TH-AB-204-12: Practical/ultrasensitive Benchtop Gold L-Shell XRF Imaging System with a Kilowatt-Range X-Ray Source and Silicon Drift Detector

Abstract

Purpose: To improve the sensitivity and throughput of a benchtop x-ray fluorescence (XRF) imaging system capable of locating and quantifying distributions of gold nanoparticles (GNPs) by detecting gold L-shell XRF photons from subcentimeter-sized ex vivo samples or superficial tumors for preclinical small animal studies. Methods: A proof-of-principle L-shell XRF imaging system has been developed previously. Originally, the system was configured with a low-power (∼50 W) x-ray source (62 kVp, 0.8 mA) and a silicon PIN diode detector. The detection limit of the system was approximately 0.02 mg/cm3 (20 ppm) for an acquisition time of 10 min. Currently, the system has been retrofitted with a high-power (∼3 kW) x-ray source. The detector was replaced with a silicon drift detector (SDD) to handle the increased photon flux at the new settings (62 kVp, 45 mA). Calibration was performed using phantoms containing water/GNPs at a wide range of concentrations (0 and 0.0001–10 mg/cm3). The acquisition time was set to 10 s (60 times shorter than that used with the former configuration, to maintain the same overall x-ray fluence and dose). The XRF/scatter spectrum at 90° was acquired from each phantom. The data were processed to extract XRF signal from background, which was then corrected for attenuation using a Compton-scatter-based normalization algorithm. The corrected XRF signals were plotted vs. GNP concentration and analyzed to determine the detection limit. Results: The detection limit with the current configuration was found to be 0.007 mg/cm3 (7 ppm), a three-fold improvement compared to the former configuration for the same dose. Conclusion: By adopting a high-power x-ray source and SDD, the material detection limit has been improved to a level comparable to that of a synchrotron-based system. Concurrently, much shorter acquisition time afforded by the higher photon flux makes the current setup practical for routine preclinical imaging tasks.