SummaryThis paper investigates how mathematical approximations and statistical errors are transmitted into computed cross sections in the analysis of experimental bremsstrahlung yield data. The resolution of bremsstrahlung experiments is defined in analogy with optical resolution and an expression for the practical evaluation of resolution is derived. Methods of cross-section computation, and smoothing and curve-fitting are discussed.
ERROR INHERENT IN THE YIELD ANALYSISThe bremsstrahlung experiment measures the yield of a photonuclear reaction. The corresponding cross section is derived by a transformation calculation. This investigation is concerned with how approximations in the transformation calculation and standard errors of the original discrete set of experimental yield ordinates are propagated into the computed cross section, and how well the computed cross section portrays the true (exact) cross section. It is assumed that the bremsstrahlung spectrum is known and that experimental yield ordinates contain no other than truly statistical errors.Yield y(E Ok ) and cross section s(E) are related through the normalized bremsstrahlung distribution function P(E,E ok ) by the integral equation
I
EOk
y(E Ok )= P(E,Eok)s(E)dE.
Eth (1)EOk is the maximum energy of the speetrum, E=hv the energy of a photon interacting with a nucleus, and E th the reaction threshold. Experimental yield data are invariably given as a discrete set of yield values only and not as a continuous function. Thus the knowledge about the yield function is restricted to that of a finite number of yield ordinates, and nothing is known a priori about the behaviour of the function between these ordinates. Ei=Eoi-ttlE;, and s(Ei ); i=l, 2, .. 'J p represents a discrete set of cross-section ordinates which are to approximate ordinates S(Ei) of the exact cross section s(E). Any approximation other than