, M is the number of moles of boron in the active volume nt thu frmnter, and "~ (v) is the capture crccc cccticr. of bcror. per niole for neutrons of velocity v. The neutron flux nv traversing the counter may be obtained by measuring y provided M is known, since for slow neutrons tTg~l/v, the product nvi^B(v) is independent of the par- ticular velocity distribution of the neutrons. For neutrons having the average velocity of the Maxwell distribution v, the value of ag (v) is taken to be 411 cm^ per mole. ♦Some doubt may be thrown on this point by the work of Maurer and Fisk, Zeit. f. Physick 112:463 (1939), who reported a group of low energy protons which might escape detection. MDDC - 879 [1 2] MDDC - 879 The standard BF3 counter which was used consisted of a nickel tube turned down to a wall thick- ness of 0.025 cm; its internal diameter was 3.62 cm and its length was 30 cm. The counter was filled with BF3 gas so that there Mere 7.31 x 10"^ moles per cm of counter length. The counter was com- pletely covered with Cd except for an opening of 10 cm along its length in the center of the counter. The absorption of the count(>r wall was determined before assembling the counter, by slipping the glass and nickel tubes over a smaller BF3 counter placed near a graphite pile from which thermal neutrons were emerging, and observing the decrease in the counting rate. In this way the absorption factor of the counter walls was found to be 1.13. The effective cross section of this counter per cm of length for neutrons having the average velocity of the thermal neutron distribution is equal to 0.0266 cm . The BFi counter was cc>nne('ted to a high gain linear amplifier and the disintegrations were counted with the use of a scale of 16 recorder. In Figure 1 is shown the variation of the counting rate as a function of voltage of the countei. Curve 1 shows the results with no Cd In front of the 10 cm window. This curve does not exhibit a very satisfactory plateau, principally due to the disintegrations taking place near the ends of the counter and due to fast neutron recoils, both of which give too small pulses at the lower voltages. The 'lifference In the counting rate taken without and with Cd in front of the window is plotted as Curve 2. These counts are due only to C neutrons* which enter through the Cd window. The plateau exnibiled by this curve is evidence for the fact that all the disintegrations due to C neutrons taking place in tl'e counter are recorded. 700 1900 2 00 2100 2200 VOLTAGE Figure 1. Variation of couwtiiig rate with voltage of BF, counter in a neutron beam. Curve I -Measurements with no Cd. Curve II-No Cd-Cd difference. * C neutrons are those slow neutrons strongly absorbed by Cadmium. The energy distribution of the C neutrons which emerge from paraffin has a more pronounced high energy tail than does the thermal neutron distribution. IviDDC - 879 B^C SHIELD-. 324 CM Figure 2. Apparatus for determining the counting rate of BF3 counter in neutron beam. In standardizing the MnOj detec- tors, the BF3 counter was removed and a detector placed 32.4 cm above the top of the paraffin in a position where the aver- age neutron was the same as that meas- ured by the counter. STANDARDIZATION OF MANGANESE 010X0)5 DETECTORS Such a Doron coimter is quite useful for the measurement of the neutron flux in a beam. For measurements of the neutror density insi'ie £. m.edium in which neutrons are diffusing, the counter may not be used J it is large, tor its presence will perturb the neutron distribution in its neighborhood. In order not to perturb the neutron distribution, we have used thin layers of MnOj as detectors of thermal neutrons inside water or graphite. The cross section of these manganese detectors was obtained by comparing their activity with the counting rate of the BF3 counter in a slow neutron beam. The detectors were prepared by pressing a thin uniform layer of about 1.05 gm of MnOj* powder in a steel die by means of a small hydraulic press. The area of the layer was 5x6 cm^. The layer was covered on both sides with scotch tape. To facilitate wrapping around our counters the foils were ruled with parallel impressions spaced 1/8 inch apart. The foils were fitted snugly around the counter for measurement. For the comparison of the MnOa detector with the standard BF3 counter, we used an arrangement which is shown in Figure 2. A 1 gram Ra+Be source was surrounded by paraffin and suitable lead protection was provided to reduce the number of y-rays. The center of the counter was 31.4 cm above the top surface of the paraffin and to enter the counter the neutrons from the paraffin had to pass through a circular opening i.i a boron carbide shield. The diameter of the hole was 13 cm. This opening could be covered with a Cd sheet and measurements were always taken as the difference no Cd-Cd so as to make sure that only C neutrons emerging from the paraffin through the 13 cm opening were being recorded. This paraffin geometry was used for reasons of intensity. It might be thought that a beam of truly thermal neutrons, such as could he obtJ^ineri usinp; granhitp wmilH hp nrpfprcihlp, Insofar as the absorption of manganese and boron both, presumably, vary according to 1/v, the com- parison would give the same result in both cases. It should be pointed out that indium detectors cannot be compared directly with a boron counter using the paraffin geometry of Figure 2 due to the presence of the strong resonance line of indium at 1.35 ev. The presence of this resonance line causes a deviation in the 1/v absorption law for indium below the Cd cutoff. For measurements inside a diffusing medium, the effect of the perturbation due * Chemical analysis of this material showed Mn content to be 58.0% due to the presence of oxides of manganese other than MnO . I 4 ] MDDC - 879 to the resonance activation of indium is much less, since inside water or graphite the ratio of the density of thermal to resonance neutrons is much greater than in a slow neutron beam. If .45 gm/ciji^ of Cd is used for determining the contribution of the resonance neutrons to the activity observed with- out Cd, then the difference :n the ■ ctivation no Cd-1.07 Cd-In-Cd is a quite reliable indication of the thermal neutron density. Irdium dstectors may be standardized against manganese detectors pro- vided the comparison is male inside the same medium in which it is desired to make absolute thermal neutron density measurements with the indium detectors. With the geometry of Figure 2, the effective length of the counter is somewhat longer than 10 cm due to the obliquity of the ncutrcns which traverse the opening in the Cd. To take this obliquity into account, we have taken ;is tlie effective length of the counter 10.46 cm. We observed a counting rate no Cd-Cd, 885 counts per minute. The manganese detectors were placed at the position of the diam- etral plane of the counter and were displaced slightly from the center of the opening in the cadmium shield of the counter, so that the average neutron intensity falling on them would correspond most closely to that falling on the opening in the cadmium. The activity of the MnOj detector was measured on a thin wall dural counter (Kanj^a).* The saturation activity no Cd-Cd was found to be 69.8 counts per minute per gram of MnCla. From these data we may calculate that the apparent cross section per gram of the MnOs detectors is given by ^Mii '■/) = ^ •■« 0-2e6 X 10.46 = .Or,2 cmVgm (2) It iS clear that this cross section refers to the particular MnOz detectors measured on the particular counter we used. The absolute value of the neutron density inside a medium in which thermal neu- trons are diffusing may 'ae ceterminec by irradiatirj; :i.n MnOa detector without and with Cd, and meas- uring its saturation activity Ajf, (Mn) = no Cd-Cd, in counts per minute per gram on the same counter as was used in siandardizmg it. Thus, Afh (M") „ „„ . ,„ , neutrons "^ = 60ir:022 = °-'6 Ath (Mn) ^^^^^. (3) A correction should be applied for the lowering of the neutron density in the neighborhood of a detector, due to the absorption by the detector. In the MnOa detectors which were used this amounts to about 2% in paraffin or water and is negligible in graphite. * The counter Kanga, which was used, was a dural counter with a wall thickness of about 0.011 cm of 2 cm diameter and a sensitive length of nbout 7 cm. Its construction will be described in a forthcoming report. The efficiency of the counter for /3-rays of Ra E and UX2 was found to be .181 and .349 respectively. 1 fiillii ^1 i