Experimental validation of Monte Carlo simulation model for X-ray security scanner

Transmission X-ray security scanners are deployed to detect smuggling of contraband articles, including weapons, narcotics, and explosives, for the purposes of homeland security. Current X-ray scanners use a fixed tube voltage (i.e., 160 kV); hence, they have a limitation in detecting thinly coated and/or low-density objects. To overcome this limitation, we are developing an X-ray scanner that applies variable tube voltage according to the physical/chemical properties of the object being inspected. To this end, in our previous study, Monte Carlo simulations with Geant4 (GEometry ANd Tracking4) and MCNP (Monte Carlo N Particle) were performed to optimize the design of the X-ray scanner for variable tube voltages. The MCNP was used to simulate the radiation generator for the X-ray source term, and the Geant4 was used to optimize the design of the dual-energy detector and to obtain the detector counts. In the present study, we experimentally validated the reliability of the Monte Carlo simulation model for the X-ray scanner. An ²⁴¹Am source and a radiation generator were employed in this validation. For the ²⁴¹Am source, the dual-energy detector signal was measured at a distance of 1 cm from the detector. In the case of the radiation generator, the source-to-object distance and the source-to-detector distance were 70 cm and 120 cm, respectively, as used in a typical X-ray security scanner. The tube voltage and the current were 140 kV and 10 mA, respectively. To obtain the X-ray images, the object was scanned while moving at a speed of 0.2 m/s on a conveyor system. The X-ray source term used in the simulation was obtained by monoenergetic-electron-beam bombardment onto the target. 4-D simulations were performed for the moving object. To validate the simulation model, we compared the simulated and measured image profiles, as well as the pixel counts of the dual-energy detector. The percent differences between the simulated and measured pixel counts and the image profiles were all within 5%. Thus, we concluded that our simulation model for the X-ray scanner can be considered to be reliable.