J. A. Pike
Westinghouse Savannah River Company
Aiken, SC 29808

Available for a processing fee to U.S. Department of Energy and its contractors, in paper, from: U.S. Department of Energy, Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN 37831-0062,  phone: (865 ) 576-8401,  fax: (865) 576-5728,  email:  reports@adonis.osti.gov

Keywords: Waste Removal, Salt Dissolution

A computer simulation model for salt dissolution of Tank 41 was constructed and run for three cases. These case runs show that eight to nine 30,000 gal batches of discharge, 240,000 to 270,000 gal total, are needed to result in the first 100,000 gallons of saltcake transferred out of Tank 41. This equates to 140,000 to 160,000 gallons of inhibited water feed solution after initially refilling the interstitial void. The process plan should allow for 140,000 to 200,000 gallons of batch feed solution in addition to the volume to fill the interstitial void to transfer enough dissolved saltcake to reach the initial 100,000 gal of saltcake goal. The initial fill will vary greatly depending on the extent of interstitial removal and refill. The refill feed volume can range from 23,000 to 230,000 gal.

Historical data from Tank 10 salt dissolution was used to compare the results predicted by the model to a real case. The results match the bulk ratios of dissolution water used to salt dissolved and discharge volumes, but consistently under predicted total volumes. This appears to be an artifact of available batch data including flow rates in and out of the tank.

SRS High Level Waste (HLW) plans to dissolve saltcake in Tank 41 to feed the low curie salt process. The low curie salt process treats high level waste saltcake for final disposal in saltstone. To prepare saltcake for final disposal, it is necessary to separate cesium, the dominant radioactive element, and soluble transuranic material from the waste stream. Since the cesium concentration is well below solubility limits in waste, all the cesium in saltcake is in the supernate and interstitial liquid. If the interstitial liquid can be removed from the saltcake, adequate seperation of cesium may be achieved for disposal in saltstone.

The process to dissolve the saltcake is simulated by an Aspen Custom Modeler™ model using Aspen Properties Plus™ to estimate chemical properties.

Saltcake may be considered composed of four parts: salt crystals, interstitial liquid, gas, and insoluble solids. When salt crystals form and accumulate in a waste tank, interstitial space between the crystals will hold supernate and, potentially, gas. Gas can consist of trapped air, water vapor, and radiolytically produced gases, primarily, oxygen, hydrogen, and nitrous oxides. Insoluble solids may consist of entrained high-level waste sludge solids or, potentially, insoluble materials that form during concentration in the evaporator. Some materials such as aluminum hydroxide as gibbsite may take a long time to precipitate and may not be present in freshly formed saltcake.

Interstial liquid and supernate probably contains all the cesium in soluble, ionic form and none in crystalline form. A fully saturated Savannah River Site (SRS) waste supernate would contain about 10,738 Ci/gallon cesium-137. Comparison with the actual radiation values for SRS salt solutions indicates the cesium levels are only 0.2 to 0.9 percent of the saturation level. This indicates that most if not all the cesium is in solution and resides in the interstitial liquids of the salt cake. Thus, removing the interstitial liquid will remove the cesium.

It is estimated that at least 75% of the interstitial liquid will need to be removed from saltcake to produce salt solution capable of meeting the Saltstone WAC. The plan for Tank 41 is to place the suction at the 24" level to drain the top portion of the saltcake as much as possible. The existing Fixed Length Jet (FLJ) will be used to decant the free supernate from the top of the saltcake. Once the free supernate is removed from the tank, the FLJ will be removed. A submersible pump will then be installed in the tank and mined to the 24" level thus forming a pump well in the saltcake. The pump will then be used to remove interstitial liquid and salt solution at a slow rate, less than 5 gpm on average. Inhibited water will be added to the tank to cover the saltcake. A relatively thin layer of liquid will be maintained above the saltcake by adding inhibited water while simultaneously transferring out of the tank at about 20 gpm. The dissolved saltcake will be sent to Tank 50 for final characterization before transferring to the Saltstone Facility. The dissolution will occur in relatively small batches, 10,000 to 30,000 gallons, to minimize the risk of sending a significant amount of dissolved saltcake with too much cesium².

The composition of the existing waste inventory accounts for radiolytic decomposition of sodium nitrate to sodium nitrite and the reactive sorption of carbon dioxide from the air to form sodium carbonate. The only other radiolytic reaction products are gaseous hydrogen and oxygen, which evolves from the waste to the tank vapor space. The vapor space vents to the atmosphere.

Free hydroxide, nitrate and nitrite concentrations are monitored in stored waste and adjusted, when necessary, by adding caustic or sodium nitrite to maintain the concentrations in a range that minimizes the corrosion rate of the carbon steel waste tanks.

An Aspen Custom Modeler™ model was developed to calculate the material balance for the Tank 41 flowsheet. The following sections describe the basis of the model.

The following assumptions are used in the modeled flow sheets:

In several instances, quiescent periods are picked to clearly delineate process steps as well as allow adequate time for the model to reach steady state conditions between steps. Though these quiescent times approximately match the technical plan, they may not completely agree or match the program schedule.

Saltcake in High Level Waste tanks consists of three phases saltcake, supernate, and gas. The gas phase is neglected in most cases since it has insignificant contribution to total mass, but does affect bulk density. Material may transfer between phases by crystallization of salts in the supernate or by dissolution of the salt cake. The dissolution rate for each salt is approximated by:

where:

In some cases, other salts control the cation or anion concentration. For example, many salts contribute to the sodium concentration in a waste solution. As such, the sodium concentration does not change significantly with a change in concentration of a specific salt. The difference between supernate concentration and the equilibrium concentration is approximately constant for an incremental change in the salt concentration. In this case, the cation term is assumed to be incorporated in the apparent K_dis, thus, the rate equation becomes:

For a salt where other salts control the supernate concentration of the anion, the rate equation becomes:

An additional term is added to avoid discontinuities that cause the modeler equation solver to fail to converge on a solution. This term smoothes the transition from solid salt to no solid salt with a ratio of

where:

This ratio is approximately 1 until the number of moles of component i approaches the value of factor. As the number of moles approaches zero, the value of the ratio will asymptotically approach zero and smooth the sharp discontinuity when a dissolving component reaches complete dissolution. Thus, the final form of the dissolution rate equations is:

Addition of this ratio prevents the model from simulating precipitation of salts if there is zero solids present initially. If no solids of a specific component are present in the initial conditions, then the rate equation always equals zero. Therefore the model will not initiate precipitation. If any amount of solids is present, then precipitation is initiated as normally expected. This essentially limits the model to use for dissolution only or very carefully setting initial conditions for other systems.

The model currently assumes that salt dissolves at a very high rate and the dissolution constant is set to an arbitrarily high value of 1000 kmol/hr / kmol/m³. Laboratory tests of salt dissolution kinetics indicate that the rate constant is very high, but data with adequate time resolution is not available to calculate the value.³ These tests clearly demonstrate that the rate limiting step is the rate of mass transfer of the concentrated anions in solution near the crystalline surface to the bulk of the solution and not a function of chemical composition in the salt cake. Diffusion from the crystalline surface can be described mathematically analogous to heat transfer. In so doing, the parameters that affect diffusion include the diffusion film thickness and surface area. The velocity of the bulk liquid above the crystalline surface controls the film thickness. The dissolution methods used will draw a significant amount of dilute solution through the salt cake, which will tend to minimize film thickness. Surface area is the same in all dissolution methods available although some methods will only affect the film thickness at the top of the salt cake. This model assumes perfect mixing in the bulk liquid, which is reasonable for dissolution methods that draw a significant amount of liquid through the salt cake, affecting the film thickness for the bulk of the salt cake. Future model improvements will incorporate limitations due to diffusion if possible.

Equilibrium concentration of soluble species in solution is based on Aspen Properties Plus™ thermodynamic solubility estimation models. Other estimation models may be used. The total tank contents are considered when calculating the equilibrium concentration. This method presumes that the supernate has a uniform composition through out the waste tank.

The true composition function returns mole fractions. In order to determine the total moles at equilibrium, the moles of water in the apparent composition is assumed equal to the moles of water in the equilibrium composition. Thus, the total moles at equilibrium can be determined by:

where:

Using Aspen Properties Plus; to estimate solution properties at equilibrium, concentration of each soluble component may be calculated by:

where:

The supernate composition is calculated by estimating actual component compositions and back calculating the corresponding stoichiometric concentration of apparent components. The model listing in section 5 indicates the stoichiometric relationships under the model "Supernate", subsection labeled "SupernateChemistry".

Mole balance for the waste tank supernate is calculated by simultaneous solution of the following:

Component balance:

where:

Total mole balance:

where:

Definition of mole fraction:

A specific phase may be defined similarly; for example, the supernate phase is as follows:

Mass balance for the waste tank supernate is calculated by the following:

where:

Mass fractions may be converted from mole fraction as follows:

where:

Individual components dissolve or precipitate until saturation levels are reached or the solids are depleted. The model assumes saturation can be achieved. Future revisions may allow user input to limit extent of saturation.

The number of components in the salt cake and supernate are limited to major components at this time. Other components can be added by expanding the chemistry model to include these components. Stoichiometry and dissolution rate constant for each additional component will need to be added to the model.

The components currently modeled are shown in Table 1.

Table 1: Waste Components Modeled

Solids of a specific component must be present in the initial conditions for precipitation to be initiated. Initial saltcake must include a small amount of all components to assure equilibration is achieved at the start of each run.

The model assumes an isothermal process, so changes in temperature during the process, exothermic reactions, or endothermic reactions are not affected. Considering that the waste tank is a large heat sink, and the dissolution process is relatively slow, the waste tank temperature is unlikely to change much during dissolution, effectively approximating an isothermal process.

Table 2 through Table 4 show the salt composition for Tank 41 as extracted from the Waste Characterization System (WCS) and used in the low curie salt process feed basis. The WCS describes the chemical composition of the salt based on an average saltcake compositions with an interstitial void space filled with supernate. The supernate is assumed to have the same composition as the supernate on top of the salt, which is sampled routinely. Saltcake is not sampled routinely.

The model design anticipates that the known starting compositions are not at equilibrium or steady state. When initiating a model run, the model starts simulating the dissolution or precipitation of salt until a steady state condition is achieved. Using the WCS composition for saltcake and supernate, the bulk saltcake level is reduced over 50" before simulating the actual process. The resulting supernate does not approximate the supernate sample data. Considering that the WCS composition for saltcake is based on an overall average and not necessarily samples from the specific tank, this result is not unexpected. Therefore, other or additional data is needed for input.

Salt sample data from Tank 41 is shown in Table 5. The results are shown for composition of the total sample. Part of the sample is supernate and part is solids. Using the measurement of water content, the amount of interstitial supernate can be approximated as shown in Table 5. The composition is estimated by assuming the interstitial supernate is the same composition as the bulk supernate in the tank. The composition of saltcake solids is calculated by subtracting the quantity of each supernate component from the total composite composition. The supernate composition is determined by averaging several supernate samples before and after the saltcake sample date. No transfers into the tank that would change the supernate composition occurred during this period.

Table 2: WCS Saltcake Composition of Tank 41

Table 3: WCS Saltcake Composition as
Equivalent Ionic Species of Major Components

Within error of measurement, the model input composition shown in Table 5 is probably at or near equilibrium. Since the model will dissolve or precipitate salts until an equilibrium condition is established, the initial calculation will bring the composition of solids and supernate to equilibrium. The amount of salt dissolved or precipitated at the start of the calculation will indicate how close the initial composition is to equilibrium estimated in the chemistry model. As seen in Figure 1 and Figure 2, the initial correction of saltcake is about 5", ~1.4%. This is within the error to measure average saltcake level as well as chemical compositions. For the high void case, the initial composition produces a saltcake level change of ~15". In order to minimize this error, the resulting chemistry after allowing equilibration time is used as initial conditions instead of the composition shown in Table 5 directly. Figure 3 shows nearly no correction after this adjustment.

Table 4: WCS Supernate Composition of Tanks 41

Table 5: Tanks 41 Salt and Supernate Data

Salt sample data from reference and supernate data from reference .

Table 6 shows the variations made for different model runs. The base case represents idealized conditions and the other cases are variations from the ideal conditions. The parameters selected for the base case represent relative maximum drainage and void fraction. The remaining two cases vary the extent of drainage and void fraction to obtain results for the expected range of operation. Using the data in Table 5, the high void fractions for each of the three samples are calculated and are shown in Table 7. The sample void fraction ranges from 0.3 to 0.5. Void faction of 0.42 is used to match the estimate made in the hydrogen evaluation for Tank 41 salt dissolution.

Table 6: Variations to Model Runs

Table 7: Tank 41 Sample Void Fraction

Table 8 through Table 10 show the volumetric results of each case. These case runs show that eight to nine 30,000 gal batches of discharge, 240,000 to 270,000 gal total, are needed to result in the first 100,000 gallons of saltcake transferred out of Tank 41. This equates to 140,000 to 160,000 gallons of inhibited water feed solution after the after initially refilling the interstitial void. The process plan should allow for 140,000 to 200,000 gallons of batch feed solution in addition to the volume to fill the interstitial void to transfer enough dissolve saltcake to reach the initial 100,000 gal of saltcake goal. The initial fill will vary greatly depending on the extent of interstitial removal and refill. The refill feed volume can range from 23,000 to 230,000 gal.

These runs show that merely measuring the change in saltcake level will not accurately indicate the quantity of saltcake transferred out of the tank. These runs show that 10,000 to 90,000 gallons of saltcake will dissolve on initial water addition before sending any to Tank 50. Further, if there is a high void fraction that is significantly larger than 0.22, then the volume of saltcake dissolved on initial filling could double.

In addition, if the drained interstitial void is not completely filled before transferring dissolved salt solution out of the tank, some amount of dissolved salt solution will continue to seep into the saltcake. In this case, the change in salt level would indicate the total of the salt transferred out of the tank and the salt that remains by filling the interstitial void.

When water is added to the top of the saltcake, the dissolution liquid begins refilling the interstitial pores. The salt around the pores dissolves to create progressively larger flow paths. As the liquid advances into the saltcake it will eventually become saturated and no longer dissolve saltcake. From this point, the liquid will still flow deeper into the saltcake, but the flow rate will be limited by the percolation rate, which is expected to be around 4 gpm.^1,2 At some point while adding dissolution water, the liquid will start to accumulate over the saltcake surface before fully saturating the interstitial void of the saltcake. When the flow stops, the liquid on the surface may appear to sorb into the saltcake, when, in fact, it is continuing to flow into the interstitial void. Planned dissolution may proceed without completely saturating the void space. During dissolution some of the dissolved saltcake is transferred into the interstitial void at the same time liquid is transferred out. A fair approximation of saltcake removed from the tank may be obtained by applying an average ratio of discharge liquid to dissolved saltcake as follows:

where

This relationship assumes perfect saturation. Adding a percent relative saturation term corrects this deficiency:

where

The ideal term R_{saltcake to discharge} can be calculated from model results as shown in Table 8 through Table 10. A model run may be run to fit the more accurately determined parameters for void fraction and/or composition resulting from the actual dissolution of salt. In order to determine percent relative saturation, composition data from at least two supernate samples from the pump suction point are needed, one at the end of the dissolution and one around the midpoint. Initial composition data already exist as shown in Table 5.

An alternate method to determine R uses the data in Table 5 some measure of total solids in the dissolved supernate as follows:

where:

Total solids in the supernate may be estimated by using a correlation of the specific gravity to total solids. This model results provides one such correlation and more specific correlation can be generated for each batch if needed.⁷

The accuracy of this estimation method can be determined by propagating the error of the parts. The equation for V_saltcake consists of three parts: measurement of transferred liquid, estimation of R, and %_sat. Relative error of V_saltcake may be estimated by:⁸

where

V_discharge is determined as follows:

where

The error of V_discharge is estimated as follows:

Therefore,

Liquid level in the receiving tanks is measured with reel tapes with a precision of 0.05 inches. For a 250,000 gallon discharge volume, (L_f – L_i) is expect to be 71.2 inches. C for the expected receiving tank is 3510 gal/inch with e_C about ± 35. The values for R_{saltcake to discharge} can range from 1.9 to 3.5. Estimating e_R at ± 0.3 for R of 2.7, and e_%sat at ± 5% for 80% relative saturation determination, then

Figure 1: Results for Min Drain Case

Table 8: Batch Summary for Min Drain Case

Batch	Feed (gal)	Salt Dissolved (gal)	Ratio Feed to Salt Dissolved	Salt Solution Discharged (gal)	Ratio Discharge to Feed	Ratio Discharge to Salt Dissolved
Initial Drain	0	0		53,580
Initial Fill	23,400	10,137	2.31	0
Batch 1	20,688	10,215	2.03	31,080	1.50	3.04
Batch 2	24,538	13,344	1.84	31,200	1.27	2.34
Batch 3	24,245	14,182	1.71	31,200	1.29	2.20
Batch 4	24,055	14,791	1.63	31,200	1.30	2.11
Batch 5	23,928	15,231	1.57	31,200	1.30	2.05
Batch 6	23,844	15,545	1.53	31,200	1.31	2.01
Batch 7	23,788	15,766	1.51	31,200	1.31	1.98
Batch 8	23,753	15,921	1.49	31,200	1.31	1.96
Batch 9	23,730	16,028	1.48	31,200	1.31	1.95
Batch 10*	47,425	32,254	1.47	62,400	1.32	1.93
Batch 11	23,704	16,188	1.46	31,200	1.32	1.93
Batch 12	23,702	16,212	1.46	31,200	1.32	1.92
Batch 13	23,295	13,712	1.70	30,653	1.32	2.24
Batch 14	23,353	8,926	2.62	31,200	1.34	3.50
Batch 15	22,794	9,465	2.41	31,200	1.37	3.30
Batch 16	22,336	9,906	2.25	31,200	1.40	3.15
Batch 17	21,974	10,255	2.14	31,200	1.42	3.04
Batch 18	21,693	10,526	2.06	31,200	1.44	2.96
Batch 19	21,480	10,733	2.00	31,200	1.45	2.91
Batch 20	21,001	10,504	2.00	30,632	1.46	2.92
Batch 21	21,566	10,507	2.05	31,200	1.45	2.97
Batch 22	21,448	10,624	2.02	31,200	1.45	2.94
Batch 23	21,361	10,702	2.00	31,200	1.46	2.92
Batch 24	21,301	10,753	1.98	31,200	1.46	2.90
Batch 25	21,262	10,791	1.97	31,200	1.47	2.89

* Batch 10 actual sums two batches. This combination is due to the model failing to output data between these two batches.

Figure 2: Results for Baseline Case

Table 9: Batch Summary for Baseline Case

Figure 3: Results for High Void Case

Table 10: Batch Summary for High Void Case

Figure 4 shows the flow of apparent component from the solid phase, saltcake solids, to the liquid phase, supernate, for the max drain case. A negative value indicates precipitation from the supernate. The initial few hours in the graph indicates that some components are dissolving and others precipitating to adjust the initial input to an equilibrium condition. Essentially all the NaAlO₂, NaCl, NaOH, and NaNO₂ in the initial composition of the solid phase dissolves. A small amount of Na₂SO₄ and Na₂CO₃ dissolves and a small amount of NaNO₃ precipitates. The final composition is still substantially similar to that shown in Table 5, e.g. primarily nitrate and carbonate salts with small amounts of sulfate, phosphate, and oxalate salts as shown in Table 11. The relatively small amount of the more soluble species identified in the salt sample are likely due to imperfect estimation of supernate composition as identified in Table 5.

Table 11: Composition after Equilibrating in Mole Fraction

Figure 4 shows additional details about which components dissolve first. The initial components that dissolve are primarily carbonate and nitrate salts with a small amount of sulfate salts. Nitrate and sulfate salts dissolve more rapidly once the carbonate salts are completely dissolved. This behavior is similar to what is observed in numerous laboratory tests.^9,10,11,12 In one laboratory study of SRS saltcake dissolution, the carbonate did not dissolve as readily as predicted and substantial quantities remained in the saltcake until most of the nitrate salts dissolved.¹³ Earlier studies¹⁴ and salt dissolution in waste tanks did not measure chemical compositions adequately to determine chemical changes occurring at intermediate steps in the dissolution.

These earlier tests and waste dissolution tracked specific gravity as an indicator of saturation. If the composition is constant, then specific gravity would indicate extent of saturation. If the composition fluctuates, the composition needs to be known to determine relative saturation. To effectively interpret the specific gravity values obtained during dissolution, the measured value can be contrasted against the predicted values at saturation for a given composition. Predicted compositions produce a relative high specific gravity peak as dissolution

progresses. This peak corresponds to the depletion of carbonate salts from the saltcake. Tank 10 dissolution history shown in Table 12 indicate an initial high specific gravity followed by a substantially lower value during last part of the cycle. Though past interpretation of this data suggests that this is due to lower levels of saturation, that may not be the case. The change in composition may be a significant contributor to this effect.

Figure 4: Flow of Components from Saltcake to Supernate for Max Drain Case

This version of the model is compared to historical dissolution process as was done for the earlier version created for Tank 37 dissolution.¹⁵ Records from Tank 10 dissolution are complete enough to reconstruct the process history and compare the actual dissolution results with the model predictions.

Tank 10 was dissolved by flooding the saltcake with liquid from Tank 23, RBOF/RRF waste. The waste is a relatively dilute salt solution and a good substitute for inhibited water. The additional salt in the waste makes only a minor difference to the amount of salt dissolved per gallon of dissolution liquid. The original Tank 23 composition is not available and inhibited water is substituted for the dissolution water in the model, 0.01 M NaOH and 0.011 M NaNO₂ solution.

Once flooded, the level was maintained while simultaneously feeding dissolution water and removing saturated solution. The removal jet was buried deep in the salt bed, about 3’ from the bottom of the tank. The flow rates in and out were maintained relatively low, about 10 gpm in and 15 gpm out. At the end of a batch, the feed was stopped, but the discharge was continued until a certain level was reached to allow inspections.

The history of salt dissolution from Tank 10 came from several documents.^{16,17,18,19,20} The compiled timeline, including flow rate, average discharge specific gravity, and tank levels, is shown in Table 12.

Historic saltcake composition is not available, so the saltcake composition used for Tank 41 dissolution modeling is assumed. This is different from what was used in reference , which used WCS composition for saltcake. This was not adequate since a significant change in saltcake level and supernate composition would occur during initial equilibration. In addition, the saltcake void fraction is assumed to be 22%.

Table 12: Tank 10 Salt Dissolutuion Historical Timeline

The Tank 10 dissolution is simulated in two ways. One method simulates the batch flow rates and tank levels as input to the model. The other method simulates the batch flow rates and flow times as input. To simulate tank levels, the input selected is based on the batch data for flow rates and ending tank levels. The batch size and time is calculated. If the tank level is not available, the batch time calculated from batch volume and flow rate is substituted. To simulate time, the model input is based on the batch time calculated from batch volume and flow rate. The levels are calculated. In both cases, the salt dissolved and batch chemistry is calculated.

Table 14 and Table 15 show the simulation results of dissolution. The simulation indicates that the historical batch flow rates and times are not consistent with the tank level data available. The exact timing of the level data may be inaccurate.

Simulating flow time produces accurate simulation of batch sizes as shown in Table 14. The model predicts volume ratios well, thus, predicts dissolved salt volume well. Simulation based on tank levels consistently under predicts the feed and discharge volumes as shown in Table 15.

This result is inherent in the available data. The input is based on the batch data available, which indicates an average flow rate, tank levels, and time. The model was designed for predicting the level and dissolution volumes based on the batch size. Since the flow rates are approximated from level changes, the historical data probably does not accurately reflect actual flow rates. If the input is restructured for batch sizes, the flow rates would need to be adjusted in order to match tank level stopping points. Given that the model predicts the important ratios well, the model would predict the salt dissolution volumes reasonably well too. The over all ratio matches well.

Predicted specific gravity is higher than measured during actual dissolution. The specific gravity drops dramatically during the last part of the dissolution, which seems to correspond to the trend observed during Tank 10 dissolution. For a given composition, the properties estimation methods used in this model over predicts the density by 4.0% to 5.6%.²¹ This amount of error accounts for a significant portion of the difference. The remaining difference probably indicates the relative saturation of the dissolution liquid.

Table 13: Tank 10 Salt Dissolution Historical Results

Table 14: Tank 10 Salt Dissolution Model Results Simulating Flow Time

Figure 5: Tank 10 Model Results

5 Model Listing

The following model listing shows example input data for the base case:

1. J. N. Brooke, K. Stahlie, and J. F. Peters, "Hydrological Methods can Separate Cesium from Nuclear Waste Saltcake", WSRC-TR-99-00358, July 1999.
   
2. L. B. Romanowski, "Low Curie Salt Processing Technical Plan", HLW-SDT-2002-00031, Rev. 0, March 18, 2002
   
3. D. L. Herting to N. W. Kirch, "Double-Shell Slurry Dissolution Kinetics", NHC Numatec Hanford Corporation, 82100-98-018, May 10, 1999.
   
4. D. T. Hobbs and C. J. Coleman, "Final Report: Analysis of Tank 41H Saltcake Sample #2 and Comparison to Sample #1 (U)", WSRC-TR-94-057, January 26, 1994.
   
5. High Level Waste Tank sample data base maintained by High Level Waste Engineering.
   
6. J. R. Hester, "High Level Waste Characterization System (WCS)", WSRC-TR-96-0264, Rev. 0, December 1996.
   
7. J. A. Pike, "Target Dilution Levels To Prevent Plugging During Tank 41 Saltcake Draining", HLW-PRE-2001-0014, Rev. 0, April 25, 2002.
   
8. "Statistical Methods of Uncertainty Analysis", National Institute of Standards and Technology, Information Technology Laboratory, http://www.itl.nist.gov/div898/carroll/u.htm.
   
9. D. G. Karraker, "Uranium Solubility Studies During Waste Evaporation (U)", WSRC-TR-93-433, Rev. 0, August 16, 1993.
   
10. D. L. Herting, "Saltcake Dissolution FY 2001 Status Report", HNF-8849, Rev 0., September 2001.
    
11. D. L. Herting, "Saltcake Dissolution FY 2000 Status Report", HNF-7031, Rev 0., September 2000.
    
12. D. L. Herting, et. el., "Saltcake Dissolution FY 1999 Status Report", HNF-5193, Rev 0., September 1999.
    
13. B. J. Wiersma, "An Investigation of Density Driven Salt Dissolution Techniques (U)", WSRC-TR-96-0160, August 1996.
    
14. D. L. Kiser, "Removal of Salt from High-Level Waste Tanks by Density-Driven Circulation or Mechanical Agitation", DP-1587, January 1981.
    
15. J. A. Pike, "Tank 37 Salt Dissolution Flowsheet Basis", WSRC-TR-2001-00599, Rev. 0, December 12, 2001.
    
16. J. C. Baily, Jr., to G. L. Albert, "Tank 10 Salt Removal Demonstration Detail", DPSP-79-17-11, April 11, 1979.
    
17. Several informal "Teletype" messages available in the HLWE File Room for Waste Removal about Tank 10 dissolution progress though most of this information is supported by the data in references 18, 19, and 20.
    
18. Waste Management Programs Report for April 1979, 79-21-4.
    
19. Waste Management Programs Report for May 1979, 79-21-5.
    
20. Waste Management Programs Report for June 1979, 79-21-6.
    
21. S. E. Aleman, "Calculation of the Mixture Density of Binary Aqueous and Simulated Waste Salt Solutions using Aspen Plus^® ElecNRTL and Aspen OLI™ Property Methods", SRT-EMS-2002-00002, January 18, 2002.