IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 53, NO. 3, JUNE 2006
Observation of 238U Photofission Products D. K. Wehe, H. Yang, and M. H. Jones Abstract—Sophisticated solutions are being developed by researchers to improve detection of clandestine fissile material. In this work, a simpler approach is investigated as an interim solution. Using the existing 9-MeV X-ray radiography units deployed for cargo inspection, delayed fission product gamma rays from induced photofission are seen minutes after probing a small sample of 238 U. By acquiring both time and energy information, a relatively large number of usable fission products can be identified. Using the radiographic images along with an efficient portable scintillator capable of detecting gamma rays above 3 MeV may provide a cost-effective improvement to existing inspection systems. Index Terms—Accelerator, inspection, radiation detectors, special nuclear material.
RELATIVELY new emphasis in radiation detection applications involves the detection of special nuclear materials (SNM) at points of commerce. What makes this problem particularly challenging is the inability to control the potential source measurement geometry and environment. The importance of the problem has drawn serious attention from a number of groups, using both active and passive approaches, and using thermal, electromagnetic, and neutron sensors. In early work, Los Alamos National Laboratory (LANL) researchers used a 10 MeV LINAC and integrated photofission product spectra from % uranium (HEU), natural uranium, highly-enriched and highly-enriched Pu between beam pulses using a heavily shielded HPGe detector . French Atomic Energy Commission researchers have used high-energy photon interrogation to stimulate photofission and used the delayed neutron count rate to determine the mass of actinides in waste packages –. U They also have compared the delayed gamma spectra for and U and demonstrated a distinct intensity distribution from the difference in the fission product distributions. Idaho National Laboratory, LANL, and Applied Research and Applications Corporation (ARACOR) researchers have tried to identify shielded SNM using high energy X-ray interrogation to view delayed neutron emissions , . To distinguish between depleted uranium (DU), HEU, and thorium, a ratio of delayed neutron count rates produced by two different beam energies is utilized. Using their detector and cargo container setup, they demonstrated detection of 4.8 kg of HEU and 4.4 kg of DU surrounded by 1.27 cm of lead or 20 cm of water. Another group of Manuscript received November 8, 2005; revised March 3, 2006. D. K. Wehe and H. Yang are with the Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor, MI 48109-2104 USA (e-mail: [email protected]
). M. H. Jones is with Ford Motor Company, NDE Lab, Livonia, MI 48150 USA (e-mail: [email protected]
). Digital Object Identifier 10.1109/TNS.2006.875154
researchers from Lawrence Berkeley National Laboratory, University of California-Berkeley, and Lawrence Livermore National Laboratory have proposed detecting the relatively high MeV that are emitted intensity of gamma rays with from short-lived fission fragments induced by neutrons , . Noting the beta-delayed gamma rays have an order of magnitude larger intensity than the delayed neutrons, the authors claim an over neutron detection for a hyincrease in sensitivity of drogenous-filled space (since the gamma rays are more likely to be transmitted than neutrons). Furthermore, energy spectra and time dependence for emission of the beta-delayed gamma rays provide unique signatures for SNM. A group at Argonne National Laboratory interrogates potential fissile material using O reaction the 6–7 MeV gamma—rays from the F to induce neutron emission by the processes of photofission and/or photoneutron production , . Since the photoneutron emission threshold in nuclear materials is less than 6 MeV, while the threshold for photoneutron emission from most benign materials is greater than 7.5 MeV, this energy range provides good discrimination. Since the photofission cross section near its threshold is small, a low gamma and neutron background is required. This approach allows counting of prompt fission neutrons, which are much more numerous than delayed neutrons. The preceding work involving active interrogation utilizes elegant hardware to accomplish the detection of SNM. In this work, we seek a simpler solution that can be applied to existing x-ray inspection units. In particular, we investigate whether the 9-MeV inspection systems currently used to scan cargo can be used not only for density measurements, but also, from the induced fission from fissile materials, as SNM-specific sensors. II. PREDICTED
U PHOTOFISSION RATE
U cylinThe expected photofission reaction rate in our drical rod (105 g, 1 cm 9 cm long, 14.8 g/cm ) was estimated using an experimental measurement of the dose rate produced by Ford’s 9-MeV LINAC and a simulation of the x-ray spectrum produced per 9 MeV electron using MCNP5 . The calculated energy spectrum of the x-rays emerging from this LINAC is shown in Fig. 1, which comes from the MCNP simulation assuming no shielding around the target head. From the spectrum, the dose per 9 MeV source electron was calcuR. At a distance of 63.5 cm (25”) lated to be from the x-ray source, the dose rate was measured using a calibrated ion chamber to be 1.56 Gy/s (177.27 R/s). The ratio of these two values gives the effective current of the LINAC to be 139 A (the LINAC works in pulse mode, with a duty cycle of 300 Hz and pulse duration of 5 s). This is larger than the 50 A reported on a similar machine . Shielding around the target head, which could account for some of this difference, should not be a significant factor when the sample is centered downstream in the beam. Since we are only interested in
0018-9499/$20.00 © 2006 IEEE
WEHE et al.: OBSERVATION OF
U PHOTOFISSION PRODUCTS
TABLE I NEUTRON-INDUCED FISSION PRODUCTS OF U THAT EMIT HIGH-ENERGY GAMMA RAYS 
Fig. 1. Calculated x-ray spectrum.
order-of-magnitude estimates, this level of agreement is acceptable. By integrating the product of the interrogating x-ray spectrum with the energy dependent U photofission cross section , and multiplying by the number of exposed U nuclei, the expected photofission reaction rate from our sample, induced – reactions per by Ford’s LINAC, was estimated to be second. An estimate of the production rate of high energy ( MeV) gamma rays was also calculated. These are particularly interesting because they can more readily penetrate most shield materials that could be used to conceal the nuclear material. U photofission, Without data on fission product yields from we used fission product yields from neutron-induced fission from LBL Nuclear Science Division’s “Isotope Explorer”, U nucleus that fissions, 2003. (For photofission, it is the U nucleus fissions in the neutron induced case). whereas the There were 253 fission products found with cumulative yields greater than 0.1% per fission, and with half-lives less than 10 minutes. Of those, there were 59 nuclides that produced reportable -ray intensities at energies above 3 MeV. The most intense of these emitters are listed in Table I. U photofission rate above and the Using the calculated yields of those neutron-induced fission products having gamma energies above 3 MeV, one can estimate the production rate (i.e., MeV) gamma rays using saturated activity) of high energy ( , where R is the reaction rate, Yi is the yield of isotope i, and fij is the branching ratio of isotope i to produce energy j. From these values, we gammas per second for predict a saturated activity of our sample or 600 delayed high-energy gamma rays per second per gram of sample exposed. An emission rate approaching this value can be achieved with an exposure time of a few effective half-lives. III. PHOTO-IRRADIATION OF AL, PB, AND
During the first experiments, a sample was placed on the axis of the x-ray beam at a distance of 63.5 cm (25”) from
TABLE II OBSERVED PHOTOFISSION PRODUCTS FROM
the bremsstrahlung target. To remove any thermal neutrons that may have originated from moderation of photoneutrons in surrounding materials, the sample was wrapped in a thick Cd sheet, with an opening facing the beam. The first experiment was a 5-minute irradiation of an Al sample to verify that no photoactivation was taking place in this common material. Since the photoneutron threshold is 13 MeV, the negligible induced activity dose rate of 50 R/h was not unexpected. The second experiment was a 5-minute irradiation of a Pb sample (with pho-
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 53, NO. 3, JUNE 2006
TABLE III U PHOTOFISSION REACTION RATE
toneutron thresholds of 7–8 MeV) that yielded a similar negligible dose rate. The third run was a 12-second irradiation on our U sample. The uranium rod was placed perpendicular to the axis of the beam, presenting a maximum thickness of 1 cm. For this case, a dose rate of 1.6 mR/h was induced with only a 12-second irradiation. An HPGe detector recorded the gamma U several minray spectrum from the photoactivation of the utes (386 seconds) later in a single measurement of 423 seconds. The isotopes that could be identified in the gamma spectrum are listed in Table II. From the net counts of gamma rays emitted by selected isotopes, the photofission reaction rate was estimated. The photofission rate R from a pulsed source can be calculated from the measured net photopeak counts C using (cf. Appendix I):
where Y is the yield fraction of this isotope from a photofission, is the period of the pulse (1/300 s), is the branching ratio of the measured photopeak, is the absolute peak efficiency of the HPGe detector, is the decay constant of this isotope, is the length of the irradiation, is the time delay between irradiation and counting, and is the counting time. As noted above, U photofission yields were approximated by the neutron-induced fission product yields. The photofission rates inferred from several observed isotopes are shown in Table III. The agreement between the mea– s reaction rate predicted surement results and the above by the MCNP simulation is acceptable since we are only seeking order-of-magnitude estimates at this stage in our investigation. IV. PHOTO-IRRADIATION OF U FOR HIGH ENERGY GAMMA PRODUCTION A second set of irradiations of U was performed to focus on the delayed high-energy ( MeV) gammas, now using both HPGe and NaI detectors. Since it would be far easier to be able to use a scintillator in the envisioned practical setting, the goal was to see whether the enhanced efficiency at higher energies of a NaI(Tl) detector could be useful despite the loss of energy resolution. This follows the work being done for Department of Homeland Security (DHS) at LLNL by Norman et al.  in
TABLE IV ADDITIONAL MEASURED FISSION PRODUCTS FROM
which they assert that gamma rays above 3 MeV are a useful signature of fission. U sample was first exposed for 24 seconds. After The U sample was a 37-second transport delay, the exposed placed on our (7.62 cm 7.62 cm) coaxial HPGe. A maximum counting rate of 3000 counts per second was measured. A sequence of 3-second measurements followed, and yielded an seconds. average half-life of After ensuring the sample had decayed to background levels, U sample was again exposed for 24 seconds. After a the 40-second transport delay, the sample was placed on a 5.08 cm (2”) tall spacer above a standard 7.62 cm (3”) by 7.62 cm (3”) NaI(Tl)/PMT detector. A maximum counting rate of 3376 counts per second was observed. The average half-life was determined to be 75 seconds. After scaling for distance, one can expect a counting rate of 12.5 cps from the high-energy gamma rays at 1 m from the same sample, using the same detector and after the same irradiation. For comparison, under the same conditions, the background was measured to be 0.7 cps. No distinct structure was observed in the spectra above 3 MeV, even from the HPGe detector. Again, as expected, no activation was observed after additional irradiations of Pb, Fe and Al samples. However, the additional fission products shown in Table IV were identified using the variable time intervals to acquire these spectra. The spectra acquired by the NaI(Tl) and HPGe detectors are compared in Fig. 2 and show the radioisotopes listed in Table IV. For the HPGe detector, spectra are shown for both before and after irradiation. The pre-irradiation count rates are scaled up by a factor of 20 in order to be visible on the same graph with the post-irradiation spectrum. The difference in the two spectra clearly indicates the occurrence of photofission, as predicted.
WEHE et al.: OBSERVATION OF
U PHOTOFISSION PRODUCTS
Fig. 2. Comparison between spectra taken by HPGe and NaI(Tl) detectors.
An efficiency-corrected ratio of the 186 keV to the 1001 keV U fraction of 0.07% (this only confirms the counts gives a sample is natural or depleted uranium without correcting for self-absorption). The relative number of counts for photopeaks from the same radioisotope decay appears to be consistent, as are the predicted half-lives. The relative number of NaI(Tl) counts is less than what might be expected on the basis of intrinsic efficiency because of the increased counting distance. Because of the small energy separation between photopeaks, the spectrum from NaI(Tl) does not provide particularly useful information on specific fission products. For the proposed application, however, the spectrum is not required. Furthermore, since there are no identifiable energy peaks shown above 3 MeV even on the HPGe-measured spectrum, the NaI(Tl) detector may be preferable due to its higher detection efficiency, portability, and relatively low cost.
TABLE V PHOTONEUTRON AND PHOTOFISSION THRESHOLDS FOR MATERIALS OF INTEREST 
V. SUMMARY AND PATH FORWARD In this work, we have observed photofission in U, and showed that result can be observed well after the irradiation, thereby avoiding complications of trying to observe fission product emissions between beam pulses with a heavily shielded detector. We have not seen significant exposure impacts on Al, Pb, or Fe samples, but we have not thoroughly explored all common materials that might be expected in practice, nor have we considered activation or fission from photoneutrons released from materials such as deuterium or carbon (cf. Table V). Since these accelerators are routinely being used in the field for inspection, we have assumed these are not significant. Another area requiring further examination is whether the count rates are adequate for practical settings and whether it would be more advantageous to look for specific radioisotopes or specific energy windows. Finally, manipulating the time evolution of
the fission product activity using multiple exposures could be examined. The extension of this work to fissile isotopes is, of course, of contemporary interest.
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 53, NO. 3, JUNE 2006
APPENDIX Assume is the number of atoms of this isotope produced during each pulse, assumed to appear instantaneously at the beginning of the pulse. (Since the isotopes are produced during ms pulse, this is a reasonable approxithe first 5 s of the mation). The number of counts under the photopeak contributed by the ith pulse is given by:
where the symbols are defined in the text. Performing the finite sum yields:
The photofission rate R is given by:
where Y is the yield fraction of this isotope from a photofission. and substituting into the Solving this latter equation for previous expression yields the desired result. REFERENCES  C. L. Hollas, D. A. Close, and C. E. Moss, “Analysis of fissionable material using delayed gamma rays from fission,” Nucl. Instrum. Methods Phys. Res. B, vol. B24/25, pp. 503–505, 1987.  N. Saurel, J. M. Capdevila, N. Huot, and M. Gmar, “Experimental and simulated assay of actinides in a real waste package,” Nucl. Instrum. Methods Phys. Res. A, vol. A550, pp. 691–699, 2005.
 M. Gmar, F. Jeanneau, F. Lainé, H. Makil, B. Poumaréde, and F. Tola, “Assessment of actinide mass embedded in large concrete waste packages by photon interrogation and photofission,” Appl. Radiat. Isot., vol. 63, no. 5–6, pp. 613–619, Nov.–Dec. 2005.  M. Gmar and J. M. Capdevila, “Use of delayed gamma spectra for detection of actinides (U,Pu) by photofission,” Nucl. Instrum. Methods Phys. Res. A, vol. A422, pp. 841–845, 1999.  J. L. Jones, W. Y. Yoon, Y. D. Harker, J. M. Hoggan, K. J. Haskell, and L. A. VanAusdeln, “Proof-of-Concept Assessment of a Photofission-Based Interrogation System for the Detection of Shielded Nuclear Material,” Idaho National Engineering and Environmental Laboratory, Idaho Falls, ID, Rep. INEEL/EXT-2000-01523, Nov. 2000.  J. L. Jones, W. Y. Yoon, K. J. Haskell, D. R. Norman, J. M. Hoggan, C. E. Moss, C. A. Goulding, C. L. Hollas, W. L. Myers, and E. Franco, “Photofission-Based, Nuclear Material Detection: Technology Demonstration,” Idaho National Engineering and Environmental Laboratory, Idaho Falls, ID, Rep. INEEL/EXT-02-01406, Dec. 2002.  E. B. Norman, S. G. Prussin, R. Larimer, H. Shugart, E. Browne, A. R. Smith, R. J. McDonald, H. Nitsche, P. Gupta, M. I. Frank, and T. B. Gosnell, “Signatures of fissile materials: high-energy gamma-rays following fission,” Nucl. Instrum. Methods Phys. Res. A, vol. A521, pp. 608–610, 2004.  D. Slaughter, M. Accatino, A. Bernstein, J. Candy, A. Dougan, J. Hall, A. Loshak, D. Manatt, A. Meyer, B. Pohl, S. Prussin, R. Walling, and D. Weirup, “Detection of Special Nuclear Material in Cargo Containers Using Neutron Interrogation,” Lawrence Livermore National Laboratory, Livermore, CA, Rep. UCRL-ID-155315, Aug. 2003.  B. J. Micklich and D. L. Smith, “Nuclear materials detection using high-energy -rays,” Nucl. Instrum. Methods Phys. Res. B, vol. B241, pp. 782–786, Dec. 2005.  B.J. Micklich, D.L. Smith, T.N. Massey, C.L. Fink, and D. Ingram, “FIGARO: Detecting nuclear materials using high-energy gamma-rays,” Nuclear Instrum. Methods Phys. Res. A, vol. A505, pp. 466–469, 2003.  F. B. Brown, “MCNP Version 5,” Trans. Amer. Nucl. Soc., vol. 87, p. 273, Nov. 2002, (LA-UR-02-3935).  R. L. Metzger, K. A. Van Riper, and M. H. Jones, “Ford motor company NDE facility shielding design,” Radiat. Protection Dosimetry, vol. 116, no. 1–4, pp. 236–238, 2005.  T-2 Information Services [Online]. Available: http://t2.lanl.gov/ LANL  Fission Product Yields Lawrence Berkeley National Laboratory [Online]. Available: http://ie.lbl.gov/fission.html  Both Sb and its isomer decay by emitting gamma rays at 696 keV. In our case, we weren’t able to tell the exact source of the peak at this energy. By presenting both results under such assumption that all the gamma rays come from either one of the two possible sources ( Sb and Sb), we actually bounded the true value.  IAEA Photonuclear Data Library [Online]. Available: http://www-nds. iaea.org/photonuclear/