Physica 137B (1986) 204- 213 North-Holland. Amsterdam THE DEVELOPMENT OF NEUTRON BEAMS FOR MATERIALS SCIENCE C.G. WINDSOR, M.T. HUTCHINGS and P. SCHOFIELD MatertaL~ Physics and Metallur,w Dwisum AERE Ilarwell, OXl l OR.4, ~'K rhe object of experiments undertaken using neutron scattering is traced as progressing from an understanding of the neutron as a probe, to an understanding of solids and liquids of importance to nuclear reactor design, through to the fundamental understanding of condenced matter and eventually to thc problems of industry. The unique information given by neutron beams in the applied field is illustrated by examples from recent research at Harwell on metallurgy, fast ion conductors, nuclear fuels and colloid chemistry. An ideal neutron instrument for materials research in the future is suggested. 1. The changing role of neutron scattering experi- meats The diffraction experiments we perform today are essentially no different from those Shull per- formed in the 40's. The instruments are better now and easier to use, but the aim of the experiments remains the understanding of the structure of the sample. What has changed, and will continue to change, is our motivation for performing the ex- periment. The original diffraction experiments of Shull and Smart in 1949 [1] were directed as much to the basic understanding of the neutron scatter- ing process as to the properties of manganous oxide. For neutron scattering this period soon passed, helped by the comprehensive theoretical framework developed by Fermi, Bloch, and Halpern and Johnson. Armed with this under- standing the object of the first experiments was the nuclear and magnetic structure of crystals, and Shull's 1951 paper ploughed the first of many furrows through this fertile field [2]. However the nature of many of the samples studied was to change surprisingly soon. The reac- tors of those early days, as now, belonged to nuclear energy centres. It was natural that the problems central to the nuclear energy industry had a growing influence on the choice of experi- ments. When the first European neutron dif- fractometer was installed by Bacon in 1949 on the BEPO reactor at Harwell, one of his first samples was graphite a moderator material for our reac- tors [3]. His measurements to determine the elec- tronic structure of graphite from differences be- tween X-ray and neutron diffraction could hardly at that time be said to be crucial to Britain's nuclear power programme. It was "'underlying research" with a high probability of being one day useful to the programme. The trend to choose "nuclear samples" for intensive study was world- wide throughout the 50's and early 60's and lead to an invaluable fund of knowledge on actmides, especially uranium and its oxides, cladding and structural alloys, radiation damage and defects in moderator materials. The need to understand the neutron modera- tion prcxzess, so fundamental to reactor design, was a cause of the growth of inelastic neutron scattering. The experiments were begun on acceler- ator sources during the war [4]. However, the possibility of burning, rather than just producing, plutonium in the early reactors led to a need to know the effect of the 0.3 eV Pu resonance on the m6derator temperature coefficient, and the design of the first power reactors with regions of modera- tors at different temperatures required knowledge of the spatial variation of the thermal neutron spectrum. Detailed calculations in this new field of 'neutron thermalisation' [5] required measurement of the 'scattering-law" for reactor materials. The need to interpolate and extrapolate these data stimulated the interpretation of the measurements