WSRC-TR-2001-00413

Flocculating, Settling, and Decanting for the
Removal of Monosodium Titanate and Simulated
High-Level Waste Sludge from Simulated Salt Supernate

C. J. Martino, M. R. Poirier, F. F. Fondeur, and S. D. Fink
Westinghouse Savannah River Company
Aiken, SC 29808

This document was prepared in conjunction with work accomplished under Contract No. DE-AC09-96SR18500 with the U.S. Department of Energy.

DISCLAIMER

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/index.html

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

Keywords: Salt Processing Program, Flocculants, Turbidity

1 Summary

We investigated addition of commercial flocculating agents to improve the gravity separation of 0.6 g/L sludge simulant and 0.55 g/L monosodium titanate (MST) solids from simulated 5.6M Na+ high level waste (HLW) salt solution. The testing provided the following conclusions:

2 Introduction

2.1 Purpose

The Salt Processing Program (SPP) evaluated three potential cesium removal technologies: caustic-side solvent extraction (CSSX), small tank tetraphenylborate precipitation (STTP), and crystalline silicotitanate (CST) ion exchange. The Department of Energy selected CSSX as the preferred cesium removal technology for Savannah River Site HLW.

A pretreatment step for the CSSX flowsheet includes contacting the incoming salt solution that contains entrained sludge with monosodium titanate (MST) to adsorb strontium and actinides followed by filtration to remove the sludge and MST. The filtrate passes through the solvent extraction system to achieve cesium removal. Testing performed by SRTC and the University of South Carolina with simulated waste showed filtration rates of 0.03 – 0.08 gpm/ft2.,,, Because of the low cross-flow filtration rates, the High Level Waste Process Engineering (HLW-PE) requested that researchers at the Savannah River Technology Center (SRTC) investigate methods to improve the separation of sludge and MST solids from high level waste salt solution. Item 6.5 of the scope of work matrix requests that SRTC evaluate alternative solid-liquid separation technologies and test promising alternatives. SRTC investigated alternative solid-liquid separation processes and recommended several for further study. Among the conclusions, SRTC resolved to investigate settling and decanting followed by polishing filtration.

This study examined enhancement of the settling rate of the bulk of the MST and entrained sludge through the addition of flocculating agents, and measured polishing filtration fluxes for the decanted supernate. The study assumed an approximate goal for insoluble solids removal, as defined by HLW-PE, of 99.5 % (see Attachment 1). For a feed with 1.15 g/L insoluble solids (0.6 g/L sludge and 0.55 g/L MST), the target solids loading in the clarified supernate equals 5.75 mg/L.

2.2 Previous Studies

Researchers previously examined the use of countercurrent decantation (CCD) as an alternative to the current extended sludge process (ESP). Various CCD studies included measurements of the settling of real and simulated high-level waste sludge.,,, Settling rates obtained in the previous studies not using chemical additives proved too slow for practical application to the SPP. Several of the previous studies examined the use of commercially available flocculants to enhance settling of real and simulated HLW sludge.,,, Ciba Alclar® flocculants improved sludge settling. One previous study includes a detailed analysis on the applicability of flocculants to the settling of WSRC sludge.

Recent SRTC studies examined use of flocculants as a filtration aids., Cytec Superfloc® hydroxamated amines and Ciba Alclar® flocculants show an approximate three times improvement in the dead-end vacuum filtration rate. Much smaller improvements (only up to 1.3 times improvement) were observed during cross-flow filtration tests, likely due to floc breakup in this high-shear system. The studies did not include information on the size and settling rate of flocs or the optimization of mixing time and rate, but did explore the effect of flocculant dose as it affected filtration of the slurry without settling and decanting.

2.3 Approach

The experimental approach involved several phases of experiments: to identify the most promising flocculants and doses, to examine the selected flocculant in an experimental configuration similar to one possible process design, and to assess the filtration rate improvements following settling to remove the majority of the insoluble solids. The single-flocculant and head-to-head tests examine the effect of the choice of flocculants and flocculant doses, the heel tests explore the influence of retaining solids in the processing vessel, and the dead-end filtration tests probe the application of polishing filtration to remove the residual solids from the supernatant solution.

3 Experimental

3.1 Simulant Preparation

Testing used the SRS average 5.6 M Na+ salt solution applied as a standard in much of the work for the SPP. Table 1 contains the composition of the feed salt simulant. The density of the unfiltered salt solution measured 1.25 g/mL. This compares well with the calculated density of 1.26 g/mL. Except when otherwise noted, we added sludge and MST at insoluble solids concentrations of 0.60 g/L and 0.55 g/L respectively, for a concentration of total insoluble solids of 1.15 g/L. We added simulated Tank 40H sludge to approximate the entrained sludge solids expected during processing. The SRS average salt solution itself contained solids prior to addition of the sludge and MST.

3.2 Flocculant Preparation

We used two families of flocculants in this study, Superfloc® HX flocculants manufactured by Cytec and Alclar® flocculants manufactured by Ciba. Both of these flocculating agents are designed for enhancing settling in the aluminum processing industry, where the treated solutions have a high pH and a high ionic strength. Alclar 600 is a homopolymer of sodium acrylate and Alclar 662 is a copolymer of acrylamide and sodium polyacrylate. The Superfloc® HX flocculants are terpolymers with acrylic acid, acrylamide, and hydroxamic acid functional groups. The main differences between the different Superfloc® HX formulations used in this study involve the polymer molecular weights, the charge densities, and the emulsifier formulation. The proprietary nature of the flocculant molecular weights and charge densities prevented Cytec from releasing these details to us.

Table 1: Salt Solution Composition

Component

Concentration (M)

Na+

5.60

K+

0.015

Cs+

0.00014

OH-

1.91

NO3-

2.14

NO2-

0.52

AlO2-

0.31

CO32-

0.16

SO42-

0.15

F-

0.032

Cl-

0.025

PO43-

0.010

SiO32-

0.004

C2O42-

0.004

MoO42-

0.0002


We received the Superfloc® HX flocculants as emulsions and the Alclar® flocculants as powders. The two forms of flocculants require different preparations. For emulsion flocculants, experimenters use a 1-mL disposable syringe to add 0.2-0.5 grams of the emulsion to 100 g of deionized (DI) water in a 200-mL beaker. At the time of addition, personnel stirred the water to provide a strong vortex and added the flocculant drop-wise to the high-shear region. We then mixed this stock solution for at least 30 minutes. We prepared the powdered flocculants by adding 0.2 – 0.5 grams of flocculant to a 200-mL glass beaker and wetting with approximately 1 mL of ethanol. We added 100 g of water and mixed the solution for at least 30 minutes. Personnel noted any lack of homogeneous appearance in the stock solution and discarded any solution that appeared irregular. Additionally, we noted that the viscosities of these concentrated solutions increased after adequate mixing. Personnel covered the beakers with laboratory film when we used longer than 30 minutes for mixing.

We prepared further dilutions from the stock solutions. Using the example of single-flocculant testing, we added five different amounts of stock solution (i.e., in the range of 4 to 50 grams) to five 100-mL beakers. We added DI water to each beaker to obtain a total solution mass of 50 g. A sixth beaker contained only DI water. Personnel stirred the beakers for at least 10 minutes, and then filled 10-mL disposable syringes with each of the six solutions. We used the syringes to introduce the flocculant into the appropriate solutions during settle and decant testing. Personnel discarded all stock and diluted flocculant solutions not consumed fully within 10 contiguous hours of testing.

3.3 Settle and Decant Tests

The flocculant-selection phase allowed us to determine the best flocculant and the optimal dose for use in settling MST and HLW sludge. We used the supernate turbidity as the criterion. The turbidity data allows estimation of the solids removal efficiency for settling and decanting. The settling time proved less important because the floc settling time is orders of magnitude shorter than the unflocculated solids settling time.

An ASTM standard for the jar testing of flocculation in water served as a basis for the flocculation and settling experiments. This method suggests simultaneously adding flocculants to several suspensions while mixing at a fast speed, followed by periods of slower mixing and quiescent settling.

Figure 1: Flocculate, Settle, and Decant Tests using a Phipps 

 and  BirdTM Model 400/500 Six-Position Stirrer and Illuminator. 

  Shown is the Slow-Mix Stage of the HX-200 Test.
Figure 1: Flocculate, Settle, and Decant Tests using a Phipps & BirdTM Model 400/500
Six-Position Stirrer and Illuminator. Shown is the Slow-Mix Stage of the HX-200 Test.

For water testing, the method recommends collecting data comparing the time of first flocculation, floc size, settling rate (or time), turbidity, color, and pH. We performed experiments for flocculant and dose comparisons using a Phipps & BirdTM Model 400/500 six-position stirrer and illuminated base (see Figure 1). This equipment allows for easy visual comparisons between performance of different flocculants and doses to corroborate the more quantitative methods also employed. The six speed-controlled agitators are 1-inch tall by 3-inch wide flat plates oriented vertically at the end of each shaft. The base illuminates the solution from below, enhancing the ability to make qualitative observations.

We performed two types of experiments during the flocculant evaluation: single-flocculant and head-to-head. The single-flocculant tests used all six of the stirrer positions to evaluate a single flocculant at different flocculant doses (from 0 to approximately 20 mg/L). The head-to-head tests compared either two or three different flocculants at concentration levels optimized in the single-flocculant tests. For both types of tests, personnel randomized stirrer positions to reduce effects of the timing of flocculant addition and supernate decanting, and any unexpected position effects. We minimized use of the illuminator to reduce the potential for solution heating and induced convection. The solution temperature remained stable at ambient temperature, 22 to 25 ° C, throughout each test.

Through this phase of experiments, researchers determined the flocculant best suited for this system, as well as the optimum dose. The reader should note that we did not optimize all of the conditions in these tests. The flash-mix and slow-mix speeds, as well as the mixing volume and geometry, remained constant in this study. The optimum condition will depend on the size of the equipment in the final application. The mixing speeds could also affect the optimum flocculant dose by affecting the growth and breakdown of flocs.

After completing the evaluation tests, we selected flocculants and doses for use in the settle and decant heel-tests and the polishing-filtration tests.

3.3.1 Single-Flocculant Tests

Personnel filled six 1 L PyrexTM No. 1003 thick-walled glass beakers with the salt/sludge/MST simulant, arranged within the multi-stirrer, and mixed at the flash-mix rate of 180 RPM for 5 minutes. We numbered the beakers, syringes, and stirrer positions from left to right. We performed all of the operations in numerical order that we could not perform simultaneously. At the start of the experiment, we quickly emptied each of the six 10-mL syringes into the vortex of each corresponding beaker. The flocculant addition to all six beakers occurred over a period of about 10 to 15 seconds, after which we started the stopwatch. We measured the mass of each syringe before and after expelling the flocculant to determine the mass of flocculant added.

Flash mixing at 180 RPM continued for one minute. During this period, the experimenters watched the beakers to determine the first appearance of flocculation. After one minute, we reduced the mixing rate to 75 RPM and mixing continued for an additional nine minutes. Approximately halfway into this period, the researchers took digital photographs to record the flocculation progress

After a total of 10 minutes of mixing, we stopped the agitation, removed the paddles from the solution, and allowed the flocs to settle for 10 minutes. As settling began, personnel estimated the typical flock size by comparison with a ruler held against the beaker exterior. The settling time proved difficult to estimate because of the lack of a tangible heel–supernate interface during settling. Thus, the definition of settling time used in this report provides a rough visual estimate of when roughly 99% of the representative flocculated material settled to the layer of solids at the bottom of the beaker. There was approximately 4.75 inches of supernatant fluid above this relatively thin (<0.25 inches) heel of solids. At approximately five minutes into the settling period, we took digital photographs to record the settling progress.

Upon completion of settling, researchers noted the visual appearance of the treated supernate. We used a transfer pump to sample fluid from a level approximately 100 – 200 mL from the bottom of the 1L beaker without disturbing the settled cake. First, we transferred approximately 100 mL of supernate to a waste reservoir to flush the pump and transfer lines adequately. We then filled two vials with approximately 30 mL of supernate for turbidity measurement. We repeated the supernate sampling for the remainder of the beakers, and measured the turbidity for each set of two samples.

3.3.2 Head-to-Head Tests

Head-to-head tests facilitated comparison of different flocculants near their optimum doses during the same experiment. The experimental procedure for the head-to-head tests is largely the same as that for the single-flocculant tests, with several key differences. We prepared flocculants in narrower ranges of concentrations based on the optimum results in the previous single-flocculant tests. We synthesized stock solutions for more than one flocculant at the beginning of the tests and sometimes a greater time passed (up to 5 hours) before using the solutions. During the head-to-head tests, researchers used a modified transfer and turbidity sampling method. Instead of the transfer pump, personnel used 25-mL pipettes to transfer aliquots of the solution from a level 100 – 200 mL from the bottom of the beakers into vials for turbidity measurement. This allowed sampling after settling periods of 2 minutes, 10 minutes, and 50 minutes. The turbidity data from these samples provide a quantitative indication of settling as a function of time and allow us to project the data collected after 10 minutes of settling to shorter or longer settling times.

3.3.3 Heel Tests

The heel test investigated the stability of the flocculant heel and evaluate the influence of a heel upon subsequent flocculation, settling, and decant cycle. Compared to the supernatant liquid after settling, the heel of settled solids proved substantially smaller in volume (about 2% of the total volume) and would be difficult to remove in these small quantities from the full-scale process. Hence a preferable process would likely accumulate solids from several cycles prior to transfer of the cumulative solids heel out of the settling tank.

Each heel test used an individual flocculant at a single dose. We discuss variations in the heel testing method in the Results and Discussion section. Personnel prepared a batch of flocculant using a dual-dilution method, in which we prepared a concentrated stock solution and further diluted to the target flocculant concentration. We stirred the solution while covered during the course of the 8-hour-long experiment to reflect actual tank operation and as a precaution to minimize any potential loss of significant volatiles.

The heel test allows a heel of solids to accumulate in a beaker over several settle and decant cycles, with this heel repeatedly mixed with new feed. Personnel added flocculant to 1L of solution and flash mixed at 180 RPM for 1 minute in a 1L beaker. We then transferred this solution into a 2L cylindrical-walled beaker and slow mixed at 80 RPM for 9 minutes. After settling for 10 minutes, we decanted the supernatant fluid with a transfer pump leaving the "heel" of solids behind. We repeated this procedure a total of 10 to 12 times (cycles), leaving the solids to accumulate in the 2L beaker. The heel test retained many of the same features of the previous tests: we monitored the time for 99% of the flocs to settle, measured the floc size against a ruler, and filled two turbidity vials while decanting.

We retained the decanted supernates from the flocculant tests in numbered 1L plastic bottles for future use in the dead-end filtration tests and residual solids analysis. At the completion of the test, personnel removed the solid heel, weighed, and retained. We gravimetrically determined the weight percent insoluble solids for the heel by driving off the free water with a microwave.

3.4 Dead-end Filtration Tests

We filtered the supernates from the settle-and-decant heel tests utilizing a small dead-end filter. The demonstration used the designed filter flux of 0.25 gpm/ft2. Operation included cleaning the filter via back-pulsing when the trans-membrane pressure (TMP) increased to 30 psi, a TMP similar to that encountered in cross-flow filtration. We expected that the time between back pulses would be significantly lengthened over that for the original salt/sludge/MST solution because of the settling of the flocculated solids from the feed supernate. We use the time between back-pulses as a comparison between the different supernates tested.

Figure 2 shows a schematic of the dead-end filtration system constructed at 679-T. The system features a Mott® 6200 Series 0.5-mm porous stainless steel filter with a 4.8 in2 surface area. The feed vessel is a 1L glass beaker on a magnetic stir plate with a stir bar. The solution is fed to the FMI® metering pump (Model QV-1) via polyethylene and stainless steel tubing. We set the stroke direction to forward, the stroke length to 50% and the stroke rate at 41.0%, corresponding to a flow rate of approximately 32 mL/min. For these filter dimensions, this corresponds to the desired filter flux of 0.25 gpm/ft2. Downstream of the pump, the design includes a relief valve set to 100 psi to ensure that the system will not over-pressurize. The Mott® filter element is a cartridge contained in a stainless steel housing. The configuration includes calibrated 100psi pressure gauges installed on the inlet and outlet ports of the filter housing. The difference between the readings of these two gauges corresponds to the pressure drop across the filter element, or the trans-membrane pressure (TMP). After the filter, the design includes an approximately 100-mL filtrate reservoir to facilitate filter cleaning through filtrate back-pulsing. We used a back-pulse duration of 3 seconds and a back-pulse pressure of 50 psi. A three-way valve allows for either the flow of filtrate to a collection vessel (a 1L polyethylene bottle) or from the back-pulse pressurization system. A rotameter included at the filtrate output assures constant flow rate and allows personnel to check for flow anomalies. Researchers determined flow rates every 15 minutes by timing the collection 10 mL of filtrate in a 10mL volumetric flask.

Figure 2: Schematic of the Dead-end Filtration Equipment.
Figure 2: Schematic of the Dead-end Filtration Equipment.

At the start of testing, we wetted the filter by passing about 100 mL of salt solution through the filter. Next, personnel poured the test solution into the feed vessel and mixed by a magnetic stirrer. We recorded the TMP every 1 to 5 minutes and checked the flow rate every 30 minutes. We examined the filtrate turbidity at least once per filtration experiment.

3.5 Analytical Methods

3.5.1 Turbidity

The turbidity, or cloudiness of the solution, is commonly used for comparison purposes. Standard methods exist for quantifying the turbidity of clear liquids. We used an Orbeco-Hellige Model 965-10A turbidity meter to evaluate the turbidity in 30 mL aliquots of solution. This turbidity meter shines visible light through the side of a cylindrical sample vial and detects the light scattered through the side of the vial 90° from the source. The turbidity, measured in nephelometric turbidity units (NTU), provides a quantitative comparison tool between samples with different amounts of solids. The turbidity correlates with the solids content for a given solution composition, but the correlation degrades when large amounts of solids cause the solution to become opaque.

We calibrated the turbidity meter daily during use. The manufacturer recommends a two-point calibration with a 0 NTU standard (0 NTU is defined as completely transparent, with no scattered light. The black, opaque 0 NTU calibration standard simulates 0 NTU by eliminating the scattered and reflected light.) and a stable 40 NTU polymer solution. Additionally, we measured the turbidity of several reference solutions to track the performance of the instrument, as seen in Figure 3. The salt 5.6 M Na+ salt solution with 0.6 g/L sludge and 0.55 g/L MST had an initial turbidity of 735 ± 12 NTU. The turbidity of the salt solution with 0.6 g/L of sludge measured 717 ± 17 NTU while the solution with 0.55 g/L MST had a turbidity of 452 ± 19 NTU. The turbidities of these reference solutions did not show any trend with time. We obtained two samples of salt solution from a 20L carboy, one taken from the agitated carboy and the other obtained after the solids settled for one week. The turbidities of the salt solution varied with time. From April 2 to May 15, the turbidity of the mixed salt solution doubled from 22.1 NTU to 44.5 NTU. Turbidities measured much lower in the salt solution in which the solids were allowed to settle, but a similar increase was noted, as seen in Figure 3. Investigation revealed that the increase in measured turbidity in the salt solutions does not result from an equipment or calibration problem, but rather provides an accurate measure of the increased cloudiness within the salt solutions. We discuss the reasons behind the steady increase in turbidity of these salt solutions in Section 0. Due to the quick settling of these solids formed in the salt solutions, they have little impact on the results of the settling and decanting tests.

A calibration plot of the turbidity of solutions with varying amounts of solids allows us to estimate the amount of solids in the supernatant fluid. Figure 4 contains a plot of the measured turbidity for several concentrations of sludge and MST. The initial solution with 0.6 g/L sludge and 0.55 g/L MST is represented by the rightmost point, 1150 mg/L solids. We measured lower solids concentrations by diluting the initial solution with settled salt solution (with an approximate turbidity of 2 NTU). For a large solids concentration range the turbidity to solids content relationship is linear. This relationship becomes nonlinear at very high solids content due to the opacity of the solution and at very low concentrations due to the turbidity of the salt solution used for dilution (approximately 2 NTU).

Figure 3: Turbidity of Reference Solutions during Settle and Decant Testing.
Figure 3: Turbidity of Reference Solutions during Settle and Decant Testing.

 

Figure 4: Turbidity Measurements of Dilutions made from 0.6 g/L Simulated Sludge and 0.55 g/L MST in 5.6 M Salt Solution.

Figure 4: Turbidity Measurements of Dilutions made from
0.6 g/L Simulated Sludge and 0.55 g/L MST in 5.6 M Salt Solution.

Using the data in Figure 4, we can relate the solids content to the turbidity assuming that the ratio of sludge to MST in the samples remains unchanged during settling (a sludge to MST ratio of 12:11). The line in Figure 4 is the relationship in Equation 1 derived from regression of the data over the turbidity range of 12 to 106 NTU (solids concentration range of 5 to 43 mg/L).

Solids concentration (mg/L) = 0.404 * Turbidity (NTU) (1)

We expect that Equation 1 would over-predict solids concentrations at turbidities below 12 NTU and under-predict solids concentrations at turbidities above 106 NTU.

3.5.2Spectroscopy

We analyzed the solids in supernates decanted from the final cycles of the heel tests. We filtered the supernates with 0.02 m m filter paper (that does not contain soluble organics or low molecular weight components). Personnel then coated the filter paper with a carbon film (by sputtering a graphite rod) to make the sample conductive. We placed the coated samples (of filter paper plus solids) into an ISI DS 130 infrared spectroscopy unit equipped with the VantageTM system (from Noran). In the instrument, we obtained a backscattered and secondary electrons image from the sample (that provides a topographical picture of the sample). Using energy dispersive spectroscopy (EDS), personnel analyzed the X-rays given off by the sample to determine the elemental composition of the sample. In a parallel study, we pressed the filter paper onto a ZnSe crystal in a ATR (attenuated reflectance mode) configuration and bombarded with modulated heat (< 0.01 microwatts). Infrared (IR) spectral analysis of the heat after interacting with the sample provided the molecular composition of the sample. In some cases, we washed the filter with water to remove excess salts. Examination after washing proved that the presence or absence of salt did not alter the conclusions.

4 Results and Discussion

4.1 Selection of Flocculants

The first phase of this study involved the comparison of the four Superfloc® HX and three Alclar® flocculants considered. We tested the performance of each of these flocculants in a series of single-flocculant tests, examining a range of five doses of the same flocculant simultaneously. We then examined the identified optimal doses in head-to-head comparison tests, testing several different flocculants within the same experiment.

Table 2 and Table 3 contain the results of the experiments using the Superfloc® HX and Alclar® flocculants, respectively. The tables contain the amount of flocculant added in both the mass of original flocculant material per volume of solution and the weight percent based on the mass of flocculant per mass of insoluble solids. These tables contain turbidity, removal efficiency, settling time, and floc size results for both the single-flocculant and head-to-head tests. In some cases, we record a settling time of zero when the bulk of the flocculated solids settled to the bottom of the beaker during the slow-mix stage. Often, the majority of the flocs settled to the bottom of the beaker in a short time (i.e., 10 seconds). A portion of the flocs required longer to reach the bottom of the beaker, and we recorded this value as the settling time. The settling times, however, prove several orders of magnitude shorter than encountered with no flocculant added.

Table 2: Data for Tests of Cytec Superfloc® HX Flocculants.

Flocculant
Name

Flocculant Amount
(mg/L)

Flocculant Amount
(wt% solids)

Turbidity
(NTU)

Removal Efficiency
(%)

Approx.
Settling Time
(s)

Approx.
Floc Size
(mm)

HX-200

1.7

0.15

50.1

98.0

140

< 1

HX-200

3.2

0.27

62.0

97.6

n.d.

1

HX-200

3.3

0.28

34.9

98.6

70

1

HX-200

5.2

0.45

33.4

98.7

62

1.5

HX-200

5.9

0.52

26.8

99.0

30

1

HX-200

7.4

0.64

33.0

98.7

n.d.

3

HX-200

9.6

0.83

39.8

98.4

15

3 to 5

HX-200

10.9

0.94

26.6

99.0

30

2

HX-200

19.8

1.72

35

98.6

30

3

HX-300

1.9

0.16

38.2

98.5

n.d.

< 1

HX-300

2.2

0.19

47.8

98.1

n.d.

< 0.5

HX-300

3.5

0.30

24.6

99.0

120

1

HX-300

4.4

0.38

30.2

98.8

n.d.

1.5

HX-300

6.4

0.55

20.9

99.2

30

1 to 2

HX-300

6.6

0.58

33.0

98.7

24

3 to 10

HX-300

8.8

0.76

30.3

98.8

15

2 to 5

HX-300

11.4

0.99

34.4

98.7

10

3 to 15

HX-300

21.0

1.82

39.0

98.5

10

3 to 15

HX-400

2.0

0.17

52.4

98.0

140

< 0.5

HX-400

3.7

0.32

42.9

98.3

140

< 1

HX-400

6.6

0.57

30.7

98.8

105

1 to 2

HX-400

9.3

0.81

28.9

98.9

23

1 to 1.5

HX-400

11.6

1.00

25.8

99.0

50

2

HX-400

12.4

1.07

27.1

98.9

45

1

HX-400

12.9

1.13

27.9

98.9

62

1.5

HX-400

15.9

1.38

32.6

98.7

35

2

HX-400

17.7

1.54

27.5

98.9

45

2

HX-400

18.3

1.59

26.4

99.0

15

2 to 3

HX-400

23.7

2.06

29.4

98.9

n.d.

3

HX-2000

2.0

0.17

47.9

98.1

60

1

HX-2000

3.6

0.31

36.0

98.6

30

1 to 2

HX-2000

6.5

0.57

50.5

98.0

0

15

HX-2000

10.7

0.93

56.5

97.8

0

15

HX-2000

22.0

1.91

50.3

98.0

0

20 to 30

n.d. = not determined

 

Table 3: Data for Tests of Ciba Alclar® Flocculants.

Flocculant
Name

Flocculant Amount
(mg/L)

Flocculant Amount
(wt% solids)

Turbidity
(NTU)

Removal Efficiency
(%)

Approx.
Settling Time
(s)

Approx.
Floc Size
(mm)

Alclar 600

1.2

0.10

64.2

97.5

0

3 to 5

Alclar 600

2.2

0.20

47.0

98.2

0

3 to 10

Alclar 600

3.4

0.30

28.6

98.9

n.d.

2

Alclar 600

3.5

0.30

63.0

97.5

0

10

Alclar 600

4.6

0.40

66.0

97.4

0

10

Alclar 600

6.8

0.59

51.5

98.0

n.d.

10

Alclar 600

13.5

1.18

125.8

95.1

n.d.

2 to 4

Alclar 662

2.1

0.18

86.4

96.6

0

large

Alclar 662

3.8

0.33

74.5

97.1

0

large

Alclar 662

7.0

0.61

67.5

97.4

0

large

Alclar 662

12.8

1.11

75.5

97.0

0

large

Alclar 662

23.1

2.01

95.8

96.3

0

large

Alclar W23

3.4

0.29

55.9

97.8

n.d.

2 to 20

Alclar W23

6.7

0.59

145.6

94.3

n.d.

2 to 4

Alclar W23

13.4

1.17

113.0

95.6

n.d.

2

n.d. = not determined

We used the supernate turbidity after 10 minutes of settling as the main criterion for comparing these tests. The turbidity of the control samples, with no added flocculant, measured 730 ± 12 NTU. As expected, this turbidity value proved indistinguishable from the turbidity of the original salt/sludge/MST solution (735 ± 12 NTU). We calculated the solid removal efficiency of the flocculate, settle, and decant method by comparing the estimated solids concentration in the decanted solid removal efficiencies for the single-flocculant and head-to-head tests. Solid removal efficiencies were typically 98% to 99% for most of the HX-200, HX-300, and HX-400 tests. We were not able to attain the required solids removal efficiency of 99.5% during the flocculant-selection phase of this study.

Figure 5 contains the supernate turbidity data for the single-flocculant tests with the HX flocculants, with the error bars being the span of the turbidity data (except for the HX-200 data, where the larger error bars reflect an analytical difficulty). Viewing the minima of the turbidity versus flocculant concentration plots, one clearly sees a difference in the optimal doses for the four HX flocculants used. The HX-2000 proved effective at the lowest dose, followed by HX-300 and HX-200. The HX-400 proved most effective at the upper end of the dose range explored during the single-flocculant tests. However, HX-2000 had the narrowest flocculant concentration range for optimal performance – as determined by turbidity of the solution – and did not remove as large a portion of the solids as the other additives. The most effective solids removal occurred using HX-300 at a dose of 6.4 mg/L (0.55 wt % based on solids).

In general, we observed expected dose phenomena for the HX flocculants. We obtained the clearest supernates (supernates with the lowest turbidity) at an intermediate flocculant doses within the range of study (usually 2 to 20 mg/L). The optimum dose is the dose that results in the minimum supernate turbidity. With less than the optimum dose added, flocs tended to be smaller, with a greater cloudiness to the bulk supernatant fluid. When the concentration exceeded an optimum dose, large flocs formed, often resulting in settling during the slow-mixing period (such as seen for HX-2000 at doses greater 6.5 mg/L). The supernates from solutions over-dosed with flocculant also had many small flocs (>0.5 mm) that did not settle within the 10 minutes.

Figure 5: Dose Optimization Tests for Cytec Superfloc(R) HX 

  Flocculants.
Figure 5: Dose Optimization Tests for Cytec Superfloc® HX Flocculants.

 

Figure 6: Head-to-Head Comparison Tests for Cytec Superfloc(R)

  HX Flocculants.
Figure 6: Head-to-Head Comparison Tests for Cytec Superfloc® HX Flocculants.

 

Figure 7: Dose Optimization Tests for Ciba Alclar(R) Flocculants.  Additional HX-400 Test Included for Comparison.
Figure 7: Dose Optimization Tests for Ciba Alclar® Flocculants.
Additional HX-400 Test Included for Comparison.

We conducted a head-to-head test of HX-200, 300, and 400 using flocculant doses near those optimized for each flocculant in the single-flocculant tests. Figure 6 contains the supernate turbidity versus flocculant dose data for this test (for 10 minutes of settling). Under the conditions of this test, HX-400 outperformed the other HX flocculants with supernate turbidities of less-than 30 NTU for the entire dose range investigated. The insensitivity of the dose on the effectiveness of the flocculant represents an advantage of HX-400 evident from this test.

The three Alclar® flocculants proved less effective than the HX flocculants for the clarification of sludge and MST from salt solution. As evident in Figure 7, only Alclar 600 had a single experiment with turbidities comparable to those attainable with the HX flocculants. This result observed for the initial tests did not occur in tests with Alclar 600 conducted over dose regime surrounding this favorable result.

Figure 8 contains turbidity data taken during the head-to-head tests. During those experiments, we used pipettes to obtain turbidity samples after three different settling times: 2 minutes, 10 minutes, and 50 minutes. The reduction in turbidity between 2 and 10 minutes and 10 and 50 minutes was small in comparison with the overall reduction from the original 735 NTU solution that occurred in the first 2 minutes. Even though all of the single-flocculant tests and heel-tests used 10 minutes of settling time, 2 minutes was adequate for the majority of floc settling. For most cases, the turbidity values measured after 10 minutes of settling adequately approximated those attained at both shorter and longer settling times.

Figure 8: Effect of Settling Time on the Turbidity Measured for 

Different Flocculant Concentrations of HX-200, HX-300, HX-400, and Alclar 600 during Head-to-Head Tests.
Figure 8: Effect of Settling Time on the Turbidity Measured for
Different Flocculant Concentrations of HX-200, HX-300, HX-400, and
Alclar 600 during Head-to-Head Tests.

Table 4 contains the results of experiments using HX-400 with the initial feeds of salt solution with 0.6 g/L sludge (without MST) and salt solution with 0.55 g/L MST (without sludge). Tests with MST only exhibited high degrees of flocculation and quick settling in comparison with MST and sludge mixtures, resulting in low supernate turbidities (<4 NTU). In contrast, supernates from tests with sludge only had higher turbidities (40 to 50 NTU) than supernates from tests with sludge and MST (typically <30 NTU). This evidence supports that the flocculation and settling of sludge from the supernate is limiting in this process. This is likely due to sludge having a smaller average particle size and a wider particle size distribution than MST, thus having a greater potential for fine particles to remain in the supernate. No average particle density information was available to evaluate the impact of this effect.

Table 4: The Flocculation of Sludge-Only and MST-Only Solutions with HX-400.

Sludge Conc.
(g/L)

MST Conc.
(g/L)

Flocculant Amount
(mg/L)

Flocculant Amount
(wt% solids)

Turbidity (NTU)

Approx. Settling Time
(s)

Approx.Floc Size
(mm)

0.6

0.00

7.1

2.6

50.0

80

1 to 2

0.6

0.00

14.2

1.2

40.2

80

1 to 2

0.0

0.55

7.1

2.4

2.8

60

1

0.0

0.55

14.2

1.3

3.6

60

1 to 1.5


As a result of the analysis of the experiments performed during this stage of testing, we identified HX-400 as the best flocculant candidate to bring forward for further study due to wider optimum flocculant concentration ranges and consistently lower decanted liquid turbidities, despite higher necessary flocculant concentrations. The HX-200 and HX-300 also showed similar potential, but had higher supernate turbidities during the head-to-head test. The potential application of HX-200 and HX-300 should not be discounted, but resources limited the number of different flocculants we could bring forward for the remainder of this study. Although it did not perform as well as three of the four HX flocculants, the researchers included Alclar 600 in some of the next set of tests due to the desire to have an option to fall back on if future testing should uncover a process incompatibility with HX flocculants.

4.2 Stability of Flocculant Heel

4.2.1 Test with HX-400

Table 5 contains the results of the heel test using HX-400 as the flocculant. The experiments demonstrated the concept of keeping the solid heel from a previous batch, contacting fresh slurry with flocculant in a separate vessel, mixing the newly flocculated slurry with the heel, and allowing the accumulated solids to settle again. The subsequent additions of new, untreated slurry mixed with flocculant caused an increase in the amount of solids in the slow-mix and settle stages in the beaker. This heel-test is closer in design to a potential full-scale flocculating, settling, and decanting process for the separation of sludge and MST. In the following discussion of the heel tests, a "cycle" indicates adding a new 1L of feed (and flocculant) to the 2L beaker, slow-mixing with the heel, settling for 10 minutes, and decanting the supernate from the solids.

As evidenced in Figure 9, a very encouraging observed trend is that the amount of solids in subsequent decanted supernatant fluid (as measured by the turbidity) decreased from that observed for the previous cycles. Retaining the solids from the previous cycle appeared beneficial, culminating with a stable turbidity value of approximately 13 NTU after the eighth cycle. Using the relation in Equation 1, 13 NTU corresponds to a insoluble solids concentration of 5.3 mg/L. This in turn corresponds to a solid removal efficiency of 99.5%, meeting the solid removal requirement estimated by HLW-PE. This holds promise in the design of a process, because once a certain amount of heel material accumulates, the data suggests that the turbidity of the next cycles will remain low. A potential reason for the reduced turbidities in the later cycles is the reduced space between the flocs during the slow-mix stage, resulting in a greater capture efficiency of fine particles for the new flocculant added.

Table 5: Results of the Heel Test using Cytec Superfloc® HX-400.

Cycle Number

Flocculant Amount
(mg/L)

Turbidity
(NTU)

Approx.
Settling Time
(s)

Approx.
Settling Rate
(ft/min)

Approx.
Floc Size
(mm)

Approx.
Solids Vol.
(mL)

1

12.7

36.2

35

25.7

0.5 to 2

60

2

12.6

30.7

37

24.3

0.5 to 2

50

3

12.7

27.3

35

25.7

0.5 to 2

60

4

12.7

23.8

72

12.5

0.5 to 2

70

5

12.8

19.5

89

10.1

0.5 to 2

90

6

12.7

18.2

90

10.0

0.5 to 2

120

7

12.6

19.7

120

7.5

0.5 to 2

140

8

12.6

12.1

120

7.5

0.5 to 2

160

9

12.7

13.8

140

6.4

0.5 to 1

180

10

12.7

12.6

120

7.5

0.5 to 1

180

11

12.8

12.2

145

6.2

0.5 to 1

210

12

12.7

13.7

135

6.7

0.5 to 1

220

 

Figure 9: Turbidity of Supernates for the Heel Test using Cytec Superfloc(R) HX-400.
Figure 9: Turbidity of Supernates for the Heel Test using Cytec Superfloc® HX-400.

Concurrent with the decrease in turbidity, the settling time increases and the settling rate decreases for subsequent addition cycles. This corresponds with a subtle decrease in the floc sizes during the course of the experiment, but could also result from the larger ratio of solids to liquid in the settling fluid. The values for settling times in Table 5, and thus the settling rates also, may not reflect an important difference in performance. As the testing progressed through the 12 cycles, the bulk supernate proved progressively clearer. That trend could have influenced our judgement of when 99% of the flocs settled. The later cycles had more small flocks that resisted settling, but the time required for the same percentage of solids settling may not have increased. Only during the last few cycles did we observe evidence of hindered settling before the majority of the flocs had settled to the solid layer. Again, the settling times reported are the settling of the final minor fraction of the flocs.

Figure 10 contains a series of images taken during the mixing and the settling periods of the second, fifth, and eleventh cycle. We obtained these images halfway (i.e., 5 minutes) into the respective mixing and settling periods. From the mixing images, the flocks appear to get progressively smaller and the space between the flocs progressively reduces. From the settling images, one notes that the supernate becomes progressively clearer and the volume of the heel of solids progressively increases.

In the later cycles, a significant quantity of particles or globules existed dispersed throughout the supernate that did not have the characteristic orange color of the sludge. These solids settled at a much slower rate than observed for the bulk of the flocculated material. These solids largely remained suspended in the supernate after the 10-minute settling period, and we decanted these solids with the supernate. We collected the solids in this supernate on a filter and analyzed by IR and EDS. Section 0 contains the analysis and discussion of these solids.

We experienced some difficulty in transferring all of the flocculated solids out of the 2L beaker. Due to the high solids content, the solid heel would not easily pour into a polyethylene bottle without leaving some in the beaker. We did not want to dilute this heel, and attempts to scrape the remaining solids into the bottle were only moderately successful, with about 10 grams of solids remaining in the beaker. Thus, we did not include these solids in the flocculant weight or weight percent. The solid fraction removed from the beaker was 256.0 grams, compared with 13.8 grams of insoluble solids in the original 12L of heel-test feed. Thus the final insoluble solids concentration of the solid fraction was approximately 5.4 wt %. The solids settled further in the storage bottle and subsequent analysis of a sample of only the resultant solid phase showed it to be 14.65 wt % insoluble solids.

4.2.2 Test with Alclar 600

We conducted the experiment for the heel test using Alclar 600 in a similar manner to the prior heel test for HX-400. Figure 11 contains images of the mixing and settling stages of several cycles of the Alclar 600 heel test. The trends resemble those during the HX-400 heel testing. As the heel volume gets progressively larger, the supernate gets progressively clearer, and the distance between flocs during mixing appears to become progressively smaller. The flocs themselves, however, do not appear to change size drastically during the course of the experiment. Comparing the images from the heel test with HX-400 (Figure 10) with those from this test (Figure 11), the supernates from the HX-400 proved much clearer in the later cycles, but the flock sizes appeared smaller using HX-400.

Table 6 and Figure 12 provide the quantitative results for the heel test with Alclar 600. As before, turbidities measured for the supernates decreased over the first several addition cycles, suggesting a trend toward the reduction in supernate turbidities as the flocculant heel formed. As expected from the single-flocculant test results, the overall clarification performance of this Alclar 600 test proved less favorable than observed for the Superfloc HX-400 test. At the dose of 2.2 mg/L Alclar 600, the settling times required remained nearly constant throughout the heel test, unlike the increase in settling times observed for the later cycles of the HX-400 heel test.

Figure 10: Mixing (Left) and Settling (Right) Solutions from the Second (Top), 

Fifth (Middle), and Eleventh (Bottom) Cycle of the Cytec Superfloc(R) HX-400 Heel Test.
Figure 10: Mixing (Left) and Settling (Right) Solutions from the Second (Top),
Fifth (Middle), and Eleventh (Bottom) Cycle of the Cytec Superfloc® HX-400 Heel Test.

Figure 11: Mixing (Left) and Settling (Right) Solutions from the Second (Top), Fifth (Middle), and Tenth (Bottom) Cycle of the Ciba Alclar(R) 600 Heel Test.
Figure 11: Mixing (Left) and Settling (Right) Solutions from the Second (Top),
Fifth (Middle), and Tenth (Bottom) Cycle of the Ciba Alclar® 600 Heel Test.

Table 6: Results of the Heel Test using Ciba Alclar® 600.

Cycle Number

Flocculant Amount
(mg/L)

Turbidity
(NTU)

Approx. Settling Time
(s)

Approx. Settling Rate
(ft/min)

Approx
Floc Size
(mm)

Approx. Solids Vol.
(mL)

1

2.2

44.0

40

22.5

1 to 3

30

2

2.2

33.4

38

23.7

1 to 3

40

3

2.2

36.3

40

22.5

0.5 to 2

60

4

2.2

34.3

30

30.0

0.5 to 2

80

5

2.2

27.8

50

18.0

0.5 to 1

90

6

2.2

24.4

56

16.1

0.5 to 1

100

7

2.2

26.6

63

14.3

0.5 to 1

104

8

2.2

27.0

55

16.4

0.5 to 1

110

9

2.2

24.9

61

14.8

0.5 to 1

140

10

2.2

23.7

54

16.7

0.5 to 1

130

11*

0.0

22.0

105

8.6

< 0.5

150

12*

2.2

19.9

52

17.3

0.5 to 2

120

*Tests using the decanted supernates from cycles 1 and 3 as the feed, respectively.

 

Figure 12: Turbidity of Supernates for the Heel Test using Ciba Alclar(R) 600.  Cycles 11 an 12 used the Decanted Supernates from Cycles 1 and 3 as the Feed, 

  Respectively.
Figure 12: Turbidity of Supernates for the Heel Test using Ciba Alclar® 600.
Cycles 11 an 12 used the Decanted Supernates from Cycles 1 and 3 as the Feed, Respectively.

We collected 160.5 g of solids fraction from 10 cycles (and two cycles without added solids), compared with the 11.5 g of sludge and MST solids in the initial 10 L of solution. This corresponds to 7.2 wt % insoluble solids in the solids fraction, measured as 11.9 wt % after additional settling.

In the final stages of this test, we investigated whether we could further treat the first few supernates by diverting them back through the process after a heel accumulated. If further solids removal occurs through such a process, it would prove beneficial in a situation where the first few supernates contained unacceptably high concentrations of solids but the remaining supernates had low solids.

After the tenth cycle, we brought the high-turbidity supernate from the first cycle into contact with the heel, slow-mixed for 9 minutes, settled for 10 minutes, and decanted. We did not add more flocculant. As evident from the data (Table 6, Cycle 11), the supernate clarified further (from 44.0 to 22.0 NTU), showing a consistent trend with that expected for an eleventh cycle. Due to the lack of additional flocculant, however, the flock size in the entire heel became much smaller and the required settling time increased. We then performed a similar cycle using the high-turbidity supernate from the third cycle. This time, we added a dose of 2.2 mg/L of Alclar 600 by flash mixing at 180 rpm with the third-cycle supernate before contacting with the heel, slow mixing, settling, and decanting. The supernate turbidity (Table 6, Cycle 12) reduced from 36.3 NTU to 19.9 NTU, which is consistent with that expected for the twelfth cycle. In contrast with the previous cycle without additional flocculant, the flocs evident during this cycle completely recovered their original larger size and faster settling rate. Thus, the process can provide treatment of the high-turbidity supernates from previous cycles, but requires additional flocculant to preserve the flocs in the heel.

4.3 Solids in Supernate

During the heel tests, we observed solids that did not have the characteristic orange color of sludge in the supernate. We obtained the contrast-enhanced photograph in Figure 13 by focusing in the middle of the beaker during the 12th cycle of the HX-400 heel test. We illuminated the supernate primarily from the bottom left rear of the beaker. While dried salt smudges and a reflection appear on the beaker surface, one clearly sees the particles or globules suspended in the supernate. We observed similar colorless particles for the heel test with Alclar 600, but the greater bulk turbidity of the supernate made them more difficult to distinguish.

From this image one notes that most the solids or flocs did not settle after about 8 minutes appeared relatively colorless. Since they lacked the typical orange tint of the sludge, we suspected that these solids did not represent a component of the sludge. As seen in Section 4.1, MST flocculates and settles relatively quickly, so it appears unlikely that the MST would persist in the supernate. The observed suspended material was initially believed to be globules of flocculant or emulsifier.

We filtered the supernates and analyzed the collected solids - washed and unwashed - with IR spectroscopy and EDS. Washing the solids did not affect the results obtained. From Figure 14, the IR spectrum of the solids did not match well those of sludge, MST, or the cured flocculant. The IR spectrum suggests that the solids contain aluminum containing compound, such as an aluminum hydroxide. As seen in and Figure 15 and Figure 16, analysis of the major component of the supernate solids agrees with that of sodium aluminosilicates examined in previous studies. Thermodynamic models suggest that practically all of the silicon in the SRS salt solution simulant will form aluminosilicate at equilibrium. The EDS of the solids confirms the presence of sodium, aluminum and silicon. A question remains why, if the colorless solids primarily aluminosilicates, is the flocculant not effective in the settling of these particles as well. As evident from Figure 17, minor components of the solids include a mixture of sludge and MST (from the figure on the left), and sodium sulfate salts (from the figure on the right).

Figure 13: Colorless Particles Floating in the Supernate after Settling. Photographed 8 Minutes into Settling Period of the 12th Cycle for the HX-400 Heel Test.
Figure 13: Colorless Particles Floating in the Supernate after Settling.
Photographed 8 Minutes into Settling Period of the 12th Cycle for the HX-400 Heel Test.

This chemistry may also influence the change the turbidity of the mixed and settled 5.6 M salt solution, as noted in Figure 3. Personnel stored these reference samples in glass turbidity vials for the duration of the testing (about 1.5 months and 1 month for the mixed and settled salt solutions, respectively). The turbidity, and thus the solids concentration, in each solution doubled over the storage period. Due to the subtle appearance changes over a long time frame, we did not note this increased salt solution cloudiness by direct visual observation during the course of the experiments. It seems likely that the increasing turbidity reflects formation of sodium aluminosilicate or aluminum hydroxide over time. Over this time frame, some additional silicon may have leached from the glass turbidity vials into the caustic solution and contributed to aluminosilicate formation. We did not analyze the solids in the turbidity samples to confirm this hypothesis.

Figure 14: IR of MST, Sludge, Cured Flocculant, and the Solids Remaining in the HX-400 Heel Test Supernate.
Figure 14: IR of MST, Sludge, Cured Flocculant,
and the Solids Remaining in the HX-400 Heel Test Supernate.

 

Figure 15: IR of the Solids Remaining in the HX-400 Heel Test Supernate Compared with Aluminosilicate Deposits from SRS HLW Evaporators.
Figure 15: IR of the Solids Remaining in the HX-400 Heel Test Supernate
Compared with Aluminosilicate Deposits from SRS HLW Evaporators.

Figure 16: EDS of the Major Component of the Solids Remaining in the Supernate after the 12th Cycle of the HX-400 Heel Test.
Figure 16: EDS of the Major Component of the Solids
Remaining in the Supernate after the 12th Cycle of the HX-400 Heel Test.

 

Figure 17A:  EDS of the Minor Components of the Solids Remaining in the Supernate after the 12th Cycle of the HX-400 Heel Test.  Figure 17B:  EDS of the Minor Components of the Solids Remaining in the Supernate after the 12th Cycle of the HX-400 Heel Test.
Figure 17: EDS of the Minor Components of the Solids
Remaining in the Supernate after the 12th Cycle of the HX-400 Heel Test.

4.4 Filtration Test

Any settle and decant operation for removal of MST and entrained sludge from waste solution would likely require polishing filtration to achieve the desired levels of solid removal. Even if the addition of flocculants led to high levels of flocculation and settling, an appreciable amount of solids may have remained in the supernate. As seen in the prior sections, several flocculant candidates worked well. A review of the flocculation and settling data showed that Cytec flocculant HX-400 yielded the best clarification, with an initial dose of about 12 mg/L. This phase of experimentation involved experiments to test the performance of filtration processes to remove the remaining solids from the supernate. Testing cross-flow filtration, the likely polishing filtration technology, on the laboratory-scale Parallel Rheology Experimental Filter (PREF) system would require a large volume (several hundred liters) of continuos fresh feed due to the low solids content of the decanted supernates. As an alternative, we used small-scale dead-end filtration to demonstrate the effectiveness of this process.

We collected approximately 10L of decanted supernate from each of the two heel tests for the filtration tests. The very low solids content and small volume did not permit the use of a cross-flow filter, such as the PREF. Instead researchers utilized the small volumes of supernate already prepared and evaluated filtration using the dead-end filter described in the experimental section.

4.4.1 Initial System Tests

To qualify the filtration system, we filtered a salt solution containing 0.6 g/L sludge and 0.55 g/L MST. At the flow rate of 0.25 gpm/ft2 (32 mL/min), salt solution with 0.6 g/L sludge and 0.55 g/L MST would create an increase in TMP on a time scale that was too short to be practical for a full-scale plant. We prepared a 20-fold volume diluted solution for this filter test using salt solution without added solids. The resulting mixture contained the 5.6 M Na+ salt solution with 30 mg/L sludge and 27.5 mg/L MST (57.5 mg/L total solids) and had a turbidity of 140.0 ± 0.2 NTU. The solids and turbidity follow the relationship of Equation 1, but the solids content is significantly greater than in the decanted supernates.

We filtered 2 L of the 20X diluted solids solution with the dead-end filter. Figure 18 contains the TMP profile for the constant feed rate of 32 mL/min. After about 55 minutes (1.7 L), the TMP increased to above 30 psi, and we performed a back-pulse sequence. After back-pulsing, the TMP started slowly increasing from about 12 psi.

4.4.2 Tests of Flocculant Supernates

Figure 18 provides the dead-end filtration data for supernates from the heel tests using HX-400 and Alclar 600 . We used a new filter element for each test. The Alclar 600 supernates caused the TMP to increase to 30 psi in about the same timeframe as the 20X diluted solids solution. The Alclar 600 supernates likely contained only about 10 mg/L solids, far less than 57.5 mg/L of solids in the 20X diluted solution. Once the filter cake accumulated during the Alclar 600 test, we performed the back-pulsing sequence several times. Back-pulsing proved ineffective in reducing the TMP during the filtration of Alclar 600 supernates, suggesting the possible irreversible filter pore blockage caused by the addition of flocculant.

The filtration of the HX-400 supernates proved more successful. After filtration of all 6.5 L of the supernate available for this test (in 3.6 hr), the TMP only increased to about 4 psi. Ideally, we would have continued the testing until we demonstrated the back-pulse sequence, but we lacked sufficient feed to challenge the filter. This result suggests the filter could operate with a flux of 0.25 gpm/ft2 for many hours (likely >10 hr) between back-pulses or filter cleanings. Future, large-scale testing should examine performance over multiple back-pulse cycles.

The filtrate turbidity was nominally 0.4 to 0.6 NTU, and was not a function of initial insoluble solids concentration or flocculant addition. Although the decanted supernate from the addition of HX-400 flocculant greatly improved the volume of feed that can be processed between back-pulses, the addition of flocculant did not affect – adversely or beneficially – the turbidity of the filtrate from the dead-end system.

Figure 18:  Dead-End Filter Tests on 57.5 mg/L Solids (20X Diluted) and Heel Test Supernates (Alclar 600 and HX-400).

Figure 18: Dead-End Filter Tests on 57.5 mg/L Solids (20X Diluted)
and Heel Test Supernates (Alclar 600 and HX-400).

5 Potential Future Studies and Applications

We recommend the following additional work to evaluate the feasibility of this process for solid-liquid separation in the SPP.

6 Conclusions

The addition of flocculant, followed by mixing, settling of the flocculated solids, and decanting the supernatant liquid proved effective for the removal of sludge and MST solids from salt solution, with estimated solids removal efficiencies of 98 to 99%. In comparing supernate turbidity, the Cytec Superfloc® HX flocculants performed better than the Ciba Alclar® flocculants for this system. HX-200, HX-300, and HX-400 performed similarly, with HX-400 showing greater consistency in head-to-head tests. Retaining a heel of solids in the processing vessel resulted in better clarification of subsequent batches of feed. Tests with Cytec Superfloc® HX-400 resulted in a reduction of supernate turbidities from above 35 NTU for the initial cycle (with no solid heel) to around 13 NTU after the eighth cycle (with the solid heel accumulated from the previous cycles). Polishing filtration of supernates after treatment with Cytec Superfloc® HX-400, settling, and decanting provided promising results. Dead-end filtration showed only a 4 psi increase in trans-membrane pressure in 3.6 hours of processing at a 0.25 gpm/ft2 filter flux. Dead-end filtration after treatment with Ciba Alclar® 600 caused the trans-membrane pressure to increase to 30 psi after 70 minutes of processing at the same flow conditions.

7 Quality Assurance

This work satisfies the requirements of the original task technical and quality assurance plan. Laboratory Notebook WSRC-NB-2001-00001 contains the experimental data.

8 Acknowledgements

The authors thank Henry Bolton for performing the bulk of the experiments and constructing the dead-end filter system, and Bob Hartley and Liz Coleman for assisting in several experiments.

9 References

  1. M. R. Poirier, F. F. Fondeur, T. L. Fellinger, and S. D. Fink, "Cross-Flow Filtration for Slurries Containing High Level Waste Sludge and Monosodium Titanate," WSRC-TR-2001-00212, Revision 0, April 11, 2001.
  2. H. H. Saito, M. R. Poirier, S. W. Rosencrance, and J. L. Siler, "Improving Filtration Rates of Mono-Sodium Titanate (MST)-Treated Sludge Slurry sith Chemical Additives," WSRC-TR-99-00343, Revision 0, September 15, 1999.
  3. 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, Revision 0, September 15, 1999.
  4. R. Haggard, T. Deal, C. Stork, and V. 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.
  5. R. A. Jacobs, Technical Task Request, HLW-SDT-TTR-99-30.0, December 1999.
  6. "Applied Technology Integration Scope of Work Matrix for Alpha Removal," HLW-SDT-2000-00047, Rev. 3.
  7. M. R. Poirier, "Evaluation of Solid-Liquid Separation Technologies to Remove Sludge and Monosodium Titanate from SRS High Level Waste," WSRC-TR-2000-00288, Revision 0, August 16, 2000.
  8. R. A. Peterson, M. S. Hay, and S. Lee, "Countercurrent Decantation Application at the Savannah River Site: Final Report (U)," WSRC-TR-97-00300, September 19, 1997.
  9. M. S. Hay, "Countercurrent Decantation Circuit Radioactive Sludge Settling Studies," WSRC-TR-97-00301, September 19, 1997.
  10. R. M. Matiea, "Counter-Current Decantation Design Progress Report I," WSRC Subcontract No. AB60289N, CMRI Project No. 962004, August 6, 1996.
  11. R. H. Cuttriss, "Final Report, Settling Tests to Develop Design and Operating Parameters," WSRC Subcontract No. AB60289N, CMRI Project No. 962004, October 23, 1996.
  12. D. T. Hobbs, "Particle Size and Settling Velocity of Tank 41H Solids," WSRC-TR-95-0249, May 30, 1995.
  13. B. Yarar and M. R. Poirier, "Evaluation of Flocculation and Filtration Procedures Applied to WSRC Sludge," WSRC-TR-2001-00213, April 16, 2001.
  14. M. R. Poirier, "Improving the Filtration of Sludge/Monosodium Titanate Slurries by the Addition of Flocculants," WSRC-TR-2001-00175, Revision 0, March 27, 2001.
  15. D. D. Walker, "Preparation of Simulated Waste Solutions," WSRC-TR-99-00116, Revision 0, April 15, 1999.
  16. "Standard Practice for Coagulation-Flocculation Jar Test of Water," ASTM Standard D2035-80.
  17. "Standard Test Method for Quantitative Test for Turbidity in Clear Liquids," ASTM Standard D5180-93.
  18. W. F. Steele, C. F. Weber, and D. A. Bostick, "Waste and Simulant Precipitation Issues," ORNL/TM-2000/348, January 2001.
  19. C. J. Martino and M. R. Poirier, "Task Technical Plan and Quality Assurance Plan for the Settle and Decant Testing of MST in Salt Simulant," WSRC-TR-2001-00030, Rev. 0, March 5, 2001.

Attachment 1: Determination of Required Insoluble Solids Removal

E-Mail

Chart