Feasibility of 106Ru peak measurement for MOX fuel burnup analysis

Feasibility of 106Ru peak measurement for MOX fuel burnup analysis

Nuclear Engineering and Design 240 (2010) 3687–3696 Contents lists available at ScienceDirect Nuclear Engineering and Design journal homepage: www.e...

1MB Sizes 16 Downloads 93 Views

Nuclear Engineering and Design 240 (2010) 3687–3696

Contents lists available at ScienceDirect

Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes

Feasibility of 106 Ru peak measurement for MOX fuel burnup analysis Matthew L. Dennis, Shoaib Usman ∗ Department of Mining and Nuclear Engineering, Missouri University of Science & Technology, Rolla, MO 65409, USA

a r t i c l e

i n f o

Article history: Received 16 October 2008 Received in revised form 27 June 2010 Accepted 6 July 2010

a b s t r a c t Simulations were performed using ORIGEN-ARP to investigate 137 Cs and 106 Ru–106 Rh as suitable fission products for non-destructive analysis of irradiated MOX. The simulations confirm that both 137 Cs and 106 Ru will provide a linear correlation with burnup when exclusively applied to MOX fuel assemblies. Moreover, 106 Ru can also be used in conjunction with cesium to form a ratio almost independent of enrichment and power history. Simulations were conducted using three different uranium enrichments and one MOX enrichment over a burnup range of 20,000–60,000 MWD/MTHM. Comparison of the three uranium enrichments indicates the 106 Ru ratio is consistent in predicting burnup with a maximum standard deviation of 0.046. Two MOX cases were simulated confirming operational history independence of 106 Ru in predicting total burnup. The 106 Ru burnup ratio also has the benefit of enabling distinction between UO2 and MOX fuel because of its significantly larger (∼11 times) fission yield from 239 Pu. To investigate the detectability of 106 Ru and other cesium peaks, data was collected using a HPGe detector at the Missouri S&T nuclear reactor (MS&TR) beam port. Gamma spectra were obtained immediately following reactor shutdown with the most promising spectrums obtained 3–5 h after shutdown. Even for relatively high enrichment (∼20%) fuel at MS&TR, cesium peaks were prominent and easily discernable from the intense Compton continuum. The 106 Ru peak was weak, though still distinguishable from the background, suggesting that with an appropriately designed collimator, suitable detector and electronics it might be feasible to reliably measure 106 Ru in even UO2 fuel. For MOX and LEU LWR fuels one would expect a more intense 106 Ru signature. © 2010 Elsevier B.V. All rights reserved.

1. Introduction To validate the initial computer simulation (Dennis and Usman, 2006), preliminary experimental data were collected for gamma emission from LEU test reactor fuel at Missouri University of Science and Technology (Missouri S&T). The primary goal of this effort is to determine if online burnup analysis of plutonium based mixed oxide (MOX) fuel is feasible using non-destructive gamma spectroscopy. Initial results are very encouraging and it seems feasible to develop techniques for determination of MOX fuel burnup, as well as for discrimination between MOX and uranium oxide (UO2 ) fuel assemblies. However, for commercial applications online, gamma spectroscopy immediately following shutdown/fuel discharge will be complicated by the extremely high activity of the irradiated fuel. As will be shown, the intense gamma radiation emanating from even the low power (200 kW) Missouri S&T Research Reactor (MS&TR) is sufficient to overwhelm a high purity Ge detector.

∗ Corresponding author. Tel.: +1 573 341 4745; fax: +1 573 341 6309. E-mail addresses: [email protected] (M.L. Dennis), [email protected] (S. Usman). 0029-5493/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.nucengdes.2010.07.003

The United States government’s temporary but enduring ban on nuclear fuel reprocessing between 1977 and 1981 crippled the United States reprocessing industry; whereas, nuclear fuel reprocessing gained momentum elsewhere. Not until recently has the U.S. taken a renewed interested in MOX fuel by initiating plans to convert U.S. and Russian weapons-grade plutonium into MOX fuel at the Department of Energy’s Savannah River Site (Shaw Areva MOX Services, 2008). However, limited effort has been made to utilize plutonium in commercial reactors, primarily due to proliferation concerns, whereas other nations are extensively using MOX as part of UO2 fueled cores without any significant safety issues. In 2005, Duke Energy commissioned AREVA to construct and deliver four MOX fuel assemblies for its Catawba nuclear power plant. This renewed commercial interest in MOX, whether using plutonium from decommissioned nuclear weapons or recycled plutonium from fuel reprocessing, indicates a radical fuel cycle change in the United States which should consequently involve a greater knowledge about MOX performance in U.S. reactors. Keeping this in mind, non-destructive techniques such as gamma spectroscopy could be employed to determine actinide burnup (e.g. plutonium, neptunium, americium and curium) in various MOX fuel compositions and discriminate these compositions from commonly used UO2 . This paper will briefly cover fuel reprocessing and ultimate MOX fuel manufacture, PWR and BWR fuel assembly specifications,

3688

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

a brief description of ORIGEN-ARP code, and the mathematics of determining burnup through non-destructive analysis (NDA). As previously mentioned, non-destructive burnup analysis using fission product gamma spectroscopy is a well-documented process for UO2 . Most of the literature reviewed suggests using 134 Cs, 137 Cs, or 154 Eu fission products either separately or in combination. Given the performance similarities between MOX and UO2 fuels, any of the previous fission products should work equally well as MOX burnup indicator and was validated through ORIGEN burnup simulation. However, due to similar fission yields, none of these isotopes are a dependable discriminator between MOX and UO2 fuels. Initial simulation results suggest that for discrimination between MOX and UO2 106 Ru–106 Rh is an appropriate choice. Moreover, 106 Ru–106 Rh can also be used for burnup analysis during outage operations. This discrimination capability can be useful for special nuclear material safeguards purposes. Finally, to demonstrate the feasibility of the concept and to uncover experimental challenges of this approach, gamma spectra of irradiated fuel were collected for the Missouri S&T research reactor using a high purity germanium detector. These initial measurements illustrate possible challenges of irradiated fuel spectroscopy and some mitigation strategies. Further development of this technique can potentially yield a useful tool for experimental validation of burnup through non-destructive analysis which maintains fuel integrity for ultimate disposal or reprocessing. Moreover, the technique can be used to validate computational burnup credit. 2. MOX fuel characteristics It should be noted that MOX could be produced from any fissionable actinide; however, this study focused solely on plutonium based MOX. MOX fuel is composed of one of either two sources of plutonium. It can be manufactured using weapons-grade plutonium, in which the MOX contains greater than 70 w/o 239 Pu in total Pu. Another option is to use reprocessed plutonium from spent UO2 fuel. Since the ultimate goal in the U.S. is to close the fuel cycle, our research was limited to recycled plutonium MOX. Therefore, the plutonium vector (this refers to the various isotopic enrichments of 238 Pu through 242 Pu) used in ORIGEN-ARP must range between 50 and 70 w/o 239 Pu. Since the MOX under investigation originates from recycled UO2 spent fuel, it begins as Pu(NO3 )4 from the PUREX process. It is then transformed into PuO2 and combined with depleted UO2 , hence the moniker “mixed oxide fuel” (the uranium vector typically consists of greater than 99 w/o 238 U) (Cochran and Tsoulfanidis, 1990a). The fabricated fuel, ready for power production, now contains typically 5–9 w/o 239 Pu. While the U.S. has not reprocessed and researched MOX fuel since the mid-70s, other countries have proceeded with extensive research and deployment efforts. European nations including France, Germany, and the United Kingdom have found that only replacing a fraction of a light water reactor (LWR) core with MOX provides the best neutronics and safety characteristics. In fact, France limits its cores to 30% MOX and different plutonium enrichments within a given assembly to flatten power peaking (Cochran and Tsoulfanidis, 1990b). Ultimately, the MOX fuel behaves differently based on multiple fissile plutonium isotopes, varying concentrations based on the level of recycling, how long the plutonium has been stored allowing the 241 Pu decay product and poison 241 Am to build-in, and neutron parameters such as absorption cross sections (Cochran and Tsoulfanidis, 1990b). As a result, along with limiting the amount of MOX in a LWR core, it is prudent to locate MOX away from boron control rods due to reduced reactivity worth concerns and using the fuel as soon as possible to limit growth of 241 Am.

Fig. 1. Cross-section of a 15 × 15 PWR MOX fuel assembly. Light circular region represents MOX fuel pins and dark region represents water.

2.1. MOX fuel assembly The fuel assembly for either a boiling water or pressurized water reactor (BWR and PWR) based on MOX fuel is physically identical to its UO2 counterpart. In the PWR case, the ceramic MOX fuel pellets are stacked in the zircaloy fuel cladding and arranged in a lattice of 14 × 14 up to 18 × 18 supported with grid plates and grid spacers. As part of the fuel pin array, specific tubes remain empty for burnable poisons or control rods. Fig. 1 is a cross-section representation of a 15 × 15 MOX assembly with the dark region representing water. This would be representative of a MOX assembly removed from the spent fuel pool for NDA. In the BWR case, the fuel is again ceramic pellets stacked within zircaloy cladding; however, the key difference between PWR and BWR assemblies is that in BWR assemblies fewer tubes are left empty for water flow, no tubes are used for control rods, and the overall lattice of the BWR assembly is smaller, on the order of 8 × 8 to 10 × 10. For our research, while NDA using gamma spectroscopy is equally applicable to both assembly types, preference was given to PWR MOX fuel assemblies which would be analogous to those potentially used at Catawba nuclear power plant. 3. ORIGEN-ARP background All computational work performed to assess fission product gamma emission peaks and correspondingly determine burnup was done using a code developed and tested by Oak Ridge National Lab (ORNL). ORIGEN-ARP uses previously developed cross-section interpolation to produce problem dependent libraries which can be used to computationally determine isotope depletion in a fraction of the time of more cumbersome computer codes (Gauld et al., 2005). ORIGEN-ARP’s ease of use, speed, and flexibility make it an excellent candidate for this research. Additionally, it also considers nuclide concentrations, radiation emissions and properties for more than 1300 actinides, fission products and activation products (Gauld et al., 2003). In fact, it has been upgraded to consider various types of MOX fueling for commercial PWRs. The gamma spectrum in ORIGEN-ARP is created using a group structure in which an energy range is divided into intervals. The most precise energy spectrum installed as a default in ORIGENARP is 45 discrete energy peaks based on ENDF5 and designated 44GrpENDF5. However, a more refined energy spectrum is required and was created using the maximum 240 discrete energy intervals. The spectrum ranges in energy from 0.1 MeV (100 keV) to 2 MeV (2000 keV) in intervals of 8.3 keV and is designated 239GrpLOW. While the current energy resolution is sufficient for this stage of

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

research, a finer resolution may be required to supplement other isotope identification techniques. Finer resolution of gamma peaks would require narrowing the energy intervals to allow for higher energy resolution (∼eV) which would effectively be the same as “zooming in” on the current spectrum. Since spent MOX fuel assemblies were not accessible during this research, and for that matter there is limited availability in the U.S., this study relied heavily on ORIGEN-ARP in the initial stages to identify potentially useful fission product gamma energies (Dennis and Usman, 2006). Two of these important fission products are 137 Cs and 106 Ru. Initial experimental data confirmed these isotopes and their respective gamma peaks at the Missouri S&T research reactor. 4. Assembly averaged burnup 4.1. NDA theory The thrust of NDA gamma spectroscopy is to characterize the isotopic composition of the fissile elements (235 U and 239 Pu) through the indirect energy signature of their respective fission products. This preserves the fuel integrity and yields useful information such as burnup, cooling time and isotopic contents. It should be noted that burnup is defined as the energy produced by the fuel in units of megawatt days per metric ton of heavy metal (MWD/MTHM). The heavy metal in the instance of UO2 is simply the uranium contents of the fuel; for MOX, heavy metal would include uranium and plutonium. When determining burnup for either safeguards purposes or operator validation, it is important to select a fission product isotope that meets certain criteria (Reilly et al., 1991; Crane et al., 1978): (1) Similar fission yield from Pu and U. This is primarily because in either MOX or UO2 , 239 Pu will be generated from absorption of fast neutrons in 238 U. Therefore, having equal fission yields alleviates the need for knowing reactor power history and initial fuel composition. (2) Low neutron capture cross-section. This ensures that the fission product is generated from fission and not subsequently consumed by parasitic neutron absorption. (3) Low fission product migration. Ideally the fission product remains where it was created from fission. Any migration of fission products will skew the burnup profile of the assembly. (4) High energy gamma emission. This ensures the fission product gamma penetrates the fuel and reaches the detector; in short, it reduces gamma attenuation. (5) Long half-life. A long half-life, or at least similar half-life compared to irradiation history, ensures that the burnup and quantity of fission product are proportional. Burnup analysis using a short-lived isotope will be highly history dependent and less desirable. If most of the prior criteria are met, then determining the fuel burnup is a two-step process accomplished through first determining the number of fission product nuclei from the gamma spectroscopy count rate and finally the burnup knowing the fission product yield. Eq. (1) is solved for number of fission product nuclei formed during irradiation as suggested by Reilly et al. (1991), and Eq. (2) is solved for burnup in MWD/MTHM. N=

I (counts/s) εkSe−T

BU (MWD/MTHM) = 1.8895 × 10−22 × E ×

(1)

N  Y

where ε is the absolute detector efficiency (depends on the intrinsic efficiency of the detector and geometry of measurement), k the gamma branching ratio (depends on the fission product physical characteristics), S the self-shielding factor (depends on the fuel geometry and gamma energy), T the fuel assembly cooling time (time of measurement after shutdown),  the decay constant (depends on the fission product physical characteristics), E the energy released per fission in MeV for plutonium (212 MeV) or uranium (202 MeV), N the number of atoms of fission product nuclei initially produced per metric ton of heavy metal (depends on the fuel type and reactor design) and Y is the fission product yield (percentage, depends on the fuel type and reactor design). Eq. (2) is a relationship between the number of burnup monitor atoms created during fission and the total fuel burnup. The expression is a modified form of the Crane et al. formula (Crane et al., 1978). In our case, the constant in (2) was determined using the sample file bm1.arp (Dennis and Usman, 2006) in ORIGENARP. When calculated burnup for this fuel was compared with the calculated burnup for oecd-ivb.arp (Dennis and Usman, 2006), the burnup error was 0.66%. Thus there is adequate agreement between simulated and directly calculated (using Eq. (2)) results. In practice, the constants of (1) can be found or determined. For example, the absolute detector efficiency for a given geometry can be obtained using a source of known intensity. For high intensity applications, detector deadtime may play a very significant role and additional corrections may be required (Patil and Usman, 2009). 4.2. Fuel assembly burnup – 137 Cs When considering the previously mentioned five fission product criteria, 137 Cs satisfies the most requirements. Its 30-year halflife, 6.26% and 6.65% fission yield for 235 U and 239 Pu respectively, 662 keV singular gamma, and 0.25 barn thermal neutron capture cross-section satisfy four of the five requirements. However, studies have shown it to preferentially migrate to the colder outer fuel pellet region creating a depressed 137 Cs concentration in the pellet core (Crane et al., 1978). Therefore, this migration must be considered when conducting NDA using 137 Cs with particular attention paid to the attenuation correction of Eq. (1). Table 1 gives the attenuation factors of both UO2 and MOX fuel for four important fission products. In MOX fuel, fission product concentration is predominantly a result of 239 Pu thermal fission. However, there will also be some transmutation of 238 U to 239 Pu and fission of 241 Pu and 235 U. It is important to point out that due to neutron energy spectrum hardening, for BWRs the rate of transmutation would be significantly different than the rate of transmutation in PWRs. Nevertheless, for this study we focused on PWR. For a PWR with UO2 or MOX, 137 Cs fission yields will be quite similar. Therefore, the total assembly average burnup can be determined taking into account fission product concentration from both U and Pu. All simulations in this research were performed using the isotopic vectors in Table 2. Using ORIGEN-ARP, the activity of 137 Cs was determined for increasing burnup up to 60,000 MWD/MTHM. Fig. 2 shows the linear relationship between fission product activity and fuel assembly burnup at 30 days after final discharge. At thirty day cooling period was chosen to replicate the “online” portion of this research since Table 1 Attenuation factors of MOX and UO2 (Sasahara et al., 2004). Nuclide

Attenuation factor

134

0.596 0.678 0.619 0.597

Cs (604.7 keV) Cs (795.8 keV) Cs (661.7 keV) 106 Ru (621.8 keV) 134 137

(2)

3689

3690

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

Table 2 Plutoniuma and uranium isotopic vectors used for various fuels. Isotope

%Pu in heavy metal 238 Pu 239 Pu 240 Pu 241 Pu 242 Pu 241 Am 234 U 235 U 236 U 238 U a

Weight percent (w/o) MOX

UO2 – 5.0 w/o

UO2 – 4.2 w/o

UO2 – 2.6 w/o

5.892 1.0137 62.329 23.781 8.928 3.9483 1.0777 0.002 0.237 0.001 99.76

– – – – – – – 0.0445 5.0 0.023 94.93

– – – – – – – 0.03738 4.2 0.01932 95.7433

– – – – – – – 0.02314 2.6 0.01196 97.3649

Plutonium vector for MOX taken from Gauld et al. (2005).

typical outage operations last approximately 30 days, allowing sufficient time for NDA measurements. Fig. 2 corroborates the assertion that MOX and UO2 fuels behave alike when used in a PWR. Thus, the 137 Cs absolute activity measurement can quantitatively describe UO2 as well as MOX fuel burnup. However, the method for obtaining a sufficiently resolved gamma spectrum from spent nuclear fuel is an entirely separate issue and will be discussed later in Section 6 as part of the results from the Missouri S&T reactor experiments. 4.3. Fuel assembly burnup – 106 Ru The fission product 106 Ru is entirely inadequate as an absolute measurement of burnup in UO2 fuel due to its drastically different fission yield from 239 Pu and 235 U. On the other hand, this difference in fission yield can be exploited for discrimination purposes, as will be discussed later. If the fuel matrix is known to be a MOX fuel assembly, 106 Ru has potential to singularly serve as a burnup indicator, whereas UO2 suffers from neutron capture of 238 U, resulting in 239 Pu production which is responsible for as much as 20% of the power. In MOX fuel there are two primary sources contributing to burnup: fission occurring in the enriched 239 Pu and fission occurring in 239 Pu produced from the conversion of 238 U. In both instances, all burnup can be traced directly back to fission arising from 239 Pu. Therefore, while direct measurement of the 106 Ru gamma signature serves as an absolute measurement susceptible to the same reproducibility issues as the 137 Cs technique, it is equally viable provided that it is known that the fuel is indeed MOX.

Fig. 3. The calculated 106 Ru activity as a function of burnup for simulated 17 × 17 MOX fuel using ORIGEN-ARP. MOX fuel total Pu initial enrichment was 5.89% and discharged after three cycles with a 30-day decay period between each cycle.

Using ORIGEN-ARP, the activity of 106 Ru was determined for increasing burnup of MOX fuel up to 60,000 MWD/MTHM. Fig. 3 shows the linear relationship between burnup and 106 Ru activity at 30 days after final discharge. The results of Fig. 3 are predicated on the assumption that NDA measurements occur shortly after discharge since after 7 years of continuous decay 106 Ru will have almost entirely disappeared. Also, the MOX assembly operational history should be known since this correlation is dependent on power history and initial enrichment. 5.

106 Ru

burnup ratio

5.1. Correlation properties The previously discussed five fission product criteria must be applied to 106 Ru to assess its effectiveness as burnup monitor. Unfortunately, it falls short in several key areas for absolute activity measurements. Its 369-day half-life (a sufficient time when compared to a one batch core) and 0.15 barn thermal neutron crosssection satisfy two requirements; however, 106 Ru migrates in the fuel similar to 137 Cs (hence the attenuation factors in Table 1) and has a drastically higher fission yield from 239 Pu as compared to 235 U, 4.48% and 0.39% respectively. This almost 11 times greater fission yield in 239 Pu makes it less accurate in describing the UO2 assembly burnup, overestimating the contribution from other fissile isotopes (239 Pu in particular) discussed earlier. Yet, this drastic fission yield difference can be utilized to ascertain if an unknown fuel assembly is either MOX or UO2 , which is a key component of nuclear material accountability and safeguards. 5.2. Burnup correlation

Fig. 2. Linear relationship of 137 Cs activity as a function of burnup for simulated 17 × 17 MOX fuel. Activity was calculated using ORIGEN-ARP with 5.89% initial total Pu enrichment and discharged after three cycles with a 30-day decay period between each cycle.

While it is not ideally suited for burnup determination on its own, 106 Ru can be used in an activity ratio to form a relative measurement almost independent of enrichment and power (Cochran and Tsoulfanidis, 1990c). Due to its short half-life relative to irradiation time, typically 1 year or more for multi-batch cores, its ultimate concentration is heavily dependent on operational history. However, if the 106 Ru activity is normalized with a nuclide, such as 134 Cs, the operational history effect can be mitigated. Eq. (3) provides the gamma activity ratio burnup correlation using the three fission products 106 Ru, 134 Cs, and 137 Cs (Cochran and Tsoulfanidis, 1990c; Simpson et al., 2007). ln(R) = a + b ln(Burnup)

(3)

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

Fig. 4. UO2 activity ratio ln(R) at 30 days after discharge for three different enrichments. Fuel was simulated in ORIGEN-ARP for three 540-day cycles and three 30-day decay periods.

where R = A(

106

Ru)

A(137 Cs) [A(134 Cs)]

2

As mentioned, Eq. (3) is insensitive to initial fuel enrichment and power history making it extremely versatile in measuring almost any fuel provided the assembly was discharged less than 7 years prior (due to 106 Ru half-life limitation). To illustrate this fact, three UO2 fuel enrichments (2.6, 4.2, and 5.0 wt%) were simulated using ORIGEN-ARP. The discharge burnup ranging from 20,000 to 60,000 MWD/MTHM was divided equally over three 540-day irradiation cycles, and two 30-day decay periods were included to emulate an operational shutdown between cycles for refueling. The activity in Becquerels for each of the three isotopes in the ratio was obtained at 30 days after discharge. Fig. 4 shows the correlation trend as a function of increasing burnup. As shown in the figure, correlation between burnup and activity ratio is nearly independent of the initial fuel enrichment. In Fig. 4, the 235 U enrichments were chosen to represent a span of typical PWR fuel enrichments and illustrate the independence of the activity ratio on initial enrichment. Fig. 4 suggests that activity ratio can be used as reliable predictor of burnup for the typical range of PWR enrichments. The ratio is fairly consistent for three enrichments over a range of burnups with a maximum and minimum standard deviation of 0.046 and 0.0041 respectively. To confirm the irradiation history independence, two MOX fuel assemblies with different irradiation histories were simulated in ORIGEN-ARP. The discharge burnup ranged from 20,000 to 60,000 MWD/MTHM, and the irradiation and decay cycles were altered, specifically:

Fig. 5. Two MOX fuel simulations of differing irradiation and decay histories using ORIGEN-ARP attaining same discharge burnup. Both Case I and II used the same Pu isotopic vector and only varied operational history.

tion equipment to resolve the peak from 106 Ru in MOX fuel due to the sizeable difference in 106 Ru fission yield from 235 U vs. 239 Pu. Initial experiments at the Missouri S&T nuclear reactor showed low intensity 106 Ru peaks from comparatively high enrichment (just less than 20%) UO2 used in Missouri S&T reactor fuel. The goal of this effort was to provide proof-of-principle type measurements to demonstrate that the 106 Ru peak is indeed observable even at low burnup for a high enrichment research reactor. These preliminary results may draw attention of those with access to spent MOX fuel to perform measurements on spent MOX fuel. 5.3. Safeguards discriminator Measurements of 106 Ru will show a clear distinction between UO2 and MOX fuel because of its significantly larger fission yield from 239 Pu. Fission of 235 U produces approximately 0.004 106 Ru, while 239 Pu produces 0.043 106 Ru. This over 10-fold increase in production of 106 Ru can facilitate discrimination between 235 U and 239 Pu. 106 Ru is a pure beta emitter (100% yield) decaying to 106 Rh with approximately 10 keV average beta energy. The beta decay scheme of 106 Ru–106 Rh is shown in Fig. 6. 106 Rh in turn decays by a complicated scheme. Most probable energies of 106 Rh are discussed in Section 6, and Table 3. A number of betas

• Case I: 540-day irradiation, 30-day decay, 540-day irradiation, 30-day decay, 540-day irradiation, 30-day final cooling. • Case II: 540-day irradiation, 600-day decay, 540-day irradiation (adjusted for final burnup), and 30-day final cooling. Fig. 5 is a plot of the activity ratio from (3) for Case I and II as a function of burnup for increasing burnup. The ratio R has a maximum and minimum standard deviation of 0.47 and 0.058 respectively, showing excellent agreement between two entirely different operational fuel cycles. This result demonstrates that the activity ratio as a burnup indicator is independent of burnup history. Figs. 4 and 5 illustrate the NDA burnup correlation from (3) works well for both MOX and UO2 irrespective of initial enrichment or operational history but also provides a unique opportunity to verify the fuel composition. In practice it may be easier using detec-

3691

Fig. 6.

106

Ru–106 Rh decay scheme.

Table 3 Photon emission energies and emission probability (Normand, 2008). Isotope

Energy (keV)

Emission probability

106

Ru

511.84 621.94

0.204 0.09935

134

Cs

604.69 795.84

0.976 0.8552

137

Csa

661.657

0.9007

a

a

Gamma energies for and 137m Ba respectively.

106

Ru and

137

Cs originate with the daughter products 106 Rh

3692

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

unknown fuel assembly is MOX or UO2 and the plutonium contents of the fuel at the time of shutdown. 6. Missouri S&T reactor acquisition experiment 6.1. Experimental setup

Fig. 7. ORIGEN-ARP emission spectra for MOX and UO2 simulated fuel assemblies. Circled regions indicate 106 Ru presence.

and gammas are emitted from 106 Rh where prominent gammas include one at 0.5119 MeV with a yield of 0.204 and another one at 0.6219 MeV with a relatively smaller yield of 0.0993 (KAERI, 2000). Because of this widely different fission yield for 239 Pu and 235 U, comparison of MOX and UO2 fuel assemblies will show a prominent 106 Ru peak for MOX fuel. To validate this difference, two samples were created in ORIGEN-ARP for a 17 × 17 nuclear fuel assembly, one for UO2 and one for MOX. Each fuel was simulated for a one-cycle discharge burnup of 16,438 MWD/MTHM after 540 days operation and 1-day cooling time. Fig. 7 illustrates the 511 and 622 keV gamma from 106 Ru–106 Rh and indicates that MOX fuel will yield a more sizeable activity compared to UO2 . Given this detectable difference in 106 Ru fission product activity and the enrichment and operational independence of the ratio R from (3), NDA measurements must show a difference between MOX and UO2 over the burnup values of 20,000–60,000 MWD/MTHM. Fig. 8 combines MOX values from Fig. 5 and the 4.2 w/o UO2 simulations from Fig. 4 to show that there is a detectable difference in the two fuel types given the same operational history using Case I. In Fig. 8, there is a clear separation between the activity ratios for MOX and UO2 where the MOX ratio is on the average a factor of 2.3 greater than the UO2 ratio. Fitting a trend line to the MOX activity ratio in Fig. 8 (Eq. (4)) resulted in a good linear fit with a (correlation coefficient) R2 = 0.9995. ln(R) = −1.71 + 19.65 ln(Burnup)

In order to ascertain the feasibility of acquiring a gamma spectrum from operational or spent nuclear fuel, specifically the detectability of 106 Ru, NDA was performed using a high pure germanium (HPGe) detector at the Missouri S&T nuclear reactor beam port. Three parts of the experimental setup will be highlighted: the Missouri S&T beam port facility, MTR reactor fuel, and the HPGe detector setup. The beam port provides a link between the reactor pool through the pool wall to the dry side sub-level basement where the detection equipment was setup. The beam port is made up of aluminum and closed at its end in the reactor pool 0.1238 m from the closest fuel element F14. The Missouri S&T reactor fuel grid is shown in Fig. 9 which also identifies fuel element F14 as the location of beam port termination. Fig. 10 shows a drawing of the beam port and associated dimensions as well as the graduated nature of the tube as it passes through the pool wall. While the beam port itself measures 15.24 cm in diameter when the shutter assembly is removed, the actual collimated beam passing through the fully assembled beam

(4)

Therefore, using the purposed 106 Ru NDA ratio, MOX fuel could be discriminated from UO2 . Of course, the practical application would involve setting a baseline using a few MOX and UO2 reference assemblies with known operational histories. Once this baseline information is obtained, it could be discerned if an

Fig. 8. The calculated ratio R for 4.2 w/o MOX and UO2 using ORIGEN-ARP at 30 days after discharge and three 540-day power cycles.

Fig. 9. Picture of the Missouri S&T reactor grid plate. Rabbits indicate experimental facilities, an “F” above the number indicates a fuel assembly, a “C” above the number indicates a control rod, and “S” represents a source holder for reactor start-up. The beam port used in these experiments was located near F14.

Fig. 10. Missouri S&T nuclear reactor beam port experimental facility. Schematic shows included shutter assembly which allows for a remotely controlled collimated neutron or gamma beam (Straka, 1998).

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

3693

The attenuating mediums between the reactor core F14 fuel element and the HPGe detector are 0.1238 m of water, 1 cm (∼0.38 in.) of 1100-H14 Aluminum, and 6.477 m of air (Schroer and King, 2008). Lead bricks are expected to collimate the beam and with 127 cm thickness, all scattered radiation will be removed from the beam allowing only the incident pin-hole beam to pass through water, aluminum, and air to interact with the detector and be recorded. An ORTEC® detector (Model GEM15P4) with 18.4% relative efficiency was used in conjunction with the associated data acquisition system and Genie 2000 ©Gamma Acquisition and Analysis software (Catalog Information, 2008). 6.2. Experimental procedure

Fig. 11. Experimental detector arrangement with lead bricks providing additional beam collimation and attenuation in front of the HPGe detector.

port and shutter assembly is rectangular with a cross-section of 7.6 cm × 5.1 cm or 3 in. × 2 in. (Straka, 1998). During normal operation, the beam port contains a 5.715 cm lead shielding plug to block unwanted gamma radiation. This shielding plug is located approximately 1.756 m from the reactor core assembly F14 and was removed during experimentation to allow gamma measurements. The reactor core is moveable within the pool walls. To maximize the gamma flux at the detector location the core was positioned closest to the beam port and locked at the distance stated. Normally in nuclear fuel NDA, the fuel assembly under investigation would be removed from the fuel pool rack or reactor core and placed near a collimator or underwater counting system to segregate any potential sources of background radiation or radiation from nearby fuel assemblies. However, for our purposes, we were only concerned with obtaining a gamma spectrum from operational nuclear fuel as proof of concept of the 106 Ru-based U/Pu discrimination scheme. Thus, no elements were removed from the reactor core, and all measurements taken were bulk measurements of the core with the most gamma contribution coming from the closest fuel element F14 with its edge facing the 45◦ face of the beam port. With regard to the individual fuel elements, the MTR fuel used at the Missouri S&T reactor is a substantially different fuel geometry than the PWR fuel discussed in Section 2.1. Nevertheless, commercial fuel assemblies were not available for experimentation and the same basic NDA theory still applies regardless of fuel geometry. The reactor core was refueled with low enrichment (∼19.9% 235 U) fuel in 1992. The fuel assemblies are approximately 7.6 cm × 7.6 cm square with an active fuel length of 60.9 cm (Straka, 1998). The detection equipment was setup in the reactor building sublevel basement positioned approximately 3.25 m from the beam port opening. This distance was originally chosen to minimize expected detector deadtime due to the intense radiation from the fuel. When an 18% efficiency HPGe detector was directly exposed to the gamma flux after shutdown the detector was completely paralyzed (∼100% deadtime). In the absence of a well-designed collimator (internally sleeved with graded shielding), a simple collimator was constructed with lead bricks. A lead brick with a pin-hole of approximately 0.318 cm was used to reduce the photons impinging upon the HPGe detector. The beam port opening, lead brick pin-hole and detector tube were aligned using a laser inserted at the beam port pool wall. The lead brick shielding was placed 2.3 m in front of the detector to minimize the amount of stray/scattered photons reaching the detector. This arrangement with lead bricks and detector is shown in Fig. 11.

Since the Missouri S&T reactor is a university research reactor primarily used for reactor operator training and various educational experiments, the reactor power level fluctuates substantially during a typical day’s operation. In fact, within a given day, the reactor rarely operates more than a few hours at its maximum power level of 200 kW. Instead, core power is extremely transient, rarely staying at a constant power. This operational characteristic is quite different from commercial power plants which strive to achieve a constant operating power over the course of 18 months. Therefore, the authors did not feel it was prudent to comment or calculate the burnup of the Missouri S&T reactor for this proof-ofprinciple type experiment. The Missouri S&T reactor is only being used as surrogate to determine if irradiated fuel measurements were feasible and the peak from 106 Ru was observable. Additionally, to date, no investigations have been performed to determine the exact burnup and associated buildup in the Missouri S&T reactor since it first went critical with the LEU core in July 1992 (Bonzer, 2008). As stated previously, initial experiments using an unshielded detector produced overwhelming count rates which produced unfavorable deadtime. Therefore, subsequent counting experiments included lead shielding producing a 0.318 cm diameter gamma beam striking the detector surface. Prior to counting, energy calibration for the detector was performed using various test sources. Data for irradiated fuel gamma spectroscopy was collected after high power runs just before the end of the day and data acquisition occurred during reactor non-operational hours, either in the evening or during the weekend. Therefore, there was little time to allow for decay of short-lived fission or activation products. As suggested by the reactor staff (Bonzer, 2008), gamma spectrum measurements were not conducted until at least an hour after reactor shutdown to allow for delayed neutron decay. This ensured no neutrons streamed through the beam port and the incident beam would consist of purely gamma radiation from the fuel. This precaution was necessary to avoid neutron activation and contamination of the detector. Moreover, the benefit of semi-immediate data acquisition following operation is that this scenario appropriately emulates the proposed “online” monitoring of the spent fuel. While this NDA technique would be equally qualified for spent nuclear fuel residing in cooling storage, its primary thrust is in the operational refueling periods between cycles lasting approximately 30 days or less depending on outage operations. Prior to gamma spectrum acquisition, the beam port was opened and health physics measurements were conducted at various locations along the beam. Measurements were conducted using an Eberline Model PIC-6B portable ion chamber. Dose rates at the beam port opening and in front of the HPGe detector were measured to be 10 mrem/h and 0 mrem/h respectively. Since gamma spectrum acquisition was conducted using acquisition software via remote control, spectrums were acquired over

3694

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

the course of 2 days with each spectrum live count time set to 1 h yielding an average MCA deadtime of 0.15%.

count rate electronics, and high efficiency detector it might be possible to reliably measure 106 Ru. Moreover, the MTR fuel under observation is UO2 , and thus the 106 Ru fission yield is substantially less than MOX. Therefore, for MOX fuel one would expect much more significant 106 Ru peaks. Finally, implementation of a Compton-suppression spectrometer could reduce the undesirably high Compton continuum. The new detector setup would consist of an HPGe detector surrounded with a NaI detector operated in anticoincidence. In this instance, Compton interactions could be subtracted from the spectrum (Cochran and Tsoulfanidis, 1990d). It is evident that the 137 Cs activity correlation purposed in Section 4.2 would work well for MST reactor fuel burnup provided the above discussed improvements are made in the detection system. The experiment presented here indicates that near online measurement of irradiated fuel is entirely feasible, and the detectability of 106 Ru in a non-MOX fuel which has been lightly irradiated (significantly less than spent fuel from commercial reactors) is possible. Detectability of these peaks can significantly improve for low enrichment uranium (commercial grade 3–5% 235 U) and MOX fuel.

6.3. Results

7. Discussion and conclusion

The thrust of this experimentation was to determine if the isotopes 106 Ru, 134 Cs, and 137 Cs could be discerned from the intense radiation spectrum emanating from operational/spent nuclear fuel. Detector efficiency calibration with this geometry was not feasible as no calibration source simulating the measurement geometry was available. Moreover, for this proof-of-concept experiment the efficiency calibration was unnecessary. Therefore, none were performed and as stated no burnup calculation of Missouri S&T reactor fuel was attempted. The gamma energies for the three isotopes and corresponding branching ratios are presented in Table 3. Five measurements were conducted following each of the two irradiation/high power run, totaling 10 measurements. Two measurements closely following reactor shutdown are presented here for Day 1 and 2 observations. Fig. 12 presents the gamma spectrum 3 h and 19 min following reactor shutdown for Day 1. Fig. 13 presents an almost identical spectrum from Day 2, 5 h following reactor shutdown. These two figures are chosen to illustrate feasibility of the technique. It is obvious from the figures that almost immediately following reactor shutdown, 106 Ru, 134 Cs and 137 Cs peaks are distinguishable in the spectrum. Many other peaks are also visible within the 400–750 keV energy range. As seen from Figs. 12 and 13, the 134 Cs and 137 Cs peaks are prominent and easily discernable from the intense Compton continuum. But the 106 Ru is weak though still distinguishable from background suggesting that even for comparatively high enrichment uranium based fuel (∼20% for the Missouri S&T reactor) with an appropriately designed collimator, suitable choice of high

As closure of the U.S. nuclear fuel cycle is becoming imminent, there is a pressing need to develop tools for reliable monitoring of spent fuel. In particular, non-destructive interrogation by gamma spectroscopy appears to be an effective means to monitor special nuclear materials. Simulation results presented here suggest that 137 Cs and 106 Ru peaks can both be utilized independently to estimate burnup of MOX fuel. Moreover, 106 Ru peak can also be used in conjunction with cesium peaks to form an activity ratio as described in Eq. (3). For several applications including special nuclear material safeguards, it may also be needed to use gamma spectroscopy data to predict isotopic content of the spent fuel. Other attempts to infer isotopic composition of the spent MOX fuel are reported in the recent literature (Willman et al., 2006). However, there are challenges associated with using 154 Eu peak(s) for this application. This manuscript is not meant to offer a universal solution for all spent fuel interrogation problems. The purpose of this manuscript is to draw attention to the possibility of using 106 Ru ratio as an alternative. This ratio is particularly effective not only for predicting burnup but also for estimating fissile material isotopic contents of the spent fuel. Results of the simulation study demonstrated that this ratio is independent of the burnup and/or burnup history. Therefore, this ratio can reliably predict the plutonium content of the spent fuel for various burnup scenarios. The capability of predicting plutonium content of the spent fuel is very useful for proliferation resistance technology and special nuclear material safeguard and accountancy. We believe that 106 Ru peak will be more effective for the application because of the following reasons:

Fig. 12. Gamma spectrum from Missouri S&T reactor NDA experiments Day 1. Spectrum was obtained approximately 3 h after reactor shutdown.

Fig. 13. Gamma spectrum from Missouri S&T reactor NDA experiments Day 2. Spectrum was obtained approximately 5 h after reactor shutdown.

(1) Half-life of 106 Ru (1.02 years) is much closer to 134 Cs half-life (2.07 years) as compared to the half-life of 154 Eu (8.6 years). Therefore, if there is any uncertainty in the cooling time (misdeclaration or otherwise) of the spent fuel, the 106 Ru-based technique will have smaller error than 154 Eu. (2) Secondly, production of 154 Eu in a commercial reactor is rather complex involving beta decay and neutron capture from five different mass chains. Moreover, 154 Eu has a rather large thermal neutron cross-section (1854 barn). This would invariable lead to strong burnup history dependence and therefore uncertainty and poor reliability in the results. (3) Thirdly, the fission yield ratio of 106 RhU-235 /106 RhPu-239 is approximately constant for thermal and fast neutrons, which means it will be a more reliable isotope for BWR applications.

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

3695

Table 4 Potential isotopes masking 106 Ru 511 keV spectrum peak (Normand, 2008). Nuclides

Energy (keV)

Emission probability

Half-life

Fission yield – 235 U

Fission yield – 239 Pu

11 Na 22 13 Al 26 23 V 48 38 Sr 85 47 Ag 106m 33 As 74 37 Rb 84 37 Rb 83 27 Co 56 77 Ir 190 27 Co 58 45 Rh 102m 52 Te 121 61 Pm 148m 63 Eu 150 30 Zn 65 64 Gd 149 53 I 126 77 Ir 190 63 Eu 150 55 Cs 136 63 Eu 148

511.00 511.00 511.00 514.01 511.85 511.00 511.00 520.48 511.00 518.50 511.00 511.00 507.60 501.30 505.60 511.00 516.55 511.00 502.50 515.80 507.20 501.31

1.80E+00 1.63E+00 1.01E+00 9.93E−01 8.77E−01 5.90E−01 4.78E−01 4.52E−01 3.91E−01 3.41E−01 3.00E−01 2.80E−01 1.77E−01 6.84E−02 4.94E−02 2.84E−02 2.69E−02 2.44E−02 1.13E−02 1.08E−02 9.90E−03 9.71E−03

2.6027 (±10) years 717.0 (±240) ×103 years 15.974 (±3) days 64.849 (±4) days 8.46 (±10) days 17.78 (±3) days 33.5 (±6) days 86.2 (±1) days 77.31 (±19) days 12.0 (±2) days 70.86 (±7) days 208 (±4) days 19.16 (±5) days 41.05 (±14) days 36.4 (±7) years 244.15 (±9) days 9.28 (±10) days 12.98 (±5) days 12.0 (±2) days 36.4 (±7) years 13.03 (±7) days 54.5 (±5) days

0 0 0 1.32E−12 0 0 4.46E−10 4.15E−12 0 0 0 0 0 1.04E−09 0 0 0 3.80E−10 0

0 0 0 4.09E−12 0 3.30E−11 1.31E−08 2.58E−10 0 0 0 1.81E−11 1.97E−12 1.18E−07 0 0 0 9.58E−09 0

However, there are practical challenges associated with the measurement of these gamma emission peaks. The two photon energies relied on for the detection of 106 Ru (511.84 and 621.94 keV) present a challenge with the most probable peak at 511 keV. A photon peak at 511 keV is possible from a number of other isotopes and also possibly from positron–negatron annihilation. Table 4 illustrates an abbreviated list of potential isotopes which could contribute a peak at 511 ± 10 keV in the spent fuel spectrum. Optimum measurement time post irradiation has to be carefully determined to maximize detectability of 106 Ru by suppressing interference from other sources. Interference from any isotopes with half-life less than 1 week can easily be avoided by collecting the gamma spectrum after sufficient cooling time. Additionally, isotopes with small fission yield and/or low emission probability of 511 keV are not likely to cause any major interference. Therefore, the list in Table 4 only includes those isotopes with an emission probability greater than 0.1. As one may notice, all of these isotopes have very small thermal neutron induced fission yield (between zero and 1.18E−07) as compared to fission yield of 106 Rh (4.19E−02 from 239 Pu and 4.10E−03 from 235 U). Moreover, the half lives of these isotopes are very small in the range of few microseconds to a couple of days as compared to 106 Ru which has a half-life of over 1 year. Nevertheless, these measurement challenges must be addressed and mitigation strategies be implemented for the proposed NDA to be effective for spent fuel monitoring. In practice, once the experimenter obtains a gamma spectrum for nuclear fuel NDA, some level of post-processing would be necessary to eliminate other isotopes contributing to the gamma spectrum obtained. Further work is required to develop these interference mitigation strategies. The purpose of this paper is to draw attention to the feasibility of using 106 Ru as a burnup indicator as well as MOX fuel discriminator. This technique by no means offers a universal solution to the gamma spectroscopy problem. In fact, multiple methods would be required for various applications. For example, for aged spent fuel (20 years or older) perhaps 154 Eu may remain the best option because of its relatively long half-life. However, for online or shorter cooling time applications 106 Ru may offer a more robust solution than 154 Eu.

Acknowledgments The authors would like to thank Missouri S&T reactor staff (Mr. William Bonzer, Brain Porter, and the Late Mr. Daniel Estel who helped us in various ways to accomplish this project) for their invaluable support during experiments, as well as their guidance and expertise on facility specifications. Also, many thanks to the Department of Energy Advanced Fuel Cycle Initiative and the WTAMU University Research Alliance for funding this work. References Bonzer, B., 2008, January. University of Missouri-Rolla Research Reactor, Rolla, MO, private communication. Catalog Information, 2008. GEM high-purity germanium (HPGe) coaxial detectors (in PopTop Capsule) [Online]. Available: http://www.orteconline.com/detectors/photon/b2 1.htm. Cochran, R.G., Tsoulfanidis, N., 1990a. The Nuclear Fuel Cycle: Analysis and Management, 2nd ed. ANS, La Grange Park, IL, pp. 100–101. Cochran, R.G., Tsoulfanidis, N., 1990b. The Nuclear Fuel Cycle: Analysis and Management, 2nd ed. ANS, La Grange Park, IL, pp. 227–325. Cochran, R.G., Tsoulfanidis, N., 1990c. The Nuclear Fuel Cycle: Analysis and Management, 2nd ed. ANS, La Grange Park, IL, p. 195. Cochran, R.G., Tsoulfanidis, N., 1990d. The Nuclear Fuel Cycle: Analysis and Management, 2nd ed. ANS, La Grange Park, IL, pp. 382–389. Crane, T.W., Hsue, S.T., Lee, J.C., Talbert Jr., W.L., 1978. Nondestructive assay methods for irradiated nuclear fuels. Los Alamos National Lab., Los Alamos, NM, LA-6923, pp. 2–8. Dennis, M.L., Usman, S., 2006. Feasibility study of MOX fuel online burnup analysis. In: 2006 Proc. ICAPP Conf., Reno, NV, USA, June 4–8, 2006 (Paper 6417). Gauld, I.C., Chare, P., Clarke, R.C., 2003, June 19. Development of ORIGEN-ARP method and data for LEU and MOX safeguards applications [Online]. Available: http://www.ornl.gov/∼webworks/cppr/y2001/pres/117710.pdf. Gauld, I.C., et al., 2005. ORIGEN-ARP: automatic rapid processing for spent fuel depletion, decay and source term analysis. Oak Ridge National Lab., Oak Ridge, TN, ORNL/TM-2005/39, Vol. I, D1.4.16. Korea Atomic Energy Research Institute. Nuclear Data Evaluation Lab, 2000. Table of nuclides [Online]. Available: http://atom.kaeri.re.kr. Normand, C., 2008, March 30. Karlsruhe Nuclide Chart Database. Nucleonica [Online]. Available: http://www.nucleonica.net/. Patil, A., Usman, S., 2009. Measurement and application of paralysis factor for improved detector dead-time characterization. Nuclear Technology 165 (February (2)), 249–256. Reilly, D., Ensslin, N., Smith, H. (Eds.), 1991. Passive nondestructive assay of nuclear materials. Los Alamo National Lab., Los Alamos, NM, NUREG/CR-5550, pp. 529–557. Sasahara, A., Matsumura, T., Nicolaou, G., Papaioannou, D., 2004. Neutron and gamma ray source evaluation of LWR high burn-up UO2 and MOX spent fuels. Journal of Nuclear Science and Technology 41 (April (4)), 448–456.

3696

M.L. Dennis, S. Usman / Nuclear Engineering and Design 240 (2010) 3687–3696

Schroer S., King J., 2008, April. Missouri Univerity of Science and Technology, Rolla, MO, Private Communication. Shaw Areva MOX Services, LLC, 2008, September 27. MOX Project [Online]. Available: http://www.moxproject.com/. Simpson, A., Marlow, T., Franco, J., Clapham, M., Chesterman, A., 2007. Operational experience in radiometric instrumentation for spent fuel monitoring [Online]. Available: http://www.pajaritoscientific.com/pdf/Final INMM 03 Spent Fuel Monitoring.pdf.

Straka, M., 1998. Safety analysis report for the University of Missouri-Rolla Reactor with low enriched uranium fuel. License R-79, Docket 50-123, University of Missouri-Rolla, Rolla, MO, Ver. 1, Sect. 4, pp. 2–5. Willman, C., Håkansson, A., Osifo, O., Bäcklin, A., 2006. A nondestructive method for discriminating MOX fuel from LEU fuel for safeguards purposes. Annals of Nuclear Energy 33, 766–773.