SUMMARYNuclear Resonance Fluorescence (NRF), which is possible for nuclei with atomic numbers greater than helium (Z=2), occurs when a nuclear level is excited by resonant absorption of a photon and subsequently decays by reemission of a photon [1]. The excited nuclear states can become readily populated, provided the incident photon's energy is within the Doppler-broadened width of the energy level being excited. Utilizing continuous energy photon spectra, as is characteristic of a bremsstrahlung photon beam, as the inspection source, ensures that at least some fraction of the impinging beam will contribute to the population of the excited energy levels in the material of interest. Upon deexcitation, either to the ground state or to a lower-energy excited state, the emitted fluorescence photon's energy will correspond to the energy difference between the excited state and the state to which it decays. As each isotope inherently contains unique nuclear energy levels, the NRF states for each isotope are also unique. By exploiting this phenomenon, NRF photon detection provides a well-defined signature for identifying the presence of individual nuclear species.This report summarizes the second year (Fiscal Year [FY] 2009) of a collaborative research effort between Idaho National Laboratory, Idaho State University's Idaho Accelerator Center, and Pacific Northwest National Laboratory. This effort focused on continuing to assess and optimize NRF-based detection techniques utilizing a slightly modified, commercially available, pulsed medical electron accelerator. During the course of FY 2008, a gain shift occurred in some data sets, resulting from radiofrequency (RF) pickup and the existence of substantial ground loops in the accelerator control system. This pickup was addressed and subsequently eliminated by replacing the legacy cables with single run signal cables, which also contained improved cable shielding. Ground loops were eliminated through an optical isolation of the data acquisition system from the accelerator control system. Overall, the RF pickup was decreased by more than a factor of 100. The FY 2009 testing configurations were further optimized with the identification and acquisition of a more appropriate nuclear material target (0.22 cm thick 238 U), which provided a significant increase in the signal-tobackground ratio and decrease in the time necessary to detect both nuclear and non-nuclear material in the presence of high-Z material.A primary focus for FY 2009 tasking included modifying/upgrading the current LINAC to support NRF operations with accelerator repetition rates up to 2 kHz. By increasing the capacitance and inductance of the pulse forming network and changing the RF driver to a long pulse mode, the LINAC's RF pulse width was extended to 20 ȝs. With the construction of a new electron gun hot deck and burst mode pulsing system, multiple electron pulses were available for injection during a single RF pulse. This multipulse capability, in conjunction with an RF repetition rate of 300 Hz, pro...