Accurately measuring the focal spot size of the electron linear accelerator is important for the optimized design and performance evaluation of high-energy industrial CT systems. However, due to the high energy, high dose rate, and strong penetration ability of X-ray generated by the electron linear accelerator, there are difficulties in accurate measurement of the focal spot size. Currently, the main test standard for the focal spot size of the electron linear accelerator is the "Sandwich" test method proposed in the IEC 62976—2021 (or stacking test method in GB/T 20129—2015). However, this method is not only cumbersome in the actual measurement, but also seriously affected by human factors during film exposure, processing, and streak counting. Meanwhile, the thicknesses of the metal and plastic sheets of the "Sandwich" test module have a significant influence on the measurement results of the focal spot size. To this end, we propose a silt translation scanning test method which is more objective, accurate, and better repeatable than the "Sandwich" test method.

The limitations of the "Sandwich" test method are analyzed by experiments and simulations. First, two measurements of the focal spot size of the 9 MeV electron linear accelerator are conducted using the "Sandwich" test method with reference to IEC 62976—2021 and GB/T 20129—2015. The differences between the two measurements are compared and the reasons are analyzed. Then we employ the Monte Carlo particle simulation software BEAMnrc to construct the electron linear accelerator and simulate the effect of different metal and plastic sheet thicknesses of the "Sandwich" test module on the focal spot size measurement results. After analyzing the disadvantages of the "Sandwich" test method, the device structure and measurement principle of the slit translation scanning test method are presented. The focal spot size of a 9 MeV electron linear accelerator is measured several times using the slit translation scanning test method for comparison with the "Sandwich" method. Finally, to further validate the accuracy of the proposed test method, we deduce the focal spot size of the electron linear accelerator by measuring the spatial resolution of the CT system and compare it with the results measured by the slit translation scanning test method.

In the actual measurement of the "Sandwich" test method, it is difficult to precisely control the exposure dose and exposure time of the film, which affects the stripe contrast of the film. Meanwhile, the developing and fixing process of the film is easily affected by the environment, developer concentration, developing and fixing time, and other factors to result in poor repeatability of the measurement results. Figs. 3 and 4 show that the measurement result error for the same electron linear accelerator can be more than 30%. Additionally, the simulation results in Fig. 5 and Table 1 indicate the measurement error of the focal spot size obtained from different thicknesses of metal and plastic sheets of the "Sandwich" test module can be up to ±12.5%. Therefore, the influence of different metal and plastic sheet thicknesses on the measurement error must be considered in the "Sandwich" test method. For the slit translation scanning test method, Table 2 and Fig. 9 reveal that the maximum error of its multiple measurements is only ±0.95%, and the measurement is little affected by the dose rate fluctuation of the electron linear accelerator and not affected by the exposure time. Thus, the measurement repeatability is good. In addition, the focal spot size of the electron linear accelerator calculated by equation (2) is the same as that measured by the slit translation scanning test method. Thus, the focal spot size measurement results of the slit translation scanning test method are more accurate and objective than the "Sandwich" test method of IEC 62976—2021 and GB/T 20129—2015.

We study and design a new set of measurement methods and devices for focal spot size, which is the slit translation scanning test method and device to address the shortcomings of the "Sandwich" test method for the focal spot size measurement of electron linear accelerator, as specified in IEC 62976—2021 and GB/T 20129—2015. The "Sandwich" test method is not only cumbersome in practice, but also greatly influenced by the environment, experimental conditions, and human subjective factors during the film exposure, processing, and streak counting. Additionally, theoretical simulation reveals that the thicknesses of the metal and plastic sheets of the "Sandwich" test module introduce a measurement error of more than ±12.5%. To verify the new measurement method, we conduct experiments such as focal spot size measurement and verification of spatial resolution of high-energy industrial CT. The experimental results demonstrate that compared with the "Sandwich" test method, this new method provides objective, accurate, and repeatable measurement results. These findings are of significance for the performance evaluation of electron linear accelerators and optimized design of high-energy industrial CT systems.