SEPTEMBER 1,1936 PHYSICAL REVIEW VOLUME 50 LETTERS TO THE EDITOR Prompt publication of brief reports of important discoveries in physics may be secured by addressing them to this department.Closing dates for this depariment are,for the first issue of the month,the eighteenth of the preceding month,for the second issue,the third of the month.The Board of Editors does not hold itself responsible for the opinions expressed by the correspondents. Communications should not in general exceed 600 words in length. Christiansen Filters as Polarizers material,viscose happened to be available,was immersed The Christiansen effect has been used for filters not only in a mixture of carbon disulphide and benzene.In due time in the visible but also in the ultraviolet and infrared.Yet it the effect was quite clear.The transmitted light was con- siderably spread out in the direction perpendicular to the apparently has not been used with doubly refracting ma- fibers,perhaps a diffraction effect due to lack of homo- terial,as such.Wood has used quartz particles in making geneity in the fibers,but the light was very nearly plane filters,but on account of convenience rather than for bire- fringence. polarized and of a quite limited spectral region. The birefringence in this case is such that it is possible, The dispersion curves for the material of the particles for example,to adjust the filter to transmit green polarized and for the immersion liquid are of course familiar.Sup horizontally and blue polarized vertically.The fact that pose however that a doubly refracting material is used. the transmitted light for one direction of polarization Then the dispersion curves will be as shown in Fig.1,where follows so definitely the behavior of the light transmitted the upper solid curve is the dispersion curve for one direc- by the Christiansen cells described by Wood and others! tion of vibration in the solid,the other is that for the indicates that this is a true Christiansen effect. perpendicular direction,T1 is the dispersion curve for the In the case of the viscose fibers used in the experiments liquid at a temperature T1,and Ta the corresponding curve the two indices of refraction were not measured with any at a somewhat higher temperature Ts. accuracy,but it was clearly shown that the index of re- Obviously there are two Christiansen wave-lengths,and fraction is greater for a vibration direction along the fiber the beams thus transmitted are polarized at right angles to than for the direction across the fiber.The same result each other.If the wave-length is just below the range of was obtained with silk,the frame being wound with white the visible spectrum,the observed transmitted beam is silk thread. monochromatic and plane polarized.On changing the As is well known,x-ray studies have shown that both index of refraction either by changing the composition of rayon and silk possess structure,but it has not been the the liquid or by changing the temperature,the changes in practice to consider their fibers as doubly refracting crys- the transmitted wave-lengths should proceed with each tals for optical work. of the two components exactly as in the case of the experi H.W.FARWELL ments in which isotropic solids have been used. Columbia University. The only difficulty then is to obtain doubly refracting Augu8t10,1936. material which is sufficiently small in particle size and still remains oriented.The birefringence of Cellophane suggested that fibers of cellulose,such as rayon,might serve the purpose.Accordingly a frame carefully wound with such a Bragg Reflection of Slow Neutrons The peak of the velocity distribution of thermal neu- tronst indicates a momentum for which the de Broglie wave-length,/m,is approximately 1.6A.If such neutrons suffer Bragg reflection,they will be regularly reflected from a magnesium oxide (MgO)crystal (2d=4.0A)when the Bragg angle is about 22(nA=2d sin 0). Sixteen well-formed single crystals of MgO,about 8X25X44 mm,were mounted in a ring with the source and detector placed on the axis for a grazing angle of 22,as shown in Fig.1. BACKGROUND,Ncd,OF HIGH SPEED NEUTRONS The detectors were (1),an ionization chamber filled with BFs in the first run,and (2),one lined with BaC in the second and third runs.The sensitivity of both of these chambers?extends to neutrons of such high velocity that it FIG.1.Dispersion curves for the solid (solid lines)and for the liquid at different temperatures (dotted lines). was quite impossible to absorb all detectable neutrons 486
LETTERS TO THE EDITOR 487 MgO CRYSTALS TABLE I.Obsersed Numbers. Cd SHIELDS Bragg Crossed Background Rate g0ve品in. Rate Run Nx/min. RSORRING 23 60.5±.4 31 55.6±.4 38 43.3±.7 2nd 11 8.6 26.5±.3 3 20.9±.4 PARAFFIN CHAMBER 3rd 12 7+23 12 37.7±，3 28.0±.4 NB-Nx= ,1士，4 ·6CM- F1G1 TABLE II.Relative Numbers. NB-NX Nx NB emerging in the direction of the chamber.These together Run NB-NCd Ncd Nca with those scattered from the general surroundings account With MgO Crystals for the number (Nca)of neutrons counted when the cad- 城 0.40±0.06 1.28±0.02 ,41409 88主83 mium screening of the chamber is completed by a sheet of 1.27±03 With Al Blocks Cd across its front. 3rd 0.01±0.04 The entire removal of the crystals made practically no 134±0.02 140.02 change in the count of these high speed (Cd penetrating) neutrons,and further,the subsequent removal of the Cd from the front of the chamber made no appreciable increase NB-Nx=0.1±0.4 in the count. indicates that the change in geometry could not have NEUTRONS SCATTERED BY SINGLE CRYSTALS changed the incoherent scattering much more than the It thus appears (1),that the crystals do not significantly statistical fluctuations. affect theamount,Ncd,of high speed neutrons counted,and (2),that the slow speed neutrons counted were scattered SUMMARY OF RESULTS from the crystals. When the crystals are in the Bragg positions,the total The greatest statistical precision was obtained for NB number NB of neutrons counted will be the background and Nx,and hence these two values in Table I are the best Nca plus both those regularly reflected and incoherently evidence for the Bragg type of reflection of neutrons.It is scattered by the crystals.The amount of incoherent also interesting to note the relative amount of reflection and scattering should be practically independent of crystal scattering as shown in Table II. orientation,so to observe this without regular reflection the The correction shown takes account of the fact that the crystals were tilted,alternately clockwise and counter- total volume of the aluminum blocks used in the third run clockwise,about 25 from the Bragg position.In this case was 1.63 times that of the crystals (due to the irregular of crossed crystals the total count Nx will be Ncd plus the outline of the crystals,as mentioned above). incoherent scattering.Hence NB-Nx should be a measure On the basis of this evidence it seems reasonable to con- of the number of slow neutrons that are regularly reflected. clude that we have in these experiments observed the In the first run,NB-Nx was eight times that accountable reflection of slow neutrons in accord with the Bragg relation on the basis of statistical fluctuations and in the second run, between the de Broglie wave-length of these neutrons and six times.These results at once indicated the Bragg reflec- the grating space of these crystals tion of slow neutrons.As a check,it seemed necessary to It should be noted that the experimental arrangement determine by actual test whether or not polycrystalline (see Fig.1)permits a sufficiently large angular divergence blocks of about the same size and scattering power would so that the Bragg conditions are satisfied for a large portion due to the change in geometric disposition,scatter more of the velocity range in the Maxwellian distribution!of the slow neutrons in the "Bragg"position than in the crossed thermal neutrons. position. Grateful acknowledgment is made to Mr.Raymond Ridgeway for supplying us with the unusually large single SCATTERING BY POLYCRYSTALLINE BLOCKS crystals of MgO used in this work. In this,the third run,aluminum metal blocks of rec DANA P.MITCHELL tangular size and thickness equal to the rectangular PHILIP N.POWERS boundary of the somewhat irregular single crystals,used in the first two runs,were mounted in place of the crystals Pupin Physics Laboratories. Aluminum was used since it has approximately the same Augu8t1,1380, effective scattering power per unit volume as MgO.The i J.R.Dunning,G.B.Pegram,G.A.Fink,D.P.Whitehall and E. result se学h，.40,453(4936. 704(1935)