Normalisation and assessment of neutron diffraction data from liquids

Abstract

A couple of normalisation criteria have been proposed to assess liquid structure data over the past decade. In this paper, a critical examination of these and other plausible criteria is made. Neutron diffraction data from liquid krypton measured by Clayton and Heaton is used to study these criteria. It is shown that if the structure factorS (Q)=1+γ (Q) satisfies the Krogh-Moe relation, namely$\rho _{0 = \frac{1}{{2\pi ^2 }}} {\text{ }}\mathop \smallint \limits_{\text{0}}^\infty {\text{ }}Q^{\text{2}} {\text{ }}\gamma {\text{ (}}Q{\text{) }}dQ$ whereQ is the wavevector transfer andρ ₀ the mean atomic number density,S (Q) is properly normalised. Subjecting such data to the other criteria would reveal, however, the quality of data in so far as systematic errors and/or termination errors are concerned. If the data is measured with reasonable accuracy to sufficiently large values of the wavevector transfer all these criteria would assess the quality of the data to almost the same extent. This is established in this paper by improving and extending the liquid krypton data by procedures similar to that of Kaplow, Strong and Averbach and subjecting this revised data to the various criteria. It is concluded, thereby, that (i) one can study the neutron diffraction data from liquids using any of the criteria mentioned in the paper to improve the quality of the data and (ii) for normalisation purposes the simple relation {ie328-2} would suffice.