Spectral measurements of 200 to 300 kV x-rays.

Abstract

The spectral distributions of x rays have been the subject of considerable investigation, and direct spectral measurements have been made by several groups, notably Cormack et al. (1) and Hettinger and Starfelt (2). We became interested in spectral measurements when we wished to compare calculated rad per roentgen factors with those measured directly by calorimetric means (3). Previous measurements by others had covered only a limited number of voltage and filter combinations; thus we decided to make spectral measurements of the beams used in our calorimetric work and to extend the program to include the entire orthovoltage range. This paper presents the results of spectral measurements on all commonly employed voltage and filter combinations in the range of 200 to 300 kV. It is also shown that a knowledge of the half-value layer is sufficient to obtain a reasonably accurate value for the rads per roentgen factor, both for soft tissue (water) and for the soft-tissue component within bone. Apparatus All the primary x-ray spectra shown here were obtained by using a scintillation counter to measure the radiation scattered by a thin sheet of low-atomic-number material placed in the beam at about the normal FSD. Measured spectra were then converted to incident primary spectra as described in the next section. Figure 1 gives the arrangement for these spectral measurements. Measurement of the once-scattered spectrum, as suggested by Scrimger and Cormack (4), offers considerable advantages over direct measurement in the high-voltage region, since x-ray generators may be operated at normal tube currents, clinical beam collimators may be employed, pinhole alignment is no longer a problem, and a large-bore collimator may be used at the crystal, thus reducing the problem of scatter effects within the collimator. The pulses from the NaI crystal were fed to a single channel analyzer with a continuous scan attachment, then to a rate meter, and the counting rate was displayed on a 12-in. linear chart recorder. Scanning speed was such that a complete spectrum was automatically plotted during a fifteen-minute x-ray exposure. Procedure Obtaining the primary spectrum from the observed spectrum required the application of two separate correction procedures. The first step was the conversion of the observed spectrum into that incident upon the crystal. This was accomplished by employing a number of gamma-emitting isotopes to obtain the crystal response to single energies. With these data, plus interpolations, the crystal response function was constructed in the form of a matrix. Matrices having various sizes and varying channel widths were experimented with; but for the spectra shown here, use was made of a 24 × 24 matrix having equal channel widths of 10 keV each. The matrix was inverted on the department's 1401 computer. Multiplication of an observed spectrum by the inverted matrix produced the spectrum incident on the crystal.