Comparison of Gamma and Passive Neutron Non-Destructive Assay Total Measurement Uncertainty Using the High Efficiency Neutron Counter

Transcription

1 Coparison o Gaa and Passive Neutron Non-Destructive Assay Total Measureent Uncertainty Using the High Eiciency Neutron Counter John M. Veilleux Los Alaos National Laboratory Los Alaos, NM 87545, USA 505/ Eail: veilleux@lanl.gov Doug Craer Canberra Industries Los Alaos National Laboratory Los Alaos, NM 87545, USA 505/ Eail: dcraer@canberra.co ABSTRACT Since 1997, the High Eiciency Neutron Counter (HENC) has been eployed by the Los Alaos National Laboratory (LANL) to assay and ship to the Waste Isolation Pilot Plant (WIPP) plutoniu waste using passive neutron ethods. During 2004 and early 2005, the HENC has undergone a odiication and recertiication to add a Canberra broad energy geraniu detector (BEGe), DSA digital spectru analyzer, and JSR-14 neutron analysis shit register so as to conduct siultaneous gaa and passive neutron easureents. The BEGe detector provides both isotopic and quantitative easureent data to copleent the neutron detectors. The assay is under the control o Canberra s NDA-2000 sotware suite that provides acquisition control and analysis. As part o the recertiication eort, the total easureent uncertainty (TMU) is being evaluated. A coparison o the gaa and the passive neutron TMU will be provided and the NDA 2000 settings used to achieve the results will be suarized and reported. INTRODUCTION The High Eiciency Neutron Counter (HENC) has been eployed by the Los Alaos National Laboratory (LANL) to assay plutoniu waste using passive neutron ethods since During 2004 and early 2005, the HENC has undergone a odiication and recertiication to add a Canberra broad energy geraniu detector (BEGe), DSA-1000 digital spectru analyzer, and JSR-14 neutron analysis shit register so as to conduct siultaneous gaa and passive neutron easureents. The BEGe detector provides both isotopic and quantitative easureent data to copleent the neutron detectors. The assay is under the control o Canberra s NDA-2000 sotware suite that provides acquisition control and analysis. As part o the recertiication eort, the total easureent uncertainty (TMU) is being re-evaluated. Unortunately, the easureents to support a ull TMU deterination could not be perored so, in lieu, data ro a siilar instruent, the MCS HENC, was used to evaluate the eect o the changes in operating ethodology 1 LANL intends to use. The ocus o this study is to evaluate how the gaa and neutron easureents, and the associated TMU, should be used or the LANL HENC in light o the operational experience o the MCS HENC. There are signiicant dierences in how the neutron data is to be used, utilizing only the add-a-source corrected doubles coincidences and neutron Ones rates. Further, while the MCS 1

2 HENC utilizes a second order calibration curve, the LANL HENC uses a linear calibration constant. These are the changes with which the MCS data are evaluated, and care should be exercised in extending these results to MCS Henc operations. METHODOLOGY 1. NDA 2000 Data ro HENC Measureent o Standards The MCS HENC has been acquiring assay data at Los Alaos since late The database or this study ranges ro data acquired beginning in January 2003 through April This data consists o assays acquired with National Institute o Standards and Technology (NIST) traceable standards in three atrices: epty dru, debris dru, and inorganic solid (sludge) dru. Source locations were varied and provide a reasonable estiate o location variation or the TMU estiate. The data o interest or this study are the neutron and gaa assay results. The neutron data acquired include the doubles coincidence rate, the Ones rate, the background doubles and ones rate, the Add-a-Source (AAS) correction actor, and the standard deviations o each o the above values. For the gaa acquisition, the Pu239 ass, Pu239 activity, Pu239 iniu detectable activity (MDA), and the related standard deviations were evaluated. 2. Total Measureent Uncertainty The TMU developed or the MCS HENC 2 is suarized in Table 1. The individual coponents o the TMU are shown expressed as relative standard deviation percentages (%RSD). A subtotal %RSD was added because the neutron assay requires an additional isotopic easureent (usually ro a gaa detector) to obtain the total Pu ass. The gaa detector in the HENC consists o an energy eiciency calibration and thus a separate isotopic easureent is not required. In the analysis to ollow, the isotopic uncertainty is not propagated since true source isotopic values are used. Consequently, the coparison is to the values identiied or the Subtotal in Table 1. Table 1. Total Measureent Uncertainty Neutron Detector Gaa Detector Debris Sludge Debris Sludge Counting Statistics 1% 1% 0.8% 1.1% AAS Correction/Matrix Non-Hoogeneity 4% 5% 5% 5% Source Distribution 2.9% 17% 9% 25% Multiplication Eects 0% 0% N/A N/A Background Eects 0% 0% N/A N/A Calibration Uncertainty 2% 2% 2.5% 2.5% SUBTOTAL 5% 18% 11% 26% Isotopic Ratio Uncertainty 12% 12% N/A N/A Noinal TMU (1σ) 13.2% 21.5% 10.6% 25.6% 3. Standards The standards used or this study are suarized in Table 2. All the sources are traceable to the NIST. The sources are either PDP 3 sources or are designed just like the PDP sources. The data 2

3 needed or this study include the total Pu ass o the source, the Pu240 eective ass raction, and the Pu239 ass raction. The source nae is included with the table. The ass raction o Pu240 eective ass to total Pu,, is deined in ters o the ass ractions o Pu238, Pu240, and Pu242 by the ollowing equation 4 : = 2.52 Pu Pu Pu242 ( 1) Table 2. Standards Noinal Source Strength Mass Pu (g) Pu240 Eective Mass Fraction Pu239 Mass Fraction Source Nae PDP NTP-0001, 0064, 0071, PDP NTP-0085, PDP1-0.1, PDP NTP PDP NTP NTP0064, 0071, 0099, PDP PDP1-3.0, PDP NTP RANT NTP RANT NTP0140, 0148, RANT25-1, 50-1, RANT50-1, 50-2, NTP0140, 0148, 0156, RANT25-1, 50-1, 50-2, Matrix Drus and Source Positioning The atrix drus are those built to PDP 5,6 speciications. The atrix drus contain tubes peritting one to place one or ore standards at varying positions in the dru. The epty and debris drus each contain three tubes, one along the central axis, one displaced 5.5 ro the center, and the third tube displaced 9 ro the center in which standards ay be placed. The height o the source can be varied ro the botto to the top o the dru. For the sludge dru, our tubes are available starting at the center with the reaining tubes located 3.5, 5.75, and 9 respectively ro the center. The requency o source placeent or the given HENC assay data is tabulated in Figure 1. As shown, there is a reasonably good distribution o sources or the epty and debris drus, but the source positions are ostly centered in tubes T3 and T4 or the sludge dru. 3

4 T1 T2 T3 T1 T2 T3 T Debris & Epty Dru Sludge Dru Figure 1. Source Location Frequency in 55 gal PDP Type Drus 5. Operations The operating conditions include a 4 Pi neutron detector with 113 neutron detectors covering the dru. The gaa detector is a single broad energy geraniu detector positioned approxiately at the idpoint o the dru. During the count, the dru is rotated and data is acquired by both the neutron and the gaa electronics. A C252 add-a-source correction is ade to the neutron assay to correct or atrix eects. During the tie the C252 source is exposed, the gaa acquisition is inhibited. The total assay tie or the gaa and neutron easureents is 30 inutes. 6. Calculation o Pu Mass or the Neutron Assay The plutoniu ass estiated ro an assay depends on whether neutron or gaa data is to be utilized. For the purpose o this study, ass estiates ro both the neutron and gaa easureents are evaluated. For the neutron assay, either the net neutron doubles rates or the ONES rates is selected depending on the cuto value. The neutron doubles rate 1 is the nuber o neutron coincidences easured over the length o the assay divided by the assay tie. The coincidences include all neutron ultiplicities ro 0 to 255. The neutrons ONES rate excludes ultiplicities higher than two to iniize the eect o cosic radiation and are thereore ore suited when the Pu ass is sall. The Neutron Doubles cuto and the Ratio o Doubles to ONES rate was originally calculated or the LANL HENC and reported at the last INMM annual eeting. The MCS HENC utilizes a calibration o the ONES rate rather than a ratio o the doubles to ONES rates that is planned or the LANL HENC. However, or the purpose o this study, the ratio ust be recalculated since it will 1 The singles rate is actually the su o all shit register triggers ro a neutron event divided by the count tie. The doubles rate is the singles rate ties the ean o the reals plus accidental neutron distribution inus the ean o the accidental neutron distribution. See reerence 4 or a coplete description. 4

5 dier ro one instruent to the next. The doubles cuto or the MCS HENC was re-evaluated based on the neutron data or the epty dru at 3 net doubles coincidences and is coparable to approxiately 1 gra (See Figure 2). Thus, at a net doubles below 3 n/s, the ONES rate is used while above that value, the Doubles rate is used. Doubles Error (%RSD) MCS HENC Doubles to Ones Cuto Cuto ~ 3 Net Doubles ~1g Pu Net Doubles Coincidences (n/s) Figure 2. Neutron Doubles to Ones Cuto The ratio o the Doubles to Ones rate is deterined ro the Doubles and Ones calibration actor. The Doubles calibration or the MCS HENC is ± 0.17 net doubles per gra o Pu240 eective while the Ones calibration is ± net doubles per gra o Pu240 eective. Consequently, the ratio o Doubles to Ones is ± This actor eectively converts the Ones rate to a Doubles rate peritting the use o the doubles calibration with the Ones rate. The ass o Pu (in gras) is given by the ollowing equations. Here, = Add-a-Source (AAS) correction actor, k AAS k DO K D N D N O = Doubles to Ones Ratio, = Doubles calibration actor, = Doubles coincidence rate (n/s), = Ones coincidence rate (n/s). Pu Pu AAS D =, or N D > 3 K k k D D N AAS DO O =, or N D 3 K k N ( 2) ( 3) The lower liit o detection in Pu ass units is calculated ro the standard deviations o the daily background and the assay as shown below. These relations are taken ro Currie 7 and Makaruk 8 and represent the error at the 95% conidence level with type I and II errors both set at 5%. Here, the ollowing sybols are used: σ BD = Standard deviation o the background doubles rate σ = Standard deviation o the background ones rate BO σ = Standard deviation o the assay doubles rate SD σ = Standard deviation o the assay ones rate SO 5

6 LLD LLD 1.65k ( σ + σ AAS BD SD =, or N D > 3 K D 1.65k AAS k DO ( σ BO + σ SO ) =, or N D 3 K D ) ( 4) ( 5) 7. Gaa Data The gaa assay Pu239 ass is taken directly ro the NDA 2000 acquisition and converted to total Pu ass ro the known Pu239 ass raction o the standard. The LLD is taken ro the NDA 2000 reported Pu239 iniu detectable activity (MDA) and converted to total Pu ass by applying the activity per unit ass o Pu239 ( Ci/g). Pu = = LLD Pu239 Pu239 MDA Pu Pu239 ( 6) ( 7) ASSAY RESULTS The results o these assays are suarized in Table 3. There are a total o 234 assays in the database. The three atrices are shown along with the true ass range. The total nuber o data points or the neutron (N) and the gaa (G) assay are shown. The nubers o neutron and gaa assays are not always identical because or soe assays either the neutron or gaa assay easureent ailed; in these cases, the neutron or gaa data was reoved ro the analysis. The average ratio o the easured to the true ass (R%), expressed as a percentage is shown next ollowed by the one percent relative standard deviation (%RSD) o the ass points. Finally, the average LLD ass is shown. Table 3. Assay Suary Neutron Gaa Matrix Pu Mass (g) Data Points (N/G) R% %RSD LLD (g) R% %RSD LLD (g) Epty / % 14.6% % 11.5% / % 1.3% % 9.5% / % 4.2% % 23.6% Overall 234/ % 15.6% % 15.6% Debris / % 12.7% % 16.7% / % 5.1% % 13.9% / % 11.6% % 29.0% Overall 106/ % 11.6% % 22.2% Sludge / % 29.2% % 37.2% All Matrices / % 17.8% % 25.7%

7 / % 11.6% % 33.5% / % 16.0% % 19.6% / % 9.1% % 23.8% Overall 229/ % 13.7% % 28.4% The accuracy or bias (%R) generally avors the neutron assay. The neutron data exhibits a high bias o less than 7.2% or the epty or debris atrix and 11.9% or the sludge atrix. By coparison, the gaa detector exhibits a low bias o less than 14.6% or the epty or debris atrices, and a high bias o 38.9% or the sludge. The data thereore suggests that the neutron data is preerred IF a good isotopic easureent can be obtained ro the gaa detector and IF the assay is above the neutron LLD. The precision (%RSD) is generally better or the neutron detectors than or the gaa detector. For all atrices, the neutron %RSD is 13.7% copared to 28.4% or the gaa. In all cases except or the range 0-1 g Pu, the precision is better or the neutron detectors than the gaa detectors. Even or the sludge atrix, the neutron detectors peror better in both %RSD (29.2 vs 37.2%) and accuracy (111.9 vs 138.9%). These data also suggest that the neutron easureent should be the easureent o choice or all atrices provided a good gaa isotopic easureent can be obtained and the assay is above the LLD. The LLD or the gaa detector is generally better than the neutron LLD as shown in the table. The gaa LLD ranges ro 25 to 82% better than the neutron LLD. This suggests that or sall quantities o Pu (e.g., below 1 g weapons grade Pu or equivalent), the gaa detector will peror better than the neutron detectors. The TMU ust include both the accuracy and precision o the errors added in quadrature. The results o this evaluation are tabulated in Table 4. The accuracy and precision errors are taken ro Table 3 and the isotopic error is taken ro Table 1 without change. The revised TMU is copared with the reported TMU in Table 1. The sludge TMU ay be understated because the source positioning did not cover the entire dru; on the other hand, the basic assuption or sludge is that the nuclear aterial is hoogeneous. The data suggests that the neutron TMU is better (saller) than the gaa TMU. Table 4. Total Measureent Uncertainty Coparison Accuracy 100-%R Precision (%RSD) Isotopic (%) Revised TMU (%) Reported TMU (%) Neutron Debris 7.2% 11.6% 12% 18.2% 13.2% Sludge 11.9% 29.2% 12% 33.7% 21.5% Gaa Debris 5.3% 22.2% NA 22.8% 10.6% Sludge 38.9% 37.2% NA 53.8% 25.6% 7

8 CONCLUSIONS The neutron assay is the preerred assay when a easurable isotopic value is available and the assay is above the systes lower liit o detection. The gaa assay generally produces a lower liit o detection that is 25 to 82% better (lower) than available ro the neutron assay and should be used whenever the neutron assay ails. The total easureent uncertainty (TMU) originally reported should be increased or both the neutron and gaa assays. REFERENCES 1 Veilleux, J.M. and D. Craer, Non-Destructive Assay (NDA) o Nuclear Waste using the High Eiciency Neutron Counter (HENC), LA-UR , 45 th Annual Meeting o the INMM, July 18-22, 2004, Orlando, FL. 2 Ceo, Robert, Total Measureent Uncertainty or the MCS HENC#1 With Integral Gaa Spectroeter, Docuent # CI-HENC-TMU-101, Rev. 2, CCP Central Records, 4021 National Parks Highway, MS: GSA 203, Carlsbad, NM, June 22, DOE/CBFO , Perorance Deonstration Progra Plan or Nondestructive Assay o Drued Wastes or the TRU Waste Characterization Progra, Revision 0.1, National Technical Inoration Service, Mar 22, Ensslin, N., W Harker, M Krick, D Langner, M Pickrell, and J Stewart, Application Guide to Neutron Multiplicity Counting, LA M, Los Alaos National Laboratory, Noveber CSM-DOC-33-03, Los Alaos National Laboratory Environental Technology Group NDA PDP Inorganic Sludge Matrix Dru Design and As-Built Data, Custo Surrogate Matrices, Inc., Idaho Falls, January CSM-DOC-32-03, Los Alaos National Laboratory Environental Technology Group NDA PDP Cobustibles Dru Design and As-Built Data, Custo Surrogate Matrices, Inc., Idaho Falls, January Currie, L.A. Liits or Qualitative Detection and Quantitative Deterination, Analytical Cheistry, 40 (3), p , March Makaruk, H.E. and R.M. Owczarek, Explanation o Matheatical Details o Currie s Methods or Detection Liit Deterination, 45 th Annual Meeting, Proceedings o the Institute o Nuclear Materials Manageent, Orlando, FL., July 19-22,