Michael R. Poirier
Westinghouse Savannah River Company
Aiken, SC 29808

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

This report has been reproduced directly from the best available copy.

Available for sale to the public, in paper, from:  U.S. Department of Commerce, National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161,  phone: (800) 553-6847,  fax: (703) 605-6900,  email:  orders@ntis.fedworld.gov   online ordering:  http://www.ntis.gov/support/ordering.htm

Available electronically at  http://www.osti.gov/bridge/

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

The proposed facility designs for the ion exchange and solvent extraction flowsheets under development to treat high level waste at the Savannah River Site use crossflow filtration to remove entrained sludge and monosodium titanate (MST). Bench-scale and pilot-scale testing performed with simulated feed streams showed much lower filtration rates than desired for the process. This report documents an investigation of the impact on filtration of using Honeywell sodium nonatitanate (ST), rather than MST, for strontium and actinide removal.

The author performed bench-scale dead-end filtration tests with 5.6 M sodium, average salt solution containing 0.6 g/L simulated sludge and 0.55 g/L MST or sodium nonatitanate. The results of the testing indicate the following.

• Testing identified no correlation between MST or sodium nonatitanate particle size and filter flux. Any potential filtration gains from differences in particle size between the MST and sodium nonatitanate appear offset by changes in filter cake porosity. The dispersion of the particle size likely contributes to this behavior.
• The sodium nonatitanate particles produced marginal improvement in filter flux (~ 30%). The improvement in filter flux proves less than previous gains by the addition of chemical additives to improve performance. The improvement does not return the flux to the target value.
• Based on this testing and the strontium and actinide adsorption testing performed by Hobbs, further filtration testing with sodium nonatitanate appears unwarranted at this time.

The Salt Disposition Systems Engineering Team selected three cesium removal technologies for further development to replace the In Tank Precipitation process: small tank tetraphenylborate (TPB) precipitation, crystalline silicotitanate (CST) ion exchange, and caustic solvent extraction.

A pretreatment step for both the CST and solvent extraction flowsheets contacts the incoming salt solution that contains entrained sludge with monosodium titanate (MST) to adsorb strontium and actinides. The resulting slurry is then filtered to remove the sludge and MST. To remove cesium, the respective processes treat the filtrate either by contacting with CST in an ion exchange column or by processing through a solvent extraction system.

The high level waste salt solution that will feed this solid-liquid separation process will contain approximately 5.6 M sodium with low levels of insoluble sludge (~600 mg/L).¹ The process will add MST or another adsorbent (~550 mg/L) to adsorb soluble strontium and actinide species. The sludge includes micron and submicron sized particles with a mean size of ~ 5 m. The MST specification requires <1 % of particles to be smaller than 1 m and < 1 % of particles to be larger than 35.5 m.² The MST samples in this testing had mean particle sizes of 5 – 10 m.

In tests performed by SRTC and the University of South Carolina, the filtration rates proved slower than desired for simulated salt solution containing varying concentrations of MST and sludge solids (0.02 – 0.08 gpm/ft² versus a goal of 0.25 gpm/ft²).^3,4,5

Honeywell continues efforts to develop commercial production of a sodium titanate material, sodium nonatitanate, which removes strontium and actinides from alkaline solutions. The vendor indicates that this material has a larger particle size than the MST. In a separate task, Hobbs will investigate the ability of sodium nonatitanate to remove strontium and actinides from high level waste salt solution.⁶ In contrast, this study satisfies a request to investigate methods to improve the separation of sludge and sorbent (i.e., sodium nonatitanate) solids from salt solution.⁷

The Kozeny-Carman model provides a simple description of colloidal fouling of microfilters.^8,9 The model is described by equation [1]

J = (-DP/L)[d_p² e³/150 m(1-e)²]      [1]

where J is the filter flux, DP is the transmembrane pressure, L is the cake thickness plus the filter thickness, d_p is the particle diameter, e is the filter cake porosity, and m is viscosity. According to equation [1], if all other parameters remain constant, an increase in particle diameter will increase the filter flux. Therefore, the larger particle size sodium nonatitanate should filter better than MST.

Table 1 shows the feed solution for these tests. The feed contained 5.6 M sodium salt solution with 0.6 g/L simulated Tank 40H sludge and 0.55 g/L MST or sodium nonatitanate.

Table 1. Feed Composition

The author performed the tests with a bench-scale dead-end vacuum filter. The tests placed a 300 mL sample of 5.6 M Na salt solution containing sludge and MST or sodium nonatitanate in a carboy and stirred using a magnetic stirrer. Personnel then poured the salt solution (100 mL) into the top of a graduated 115 mL capacity, 0.45 m m pore-size Nalgene disposable dead-end filter (Cat. No. 245-0045) connected to a rotary vacuum pump (Figure 1). The researcher started the pump and measured the filtrate volume as a function of time.

Table 2 shows results from previous filter tests, which filtered the feed solution with the dead-end vacuum filter and with a crossflow filter.^3,10 The baseline feed contained 6.4 M sodium salt solution with 0.6 g/L sludge, and 0.55 g/L MST. In some of the tests, personnel added bentonite, proprietary agent SRTC1, or proprietary agent SRTC2 to the feed to try to improve filter flux. In other tests, personnel processed KTPB slurries containing 4.7 M sodium, sludge, and MST to compare filter performance with that for the tetraphenylborate flowsheet. The results show the dead-end filter fluxes correlate well with crossflow filter fluxes, and the dead-end filter serves as a useful screening tool to evaluate the impact of changes in feed composition on crossflow filter flux.

The researchers collected samples of the MST and sodium nonatitanate used in this testing and submitted them for particle size analysis by Microtrac-SRA150. Table 3 and Figure 2 show results from the particle size analyses. All three of the sodium nonatitanate samples appear to contain approximately the same size particles. The MST-96QAB281-Old sample had a significantly larger particle size than the other two MST samples. It also had a bimodal particle size distribution. This sample came from archived material that had dried, with no free standing liquid. The drying caused agglomeration. This sample does not represent the MST the solid-liquid separation process will process, but the filtration results provide insight into the effect of particle size on filter flux. Table 3 and Figure 2 also show the particle size of the simulated Tank 40H sludge used in this testing. The mean sludge particle size measured 2 m.

Table 4 and Figure 3 show the results from the dead-end filter tests. Personnel performed each test three times, with standard deviations of less than 6%.

Although samples ST-73A, ST-73B, and ST-01520 had the same particle size, ST-73B produced the largest filter flux of all slurries tested. MST-33180 has a smaller particle size than MST-96QAB281, but the filter flux proved much higher. The average filter flux with the three MST-containing feeds measured 0.68 mL/min. The average filter flux with the three sodium nonatitanate-containing feeds measured 0.88 mL/min, a 30% improvement.

The variation in filter flux between different feed solutions appears larger than the variations in filter flux due to experimental uncertainty, which means the difference in filter flux between feeds is real. To evaluate that hypothesis, the author performed an analysis of variances (ANOVA) to compare the differences in filter flux between different feed solutions with the differences in filter flux measured during replicate tests with the same feed. Table 5 shows the ANOVA results. In the table, SS is the sum of the squares of the deviation from the mean, df is the degrees of freedom, MS is the mean square deviation, F is the ratio of the mean square between feed solutions to the mean square within feed solutions, and F_crit is the value of F which determines whether the differences in filter fluxes from different feed solutions prove significant.¹³ If the errors remain independent and identically distributed according to a normal distribution with zero mean and fixed variance, the F-ratio has a distribution that varies with the number of degrees of freedom. If F < F_crit, the differences in filter flux for different feed solutions are not significant. If F > F_crit, the differences in filter flux for different feed solutions are significant. Since F is 104.3 and F_crit is 2.85, the differences in filter flux measured with the different feed solutions are significant.

Figure 4 shows the filter flux as a function of particle size. From the figure, no relationship appears to exist between particle size and filter flux. To evaluate that hypothesis, the author also performed a regression analysis to determine if any relationship exists between particle size and filter flux.¹³ Table 6 shows the results.

Table 6 compares the variance due to the deviations predicted by the regression model with variances due to differences between predicted filter fluxes and measured filter fluxes (residual). The table shows the variance due to differences between the predicted and measured filter fluxes are much larger than variance in flux due to the regression model. Therefore, no correlation exists between the MST or sodium nonatitanate mean particle size and filter flux.

This finding conflicts with the Kozeny-Carman model (equation [1]) which predicts filter flux increases with increasing particle size, assuming all other parameters remain constant. The model’s assumptions may not be valid. Specifically, the model assumes the cake formed by each feed solution has the same porosity or void fraction. The different MST and sodium nonatitanate samples might pack differently and have different void fractions due to the dispersion in particle sizes. According to the Kozeny-Carman model, the filter flux varies with porosity to the third power (J a e³). If the particle size increased by x and the porosity decreased by x^2/3, one would observe no change in filter flux. The model also assumes monodisperse particle sizes. The tested particles have a range of sizes. Because of differences in particle size, the packing may differ from that observed for a single particle size thus changing the filter cake porosity. The small fines may become trapped in the interstitial volume between the larger particles, which would reduce cake porosity and filter flux. The Kozeny-Carman model assumes hard spheres for the particles. The tested particles are not spheres. If the particles interact or compress, the model would not account for differences in particle-particle interactions and cake compressibility with the different particles.

The testing provides the following conclusions.

• Testing identified no correlation between MST or sodium nonatitanate particle size and filter flux. Any potential filtration gains from differences in particle size between the MST and sodium nonatitanate appear offset by changes in filter cake porosity. The dispersion of the particle size likely contributes to this behavior.
• The sodium nonatitanate particles produced marginal improvement in filter flux (~ 30%). The improvement in filter flux proves less than previous gains by the addition of chemical additives to improve performance. The improvement does not return the flux to the target value.
• Based on this testing and the strontium and actinide adsorption tests performed by David Hobbs, further testing of sodium nonatitanate filtration appears unwarranted at this time.

1. R. A. Dimenna, O. E. Duarte, H. H. Elder, J. R. Fowler, R. C. Fowler, M. V. Gregory, T. Hang, R. A. Jacobs, P. K. Paul, J. A. Pike, P. L. Rutland, F. G. Smith III, S. G. Subosits, and G. A. Taylor, "Bases, Assumptions, and Results of he Flowsheet Calculations for h Decision Phase Salt Disposition Alternatives", WSRC-RP-99-00006, Rev. 0.
2. "Procurement Specification for Monosodium Titanate", Specification No. Z-SPP-H-00001, Rev. 2, May 1992.
3. H. H. Saito, M. R. Poirier, S. W. Rosencrance, and J. L. Siler, "Improving Filtration Rates of Mono-Sodium Titanate (MST)-Treated Sludge Slurry with Chemical Additives", WSRC-TR-99-00343, September 15, 1999.
4. H. H. Saito, M. R. Poirier, and J. L. Siler, "Effect of Sludge Solids to Mono-Sodium Titanate (MST) Ratio on MST-Treated Sludge Slurry Cross-Flow Filtration Rates", WSRC-TR-99-00342, September 15, 1999.
5. Ralph Haggard, Carol Stork, Travis Deal, and Vince van Brunt, "Final Report on the Crossflow Filter Testing for the Salt Disposition Alternative", USC-FRED-PSP-RPT-09-0-010, Rev. 0, December 4, 1998.
6. D. T. Hobbs, "Task Technical and Quality Assurance Plan for Phase V Strontium and Actinide Removal Testing", WSRC-RP-99-01080, December 20, 1999.
7. R. A. Jacobs, Technical Task Request, HLW-SDT-TTR-99-30.0, December 1999.
8. C. O. Bennett and J. E. Myers, Momentum, Heat, and Mass Transfer, 3^rd Ed., New York: McGraw-Hill, 1982, p. 229.
9. T. Carroll, S. King, S. R. Gray, B. A. Bolto, and N. A. Booker, "The Fouling of Microfiltration Membranes by NOM after Coagulation Treatment", Wat. Res. Vol. 34, no. 11, pp. 2861-2868, 2000.
10. M. R. Poirier and J. L. Siler, "Continuous Concentration and Constant Volume Washing of Tetraphenylborate Slurries", WSRC-TR-99-00243, August 5, 1999.
11. D. T. Hobbs, Laboratory Notebook, WSRC-NB-2000-00063.
12. M. R. Poirier, Laboratory Notebook, WSRC-NB-2000-00044.
13. G. E. P. Box, W. G. Hunter, and J. S. Hunter, Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building, New York: Wiley, 1978.