ML23104A050

DOE-OR-01-2513aD1 Gamma Spec of Salt

ML23104A050
Person / Time
Site:	Abilene Christian University
Issue date:	08/26/2011
From:	Oak Ridge
To:	Office of Nuclear Material Safety and Safeguards, US Dept of Energy, Office of Environmental Management
References
DE-SC-0004645 DOE/OR/01-2513&D1
Download: ML23104A050 (1)

Text

DOE/OR/01-2513&D1

Nondestructive Assay Measurements of Defueled Salts at the Molten Salt Reactor Experiment, Oak Ridge, Tennessee

DOE/OR/01-2513&D1

Nondestructive Assay Measurements of Defueled Salts at the Molten Salt Reactor Experiment Oak Ridge, Tennessee

Date Issued: August 2011

Prepared for the U.S. Department of Energy Office of Environmental Management

URS l CH2M Oak Ridge LLC Managing and Safely Delivering the Department of Energys Vision for the East Tennessee Technology Park Mission under contract DE-SC-0004645

This page intentionally left blank.

CONTENTS

TABLES....................................................................................................................................................... v

FIGURES...................................................................................................................................................... v

APPENDICES.............................................................................................................................................. v

ACRONYMS.............................................................................................................................................. vii

EXECUTIVE

SUMMARY

......................................................................................................................... ix

1. Background............................................................................................................................................ 1

2. Purpose and Scope.................................................................................................................................. 5

3. Approach................................................................................................................................................ 6

4. RESULTS............................................................................................................................................... 6 4.1 GROSS GAMMA PROFILE OF FEBRUARY 17, 2011............................................................. 6 4.2 GAMMA SPECTRUM OF JUNE 9, 2011................................................................................. 15

5. REFERENCES..................................................................................................................................... 18

iv TABLES

Table 1. Characteristics of MSRE Salt Tanks and Contents......................................................................... 1 Table 2. Comparison of Gross Gamma Profiles of 1998 and 2011............................................................. 8 Table 3. Melting Temperatures for Major Fuel Salt Components As Shown in ORNL 2002.................... 12 Table 4. Calculated U Masses in FDT-1 Based on Inventory Records and Engineering Estimates........... 17 Table 5. Comparison of Extent of Defueli ng of MSRE Salt Based on Various Methods.......................... 17

FIGURES

Figure 1. Fuel Drain Tank............................................................................................................................. 3 Figure 2. Sketch of Maintenance Shield....................................................................................................... 7 Figure 3. Workers Assembling Detector Collimation Shielding On the MSRE Maintenance Shield.......... 7 Figure 4. GM Detector in Thermal Well....................................................................................................... 9 Figure 5. Location of Gross Gamma Profile Data Acquisition................................................................... 10 Figure 6. Gross Gamma Detector Positioning through the MSRE Maintenance Shield............................ 10 Figure 7. Gross Gamma Profile Outside FDT-1 Compared to Profile Predicted for Uniform 137Cs.......... 11 Figure 8. Conception of Location of Fuel Salt Ph ases Having Different Melting Temperatures As Shown in ORNL 2002............................................................................................................................................. 13 Figure 9. Positioning the Germanium Crystal Detector on the Lead Collimator........................................ 13 Figure 10. Probe illustrating how fluorine was fed from a single point...................................................... 14 Figure 11. Geometry for NDA of June 9, 2011.......................................................................................... 15 Figure 12. Assembly Ready for Gamma Spectrocopy Through the Maintenance Shield.......................... 16

APPENDICES

A - FTIR BEST EXPONENTIAL CURVE FITS FOR FDT-1 FINAL PPM U

B - EVALUATION OF GAMMA SPECTROSCOPY RESULTS Tables in Appendix B:

Table B-1. Peak Search Results for Spectrum of June 9, 2011.................................................... B-3 Table B-2. Normalized Relative Efficiency for Spectrum of June 9, 2011................................. B-4 Table B-3. Fit to Normalized Efficiency for Spectrum of June 9, 2011...................................... B-6 Table B-4. Decay Parameters for Isotopes of Interest................................................................. B-9 Table B-5. Three Common Constants and Conversions.............................................................. B-9 Table B-6. Calculated Pre-2008-defueling U Masses in FDT-1................................................ B-10 Table B-7. Uranium Remaining in FDT-1 Based on Evaluation Using Gamma Spectroscopy Results........................................................................................................................................ B-11 Figures in Appendix B Figure B-1. Spectrum of June 9, 2011 - C ounts (YAxis) vs. Energy (X Axis)........................... B-2 Figure B-2. Normalized Relative Efficiency vs. Energy............................................................. B-5

v This page intentionally left blank.

vi ACRONYMS

BJC Bechtel Jacobs Company LLC CERCLA Comprehensive Environmental Respon se, Compensation, and Liability Act D&D Decontamination and Decommissioning DOE U.S. Department of Energy DTC Drain Tank Cell FDT Fuel Drain Tank FFT Fuel Flush Tank FTIR Fourier Transform Infrared Spectrometer GM Geiger-Mueller MSRE Molten Salt Reactor Experiment NDA Nondestructive Assay ORIGEN Oak Ridge Isotope Generation ORNL Oak Ridge National Laboratory WAC Waste Acceptance Criteria

vii This page intentionally left blank.

viii EXECUTIVE

SUMMARY

The primary purpose of this NDA was to attempt to provide a more accurate measure of the uranium remaining in the salt tanks. This would support the cu rrent efforts of DOE to declare that the salt was defueled and make WIPP a disposal option. The fuel salt was drained from the reactor loop as a liquid on December 1, 1969 containing 37,557 g U and collected in salt tanks FDT-1 and FDT-2. Of the total mass, 22,916 g U were estimated to have been re moved in 1998 by defueling/reactive gas removal operation. After the 1998 defueling, and estimating the uranium that had migrated out of the salt and deposited in other MSRE systems, an estimated 4999 g U remained in FDT-1. However, the NDA evaluation in this report indicates that figure was more likely to have been about 4,193 g and that the predefueling U in FDT-1 could have been overstated in previous inventory estimates. In 2008, because DOE was committed to defuel the salt to the extent practical, an additional 3449 g U were removed from FDT-1 by final defueling. Inventory records would indicate that 1,548 g U remained in FDT-1 after the final defueling, but the evaluation described in this report based on the gamma spectroscopy measurement indicates only 713 g (280 ppm) are most likely what remains.

The latter result (713 g U or 280 ppm U) is believed to be far more accurate than the inventory records, which means over 98 wt% of the original fuel has now been removed. At 280 ppm uranium the fissile content of the salt, including 650 g of 239Pu, cannot support criticality and the salt should now be considered defueled.

This estimate for the current residual U was calculated to be about 280 ppm in FDT-1 based on 2011 gamma spectroscopy measurements in conjunction with the estimate of the total U removed in 2008. The 2008 U transferred from MSRE is the most accurate of the historic quantities known because that quantity was based on careful measurement and tracking involv ing sensitive direct mass measurements, FTIR, and gamma spectroscopy measurements together. This result is compared to other approaches in Table 5 of this report. This result is considered representative of all of the drain tank salt.

While the gamma spectroscopy result cannot identify a specific concentration value with high precision, by validating the existing FTIR results it greatly supports the cumulative body of evidence that the salt has been defueled and is now characterized adequately for disposal. The fissile content can not support criticality and this makes the salt more acceptable for permanent disposal.

A secondary purpose of the NDA was to study the distribution of 137Cs in the salt to aid in the design of storage and transportation containers. The NDA wo rk showed that the dominant gamma source, 137Cs, is not uniformly deposited throughout the salt and do es not tend to uniformly deposit based on the distributions remaining from 1969 and 2008 freezings. 137Cs was more concentrated toward the bottom of the FDT-1 salt in 1998 and near the top of FDT-1 sa lt and to a lesser extent near the bottom in 2011.

Greater detail for transport packaging shielding d esign and process design considerations is provided regarding the degree of nonhomogeneity in 137Cs later in this report. Note: the gamma profile does not call into question the adequacy of the gamma spectroscopy for residual uranium determination because the gamma level where the gamma spectroscopy measurement was focused appears to fairly represent the average gamma for the entire tank. A conservative location for sampling the very highest levels of 137Cs would be about six inches deep in the salt as it is currently frozen.

ix This page intentionally left blank.

x

1. BACKGROUND

The Molten Salt Reactor Experiment (MSRE) site is located at Oak Ridge National Laboratory (ORNL),

a U.S. Department of Energy (DOE) facility in Oak Ridge, Tennessee. The MSRE facility included a graphite-moderated, liquid-fueled molten salt reactor operated from June 1965 through December 1969.

The reactor employed a liquid molten fuel salt composed of UF 4 as the fuel. Pu was primarily produced in the salt during 235U/238U operations, but a smaller amount of Pu, as 239PuF3, was added as part of an experiment, all dissolved in a carrier salt mixture of LiF, BeF 2 and ZrF4. The UF4 was initially 235UF4 but in 1965 the MSR was defueled with a target of less than 50 ppm U and refueled with 233UF4 which was burned until shutdown in 1969. Defueling U by fluorination to UF 6 results in negligible Pu removal because of the instability of PuF6 gas under the process conditions. The salt resides today in three tanks fabricated from Hastelloy-N alloy. The general characteristics of the MSRE salt tanks and their salt are summarized in Table 1.

Table 1. Characteristics of MSRE Salt Tanks and Contents

Parameter Fuel Drain Tank 1 Fuel Drain Tank 2 Flush Drain Tank (FD1 or FDT-1) (FD2 or FDT-2) (FFT)

Height (in) 86.125 (86-1/8)a 79.8125 84.125 (84-1/8)

Height less hemispherical top and bottom (58-5/8)c 59.0d (in) 64.6875 58.625 Height Hemispherical Top (in) 12.5 (12-1/2)a 12.25 (12-1/4)c 12.5625 (12-9/16)d Height Hemispherical Bottom (in) 8.9375 (8-15/16)a 8.9375 (8-15/16)c 12.5625 (12-9/16)d Tank Wall Thickness (in) 0.75 (3/4)b 0.75 (3/4)c 0.75 (3/4)d Tank Diameter (O.D., in) 49.993a 49.990 50.051d Tank Volume ~80 ft3 (2,265-L) ~80 ft3 (2,265-L) ~80 ft3 (2,265-L)

Salt Volume 35.3 ft3 (1,000 L) 30.9 ft3 (875 L) 67.9 ft3 (1,923 L)

a Fuel Drain Tank 1 Cooling System Assembly, M20794RF003E10, 6/30/1961.

b Fuel Drain Tank Details, M20794RF007D10, 6/28/1961.

c Fuel Drain Tank Cooling System Assembly, FDT2 M20794RF016D2, 6/30/1961.

d Fuel Salt System Flush Tank Assembly and Details, D-FFA40462, M20794RF010D11, 12/29/1961.

The three carrier salt components have different melting points and the degree of homogeneity has not been well determined as the salts freeze during cooldown. During criticality, the fuel salt was circulated by a fuel salt pump through the reactor vessel and a pr imary heat exchanger when the reactor operated.

When the reactor was permanently shut down, the salt was drained into two fuel salt drain tanks (FDTs) located within the stainless steel-lined, concrete-shielded drain tank cell (DTC) adjacent to the reactor cell. See the sketch of an FDT in Figure 1.

A flush salt similar in composition to the fuel salt, but without the uranium fuel, was used to flush the reactor system to remove residual pockets of fuel salt. The flush salt, which became contaminated with a small amount of residual fuel salt after each flush, was drained to the Fuel Flush Tank (FFT), a third tank located in the DTC. The FDTs contain a large array of cooling thimbles or thermal wells and the FFT

1 contains none. FDT 1, FDT 2, and FFT are appr oximately 44%, 40% and 85% filled with salt, respectively. The radioactivity of salts within th e FFT accounts for approximately 1 to 2% of the total radioactivity of the three tanks. Between 2004 and 2008, the final uranium fuel was removed from the salts in the tanks, the FFT in 2005, FDT-2 in 2006 and FDT-1 in 2008, converting all fuel salt to a defueled carrier (or coolant) salt. This report adds to the evidence that defines the extent to which the salt was defueled.

In December 2010, Bechtel Jacobs Company, LLC (BJC) prepared Engineering Evaluation of Options for Molten Salt Reactor Experiment Defueled Coolant Salt (BJC2010) evaluating the salt disposition options for MSRE. The Engineering Evaluation describes sa lt removal as the key to decontamination and decommissioning (D&D) of the MSRE facilities because other D&D work cannot be done without remote tools until the extremely high dose-rate salt is addressed. The Engineering Evaluation points out that whether stratification exists in the frozen so lid salt is important for performing exposure rate calculations and for the developing methods for salt removal and transfer to new containers. Exposure rate calculations normally assume that the large gamma contributor, 137Cs, is distributed uniformly throughout the salt. (Please note that while this re port refers to the dominant gamma emission in MSRE salt as being associated with 137Cs as has become traditional in nonrigorous reference, the gamma emissions are actually associated with the isomeric transition of 137Cs metastable daughter 137mBa immediately following the 137mBa beta decay that follows the 137Cs beta decay.)

In contrast to a uniform distribution, theory offers the possibility that the components of the pure salt itself will separate and freeze to a solid in a very nonuniform predictable manner as the molten salt cools, based on their different melting points of the salt components (ORNL 2002). The chemical form, melting point, and affinity of the 137Cs for any of these various non-homogeneous salt components are unknown.

The melting point of CsF is above any of the salt components (ScienceLab 2011). If the 137Cs is far higher in concentration in a particular location or stra ta in the salt, this could strongly impact the exposure rate in or near specific locations around a container, control the shielding design, and favor process designs that employ mixing and container filling methods that more uniformly distribute the 137Cs (BJC2010).

This report provides confidence in defueling, which theoretically could be provided by either in-situ radiation measurements or by laboratory analyses of samples, if representative salt samples could be obtained. Salt collected from Probe 5 at the cessation of defueling FDT-1 indicated a large concentration of U remaining, documented as Probe 5 of FDT1 Sa mple 4, FDT1-P5-TS-040108, sent for analysis April 4, 2008 (ORNL2008). The probe sample is nonrepresenta tive of the defueled salt in the tank because the result is far outside anything physically possible based on inventory and process knowledge. That result illustrated how obtaining a representative sample coul d be difficult and is compounded by the high dose rate mandating a very small sample size for acceptability by laboratories.

2 Figure 1. Fuel Drain Tank

3 Significant uncertainties in U concentration have been identified, especially regarding quantities of uranium outside the tanks that impact the predefue ling tank inventories, including significant shipper vs.

receiver quantity discrepancies, estimation of migration quantities of uranium into the ducts, pipes and charcoal, dependence on the ORIGEN Code, visual estim ation of the quantity lost in a spill from a piping rupture, and hypothetical precipitation on the tank wa lls in the tank head space above the frozen salt (BJC 2002 and BJC 2005). Since the salt tanks acted as the central salt repository for the fuel, when large uncertainties are stated for quantities deducted from these tanks, these deductions lead necessarily to uncertainties with the remaining tank inventory. The quantity associated with FDT-1that is considered to have the greatest precision is the amount of U rem oved during defueling in 2008. This result was rigorously tracked and was published soon after the re moval (BJC2008) and is relied upon as part of the evaluation described in this report.

Removals intended to clean headspace and piping of migrated U took place in 1997 and 1998. Defueling removed the U by sparging fluorine through the molten salt in 2005, 2006, and in 2008. There was no standard defined for the level of uranium removal th at would qualify as a defueling end point for carrier or coolant salt. A 50 ppm goal was set based on the 235U MSRE defueling history from 1965 published data, the only other time a molten salt reactor had been defueled. In 2008, the 50 ppm goal was declared to have been be achieved for the final tank, FDT-1, based on visual extrapolation of Fourier Transform Infrared (FTIR) data that measured uranium removal rate in gas leaving the salt, but did not measure uranium remaining in the salt itself. All three salt tanks employed exactly the same methodology of arriving at an end point, so while the method could be questioned, the end concentrations are believed to be the same. The visual curve fit extrapolations i ndicated that less than 50 ppm U remained in all three tanks.

Later standard curve fit computer algorithms were applied to reevaluate the same FTIR data for FDT-1.

Use of these algorithms indicate that various alternative conclusions could be reached as well, and more significantly, scrutiny of the data pattern reveals an ap parent change in the reaction rate controlling step, that is expressed as a distinct change in slope, so that two curves appear to fit the data far better than one overall curve. The best exponential fits for the alte rnative interpretation recognizing the rate controlling step change at roughly 3.5 hrs into the final FTIR da ta set, are shown in Appendix A. This algorithmic maximized fit interpretation would result in a final concentration of 78 ppm rather than less than 50 ppm, either of which should be equally acceptable given ther e is no molten salt defueling standard and either concentration, or even an order of magnitude gr eater concentration, would be acceptable, so the differences have no meaningful consequence.

Because the FTIR was measuring removal rates from the exhausted gas and not from U residuals in the salt itself, and because fluorine was sparged from a central point with no mechanical mixing underway during defueling (see Figure 10), this extrapolation ap proach lent itself to questions. Using the FTIR measurements on exhaust gas to de termine remaining tank content woul d presume that the tank is mixed sufficiently through convection plus the single point quiescent sparging for the fluorination reaction contact to be relatively uniform for any given sub unit of salt over short periods of time. Without sufficient circulation, the FTIR measurements would only represent the region of molten salt contacted by fluorine rather than a representative average of a whol e tank. To verify the conclusion reached through FTIR exhaust gas extrapolation, the NDA by direct gamma spectroscopy of a salt tank described here was completed, and these direct measurements provide a totally different form of evidence than FTIR that the salt has been defueled. The NDA data do positively support the FTIR data in indicating that salt defueling is adequate.

4

2. PURPOSE AND SCOPE

Because a probe sample laboratory analysis has been shown not representative of the tank content, there are published uncertainties regarding the inventory track ing data, and there are questions of interpretation regarding the FTIR extrapolation method employed, the sc ope of this work was to make external radiation measurements that could provide additional evidence regarding the fact and extent of defueling of the MSRE salt tanks. Additional sampling and laboratory analysis of salt drilled directly from the tank salt was also considered, but for the following reasons, external nondestructive assay radiation measurements were elected rather than any further sampling.

There was insufficient basis to know where a single sample might be taken that would be representative of the entire tank inventory, especially a sample that could be obtained with reasonable effort through the single 2-inch ball-valve inlet opening on the top of each drain tank.

To drill through a tank wall or thermowell and then plug the hole to obtain a sample might limit future salt and tank disposition options. The wall thickness is unknown, extensive fluorine attack may have taken place during fl uorination, and the degraded wall condition might exacerbate problems that could arise when plugging sampling holes.

There is no reason to expect salt on the tips of probes previously used to feed fluorine would be representative of the salt in the tanks. There is contrary evidence that the probe salt is not representative in that uranium results are unbelievably high from Probe 5 of FDT1 Sample 4, FDT1-P5-TS-040108, sent for analysis April 4, 2008 (ORNL2008) because the uranium results from the probe sample are far outside the uncertainty range of any FTIR or inventory mass-balance estimates.

Unlike a single sample, external radiation measurements might be able to assess the spatial distribution of activity and provide confidence in whether blending or mixing is appropriate in some disposal designs, and especially for the design of shielding for transport containers.

1998 predefueling external measurements of gross gamma in inner row thermal wells indicated that gamma continues to increase with depth to the bottom of the outer thermal wells and measurements went offscale in the inner thermal wells so the previous effort at the time of defueling to collect maximum gro ss gamma data in this central region was unsuccessful.

If successful, in-situ gamma spectroscopy measurements could provide sufficient confirmation for the disposal site that the extent of defueling is within the expected range or is reasonable.

The FFT only contained fuel as a slight contaminant while the two FDTs contain identical material drained from the same reactor loop. Defueling of FDT-1 and FDT-2 was performed identically, using identical equipment and processes and using the same method of visual extrapolation of FTIR data as a means to declare a defueling endpoint. The current configuration of the maintenance shield made measurements at FDT-1 easier than measurements at FDT-2, and more historic data was readily available for FDT-1 because FDT-1 was the last tank defueled. Therefore FDT-1 was selected for the NDA measurements made for this report.

5

3. APPROACH

NDA measurements were made on FDT-1 in 2011 as follows:

(1) gross gamma measurements were made on FDT-1 inside thermal wells and beside the tank on February 17, 2011 with an instrument employi ng a small Geiger-Mueller (GM) tube capable of withstanding thousands of R/hr without going o ff-scale or saturating-- one calibrated to operate in that very high radiation range.

(2) on June 9, 2011 a gamma spectrum was obtained with a lead collimator focused through the maintenance shield and used to establish 208Tl and 212Bi to 137Cs ratios.

4. RESULTS

4.1 GROSS GAMMA PROFILE OF FEBRUARY 17, 2011

The gross gamma measurements were made with an AMP-200 detector on February 17, 2011. Thermal wells, also called thimbles, are vertically standing blin d flanged tubes arranged in two concentric rings in the salt tanks (See figures 4 and 5). Data was collected from an inside thermal well, an outside thermal well, and outside the tank about a foot away from the tank surface.

Mirion Technologies, the manufacturer of the Amp-200 survey instrument, states in their product data sheet that the AMP-200 features linear response from 5 R/hr to 10,000 R/hr and employs an energy-compensated Geiger-Muller detector with a detection range (+/- 10%) for 0.5R/hr to 10,000 R/hr and an energy range sensitivity: 11 cps per R/hr. All NDA measurements on the salt tanks, including any gamma spectroscopy or gross gamma, are made through a port in the six inch thick steel maintenance shield that is on top of the DTC that houses the salt tanks. A sketch of the maintenance shield is shown in Figure 2.

Workers are shown in Figure 3 assembling the collim ator on top of the maintenance shield in 2011.

Table 2 compares the 1998 data in Thermal Well 19, decayed as 100% 137Cs to 2011, to the February 17, 2011 acquired data for Thermal Wells 19 and 32. There is fairly close agreement, but an obvious pattern indicating a larger radiation source is present about a foot above the bottom of the thermal well, which is about six inches beneath the salt surface. The 137Cs appears to have frozen much higher in the salt after defueling than before. Both data sets have zones th at are not statistically different and this suggests a nonuniform distribution (a uniform distribution would r esult in a bell-shaped curve as shown in Figure 7).

Given no better information, shield ing designers would expect and design for the bell curve. In October 1998 the salt had been solidified for almost 29 year s and contained much more uranium. In 2011 it had been solidified almost three years after defueling and fluorine sparging took place until the cool-down began, so it is possible the rising fluorine bubbles also could have had a buoyant effect on the CsF.

Sparging was from a single point (see Figure 10 depicti ng the probe used for fluorination). CsF would not melt at defueling temperatures. Any differences in ambient conditions between 1969 and 2008 that might have affected tank cooldown behavior are unknown. In Tables 2 and 3 the numbers in bold italics (those taken up to 18 from the thermal well bottoms) are measurements taken below the top of salt (see Figure 4).

6 Figure 2. Sketch of Maintenance Shield

Figure 3. Workers Assembling Detector Collimation Shielding On the MSRE Maintenance Shield

7 Table 2. Comparison of Gross Gamma Profiles of 1998 and 2011 Thermal Dist from Thermal Well 19 Thermal Thermal Bottom of Well 19 1998 Data Well 19 Well 32 Well (in) Oct 1998 Decayed 2011 Data 2011 Data R/hr to 2011 (mr/hr) (mr/hr)

(mr/hr) 78 13 10229 5213 5788 75 5377 7137 72 13 10229 6417 11270 69 12760 19340 66 35 27539 19360 21360 63 21280 24610 60 43 33833 23050 26450 57 26570 32320 54 53 41702 30910 36520 51 35350 45280 48 72 56651 41920 50750 45 49670 60400 42 106 83403 59710 73790 39 70330 88740 36 158 124318 87280 111900 33 104800 132000 30 280 220310 140600 174400 27 202500 232400 24 514 404426 360000 409600 Above 21 450600 544400 Salt 18 685 538972 532900 721300 Salt 15 632500 985500 12 774 609000 685700 1174000 9 685700 1140000 6 809 636538 663000 1020000 3 643300 925000 0834 656209 619700 825200 Salt

8 Figure 4. GM Detector in Thermal Well

9 Figure 5. Location of Gross Gamma Profile Data Acquisition

Figure 6. Gross Gamma Detector Positioning through the MSRE Maintenance Shield

10

Figure 7 indicates that the 137Cs is not uniformly distributed, but is more concentrated at the bottom and even more so near the top of the salt. The shape of curve that would be expected for a uniform distribution of Cs-137 over the span of the salt is modeled for comparison as the green line in Figure 7 taken from Microshield 6.02 results (see Grove2011 for a description of MicroShield). While the concentration is not uniform, the 137Cs gamma field as viewed from above the tank appears to very reasonably represent the average for the tank, so th e location of the subsequent gamma spectroscopy measurement (from above) appears to be a fairly representative geometry.

The melting temperature of the expected form CsF is 682ºC. The melting temperature of some major components of the FDT-1 and 2 salt are provide d in Table 3 taken directly from ORNL 2002.

Table 3. Melting Temperatures for Major Fuel Salt Components As Shown in ORNL 2002

Also from ORNL 2002 is a diagram presenting a best idea of how the phases might migrate during freezing based on their different melting points. The isotopes are not addressed, so the primary author was asked, and stated that Solidification is generally expected to occur first at the outer salt surfaces contacting the tank and near the bottom of the tank and in contact with the thermal wells, with the last salt to freeze being in the upper, central portion of the mass. Concentration gradients may arise under conditions where heat is removed very slowly and th e system achieves equilibrium at every point along the temperature-time curve. Rapid cooling, or quenching, would tend to freeze everything in place resulting in microcrystalline phases that are well dispersed. (ORNL 2010). The author also provided an analysis that supported the preferential solidification of thorium last in the upper, central portion of the salt mass, and that europium and cesium would prefer the lower and outside early cooling regions and the thallium might be in the high or low melting point re gions depending on free fluorine and the extent of fluorination. These are all best guesses depending strongly on the cool down time.

Based on Figures 8 and 11 the gamma spectroscopy of J une 9, 2011 from above FDT-1 appears to include both low freezing point (high melting point) areas adjacen t to the thermal wells and on the salt surface as well as the central late freezing (low melting point) zone just beneath the surface. Moreover, there is a theoretical basis in the freezing theory for concluding the zone focused on for gamma spectroscopy would be acceptable or conservative in regard to 208Th and 212Bi (the primary target isotopes with which uranium content is determined). Based on freezing theory, we expect our gamma spectroscopy from above to produce a representative to conservative result.

12 Figure 8. Conception of Location of Fuel Salt Phases Having Different Melting Temperatures As Shown in ORNL 2002

Figure 9. Positioning the Germanium Crys tal Detector on the Lead Collimator

13 Figure 10. Probe illustrating how fluorine was fed from a single point

14 4.2 GAMMA SPECTRUM OF JUNE 9, 2011

Figure 11. Geometry for NDA of June 9, 2011

15 Figure 12. Assembly Ready for Gamma Spectrocopy Through the Maintenance Shield

The details of evaluation of the 2011 gamma spectrosc opy results are provided in Appendix B. The point where the data mathematically supports the quantit y of U removed during final defueling, a value accepted as being of the highest precision of historic values, is highlighted in Table B-7 in Appendix B and leads to the most accepted conclusion of 280 ppm.

For comparison, we can estimate the ppm U from inventory records along with BJC2002. BJC2008 indicates that 3449 g of U were removed during the 2 008 defueling and this number is accepted to have high precision. Deducting this from the Table B-7 predefueling U grams (taken from BJC2002) and converting to ppm using the most probable current FD T-1 salt mass of 2,548,551 g, we have this result:

16 Table 4. Calculated U Masses in FDT-1 Based on Inventory Records and Engineering Estimates Predefueling U Postdefueling U Postdefueling g g ppm U Most Probable 4999 1550 608 Minimum 3673 224 88 Maximum 5889 2440 957

In summary we can compare the most probable results from the various approaches that have been used to our gamma spectroscopy evaluation done in this report, as follows:

Table 5. Comparison of Extent of Defueling of MSRE Salt Based on Various Methods Method ppm U

Evaluation of 2011 Gamma Spectrum Above Based on U Removed in 2008 280

FTIR Extrapolation of 2008 Defueling Data by Machine Fit Algorithm 78

FTIR Extrapolation at Termination of 2008 Defueling by Visual Fit < 50

Calculated From Inventory Records and Engineering Estimates 608

Note that these estimates represent three totally different approaches (gamma spectrum, FTIR, and inventory records) and all three approaches produce r esults that indicate the salt has been adequately defueled.

17

5. REFERENCES

Allentum2011 Advanced Grapher Version 2.2, Allentum Software, Ltd., Ramat-Gan, Israel, 2011.

Baglin1986 Baglin, C.M., S.Y.F. Chu, and J. Zipkin, Table of Radioactive Isotopes, Eds.: R.B. Firestone and V.S. Shirley, John Wiley and Sons, New York, 1986.

BJC2010 Engineering Evaluation of Options for Molten Salt Reactor Experiment Defueled Coolant Salts,

Oak Ridge, Tennessee, DOE/OR/01-2496&D1, Bechtel Jacobs Company LLC, December, 2010.

BJC2008 Phased Construction Completion Report for th e Removal and Transfer of Uranium from the Molten Salt Reactor Experiment Facility at th e Oak Ridge National Laboratory, Oak Ridge, Tennessee, DOE/OR/01-2256&D1, Bechtel Jacobs Company LLC, December, 2010.

BJC2005 Molten Salt Reactor Experiment Post -Fuel Salt Disposition Inventory for Stabilization/Deactivation, Oak Ridge National Laboratory, Oak Ridge, Tennessee, BJC-OR/2011 Revision 1, Bechtel Jacobs Company LLC, April 2005

BJC2003 Salt Canister Radionuclide Inventory and Neutron/Gamma Source Term Engineering Calculation, CAN-02MSRE-A010 Revision 1, Bechtel Jacobs Company LLC, Nov. 3, 2003.

BJC2002 Inventory of the MSRE Fuel and Flush Salt Constituents, Engineering Calculation CAJ-02MSRE-A014 Revision 2, Bechtel Jacobs Company LLC, Nov. 1, 2002.

Grove2011 Microshield Version 6.02, Grove Software, Inc. Lynchburg, VA, 2011

Mirion 2011 Data Sheet for AMP-200 GM Tube-Based Ratemeter and Area Monitor, Health Physics Division, Mirion Technologies, Smyrna, GA, 2011.

ORNL2010 Letter from Barry Spencer to G.R. Wilson: Response to Questions in E-Mail Dated December 3, 2010, UT-Battelle, Oak Ridge, TN, December 15, 2010

ORNL2008. Results of Analyses for RMAL951 (April 08), Chemical Sciences Division, Radioactive Materials Analysis Laboratory, Chemical Sciences Division, UT-Battelle, Oak Ridge, TN.

ORNL2002 Consideration of Operating Temperatures for Fluorination ( Jul-02), Nuclear Science and Technology Division, UT-Battelle, Oak Ridge, TN, 2011.

ScienceLab2011 Materials Safety Data Sheet; Cesium Fluoride MSDS, Section 9 Physical and Chemical Properties, Sciencelab.com, Inc., Houston, Texas, 2011.

18 Drawings

Drawing J3E020794-A005. Melting - Hydrofluorination and Fluorination P&I Diagram, Bechtel Jacobs Company LLC, Oak Ridge, TN.

Drawing M20794RF003E10. Fuel Drain Tank 1 Cooling System Assembly (6/30/61)

Drawing M20794RF007D10. Fuel Drain Tank Details (6/28/61).

Drawing M20794RF016D2, Fuel Drain Tank Cooling System Assembly, FDT2 (6/30/61).

Drawing M20794RF010D11, Fuel Salt System Flush Tank Assembly and Details, D-FFA40462 (12/29/61).

19 APPENDIX A.

FTIR BEST EXPONENTIAL CURVE FITS FOR FDT-1 FINAL PPM U

APPENDIX B.

EVALUATION OF GAMMA SPECTROSCOPY RESULTS

Table B-1. Peak Search Results for Spectrum of June 9, 2011 Li brary E nergy M eas ured Net P eak A rea (k eV ) E nergy (k eV ) Area Uncert.

75. 04 1. 27E + 07 11976. 3 P b x -ray (K1)

84. 89 4. 88E + 06 12098. 2 P b x -ray (K1) 511. 04 2. 98E + 05 5437. 7 A nnihi lat i on 583. 191 583. 52 8. 62E + 04 7110. 1 Tl-208 661. 66 662. 13 1. 00E + 08 10850. 2 Cs -137 727. 33 727. 77 2. 34E + 04 446. 62 B i-212 763. 13 763. 64 1. 63E + 03 328. 53 Tl-208 785. 37 785. 91 5. 00E + 03 346. 17 B i-212 860. 564 861. 12 2. 22E + 04 984. 38 Tl-208 867. 388 868. 65 1. 47E + 03 287. 04 E u-152 873. 19 873. 94 9. 35E + 03 315. 29 E u-154 893. 408 894. 2 2. 65E + 03 721. 35 B i-212 952. 12 952. 52 1. 32E + 03 279. 99 B i-212 964. 131 964. 99 4. 65E + 03 329. 64 E u-152 1004. 725 1005. 64 1. 95E + 04 1605. 63 E u-154 1078. 62 1079. 47 3. 74E + 03 279. 97 B i-212 1085. 914 1086. 85 3. 40E + 03 270. 05 E u-152 1093. 9 1094. 22 9. 26E + 02 244. 87 Tl-208 1274. 436 1275. 85 3. 77E + 04 993. 95 E u-154 1408. 011 1409. 68 8. 85E + 03 889. 95 E u-152 1512. 7 1514. 5 4. 48E + 03 468. 5 B i-212 1592. 533 1594. 08 1. 10E + 05 1194. 18 Doubl e E s c ape (2616. 55k eV )

1620. 5 1622. 56 2. 08E + 04 890. 15 B i-212 1806 1808. 39 1. 61E + 03 487. 77 B i-212 2103. 533 2106. 39 8. 73E + 04 1032. 03 S ingle E s c ape (2616. 55k eV )

2614. 533 2619. 08 5. 98E + 05 804. 23 Tl-208

B-3

Table B-3. Fit to Normalized Efficiency for Spectrum of June 9, 2011

Radionuclide Relative Activity Bi-212 1.0000 Tl-208 0.3594 Cs-137 352.6000 Eu-152 0.0400 Eu-154 0.1200

Energy Fit to normalized relative efficiency

500 0.0451 600 0.1770 700 0.2886 800 0.3853 900 0.4705 1000 0.5468 1100 0.6158 1200 0.6788 1300 0.7367 1400 0.7903 1500 0.8403 1600 0.8870 1700 0.9309 1800 0.9723 1900 1.0114 2000 1.0485 2100 1.0838 2200 1.1175 2300 1.1497 2400 1.1805 2500 1.2100 2600 1.2384 2700 1.2657

B-6 The following reasonable assumptions with supporting ju stification enable the subsequent evaluation.

1. The salt in the two drain tanks, FDT-1 and FDT-2, contain the same isotopic concentrations. This assumption is supported through the origin of sa lt being from the same reactor loop, identical maintenance, identical defueling process terminated by the same method of FTIR data extrapolation.

2. The ratio of the 137Cs, 212Bi, and 208Tl measurements derived from the gamma spectrum acquired June 9, 2011 represents the average ratios for these three isotopes in the salt itself. Gross gamma activity high in the tank as shown in Figure 7 s upports this assumption, with the weakness that 137Cs is the dominant contributor to gross ga mma and the isotopes are undifferentiable through gross gamma profiling. Because 212Bi has a short half life, its behavior in molten salt as well as that of all parent isotopes back to 228Th is irrelevant. The location of 212Bi is totally determined by the controlling grandparent 228Th since that is the first isotope up the chain with a sufficient half-life so atoms present at the time of gamma spectroscopy could have existed when the salt last cooled and froze to a solid in April 2008. Th ere is no electrochemical rationale that indicates thorium would freeze low in the tank and there is subject matter expert opinion to the contrary, which says the spectrum from above would be very representative to conservative (ORNL2010).

3. The net effect of cold fluorination occurring due to radiolytic decay (which robs fluorine atoms from the fluoride compounds that make up the salt) and hot fluorination during defueling (which replaces the fluorine atoms on the salt compounds and al so adds the F atoms that make the U into UF6 gas to facilitate removal) in concert have a negligible effect on the cumulative salt mass over time.

4. The salt was drained from the reactor loop as a liquid on December 1, 1969 (37,557 g) into FDT-1 and FDT-2. Of that total mass, 22,916 g U were removed by defueling/reactive gas removal (taken as August 1, 1998 when the removal effec tively terminated based on mass flow) and the amount of U to be determined by this report was effectively all removed from FDT-1 on April 1, 2008. An otherwise unaccounted for 5,262 g of U migrated from the tanks into pipes, vessels, ducts, and media on December 1, 1969. These mass quantities are all based on BJC2002. These quantities are not believed to be highly accurate, but are only employed to estimate undecayed 228Th daughter emissions in a conservative manner. The dates of defueling transfers chosen are

near the end of each removal cycle and mass transfers are assigned to one specific day to facilitate otherwise exceedingly complex decay calculations. The migrations of material may have occurred in the late 1980s to mid 1990s but are assumed to have all occurred instantaneously on December 1, 1969 because the actual time of migration is unknown and the assumption of an earlier date is conservative, pr oducing a higher ppm U final result through the calculations below. This assumption is reasonable as well as conservative because even if the migrations occurred in 1994 just before the discove ry of a leaking valve that failed to seal the salt tank headspace, the migration-orphaned 228Th left in FDT-1 would have experienced eight half lives by the time of the 2011 gamma spectroscopy measurements, so the effect of making this conservative assumption on the calculation is very small.

5. The 137Cs content of the salt is accurately calculated by the SCALE module of the ORIGEN computer code (BJC2003) and then corrected fo r decay for use in this evaluation. Because 137Cs is a very dominant isotope in all mixed fission product distributions from a spectrum of reactor designs and should be a calibration standard for so ftware, there is added confidence in that the 137Cs prediction by ORIGEN is accurate. To add even greater confidence in this specific

instance, the range of exposure rates that might be expected of this quantity of 137Cs is reasonable

B-7 based on the comparison of actual gross gamma measurements to MicroShield uniform concentration modeling results shown previously.

6. The 222 ppm 232U value stated in BJC2002 was accurate and the ratio of 232U to U was unaffected by the nuclear reactor operations in 1968 and 1969. Note that the interest in 232U concentration as well as the nonuranium isotopes is merely as tracers or parameters used to derive total U. While total U in the salt is the objective of the gamma sp ectroscopy measurement, yotal U is not directly detectable but must be derived from othe r measurements and known quantities.

7. The minimum and maximum probable predefue ling U mass values provided in BJC2002 are accurate enough to determine a range over which the evaluation of gamma spectroscopy data should be made.

8. The mass of U removed in 2008 (from BJC2008) is the most accurate historic measurement known that can provide a sound basis to calculate a final residual U post defueling result. This is the quantity best known and therefore most depended on in this evaluation.

Because the top dead center appears representative of 137Cs based on gamma profiling and because 228Th is expected to be more concentrated in this area based on freezing theory, the use of 137Cs for prorating and of 208Tl and 212Bi for prorating from the gamma spectrum is conservative. Other than simple four-function math to convert units and sum quantities, two equations were used to determine the total uranium in FDT-1 as follows. It is well known that the decay of a sufficiently large sample of radioactive atoms will be observed to be proportionate to a constant rate, so that at any time t, pN parent atoms can be observed to undergo radioactive decay to a daughter nuclide at rate p :

dN p N dt p p

while the number of daughter atoms dN will be undergoing change with respect to time as follows:

dN d N N dt d d p p

The activity A (as measured in Ci) of both parent and daughter is proportionate to their number of respective atoms remaining, or:

Ap Npp and dAd Nd

From this, the daughter activity dA at any time t is concluded to be:

A ( t) )A(0)d(e pted et)A(0dt d p d d p

The daughter equation above is key to this evaluation because by knowing the 228Th daughter activity Ad at time t when gamma spectroscopy measurements were obtained, along with the initial (t) predefueling 228Th (daughter of 232U) activity )A at time 0, d(0

B-8 the activity of the parent nuclide 232U predefueling )A can be determined, and this result can be p(0 decayed to the time of gamma spectroscopy using the ordinary decay equation:

A ( t) )A(0ept p p

and once the present 232U parent has been established, the U ppm in the salt can be determined through the 232U to total U ratio. The last two equations described above (the daughter equation and the decay equation) are the only equations required for performing this evaluation, and these were used to set up some of the functions in cells of a spreadsheet using Excel so that an array of values for predefueling 232U could be evaluated. The results are shown in Tabl e B-7, but to perform the evaluation, several key parameters must be known, as follows.

Table B-4 shows isotopic decay parameters from Bag lin1986 used in the evaluation and important to the discussion below.

Table B-4. Decay Parameters for Isotopes of Interest Specific Nuclide Atomic mass T1/2 (yr) (yr-1) T1/2 (sec) (sec-1)Activity (Ci/g)

U232 232 68.9 0.010060191 2.17E+09 3.19E-10 2.24E+01 Th228 228 1.9131 0.36231623 6.04E+07 1.14809E-08 8.20E+02 Cs137 137 30.07 0.02305112 9.49E+08 7.30433E-10 8.68E+01

The following three values were also used in the spreadsheet:

Table B-5. Three Common Constants and Conversions Avogadro's # 6.02E+23 sec/yr 3.16E+07 Bq/Ci 3.70E+10

Table B-3 indicates that in terms of relative activity, for every Ci of 212Bi FDT-1 also contains 352.6 Ci of 137Cs. Because 212Bi and 228Th are in secular equilibrium in FDT-1, their activities are roughly equal so for each Ci of 228Th, FDT-1 contains 352.6 Ci of 137Cs. The 137Cs reported based on results of the SCALE module of the ORIGEN computer code (BJC2003) in FDT-1 is 2.74E+03 which decayed 8.512 years using the decay equation to June 1, 2011 becomes 2.252E+03 Ci. Pro rating we can expect 2.252E+03 / 352.6 = 6.386 Ci of 228Th to also be in FDT-1. The interest in all these non-U isotopes is merely as a method to get to total U, which cannot be measured directly.

B-9 At the instant before defueling in late March 2008, the 228Th activity in FDT-1 originated from two sources:

1. the 232U that remains present in the salt that has continually decayed to 228Th from the point of creating the fuel by chemical separation of elements.

2. the 1998 orphaned 228Th that has not completely decayed after being left without its 232U parent mass at the time of the 1998 defueling. Using the decay equation, we can decay the 222 ppm 232U reported by BJC2002 from August 1, 1968, the time of fueling, to the time of initial defueling, taken as August 1, 1998 (29.9 years) and that w ill result in 164.2 ppm in 1998 before defueling.

The latter reference also reports that 22,916 g of U were removed by initial defueling in 1998, so the product of these three values indicates 2.005 g (or 44.84 Ci) of 232U was removed in 1998 and per the decay equations, the orphaned 228Th will produce about 0.4413 Ci that will remain as undecayed 228Th by the time the gamma spectrum of FDT-1 was acquired in June 2011 (this time is 42.831 years after the measurement of 232U concentration on August 1, 1968).

At the instant after defueling in 2008, the activity of 232U was again reduced, so that from that point forward, the remaining 228Th activity was composed of both components above plus a third component, which is:

3. the April, 2008 orphaned 228Th daughter mass that existed at the instant the final defueling was completed.

Since the 228Th is derived from these three different sources and includes two orphaned residual sources, we cannot determine the 232U content with absolute confidence without knowing the predefueling inventory. BJC2002 provides a best estimate, minimum, and maximum predefueling 232U mass inventory for FDT-1 and provides the relative fraction fo r each U isotope. As shown previously, the 232U has decayed to 145 ppm parts U in 2011so we can use this result to adjust the values in BJC2002 to account for that decay and we have the following most probable, minimum, and maximum predefueling 232U masses in FDT-1:

Table B-6. Calculated Pre-2008-defueling U Masses in FDT-1 U g U232 g Most Probable 4999 7.213E-01 Minimum 3673 5.300E-01 Maximum 5889 8.497E-01

Using only the decay equation, daughter equation, and four function arithmetic for conversion and combining quantities, Table B-7 below displays the results over the range of interest. A method selected as the very best way for arriving at a defueling U concentration is by finding the point where the 2011 gamma spectroscopy result corresponds to the quantity of U known to be removed in 2008, which is accepted as a highly accurate quantity.

B-10

DOE/OR/01-2513&D1

RECORD COPY DISTRIBUTION

FileDMCRC