USF Libraries
USF Digital Collections

Feasibility of using nanofiltration as a polishing process for removal of cyanobacterial exudates from treated surface water

MISSING IMAGE

Material Information

Title:
Feasibility of using nanofiltration as a polishing process for removal of cyanobacterial exudates from treated surface water
Physical Description:
Book
Language:
English
Creator:
Mody, Anand J
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
TOC
disinfection by-products
Microcystin-LR
Geosmin
2-Methylisoborneol
Dissertations, Academic -- Environmental Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Nanofiltration (NF) membrane technology is effective for removal of natural organic matter (NOM) and Disinfection By-Product (DBP) precursors from treated surface water (Allgeier et al., 1995, Chellam et al., 2000, Smith et al., 2002). However, there is a need to control other micropollutants, such as compounds released from algal blooms. In this research, the feasibility of using NF for removal of cyanobacterial exudates was evaluated as a polishing process for conventionally treated surface water. Screening tests were conducted to compare the performance of four NF membranes, Filmtec's NF90 and NF270, and Hydranautics's LFC1 and NTR7450, for removal of NOM and cyanobacterial exudates. The source water for the experiments was derived from Lake Manatee (FL) following full scale treatment by enhanced coagulation and dual media filtration. Water samples were amended with low levels of three cyanobacterial exudates: microcystin-LR, geosmin and 2-Methylisoborneol (MIB). The rapid bench scale membrane test (RBSMT) protocol was used to test NF at four recoveries of 50%, 70%, 85% and 95%. Bulk organics (TOC and UV₂54) and inorganics (conductivity, total and calcium hardness) were monitored along with other operating parameters during the setting and recovery tests. Spike tests were performed by spiking microcystin-LR (9.5 to 12.0 micro g/L), geosmin (45 to 220 ng/L) and MIB (45 to 225 ng/L). Three NF membranes (NF90, NF270 and LFC1) were effective for over 90% rejection of TOC and associated disinfection by-product formation potential (DBPFP). Due to NF treatment, the bromide:TOC ratio increased resulting in a shift towards higher levels of brominated DBPFPs. Similarly, these three NF membranes (NF90, NF270 and LFC1) were effective for removal of microcystin-LR to below the World Health Organization (WHO) guideline of 1 micro g/L. Only two of the NF membranes tested (NF90 and LFC1), were capable of removing geosmin and MIB to levels below the taste and odor threshold. These membranes removed greater than 92% of the geosmin and MIB. Based on these bench scale tests, further testing of NF on a pilot scale is warranted.
Thesis:
Thesis (M.S.Env.E.)--University of South Florida, 2004.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Anand J. Mody.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 133 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 56563612
notis - AJS2478
usfldc doi - E14-SFE0000432
usfldc handle - e14.432
System ID:
SFS0025124:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200421Ka 4500
controlfield tag 006 m d s
007 cr mn
008 040811s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000432
035
(OCoLC)56563612
9
AJS2478
b SE
040
FHM
c FHM
090
TA170 (ONLINE)
1 100
Mody, Anand J.
0 245
Feasibility of using nanofiltration as a polishing process for removal of cyanobacterial exudates from treated surface water
h [electronic resource] /
by Anand J. Mody.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (M.S.Env.E.)--University of South Florida, 2004.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 133 pages.
520
ABSTRACT: Nanofiltration (NF) membrane technology is effective for removal of natural organic matter (NOM) and Disinfection By-Product (DBP) precursors from treated surface water (Allgeier et al., 1995, Chellam et al., 2000, Smith et al., 2002). However, there is a need to control other micropollutants, such as compounds released from algal blooms. In this research, the feasibility of using NF for removal of cyanobacterial exudates was evaluated as a polishing process for conventionally treated surface water. Screening tests were conducted to compare the performance of four NF membranes, Filmtec's NF90 and NF270, and Hydranautics's LFC1 and NTR7450, for removal of NOM and cyanobacterial exudates. The source water for the experiments was derived from Lake Manatee (FL) following full scale treatment by enhanced coagulation and dual media filtration. Water samples were amended with low levels of three cyanobacterial exudates: microcystin-LR, geosmin and 2-Methylisoborneol (MIB). The rapid bench scale membrane test (RBSMT) protocol was used to test NF at four recoveries of 50%, 70%, 85% and 95%. Bulk organics (TOC and UV54) and inorganics (conductivity, total and calcium hardness) were monitored along with other operating parameters during the setting and recovery tests. Spike tests were performed by spiking microcystin-LR (9.5 to 12.0 micro g/L), geosmin (45 to 220 ng/L) and MIB (45 to 225 ng/L). Three NF membranes (NF90, NF270 and LFC1) were effective for over 90% rejection of TOC and associated disinfection by-product formation potential (DBPFP). Due to NF treatment, the bromide:TOC ratio increased resulting in a shift towards higher levels of brominated DBPFPs. Similarly, these three NF membranes (NF90, NF270 and LFC1) were effective for removal of microcystin-LR to below the World Health Organization (WHO) guideline of 1 micro g/L. Only two of the NF membranes tested (NF90 and LFC1), were capable of removing geosmin and MIB to levels below the taste and odor threshold. These membranes removed greater than 92% of the geosmin and MIB. Based on these bench scale tests, further testing of NF on a pilot scale is warranted.
590
Adviser: Daniel P. Smith.
653
TOC.
disinfection by-products.
Microcystin-LR.
Geosmin.
2-Methylisoborneol.
690
Dissertations, Academic
z USF
x Environmental Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
949
FTS
SFERS
ETD
TA170 (ONLINE)
nkt 9/1/04
4 856
u http://digital.lib.usf.edu/?e14.432



PAGE 1

Feasibility Of Using Nanof iltration As A Polishing Process For Removal Of Cyanobacterial Exudates Fr om Treated Surface Water by Anand J. Mody A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Environmental Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Daniel P. Smith, Ph.D. Audrey D. Levine, Ph.D. Tory Champlin, Ph.D. Date of Approval: July 9, 2004 Keywords: toc, disinfection by-products, mi crocystin-lr, geosmin, 2-methylisoborneol Copyright 2004 Anand J. Mody

PAGE 2

Acknowledgements A special thanks goes to Manatee Count y Water Treatment Plant for funding the project and providing me the opportunity to ear n the Masters of Science in Environmental Engineering. Additional thanks goes to Film tec and Hydranautics for providing flat sheet membranes, to Naltex for supplying feed spacer and permeate carrier materials, to my thesis committee, and to the staff at the Manatee Coun ty Water Treatment Plant. A special thanks is also given to my family for their support. And last but not least, I would like to thank my wife, Dhara, for her constant support and encouragement.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES iii LIST OF FIGURES v LIST OF ABBREVIATIONS AND TERMS x ABSTRACT xii INTRODUCTION 1 Purpose 2 Research Objectives 2 Significance 2 BACKGROUND 4 Nanofiltration Membranes 4 Nanofiltration Treatment of Natural Organic Matter and DBP Precursors 4 Nanofiltration Treatment of Cyanobacterial Toxin and Taste and Odor Compounds 7 Rapid Bench Scale Membrane Test (RBSMT) to Evaluate NF Membranes 9 METHODOLOGY 10 Experimental Design 10 Membrane Experiments 11 Nanofiltration Membranes 11 Rapid Bench Scale Membrane Test 13 Experimental Setup 13 Operating Parameters 17 Test Sequence 19 System Start Up 21 Membrane Setting 22 Recovery Test 22 Spike Test 24 System Shutdown 25 System Monitoring 25 Sample Collection 26 Source Water 28 Analytical Methods 30 Statistical Analysis 32

PAGE 4

ii RESULTS 34 Setting and Recovery Tests 34 Mass Transfer Coefficient 36 Nanofiltration Membrane Rejection Performances 37 Comparison of Membranes for Rejection of TOC and UV254 37 Relationship of TOC with Recovery 41 Comparison of Membranes for Rejection of DBP Precursors 42 Comparison of Membranes for Rejection of Algal Exudates 47 Relationship of TOC with Algal Exudates Removal 53 DISCUSSION 55 Comparison of Membranes Performance 55 Operational Parameters 56 Concentrate Disposal 57 CONCLUSIONS 58 RECOMMENDATIONS FOR FUTURE RESEARCH 60 Bench Scale Testing 60 Pilot Testing 62 REFERENCES 63 APPENDICES 66 Appendix A: Equipment List and Experimental Photos 67 Appendix B: Chlorine Demand of Source Water 69 Appendix C: Comparison of Membranes for Rejection of Inorganics 71 Appendix D: Microcystin-LR Analyzed by ELISA 73 Appendix E: Geosmi n and MIB: Closed-Loop, GC/FID Protocol 79 Appendix F: Setting and Recovery Test Graphs 85 Appendix G: NF Membrane Tests Overall Performance 110

PAGE 5

iii LIST OF TABLES Table 1 Summary of NOM Rejection Using Nanofiltration Membranes Table 2 Experimental Design Table 3 List of RBSMT Experiments Table 4 Characteristics of Nanofiltra tion Membranes Used Table 5 Operating Parameters Equations for RBSMT Table 6 Sequence of Steps Used in RBSMT Experiment Table 7 Water Quality Samples Single Membrane Recovery Test Table 8 Number of Samples Si ngle Membrane Spike Test Table 9 Monitoring Frequency Duri ng Setting and Recovery Test Table 10 Sample Collection Matrix Table 11 Lake Manatee a nd Post Filtration Water Characteristics Table 12 Analytical Methods Table 13 Randomized Block Design Table for TTHMs Table 14 Mass Transfer Coefficient for all Membrane Tests Table 15 TOC and UV254 Data for NF Membranes Tested Table 16 Linear Regressi on Values for Comparison of Permeate TOC Correlation with Recovery Table 17 THMFP & HAAFP Data for NF Membranes Tested Table 18 Microcystin-LR, Geosmin and 2-Methylisoborneol Data for NF Membranes 6 10 11 12 18 20 23 25 26 27 29 31 33 36 39 42 44 49

PAGE 6

iv Table 19 Comparison of Memb rane Performance based on Percent Removal at 70% Recovery Table B-1 Post Filtration La ke Manatee Chlorine Demand Test Results Table C-1 Total Hardness, Calc ium Hardness and Conductivity Data for NF Membranes Tested Table D-1 Limit of Detection for Microcystins using ELISA Tube Kit Table G-1 Summary of NF 90 Treatme nt Performance Table G-2 Summary of NF 270 (a) Treatment Performance Table G-3 Summary of NF 270 (b) Treatment Performance Table G-4 Summary of NF 270 (c) Treatment Performance Table G-5 Summary of LFC 1 Treatment Performance Table G-6 Summary of NTR 7450 Treatment Performance 55 69 72 77 111 112 114 115 117 118

PAGE 7

v LIST OF FIGURES Figure 1. Chemical Structure of Microcystin-LR Figure 2. Chemical Structure of Geosmin and MIB Figure 3. RBSMT Membrane Cell Set up Figure 4. Flow Schematic of Rapid Bench Scale Membrane System Figure 5. Flow Rate, Pressu re, and Concentration Measurement Locations Figure 6. Surface Water Treatment Process Train Manatee County Water Treatment Plant Figure 7. Setting Te st Permeate Flux versus Time Figure 8. Flow-Rate versus Time Recovery Tests for NF270 Membrane Test at 19 gfd Figure 9. TOC & UV254 versus Recovery for NF270 Membrane Figure 10. TOC Rejection for NF Membranes at Tested Recoveries Figure 11. Time Variati on of Permeate Conductivity and UV254 for NF270 at 19 gfd Figure 12. THMFP Reduction Per centages for Nanofiltration Membranes Tested Figure 13. Feed and Permeate THMFP Concentration with Speciation Figure 14. SUVA and THMFP Permeate Conc entration versus Recoveries for NF90 Membrane Test Figure 15 NF270 Microcystin-LR Rej ection Function of Recovery and Permeate Flux Figure 16. NF 270 Geosmin rejection Function of Recovery and Permeate Flux 8 8 14 16 17 28 34 35 38 40 40 43 45 46 48 48

PAGE 8

vi Figure 17. Microcystin-LR Concentration in Permeate Water of Each Nanofiltration Membrane at Recoveries Tested for Water Spiked With 8.5 12.0 g/L Figure 18. 2-Methylisoborneol & Geosmin Con centration in Permeate Water of Each Nanofiltration Membrane at Recoveries Tested for Water Spiked with 40-215 ng/L Figure 19. NF 270 MIB Rejection Func tion of Recovery and Permeate Flux Figure 20. TOC Rejection vers us Microcystin-LR, Geosmin and MIB Reject ion for NF90 Membrane at F our Different Recoveries Figure A-1. Osmonics Sepa Test Cell Figure B-1. Chlorine Demand Test Conducted on Lake Manatee Post Filtered Water Figure C-1. Total Hardness Rejection fo r NF270 at Different Flux Rates and Recoveries Figure D-1. Absorbance 450nm vs Microc ystin-LR Concentration (ppb) Figure F-1. Flow Rates and Permeate Flux vs Time, Filmtec NF90, Setting Test Figure F-2. MTC vs. Time, Film tec NF 90, Setting Test Figure F-3. MTC vs. Time, Filmtec NF90, Setting Test Figure F-4. MTC vs. Time, Filmtec NF90, Recovery Test Figure F-5. Flow Rates vs Time, Film tec NF90, Recovery Test Figure F-6. Permeate Flux vs Time, Film tec NF90, Recovery Test Figure F-7. Permeate Conductivity & UV254 vs. Time, Filmtec NF90, Recovery Test Figure F-8. Pressure vs. Time, Filmt ec NF90, Recovery Test Figure F-9. Flow Rates and Permeate Fl ux vs Time, Filmtec NF270(a), Setting Test Figure F-10. MTC vs. Time, Filmtec NF270 (a), Setting Test 50 51 52 54 68 70 71 78 86 86 87 87 88 88 89 89 90 90

PAGE 9

vii Figure F-11. MTC vs. Time, Filmtec NF270 (a), Setting Test Figure F-12. MTC vs. Time, Filmtec NF270 (a), Recovery Test Figure F-13. Flow Rates vs Time, Filmt ec NF270(a), Recovery Test Figure F-14. Permeate Flux vs Time, Filmt ec NF270(a), Recovery Test Figure F-15. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(a), Recovery Test Figure F-16. Pressure vs. Time, Film tec NF270(b), Recovery Test Figure F-17. Flow Rates and Permeate Fl ux vs Time, Filmtec NF270(b), Setting Test Figure F-18. MTC vs. Time, Filmtec NF270 (b), Setting Test Figure F-19. MTC vs. Time, Filmtec NF270 (b), Setting Test Figure F-20. MTC vs. Time, Filmtec NF270 (b), Recovery Test Figure F-21. Flow Rates vs Time, Filmt ec NF270(b), Recovery Test Figure F-22. Permeate Flux vs Time, Filmt ec NF270(b), Recovery Test Figure F-23. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(b), Recovery Test Figure F-24. Pressure vs. Time, Film tec NF270(b), Recovery Test Figure F-25. Flow Rates and Permeate Fl ux vs Time, Filmtec NF270(c), Setting Test Figure F-26. MTC vs. Time, Filmtec NF270 (c), Setting Test Figure F-27. MTC vs. Time, Filmtec NF270 (c), Setting Test Figure F-28. MTC vs. Time, Filmtec NF270 (c), Recovery Test Figure F-29. Flow Rates vs Time, Film tec NF270(c), Recovery Test Figure F-30. Permeate Flux vs Time, Filmt ec NF270(c), Recovery Test Figure F-31. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(c), Recovery Test 91 91 92 92 93 93 94 94 95 95 96 96 97 97 98 98 99 99 100 100 101

PAGE 10

viii Figure F-32. Pressure vs. Time, Film tec NF270(c), Recovery Test Figure F-33. Flow Rates and Permeate Flux vs Time, Hydranautics LFC1, Setting Test Figure F-34. MTC vs. Time, Hydranautic s LFC1, Setting Test Figure F-35. MTC vs. Time, Hydranautic s LFC1, Setting Test Figure F-36. MTC vs. Time, Hydranau tics LFC1, Recovery Test Figure F-37. Flow Rates vs Time, Hydranau tics LFC1, Recovery Test Figure F-38. Permeate Flux vs Time, Hydran autics LFC1, Recovery Test Figure F-39. Permeate Conductivity & UV254 vs. Time, Hydranautics LFC1, Recovery Test Figure F-40. Pressure vs. Time, Hydran autics LFC1, Recovery Test Figure F-41. Flow Rates and Permeate Flux vs Time, Hydranautics NTR 7450, Setting Test Figure F-42. MTC vs. Time, Hydranautics NTR 7450, Setting Test Figure F-43. MTC vs. Time, Hydranautics NTR 7450, Setting Test Figure F-44. MTC vs. Time, Hydranautic s NTR 7450, Recovery Test Figure F-45. Flow Rates vs Time, Hydran autics NTR 7450, Recovery Test Figure F-46. Permeate Flux vs Time, Hydranau tics NTR 7450, Recove ry Test Figure F-47. Permeate Conductivity & UV254 vs. Time, Hydranautics NTR 7450, Recovery Test Figure F-48. Pressure vs. Time, Hydran autics NTR 7450, Recovery Test Figure G-1. Summary of NF 90 Membra ne Performance at 19 gfd Figure G-2. Summary of NF 270 Membra ne Performance at 32 gfd Figure G-3. Summary of NF 270 Membra ne Performance at 19 gfd Figure G-4. Summary of NF 270 Membra ne Performance at 12 gfd 101 102 102 103 103 104 104 105 105 106 106 107 107 108 108 109 109 110 110 113 113

PAGE 11

ix Figure G-5. Summary of LFC 1 Membra ne Performance at 10 gfd Figure G-6. Summary of NT R 7450 Membrane Performance at 19 gfd 116 116

PAGE 12

x LIST OF ABBREVIA TIONS AND TERMS ANOVA Statistical Analysis of Variance CE Capillary Electrophoresis cfs Cubic Feet per Second Concentrate Membrane reject stream Da Daltons D/DBP Disinfectants/Dis infection By-Products EERL Environmental Engineering Resear ch Laboratory, University of South Florida ELISA Enzyme Linked Immunosorbent Assay Experiment Sequence of tests performed on a single membrane setting, recovery and spike Feed Inlet Stream to the Membrane System Flux Volume rate of transfer through membrane surface (gfd) fps Feed per Second gfd Gallons per Square Foot per Day gpd Gallons per Day HAAs Haloacetic Acids HPLC High Performance Liquid Chromatography ICR Information Collection Rule Influent Input stream to membrane after r ecycle stream combines with feed stream MIB 2-Methylisoborneol MCL Maximum Contaminant Level MCWTP Manatee County Water Treatment Plant, Florida MTC Mass Transfer Coefficient (gfd/psi) MW Molecular Weight MWCO Molecular Weight Cut Off MYCST Microcystin NF Nanofiltration nm Nanometer NOM Natural Organic Matter Permeate Output Stream of the Membrane System pH -log[H+] psi Pounds per Square Inch RBSMT Rapid Bench Scale Membrane Test Rejection Percent solute concentration reduction of permeate stream relative to the feed stream Recovery Ratio of permeate flow to feed flow Scaling Precipitation of solids due to increased solute concentration of the feed stream on the membrane

PAGE 13

xi Solute Dissolved constituents in the feed stream Test Membrane run at a specific recove ry and/or spike. Membrane runs at 50%, 70%, 85% and 95% recove ry are different tests TFC Thin Film Composite THMs Trihalomethanes TOC Total Organic Carbon UV254 Ultraviolet abso rbance at 254 nm Waste Fraction of the concentrate stream that is removed from the system. This stream is also sometimes referred as Retentate

PAGE 14

xii FEASIBILITY OF USING NANOFILTRATION AS A POLISHING PROCESS FOR REMOVAL OF CYANOBACTERIAL EXU DATES FROM TREATED SURFACE WATER ANAND J. MODY ABSTRACT Nanofiltration (NF) membrane technology is effective for removal of natural organic matter (NOM) and Disinfection By -Product (DBP) precursors from treated surface water (Allgeier et al., 1995, Chellam et al., 2000, Smith et al., 2002). However, there is a need to control other micropolluta nts, such as compounds released from algal blooms. In this research, the feasibility of using NF for removal of cyanobacterial exudates was evaluated as a polishing process fo r conventionally treated surface water. Screening tests were conducted to compare the performance of four NF membranes, Filmtecs NF90 and NF270, and Hydranauticss LFC1 and NTR7450, for removal of NOM and cyanobacterial exudates. The source water for the experiments was derived from Lake Manatee (F L) following full scale treatm ent by enhanced coagulation and dual media filtration. Water samples were amended with low levels of three cyanobacterial exudates: microcystin-LR, ge osmin and 2-Methylisoborneol (MIB). The rapid bench scale membrane test (RBS MT) protocol was used to test NF at four recoveries of 50%, 70%, 85% a nd 95%. Bulk organics (TOC and UV254) and inorganics (conductivity, total and calcium ha rdness) were monitored along with other operating parameters during the setting and reco very tests. Spike tests were performed

PAGE 15

xiii by spiking microcystin-LR (9.5 to 12.0 g/L), geosmin (45 to 220 ng/L) and MIB (45 to 225 ng/L). Three NF membranes (NF90, NF270 and LFC1) were effective for over 90% rejection of TOC and associated disinfec tion by-product formation potential (DBPFP). Due to NF treatment, the bromide:TOC ratio increased resulting in a shift towards higher levels of brominated DBPFPs. Similarly, these three NF membranes (N F90, NF270 and LFC1) were effective for removal of microcystin-LR to belo w the World Health Organization (WHO) guideline of 1 g/L. Only two of the NF membranes tested (NF90 and LFC1), were capable of removing geosmin and MIB to le vels below the taste and odor threshold. These membranes removed greater than 92% of the geosmin and MIB. Based on these bench scale tests, further testing of NF on a pilot scale is warranted.

PAGE 16

1 INTRODUCTION Disinfection of drinking water using chlo rine can result in the formation of disinfection by-products (DBPs) such as tr ihalomethanes (THMs) and haloacetic acid (HAAs). The Environmental Protection Agency (EPA), through the St age I Disinfectants / Disinfection By-Products (D/D BPs) Rule, has established maximum contaminant levels (MCLs) for total TTHMs (chloroform, dich lorobromoform, dibromochloroform and bromoform) of 80 micrograms per liter ( g/L) and 60 g/L for five HAAs (monochloroacetic acid, dichloroacetic acid, trichloroacetic acid, monobromoacetic acid, and dibromoacetic acid) as these have been associated with health effects including increased incidence of rectal and bladde r cancer (Morris, R.D., et al. 1992) and miscarriages (Waller, K., et al 1998). To meet th ese MCLs, utilities must either provide improved removal of DBP precursors or redu ce reliance on chlorine by using alternative disinfectants such as ozone, chlorine di oxide or chloramines. For surface water sources, other contamin ants of concern include exudates from algae and cyanabacteria. U nder certain conditions, cyanoba cteria, can release toxic and/or nuisance compounds. The World Heal th Organization (WHO) issued a drinking water guideline of 1 g/L standard for microcystin-LR, one type of cyanobacterial toxin (Chorus and Bartram, 1999). Cyanobacterial bl ooms can also result in the release of algal exudates such as geosmin and 2-Me thylisoborneol (MIB) which can impart unpleasant taste and odor compounds in potable water. Although these contaminants are

PAGE 17

2 not regulated by the EPA, they have been placed on the Contaminant Candidate List (CCL) by the EPA (EPA Draft CCL List Volume 69, Number 64, April 2004). Membrane technologies such as ultraf iltration (UF), nanof iltration (NF) and reverse osmosis (RO) have the potential to reduce the levels of DBP precursors (Fu et al. 1994). NF has been shown to remove up to 90% of the DBP precursors, from conventionally treated surface water (Allgeier and Summers, 1995) and 50-90% of algal exudates from surface water (M unitsov and Trimboli, 1996). Purpose The purpose of this research was to evaluate the potential of NF to remove DBP precursors, toxins released by cyanobacter ia, and taste and odor compounds from conventionally treated surface water. Research Objectives The objectives of this research were: 1. Quantify the effectiveness of NF to re move DBP precursors from conventionally treated surface water. 2. Quantify the effectiveness of NF to rem ove microcystin-LR, geosmin and MIB. 3. Compare the performance of selected co mmercially available NF membranes for removal of DBP precursors, micr ocystin-LR, geosmin and MIB. Significance Surface water impoundments are vulnerable to seasonal algal blooms, particularly in tropical climates such as in Florida. Treatment alternatives are needed to provide reliable control of the nuis ance and toxic compounds that may be released from algal blooms.

PAGE 18

3 Due to variations in membrane character istics, the performance of NF membranes for control of DBP precursors, cyanobact erial toxins, and taste and odor compounds cannot be predicted from manufacturer’s inform ation alone. The use of bench scale tests can provide an approach to screen and compar e membranes prior to pilot testing. The use of NF as a polishing treatmen t for control of DBP precursors, cyanobacterial toxins, and taste and odor compounds can provide a means for utilities to impr ove water quality and provide intermittent treatment to address seasonal variations in water quality. This research provides an approach that can be used to screen alternative membranes and to guide pilot plant testing.

PAGE 19

4 BACKGROUND A brief summary of Nanofiltration (NF) membrane technology is presented in this chapter. Information on natural organic ma tter (NOM), disinfecti on by-products (DBPs), and algal exudates is also pr esented. Finally, the rapid be nch scale membrane test is briefly described. Nanofiltration Membranes NF is a pressure driven membrane pro cess that has the potential of rejecting solutes and particles down to 1 nanometer in size (Letterman, 1999). These membranes are classified by molecular weight cut off (MWCO) and solute rejection capability. MWCOs for NF membranes range from 200 Dalt ons (Da) to 1000 Da, where Dalton is a unit of mass for expressing masses of atoms, mo lecules, or nuclear pa rticles equal to 1/12 the mass of a single atom of the most abundant carbon isotope 12C (Webster online, 2004). NF membranes separate water const ituents by a combinati on of molecular size sieving and diffusion. Nanofiltration Treatment of Natura l Organic Matter and DBP Precursors NOM in surface water consists of a co mplex mixture of humic and non humic substances (Roalson et al., 2003) that can be quantified using total or dissolved organic carbon (TOC or DOC) and UV254. Organic DBPs are formed when chemical disinfectants react with NOM. The type and concentration of DBPs are influenced by treatment conditions and water quality variables such as TOC, bromide, pH, temperature, ammonia, and carbonate alkalinity (Garvey et al., 2003).

PAGE 20

5 The effectiveness of NF membrane t echnology for removal of the fraction of NOM that acts as a DBP precursor is well documented (Tan et al ., 1991, Allgeier et al., 1995, Chellam et al., 2000, Smith et al., 2002). TOC removal is controlled by a combination of size exclusion, electrostat ic repulsion, and hydr ophobic interactions between the solutes and the membrane surf ace and pores (Cho et al. 1999). A brief summary of research relating to NF removal of NOM and DBP precursors is presented in Table 1. Taylor and Chellam (2000) tested several water sources in Florida, Virginia, and Texas with TOC levels rangi ng from 3.3-13.1. The rejection of TOC and THMFP ranged from 71% (Hydranautics, NTR7450) to 94% (Koch, TFC-SR) depending on the water source and membrane. The City of Fort Meye rs, Florida operated a NF plant to remove THMFPs and color from well water containi ng about 10-20 mg/L of TOC. Hydranautics NCM-1 and PVD-1 membranes were able to reduce the THMFPs from 537 g/L to 15 g/L (Taylor and Jacobs, 1996). Taylor et al (1987) demonstrated that THM precursors could be removed from surface waters usi ng a Filmtec NF membrane (N50) with a MWCO of 400, a recovery of 65%, and a pressu re of 60 psi. Rejection in the range of 97% to 98% for THMFP was observed during these tests. Visvanathan et al. (1998) tested reject ion of THM precursors from surface water using negatively charged Filmtec’s NF45 a nd DK membranes. Tests were conducted using a small scale flat sheet in a test unit. Compaction of the membrane resulted in a reduced pore size and an increased charge density, increasing THMFP rejection and electrostatic repulsion (Visva nathan et al. 1998).

PAGE 21

6Table 1 Summary of NOM Rejection Us ing Nanofiltration Membranes TOC THMFP Feed Permeate Percent Feed Permeate Percent Reference Membrane Manufacturer Membrane Trade Name Water Source Pretreament mg/L mg/L Removal g/L g/L Removal DS-5-DK 4.7 0.4-0.9 81-91 325 39 88 Osmonics DS-5-DL 4.7 0.2-0.3 95-96 325 17 95 NF90 4.0 0.1-0.3 95-98 154 17 89 NF200 4.0 0.1-0.3 95-98 154 15 90 Filmtec NF270 3.8 0.1-0.3 94-96 154 17 88 Hydranautics ESNA 3.8 0.4-0.9 77-90 154 56 64 Koch TFC-S 3.7 0.2-0.5 86-94 154 33 79 XN40 3.7 0.7-0.9 76-82 154 88 43 Falls, 2002 Trisep TS80 Lake Manatee, FL Conventionally treated surface water with enhanced coagulation 3.4 0.1-0.2 95-96 154 7 95 NF200B 66.5 15 Filmtec NF200 27 30 Hydranautics LFC1 93 72 Veerapaneni et al., 2001 Koch TFC-S Ohio River, KY Conventionally treated surface water 58 7 NF45 Biscayne Aquifer, FL 12.1 1.1 93 342 38 89 Filmtec NF200B Lake Meade, VA 3.3 0.3 91 75 10 87 Hydranautics NTR7450 Caloosahatch ee River, FL 7.1 2.1 71 252 80 68 Taylor et al.., 2000 Koch TFC-S Rio Grande River, TX Conventionally treated surface water and cartridge filter 2.9 0.18 94 214 18 93 Hydranautics LFC1 2-16 <0.5 46-1062 3.3-6 Reiss et al., 1999 Koch CALP Hillisborough River, FL Microfiltration 3-15 0.5-0.8 124-1037 34-60 NF50 Village of Golf, FL 93 98 Taylor et al.., 1987 Filmtec NF50 Acme District, FL 90 97

PAGE 22

7 Nanofiltration Treatment of Cyanobacterial Toxins and Taste & Odor Compounds Blue-green algae, otherwise known as cyanobacteria, can produce hepatotoxic cyclic peptides of the microcystin and nodular in family and neurotoxic alkaloids of the anatoxin family (Chorus and Bartram, 1999) Microcystins and nodularins are liver toxins while anatoxins affect the nerve syna pse usually leading to respiratory arrest (Westrick, 2003). One of the cyclic peptides is Microc ystin-LR (MCYST-LR). MCYST-LR is the most commonly monitored microcystin. It has a molecular weight (MW) of 980, is soluble in water, has a high toxicity, and is suspected of causing liver toxicity (Chorus and Bartram, 1999). Microcystins can be oxi dized using ozone, UV light, or other strong oxidants (Chorus and Bartram, 1999) and c ould be inactivated at pH below 8 and chlorine residual of above 0.5mg/L (Drikas et al, 2001). Taste and odor compounds produced by cyanobacterial blooms include geosmin (MW of 182) and MIB (MW of 168). These two compounds have a stable ring formation that is resistant to oxidati on. Limited studies have been conducted to assess the removal of cyanobacterial exudates using NF. In one study, NF effectively removed micr ocystin (>90% rejection) from surface water pretreated by microfiltration (Mun itsov and Trimboli, 1996) along with 60% rejection of geosmin and 45% rejection of MIB. This pretreated water was spiked with MCYST-LR (8.4 g/L), geosmin (52 ng/L) a nd MIB (38 ng/L) and was treated using a Hydranautics PVD1 NF membrane.

PAGE 23

8 Figure 1. Chemical Structure of Micr ocystin-LR (James and James, 1981) (a) (b) Figure 2. Chemical Structure of (a ) Geosmin and (b) MIB (Montgomery, 1985) In another study, Taylor et al. (1999) were able to re move 98% of both geosmin and MIB using an NF membrane (LFC 1) with a 200 MWCO. Falls et al. (2002), assessed the capability of NF membranes fo r removal of microcystin-LR, geosmin and CH3 CH3OH H3C H3C H3C CH3 OH

PAGE 24

9 MIB. Excellent removal (>95%) of microc ystin-LR was observed, however, rejection of microcystin-LR did not always correlate to the rejection of geosmin and MIB. Rapid Bench Scale Membrane Test (R BSMT) to Evaluate NF Membranes Allgeier and Summers developed the ra pid bench scale membrane test (RBSMT) to simulate the performance of full-scale spiral wound NF elements using a tangential flow flat sheet membrane cell with retentat e recycle. The flat membrane sheet is a differential element of a full-scale spiral wound NF element (Allgeier et al., 1995). The membrane cell requires a 24 square inch membrane sheet that is operated with feed spacer and permeate carrier used in full scale elements (or replicates) to make the test hydraulically similar to full-scal e operation. The recycle loop allows for representative recoveries and cross-flow ve locities of a full-scale system. The RBSMT can be used to compare NF membrane performa nce and flux requirements. The results of the RBSMT can be used as a screening tool for selection of membranes to be used in pilot plant studies. However, the results from RBSMTs cannot be used to predict performance due to seasonal variations in feedwater, biofouling, long-term cleaning requirements, and stressed or upset proce sses (Allgeier and Summers, 1995).

PAGE 25

10 METHODOLOGY This research project was develope d to compare the effectiveness of commercially available NF membranes usi ng the Rapid Bench Scale Membrane Test (RBSMT) at various recoveries and fluxes for removal of total organic carbon (TOC), microcystin-LR, geosmin, and MIB from conventionally treated surface water. The experimental design, membrane experiment s, specific NF membranes tested, RBSMT experimental setup and procedur e, source water used and its characteristics, and finally the analytical methods are described in this chapter. Experimental Design An overview of the experiment design incl uding the objectives of this research, the approach implemented, analysis, and key measurements are presented in Table 2. Table 2 Experimental Design Objective Approach Analysis Quantify removal of NOM & DBP precursor Test 4 membranes using RBSMT at 50%, 70%, 85% and 95% permeate recoveries Compare permeate concentrations of THMFP and HAAFP to EPA Stage 1 MCLs and evaluate removal efficiencies of TOC and UV254 Quantify removal of Algal Exudates Spike test with microcystin-LR, geosmin and MIB and test 4 membranes at 50%, 70%, 85% and 95% permeate recoveries Compare permeate concentrations of microcystinLR, geosmin and MIB to WHO guideline and odor threshold values Compare removal performance of the selected membranes Compare results from the six membrane experiments. Compare removal efficiencies of TOC, UV254, THMFP, HAAFP, hardness, microcystin-LR, geosmin and MIB.

PAGE 26

11 Membrane Experiments Four NF membranes were evaluated in this research, each at four different recoveries. Three membranes (NF90, NF270 and NTR7450) were tested at 19 gallons per square feet per day (gfd), NF270 was also tested at 12 and 32 gfd flux rates. LFC1 membrane was tested at a flux rate of 10 gfd. Six RBSMT experiments were conducted as listed in Table 3. Each experiment consis ted of three steps: a setting procedure, a recovery test for quantifying the rejection of NOM, and a spike test for quantifying the rejection of cyanobacterial toxin and exudates. Table 3 List of RBSMT Experiments Experiment Membrane Manufacturer Membrane Trade Name Flux Rate (gfd) 1 Filmtec NF90 19 2 Filmtec NF270 32 3 Filmtec NF270 19 4 Filmtec NF270 12 5 Hydranautics LFC1 10 6 Hydranautics NTR7450 19 Nanofiltration Membranes The NF membranes used in this study are listed in Table 4 along with the details on manufacturers, trade names, membrane material, nominal MWCO, and operating characteristics. These membranes were obtai ned from the manufactur ers as flat sheets and then cut to 24 square inches (sq.in.) (6 inches x 4 inches), needed for bench scale testing. Additional data on th e feed spacer and permeate carrier material, clean water mass transfer coefficient (MTC), and cro ss flow velocity were obtained from the manufacturers. For consistency, an Osm onics 34 mil (mil=1/1000 inch) feed spacer was used for all the membranes tested in this study.

PAGE 27

12 The NF90 and NF270 from Dow-Filmtec and LFC1 and NTR7450 from Hydranautics were selected based previous testing reported by Falls (2002), Taylor (2000), and Reiss (1999). The Dow-Filmtec NF90 membrane was selected for its reported high rejection of TOC (> 85 %) wh ereas the NF270 membrane was selected due to its high flux (~ 20 gfd) and mode rate TOC rejection (> 75 %). Table 4 Characteristics of Nanofi ltration Membranes Used Membrane Trade Name NF90 NF270 LFC1 NTR 7450 Manufacturer Filmtec Filmtec Hydranautics Hydranautics Material Polyamide Polyamide Composite Polyamide Sulphonated Polyether Sulphone Structure Thin Film Composite Thin Film Composite Thin Film Composite Thin Film Composite Nominal MWCO, Daltons 200-3001 200-3001 200 600-800 Operating pH Range 2 – 10 2 – 10 3-10 2-12 Maximum Operating Flux (gpd/ft2) 18.0 35.5 27.0 22.5 Solute Rejection (%) CaCl2 85-952 40-602 >99.2 >50 MgSO4 >973 >973 >99.2 >50 Chlorine Tolerance <0.1ppm <0.1ppm <0.1ppm <10 ppm 1 Dow Filmtec: “NF has (MWCO of) 200 to 300 Da” 2 500 ppm CaCl2, 70 psi, 15% recovery 3 2000 ppm MgSO4, 70 psi, 15% recovery The Hydranautics LFC1 membrane was selected for its reported high TOC rejection (>95%) and 200 MW CO for the removal of algal exudates. The NTR 7450 membrane was selected because of its surf ace charge. It is a sulphonated polyether sulphone membrane with a nominal MWCO of 600-800 Daltons (Da) that is capable of

PAGE 28

13 increased rejection of negatively char ged molecules (Spangenberg et al., 2002). The NTR7450 membrane also has a higher chlorine tolerance, up to 10 mg/L which could be beneficial for intermittent injections of chlori ne to control organic and biological fouling. The MWCO’s for the NF membranes tested in this research varied from 200 to 800 Da. Rapid Bench Scale Membrane Test The objective of this research was to evaluate the ability of commercially available NF membranes to be used as a polishing treatment for removal of TOC and algal exudates from conventionally treated su rface water using the RBSMT. This section describes the RBSMT design including the ex perimental setup, operating parameters, start-up and shut-down procedures along with the description of th e membrane setting, recovery and spike tests. Experimental Setup The experimental setup and procedure of the RBSMT were adapted from the ICR Manual for Bench-and Pilot-scale Treatment Studies (EPA, 1996). The RBSMT test cell was manufactured by Osmonics (Sepa CF Membrane Cell) and is shown in Figure 3. The cell body and cell holder were 6.50 x 8.38 x 2.04 inches and 7.9 x 11.0 x 7.9 inches respectively in size. The experimental set up consisted of an RBSMT test cell and ” OD 316 stainless steel tubing fo r connecting the system. Each test membrane was cut to the appr opriate 24 sq. in. size using the membrane cell as a template to ensure proper fit with in the membrane cell. The 34 mil (mil=1/1000 inch) feed spacer and permeate carrier suppl ied by Osmonics were also cut to the required size. The membrane cell was assembled according to manufacturer’s instructions. Inside the cell, a permeate ca rrier was placed on top of the membrane, a

PAGE 29

14 feed spacer below the membrane, and O-rings were used to seal the cell. The active side of the membrane faced the feed spacer. Th e permeate carrier and the feed spacer were pre-wetted with laboratory Nanopure water and placed in the membrane cell. The permeate carrier, feed spacer, and NF membrane were placed in the cell body and compacted using compressed air or nitrogen a pplied between the two chemically resistant stainless steel halves of the cell body to pr essurize the pneumatic ram to seal the cell body. The feed, permeate, and concen trate lines were then connected. Figure 3. RBSMT Membrane Cell Set-up (Falls, 2002)

PAGE 30

15 The experimental setup of the RBSMT sy stem is shown in Figure 4. A variable speed feed pump (Cole Parmer, Model 75211-10, gear pump drive fitted with a micropump, Model 200.150-000, Serial number 374302 gear pump head) was used to pump the source water from the 16 gallon feed tank. The feed pump was connected to a tee with one side connected to an adjustable pressure relief valve and the othe r to another tee. The adjustab le pressure relief valve (set to operate at 100 psi) was used to control the system pressu re and minimize the risk of over-pressurizing the RBSMT. The second tee re ceived flow from the recycle line and the influent line to the membrane cell whic h passed through a BelArt flowmeter (size #5, serial number 59778-5). This flow rate was used to calculate the infl uent flow rate and cross flow velocity of the membrane cell. Pressure and temperature gauges were installed inline to the membrane cell to monitor the inlet pressure and temperature. Temperature readings were used to standardize the flux rates to a temperature of 20C. Pressure readings were used to calculate water mass transfer coefficients as a means to normalize variations in flux and pressure. The influent water was pumped tangentia lly to the membrane feed spacer and the membrane. Permeate passing through the me mbrane was collected in the permeate carrier where it discharged the RBSMT thr ough the permeate outlet port. A pressure gauge was installed and the permeate flow rate was measured in the permeate line. The membrane reject water discharged the RBS MT through the concentr ate line port. The concentrate line was provided with a pressure gauge and a needle valve to control the back pressure on the RBSMT which also controlled the system recovery.

PAGE 31

16 Feed Tank Feed water tank (Nanopure wa ter or Treated Lake Manatee water) Influent Sum of the feed flow and the recycle flow (actual influent to membrane cell) Permeate Fraction of influent flow that passes through the membrane Concentrate Fraction of influent flow that does not pass through the membrane Waste Fraction of the concentrate flow that is removed from the system Recycle Fraction of the concentrate fl ow that is recycled to the influent P1 Variable speed feed pump P2 Recycle pump PRF Pressure Relief Valve (set @ 100 psi) FI Flowmeter Influent PI Pressure gauge Influent TI Temperature gauge Influent PS System Pressure gauge (pressure app lied to cell holder to prevent leaks – always greater than influent pressure) PP Pressure gauge Permeate SP Sampling Port Permeate PC Pressure gauge Concentrate SW Sampling port Waste NVC Control valve on the concentrate outlet line NVR Needle valve on the recycle line CV Check valve (prevent b ackflow into the recycle line) WV Waste valve (16 turn metering valve) Figure 4. Flow Schematic of Rapid Bench Scale Membrane System (Falls, 2002) The concentrate line was routed to a tee with one line connected to the waste line and the other line routed back to the RBSMT influent line as a recycle line. A 16 turn P1P2Feed Tank PRV FITIP S Pc PIPPOsmonics Se p a Cell Recycle (Concentrate) Permeate Waste WV NVR CV Influent NVC S W Compressed Gas Tan g ential-Flow Membrane Cell SP

PAGE 32

17 metering valve (Cole Parmer, part number V 07638) was used to control the waste flow rate. The recycle line flowed to a variab le speed recycle pump (Cole Parmer, 75211-10, Gear Pump Drive fitted with a Micr opump, Model 200.150-000, Serial number 377078 gear pump head). The recycle pump of the RBSMT system controls the influent flow rate. A one way check valve (Watts Regulat or) was provided downstr eam of the recycle pump prior to the line joining the feed line. Operating Parameters The RBSMT test cell was operated at the permeate flux and pressure specified by each membrane manufacturer. The crossflow velocity was kept constant throughout all the membrane experiments. The importa nt parameters of flow, pressure, and concentrations were measured at the locati ons shown in Figure 5. Equations used to compute fluxes, flow rates, recoveries and rejection are listed in Table 5. QF – Feed Flow rate PF – Feed Pressure CF – Feed Concentration QP – Permeate Flow rate PP – Permeate Pressure CP – Permeate Concentration QR – Recycle Flow rate PR – Recycle Pressure CR – Recycle Concentration QI – Influent Flow rate PI – Influent Pressure CI – Influent Concentration QW – Waste Flow rate PW – Waste Pressure CW – Waste Concentration Figure 5. Flow Rate, Pressure, and Concentration Measurement Locations Feed Tank QF, PF & CF RBSMT Experiment Cell Influent QI, PI & CI Recycle QR, PR & CRFeed Permeate QP, PP & CP Waste QW, PW & CWRecycle Pump Feed Pump Flowmeter

PAGE 33

18 Table 5 Operating Parameters Equations for RBSMT Operating Parameter Equation No. Water Flux F w wP MTC F 1 Permeate Flow cell w pA F Q 2 Waste Flow P F wQ Q Q 3 Total Volume of Feed Water Required 22 8 A F 785 3 Vcell w 4 Net Driving Pressure 2 / ) P P ( NDPe concentrat inlet 5 Mass Transfer Coefficient NDP / F MTCw w 6 Permeate Recovery % 100 Q / Q RF p 7 Cross Flow Velocity cell I cW T Q V 8 pH Adjustment pH feed water < Feed Feed 12ALK CH 10 4 2 log 9 Rejection 100 )) C / C ( 1 ( R %F p 10 Where, Fw = Water flux (gpd/ft2) MTCw = Mass Transfer Co efficient (gpd/ft2-psi) Qp = Permeate Flow (mL/min) Acell = Surface area of membrane (ft2)=(0.167 ft2) Qw = Waste Flow (mL/min) QI = Influent Flow (mL/min) CHFeed = Calcium Hardness of Feed as CaCO3 (mg/L) ALKFeed = Alkalinity of Feed as CaCO3 (mg/L) QF = Feed Flow (mL/min) V = Volume of Feed Water required (L) NDP = Net Driving Pressure (psi) Pinlet = Inlet Pressure (psi) Pconcentrate = Concentrate Pressure (psi) R = Permeate Recovery (%) Vc = Cross Flow Velocity (ft/sec) T = Feed Spacer Thickness (ft) Wcell = Width of Membrane Thickness (ft) Cp = Permeate Concentration (mg/L) CF = Feed Concentration (mg/L) The influent flow rate to the cell (QI) was calculated from the design cross flow velocity (vc) as reported by the manufacturer, the thickness of th e feed spacer (T=0.0025ft, 0.76mm), and the width of the cell (wcell=0.33ft): QI = vc x T x wcell (11) At a vc of 0.6 ft/sec, QI is 319.9 g/d or 840.9 mL/min. The Osmonics membranes are designed with a vc ranging from 0.5 to 1 ft/sec.

PAGE 34

19 The one pass recovery was estima ted using the following equation: One Pass Recovery = (QP/QI)*100 (12) where, QP is permeate flow rate (mL/min) and QI is influent flow rate (mL/min). The volume of test water needed pe r membrane was estimated through the equation: V=3.785 x Fw x Acell x [(t1/R1) + (t2/R2) + (t3/R3) + (t4/R4)] (13) where, V is the test water vol ume requirement (liters) and tn is the number of days which the cell is operated at the recovery of Rn. Using the specified recoveries and time estimates from the EPA ICR manua l, equation (3) reduces to: V=3.785 x Fw x Acell x 8.22 (14) where, the area of the test cell is 0.167 ft2 and Fw is water flux given in gfd (EPA, 1996). Test Sequence Each RBSMT experiment consisted of setting, recovery and spiked recovery tests. A summary of the sequence of steps used in th e RBSMT is listed in Table 6. The setting procedure involved running th e RBSMT with laboratory Na nopure water, the recovery tests used post filtered Lake Manatee water and the spike tests used post filtered Lake Manatee water spiked with microcystin-LR, geosmin and MIB. The same membrane was used for all three membrane tests of each RBSMT experiment. The setting test was run until a steady st ate mass transfer coefficient (MTC) was achieved, based on a maximum of a 4% change over 12 hours of continuous operation in accordance with ICR protocol at 70% recovery. The recovery test at 70% recovery was then initiated and run for at least 78 hours until the permeate UV254 was stable (less than 3% change over 10 hours).

PAGE 35

20 Table 6 Sequence of Steps Used in RBSMT Experiment Procedure Recovery Water Operate until Setting 70% USF Lab Nanopure Water < 4% change over 12 hours in MTC & samples collected Recovery 70% Manatee Post Filtration Water < 3% change over 10 hours (after 78 hours of operation) in Permeate UV254, conductivity & samples collected Recovery 70% Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254, conductivity & 1 – 1.5 liter samples of permeate collected Recovery 50% Manatee Post Filtration Water <2% change over 1 hour in Permeate UV254, conductivity & 1 – 1.5 liter samples of permeate collected Recovery 85% Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254, conductivity & 1 – 1.5 liter samples of permeate collected Recovery 95% Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254, conductivity & 1 – 1.5 liter samples of permeate collected Spiked 70% Spiked Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254 conductivity & 2 1 liter samples of permeate collected Spiked 50% Spiked Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254 conductivity & 2 1 liter samples of permeate collected Spiked 85% Spiked Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254 conductivity & 2 1 liter samples of permeate collected Spiked 95% Spiked Manatee Post Filtration Water <2% change over 1 hour in Permeate UV-254 conductivity & 2 1 liter samples of permeate collected

PAGE 36

21 The recovery and spike tests were each r un at a sequence of four recoveries of 50%, 70%, 85% and 95%. The predetermined flux rate and cross flow velocity were kept constant over the four recovery tests to isolate the effects of recovery on permeate quality. Sampling was done only when change in permeate UV254 was less than 2% over 1 hour. One membrane, NF270, was tested at three different fluxes of 12, 19, and 32 gfd to quantify the effects of flux on permeate water quality. NF90 and NTR7450 membrane were tested at 19 gfd whereas LFC1 membrane was tested at 10 gfd. Immediately upon completion of recovery te sts, the spike tests were started and operated at the same four recoveries of 50%, 70%, 85% and 95%. Spike tests were performed last in the sequence due to the in terference of methanol which was used to extract and spike microcystin-LR, geosmin a nd MIB. The spike tests were performed by spiking microcystin-LR (9.2 to 12.1 ug/L), ge osmin (45 to 220 ng/L) and MIB (45 to 225 ng/L). Methanol increased the organic car bon content of feed water and consequently increased feed water TOC. Six membrane e xperiments, each run at four recovery tests produced 24 discrete operationa l endpoints for which water quality were analyzed. The results of the membrane experiments were then compared to evaluate membrane performance. System Start Up The RBSMT test cell start up procedure was adapted from the EPA ICR manual (EPA, 1996). Initially, the test apparatus wa s assembled keeping the pressure relief and recovery valves open to minimize system pressure upon start-up. The system was checked for leaks. Then, the feed line was placed into the feed tank and the feed pump turned on.

PAGE 37

22 The pump feed rate was set for constant flux operation, and the pressure relief valve was opened at a pressure below the maximum operating pressure for the system. The recycle pump was turned on and the recovery valve adjusted until the waste flow rate matched the desired setting. The flow rate was adjusted by varying the speed of the recycle pump and adjusting the needle valve to desired influent flow rate. It must be noted that the recycle flow rate was mu ch greater than the feed flow rate. Membrane Setting The membrane setting test was operate d at 70% recovery. This test was performed using lab Nanopure water as the f eed water. The RBSMT was operated with a new membrane and laboratory Nanopure water until the change in the mass transfer coefficient (MTC) over a 12-hour period was le ss than 4%. The pressure and flow readings were monitored during e ach test. The MTC was taken as: MTC = [((Pinlet/Pconcentrate)/2) – (Ppermeate – )] (15) where is the osmotic pressure gradient, assu med to be 0 psi for lab nanopure water with a total dissolved solids (TDS) of 0 mg/L. The permeate flux was calculated using permeate flow (Qp), membrane area (A) of 0.167 and the following equation (9): Fw(gfd)=(Qp mL/min)*(1440min/day)*(1L/ 1000mL)*(1gal/3.785L)*(1/0.167ft2) Recovery Test The recovery tests were in itialized at 70% recovery. The recycle valve, waste valve and the pump speed were adjusted, if re quired, to maintain the constant flux and required recovery. The permeate samples were analyzed for pH, alkalinity, conductivity, turbidity, temperature, calcium hardness, total hardness, TOC, UV254, THMFP and

PAGE 38

23 HAAFP. Similarly, the concentr ate sample was analyzed for UV254, conductivity, pH, calcium hardness and alkalinity. The number of samples taken and analyzed during the recovery tests for a single membrane is summarized in Table 7. Table 7 Water Quality Samples Single Membrane Recovery Test Analysis Parameter Location Recovery % UV254 ConductivityTOCTHMFPHAAFP Ca & Total Hardness Alkalinity Feedwater 2 2 2 3 3 2 2 Permeate 70 2 2 2 3 3 2 2 Retentate 70 2 2 2 2 2 Permeate 50 2 2 2 3 3 2 2 Retentate 50 2 2 2 2 2 Permeate 85 2 2 2 3 3 2 2 Retentate 85 2 2 2 2 2 Permeate 95 2 2 2 3 3 2 2 Retentate 95 2 2 2 2 2 Feedwater 2 2 2 3 3 2 2 After collection of samples at 70% recovery, the membrane system was operated at the next set of operating parameters for r ecovery and flux and all steps were repeated for each test. The remaining three recovery tests then were evaluated over a 24 to 48 hour period. The permeate UV254 was monitored at 20 minute intervals over the first 2 hours to verify stable permeate quality (le ss than a 2% change over 1 hour) before collecting one gallon of permeate and one liter grab concentrate samples. Typically, the system stabilized within two hours.

PAGE 39

24Spike Test Upon completion of the recovery tests, spike tests were conducted by spiking Lake Manatee post filtration water with micr ocystin-LR (8 to 12 g/L), geosmin (45 to 220 ng/L), and MIB (45 to 225 ng/L). Mi crocystin-LR was spiked at 10 g/L concentration into the Lake Manatee post f iltration water. The microcystin-LR was purchased from Alexis Corporation (Catal og number 350-12-C500 and Biomol, Catalog number EI-193) as a white powder residue on th e side of a 2mL gla ss amber vial and was extracted using methanol. Geosmin and MIB were also added at 50 ng/L from a stock solution containing 100 g/mL of each compound in methanol (Supelco, Catalog number 4-7525-U LA86204). The number of samples taken for the cyanobacterial toxin, geosmin and MIB, for the spike tests is summarized in Table 8. Spiked feed water samples were collected and analyzed for initial concentrations of microcystin-LR, geosmin and MIB. The system was allowed to stabilize by operating for at least 1 hour before taking permeate and grab samples. Two 1-liter samples of permeate and one 1-liter grab sample of c oncentrate were collected and analyzed for geosmin and MIB. Similarly, two 1-liter samples of spiked feed water were also collected after the end of spike tests at all rec overies and analyzed for geosmi n and MIB. Triplicate 25 mL samples of initial feed water, permeate, concentrate and final feed water were collected and analyzed for microcystin-LR.

PAGE 40

25 Table 8 Number of Samples Single Membrane Spike Test Analysis Parameter Location Recovery % Microcystin-LR Geosmin MIB Feedwater 3 2 2 Permeate 70 3 2 2 Retentate 70 3 2 2 Permeate 50 3 2 2 Retentate 50 3 2 2 Permeate 85 3 2 2 Retentate 85 3 2 2 Permeate 95 3 2 2 Retentate 95 3 2 2 Feedwater 3 2 2 Note: For the spiked test, microcystin-LR, geosmin and MIB samples analyses were performed by the Manatee County MCWTP. Microcystin-LR analyses we re performed at USF, Environmental Engineering Lab. System Shut-Down The membrane system was shut down by opening the recovery valve. The pressure relief valve was then slowly opene d to relieve the system pressure and the recycle pump was turned off. The feed pump was turned off after turning off the recycle pump. The system pressure was allowed to reach zero and then the concentrate, permeate and influent lines were disconnected from the RBSMT cell. The pressure in the pneumatic cell holder was slowly rel eased and the cell was removed. System Monitoring Operating parameters that were monitored to assess the membrane performance included permeate, waste and feed flow rates, feed, concentrate and permeate pressures,

PAGE 41

26 and influent temperature. These parameters were monitored approximately six times a day at four hour intervals. The membrane feed, permeate, and concentrate were sampled for conductivity, pH, and UV254 approximately four times a day. The conductivity and UV254 measurement were used to monitor the vari ability of permeate quality. The frequency and sample location for each monitoring parameter are given in Table 9. Table 9 Monitoring Frequency During Se tting and Recovery Test Test Operating Parameter Feed Influent Permeate Waste Flow 6 per day 6 per day 6 per day 6 per day Temperature 6 per day Setting Pressure 6 per day 6 per day 6 per day 6 per day Flow 6 per day 6 per day 6 per day 6 per day Pressure 6 per day 6 per day 6 per day 6 per day pH 6 per day 6 per day 6 per day UV254 Absorbance 6 per day 6 per day 6 per day Recovery Conductivity 6 per day 6 per day 6 per day Sample Collection The sampling locations, frequency, and an alytes measured are given in Table 10 for each test and process stream. The membrane feed was sampled twice, once immediately before the test and once at the end of the test, for the following: alkalinity, conductivity, total a nd calcium hardness, pH, TOC, UV254, THMFP, HAAFP, microcystin-LR, geosmin, and MIB. One gallon of permeate was collected from each test at each recovery for the following analyses: pH, alka linity, conductivity total and calcium hardness, TOC, UV254,

PAGE 42

27 THMFP, HAAFP, chlorine demand, microcysti n-LR, geosmin, and MIB. One gallon of concentrate was collected for each test and analyzed for TOC, UV254, conductivity, pH, alkalinity, total and calcium hardness, microcystin-LR, ge osmin, and MIB. Sample holding times, preservation, and sampling techniques were adapted from the “ICR Sampling Manual” (EPA 814-B-96-001). Table 10 Sample Collection Matrix Location Frequency Analyte Setting Test – New Membrane Permeate Once every 4 hours until stable (12-hour period with less than 4% change) UV254 and conductivity Recovery and Spike Tests Permeate Every four hours until stable UV254, conductivity, pH, temperature, hardness and alkalinity Concentrate Once a day until stable UV254, conductivity, pH, temperature, hardness and alkalinity Permeate Once permeate is stable / two samples of one liter pH, alkalinity, turbidity, temperature, TDS, calcium hardness, total hardness, TOC, UV254, THMFP, HAAFP, MYCST-LR, MIB and geosmin Concentrate Once permeate is stable / two samples of one liter TOC, UV254, turbidity, pH, alkalinity, TDS and calcium hardness, MYCST-LR, MIB and geosmin Permeate and concentrate samples from spiked tests (i.e. containing microcystinLR) were collected in 40 mL glass vials a nd stored at 4C and analyzed either at

PAGE 43

28 University of South Florida – Environmenta l Engineering Research lab (EERL) or at MCWTP using Enzyme Linked Immunosorbent Assay (ELISA) tube or plate kit, respectively. For spiked test s, the permeate samples were co llected in a one liter glass bottle with 1mL of 40gm/L of me rcuric chloride as a preser vative. These samples were collected in duplicate. Concentrate sample s were collected in either 250 or 500 mL sample bottles with the same preservative. Source Water The source water used for this project was obtained from the surface water treatment train at the Mana tee County Water Treatment Plant (MCWTP) located in Manatee County, Florida. The surface wate r derived from Lake Manatee, is treated using enhanced coagulation with alum and pol yelectrolytes, flocculation, sedimentation, lime and chlorine addition and granular me dia filtration for rem oval of turbidity and organics as shown in Figure 6. Figure 6. Surface Water Treatment Process Train Manatee County Water Treatment Plant Pre mix Coagulation Flocculation Sedimentation PostMix Filtration Carbon Polymer To Blend Chamber Test Water Sample Site Alum & Lime Chlorine & Lime

PAGE 44

29 Several batch samples of water were collected between D ecember 2002 and April 2003 from MCWTP and transported to the En vironmental Engineer ing Laboratory at University of South Florida for RBSMT analys is. The characteristics of untreated and treated water (enhanced coagulation, floccula tion, and filtration) from Lake Manatee are summarized in Table 11. The addition of lime in the treatment process results in increases in the total and calcium hardness in the post filtration water. TOC reductions of up to 75% are seen through the conv entional surface water treatment. Table 11 Lake Manatee and Post Filtration Water Characteristics Parameters Lake Manatee Water Post Filtration Water Alkalinity, mg/L as CaCO3 25 – 40 3 – 23 Color, as Pt Co 170 – 210 6 – 12 pH 6.9 – 7.5 6.2 – 7.0 Total Organic Carbon, mg/L 12 – 30 3.5 – 4.6 Ultraviolet Abso rbance @ 254 nm, cm-1 0.620 – 0.750 0.064 – 0.120 Total Hardness, mg/L as CaCO3 50 – 70 85 – 120 Calcium Hardness, mg/L as CaCO3 35 – 55 60 – 85 Conductivity, S/cm 80 – 120 190 – 250 Over the course of this pr oject, alum dosages at the plant ranged from 90 to 120 mg/L (dry hydrated alum) with a typical dose of 75 mg/L. Typical ly, the TOC levels in Lake Manatee range from 12-30 mg/L dependi ng on the season and weather conditions. The surface water treatment tr ain, as shown in Figure 6, treats about 40 MGD and removes about 50-75% of the TOC, yielding fi nished water TOC ranging from 3.5 to 4.5 mg/L. Powder Activated Carbon (PAC) is used to control algal taste and odor compounds when needed. The source water for this project was taken after filtration and

PAGE 45

30 before disinfection. Although chlorine was a dded prior to filtration, no residual chlorine was detected in the source water. Analytical Methods The analyses that were performed at EERL included alkalinity, calcium hardness, total hardness, color, UV254, pH, turbidity, conductivity, total dissolved solids, TOC, and microcystin-LR. PeLa Moreaux & Associat es analyzed the total trihalomethane formation potentials (THMFPs) and haloacetic acid formation potentials (HAAFPs). Microcystin-LR, geosmin and MIB were analyzed either at MCWTP. The analytical parameters that were mon itored during the experiment are listed in Table 12. Most of the methods followe d were derived from Standard Methods, Examination of Water and Wastewater, 20th Edition (1998). TOC samples were collected in 40mL vi als. Samples were preserved using phosphoric acid to bring the pH of sample to less than 2 and stored at 4C. Batch analysis of samples was performed using the Siev ers 800 Portable TOC Analyzer (Model # TOC800120V). Nanopure water was placed between each sample to flush the instrument between samples. Ratio of UV254 to TOC is defined as specific UV absorbance (SUVA). Incubation of the samples for THMFPs and HAAFPs test was performed at the EERL according to SM5710 before samples were shipped to the PELA lab for analysis. Permeate samples spiked with sodium hypochlor ite were incubated at 25C so that the free chlorine at the end of seven days was be tween 3 mg/L to 5 mg/L. Sodium thiosulfate was used to quench the chlorine in TH MFP and HAA(6)FP samples. Samples were shipped to PELA labs on ice.

PAGE 46

31 Table 12 Analytical Methods Analyte Method Equipment Detection Limits Alkalinity SM 2320 B H2SO4 titration 20 mg/L as CaCO3 Calcium Hardness SM 3500-Ca Perkin Elmer, Analyst 100, Atomic Absorbtion Spectrophotometer 2 g/L as CaCO3 Total Hardness SM 2340 C Perkin Elmer, Analyst 100, Atomic Absorbtion Spectrophotometer 2 g/L as CaCO3 Color HACH 2120 Hach 4000 3 Pt-Co UV254 HACH 2640 Hach 4000 0.005 cm-1 pH SM 4500-H+ Fisher Scientific Accumet (-)2.000 Temperature SM 2550 Fisher Sc ientific Accumet (-)5.0 C Turbidity SM 2130 B WTW Turb 550 0.01 NTU Conductivity SM 2510 WTW Multiline P4 0.01 S/cm Total Dissolved Solids SM 2540 Gravimetric 10 mg/L TOC SM 5310 Sievers 800 Portable TOC analyzer 0.5 g/L THMFPs EPA 502.2 Gas Chromotography / Mass Spectrophotometer 0.1 g/L HAA(6)FPs EPA 522 Gas Chromotography / Mass Spectrophotometer 0.01 g/L 2-Methylisoborneol SM 6040B Ga s Chromotography / FID 1 ng/L Geosmin SM 6040B Gas Chro motography / FID 1 ng/L Microcystin-LR ELISA E nvirologix Tube/Plate Kit 0.01 g/L SM = Standard Methods for the Exam ination of Water and Wastewater, 20th Edition

PAGE 47

32 Cyanobacterial toxins were analyzed at EERL and MCWTP using the plate / tube kit from Envirologix. The microcystin-LR samples were preserved by storing them at 4C for a maximum of fourteen days before analysis. A detailed description of the method used is described in Appendix D. Geosmin and MIB were analyzed at MCWTP using the method described in Appendix E. The lower detection limit of geosmin and MIB was 1ng/L. Any concentrations of MI B lower that this was reported as below detection limits. Gas chromatography in c onjunction with mass spectrophotometer was used for detection of both MIB and geosmin. Statistical Analysis Statistical analysis using the analysis of variance (ANOVA) test was performed on TOC, DBPFP, microcystin-LR, geosmin, and MI B results to verify if there were any significant differences between each membra ne experiment. ANOVA tests were also used to compare membrane performance at different recoveries. A randomized complete block was de signed and all data, i.e. permeate concentrations for each recovery and each me mbrane experiment, were entered into a Microsoft Excel spreadsheet. This block design was an extension of the paired t-test to compare more than two NF treatments. A sample block design table is presented in Table 13. The data were first checked fo r normal distribution. A null hypothesis was then assumed that all membranes experiments performed the same and statistically there were no differences (Montgomery, 2004). Similar ANOVA tests were performed to verify if there was any significant difference between each membrane test. ANOV A tests indicated whether there existed a difference in NF membrane/te st performance. Follow up tests were performed using

PAGE 48

33 normal distribution to isolate the specific differences between each NF membrane experiment and test performance. Table 13 Randomized Block Design Table for TTHMs Recovery Membrane 50 70 85 95 Total NF90 13.7 15.6 17.0 20.3 66.6 NF270( a ) 23.7 23.4 23.9 26.6 97.6 NF270 ( b ) 22.7 22.9 23.2 25.1 93.9 NF270( c ) 30.9 34.8 27.3 28.1 121.1 LFC1 28.7 25.1 31.3 41.7 126.8 NTR7450 75.1 87.1 108.1 123.3 393.6 Totals 194.8 208.9 230.8 265.1 899.6 Averages 32.5 34.8 38.5 44.2

PAGE 49

34 RESULTS This section summarizes the results obta ined testing six membranes for TOC, UV254, DBP precursors, and algal exudates reje ction. Detailed results showing time variations of the permeate flux, permeate and waste flow, influent and concentrate pressure, MTC, temperature corrected MTC, UV254, conductivity, pH, recovery, and one pass recovery are given in Appendix F. Setting and Recovery Tests A summary of the permeate flux versus tim e for the setting test is presented in Figure 7. As shown the flux rates for all th e membranes decreased with time during the setting procedure. 0 5 10 15 20 25 30 35 40 45 50 55 60 051015202530Time (hrs)Permeate Flux (gfd) LFC1 10 gfd NTR 7450 19 gfd NF270 19 gfd NF270 32 gfd NF90 19 gfd NF270 12 gfd Figure 7. Setting Test – Pe rmeate Flux Versus Time

PAGE 50

35 At the beginning of the membrane tests, the membranes are free of contaminants and should achieve higher flux ra tes. The flux rate for NF270 membrane test at 32 gfd increased initially for one hour and then star ted to gradually decrease over time. Over time, contaminants accumulate on the membrane surface resulting in lower flux rates. The differential pressure, defined as the pre ssure difference between the influent and the concentrate, also increases with time if any membrane fouling is observed. Recovery tests were performed to qua ntify the rejection capabilities of each membrane. The NF270 membrane was tested at three permeate flux rates: 32 gpd/ft2, 19 gpd/ft2, and 12 gpd/ft2. The results from testing at 19 gpd/ft2 are discussed here. The permeate flux rate was maintained at 19 gpd/ft2 and the mass transfer coefficient ranged between 0.32 and 0.36 gpd/ft2-psi during the recovery test. Permeate and waste flow rates for the recovery test are presented in Figure 8. Figure 8. Flow-Rate versus Time, Recovery Tests for NF 270 Membrane Test at 19gfd

PAGE 51

36Mass Transfer Coefficient MTC is the amount of water produced as permeate per unit area of membrane per unit of applied pressure. It is one of the most important calculated operational parameter because it is one of the first parameter that will indicate any membrane fouling. The results of the average MTCs for each membrane experiment are presented in Table 14. The MTC for each membrane experiment re mained constant during the entirety of the test. This suggests that there was little or almost no f ouling over the time period of the tests as the flow was kept constant and the pressures varied s lightly. RBSMT for a longer period of time (minimum of 12 weeks) is require d to observe any membrane fouling. Higher MTCs indicate more water production per area of membrane and applied pressure than membrane with lower MTCs LFC1 membrane had the lowest MTC indicating that this membrane would require the maximum pr essure for achieving target flux rate. Similarly, the NF270 membrane ha d the highest MTC indi cating that it would require the least pressure for achieving target flux rate. Table 14 Mass Transfer Coefficient for all Membrane Tests Membrane Operating Flux gpd/ft2 Mass Transfer Coefficient gpd/ft2-psi NF90 19 0.29 NF270 (a) 32 0.41 NF270 (b) 19 0.32 NF270 (c) 12 0.29 LFC1 10 0.15 NTR7450 19 0.32

PAGE 52

37Nanofiltration Membrane Rejection Performances The section of the chapter is divide d into three subsections: (1) TOC and UV254 rejection (2) DBP precursor rejection, a nd (3) Microcystin-LR, geosmin, and MIB rejection. Comparison of Membranes for Rejection of TOC and UV254 The feed water TOC ranged from 3.5 to 4.5 mg/L in the post filtration Lake Manatee water. NF permeate TOC concen trations ranged from 0.154 to 2.02 mg/L depending on the membrane, flux rate and overall system recovery. Higher TOC rejections corresponded with lowe r recoveries. The feedwater (CF), permeate (CP), and the percent removal of UV254 based on permeate recovery and membrane are listed in Table 15. SUVA values are also presented. TOC rejections of tested membranes versus recovery are shown in Figure 10. The TOC percent rejections ranged from 46 to 96% based on membranes and permeate recovery. Both NF90 and NF270 showed greater than 90% rejection of TOC, even at the higher recoveries of 95%. The NF270 in comp arison to NF90 at same flux rate of 19 gfd, ran at lower inlet pressure. This coul d impact full-scale operations by decreasing operational costs. In contrast, the NT R7450 membrane rejected 70% TOC at 50% recovery which decreased to 46% at 95% recovery. The NTR7450 had the let rejection whereas LFC 1 showed the highest rejection of TOC. All membranes except for NTR 7450 showed a TOC rejection greater than 90 % for all reco veries tested. A similar trend was observed for the UV254 rejection at the va rious recoveries and fluxes tested. The average influent UV254 was 0.08 cm-1 whereas the permeate UV254 varied from 0.0006 to 0.042 cm-1. UV254 rejection correlated well with TOC rejection for

PAGE 53

38 all membrane experiments. TOC and UV254 rejection for the three membrane experiments for NF270 at four recoveries tested are presented in Figure 9. 80 82 84 86 88 90 92 94 96 98 100 405060708090100 Recovery (%)TOC and UV-254 Rejection (%) TOC 32 gfd TOC 19 gfd TOC 12 gfd UV-254 32 gfd UV-254 19 gfd UV-254 12 gfd Linear (UV-254 19 gfd) Linear (TOC 12 gfd) Linear (TOC 19 gfd) Linear (TOC 32 gfd) Linear (UV-254 12 gfd) Linear (UV-254 32 gfd) Parameter Slope Regression (R2) TOC 32 gfd -0.0678 0.94 TOC 19 gfd -0.0661 0.87 TOC 12 gfd -0.0687 0.87 UV254 32 gfd -0.1013 0.94 UV254 19 gfd -0.1301 0.87 UV254 12 gfd -0.0626 0.99 Figure 9. TOC & UV254 versus Recovery for NF270 Membrane As the permeate recoveries and flux ra tes increased, rejection of target contaminants decreased. The decrease in reje ction may reflect that the rejection by NF is diffusion controlled. Also, the slope of decr ease in rejection as recovery increases is more significant for NTR 7450 than other membranes. SUVA for the feed water ranged from 1.7 to 2.3 L/mg-m which is characteristi c of nonhumic material (Volk et al., 2002). Also, water pretreated by chemical coagulat ion and conventional filtration are expected to have these values of SU VA (Volk et al., 2002). The trend for rejection of UV254 and conductivity at diffe rent recoveries along with the permeate concentrations are shown in Figure 11. The feed concentration of UV254 ranged from 0.068 to 0.081 cm-1 and the feed concentrati on of conductivity ranged from 220 to 235 S/cm.

PAGE 54

39 Table 15 TOC, UV254 and SUVA Data for NF Membranes Tested TOC (mg/L) UV254 (cm-1) SUVA (L/mg-m) Membrane Recovery CF 2 CP 3 Rej.1 CF 2 CP 3 Rej.1 CF 2 CP 3 NF90 50 4.52 0.227 95.0 0.080 0.0006 99.3 1.76 0.26 70 4.52 0.268 94.0 0.080 0.0011 98.6 1.76 0.41 85 4.52 0.287 93.6 0.080 0.0021 97.7 1.76 0.73 95 4.52 0.319 92.9 0.080 0.0026 96.7 1.76 0.82 NF270 (a) 50 3.81 0.242 93.6 0.079 0.0014 98.3 2.07 0.58 70 3.81 0.288 92.4 0.076 0.0021 97.3 2.07 0.73 85 3.81 0.311 91.8 0.076 0.0029 96.2 2.07 0.93 95 3.81 0.369 90.3 0.074 0.0034 95.5 2.07 0.92 NF270 (b) 50 3.62 0.183 94.9 0.075 0.0011 98.6 2.07 0.60 70 3.62 0.198 94.5 0.078 0.0019 97.7 2.07 0.96 85 3.62 0.245 93.2 0.071 0.0033 95.4 2.07 1.34 95 3.62 0.302 91.7 0.068 0.0051 92.5 2.07 1.69 NF270 (c) 50 3.55 0.179 95.0 0.081 0.0011 98.8 2.28 0.61 70 3.55 0.192 94.6 0.076 0.0020 97.4 2.28 1.04 85 3.55 0.233 93.4 0.071 0.0033 95.4 2.28 1.41 95 3.55 0.287 91.9 0.073 0.0045 93.9 2.28 1.56 LFC1 50 3.77 0.154 95.5 0.078 0.0009 98.9 2.07 0.58 70 3.77 0.171 95.9 0.080 0.0017 97.9 2.07 0.94 85 3.77 0.185 95.1 0.071 0.0031 95.7 2.07 1.67 95 3.77 0.225 94.0 0.068 0.0039 94.3 2.07 1.73 NTR 7450 50 3.75 1.14 69.6 0.076 0.0250 67.6 2.03 2.19 70 3.75 1.41 62.4 0.078 0.0330 57.7 2.03 2.34 85 3.75 1.68 55.2 0.076 0.0360 52.7 2.03 2.14 95 3.75 2.02 46.1 0.079 0.0420 46.9 2.03 2.08 1Rej. – Percent Rejection 2CF-feed concentration 3 CP-permeate concentration

PAGE 55

40 0 10 20 30 40 50 60 70 80 90 100 405060708090100Recovery (%)TOC Rejection (%) NF 270 32 gfd NF 270 19 gfd NF 270 12 gfd NF90 19 gfd NTR 7450 19 gfd LFC1 10 gfd Figure 10. TOC Rejection for NF Me mbranes at Tested Recoveries 0 10 20 30 40 50 60 70 80 90 100 03102647678084859294Time (hours)Conductivity (uS/cm)0 0.001 0.002 0.003 0.004 0.005 0.006UV-254 (cm-1) Permeate Conductivity Permeate UV-254 70 % 85 % 50 % 95 % Figure 11. Time Variation of Permeate Conductivity and UV254 for NF270 at 19gfd

PAGE 56

41 The permeate concentration of UV254 for NF270 membrane at 19 gfd remained almost constant throughout th e stabilization period of r ecovery test at about 0.018 cm-1. Once the recovery changed the UV254 permeate concentration varied from 0.011 cm-1 to 0.048 cm-1 depending on the recovery. As the recovery increased the permeate water quality decreased. Similar trends were observed for conductivity. The permeate concentration of conductivity ranged from 35 S/cm at 50% recovery to 85 S/cm at 95% recovery. Statistical analysis using ANOVA concl uded that the NF90, LFC1, and NTR7450 membranes behaved significantly differently The NF90 and LFC1 membranes produced significantly lower permeate concentrations of TOC whereas NTR7450 membrane produced significantly higher permeate concen trations of TOC when compared to the mean concentrations of all the membrane expe riments. A similar analysis for recoveries was also performed, but there was not enough evidence statistically to state that the membrane tests at different recoveries were si gnificantly different. However, there exists a linear relationship of membrane recovery to permeate TOC concentration as seen in Table 16. Relationship of TOC with Recovery A linear relationship of TOC rejection versus recovery was observed for all NF membranes and correlation coefficients are pr esented in Table 16. The linear regression value (R2) for the NF90 membrane experiment was calculated as 0.98. Good correlations were observed for all the membrane tests as seen in Table 16. A lthough statistically, the test runs at different recoveries were not significant different, there exists a linear relationship of membrane rejection with recovery.

PAGE 57

42 Table 16 Linear Regression Values for Comparison of Permeate TOC Correlation with Recovery Membrane Slope Intercept Correlation Coefficient R2 NF90 0.002 0.13 0.98 NF270(a) 0.003 0.11 0.94 NF270(b) 0.003 0.04 0.86 NF270(c) 0.002 0.05 0.85 LFC1 0.001 0.08 0.86 NTR7450 0.019 0.15 0.96 Comparison of Membranes for Rejection of DBP Precursors THMFPs and HAA(6)FPs were monitored in this project. The range of THMFP rejection ranged from 17 % to 91 % for all membranes. The feedwater concentrations (CF) and the permeate concentrations (CP) along with the percen tage rejection for THMFP and HAAFP are listed in Table 17. The feed water THMFP varied from 113 to 154 g/L. The THM precursor rejection achieved for all NF membranes tested except fo r NTR7450 was greater than 70 % at the recoveries tested. A summary of the results for THM precursor rejection by all membranes is presented in Figure 13 and is also compared to the Stage 1 EPA MCL for THM. It also presents the concentration of the individual species of THMFP in the feed and permeate water. The NF270 membrane was tested at three different recoveries of 12, 19 and 32 gfd. As the flux and the recoveries increas ed, the THM precursor re jection decreased. NF90 showed the best rejection (86 to 91%) capabilities whereas NTR7450 membrane showed lower (17 to 48%) rejections for THMFP. Graphical re presentation of THMFP rejections for all NF membranes tested in th is research is presented in Figure 12.

PAGE 58

43 10 20 30 40 50 60 70 80 90 100 405060708090100Recovery (%)THMFP Reduction (%) NF270 32 gfd NF270 19 gfd NF270 12 gfd NF90 19 gfd NTR7450 19 gfd LFC1 19 gfd Figure 12. THMFP Reduction Percentage s for Nanofiltration Membranes Tested The TTHMs and HAA(5)s formed due to chlorination before filtration at MCWTP were in the range of 5-7 g/L and le ss than 1 g/L respectively (Mark Simpson, MCWTP, April 2004). These concentrations have not been considered for THMFP and HAAFP rejections in this research. If these concentrations are considered, the rejections would either be the same or higher than the rejections mentioned in this section. ANOVA analysis suggested that NF90 and NTR7450 membranes behaved significantly differently for DBPFP reduction. The NF90 membrane produced lower permeate concentrations of THMFP whereas NTR7450 membrane produced much higher permeate concentrations of THMFP when comp ared to the mean concentrations of all membrane tests. A sample matrix was then made to evaluate if there was any significant difference in the membrane tests at differe nt recoveries. Ther e was not enough evidence statistically to state that the membrane test s at different recoveries were significantly different. Again, similar to TOC rejection, a linear relationship of membrane recovery to permeate THMFP concentration exists.

PAGE 59

44 Table 17 THMFP & HAAFP Data for NF Membranes Tested THMFP4 HAA(6)FP5 CF 2 CP 3 Rejection1 CF 2 CP 3 Rejection1 Membrane Recovery g/L g/L % g/L g/L % NF90 50 142.5 13.7 90.4 283.6 ND 70 142.5 15.6 89.1 283.6 ND 85 142.5 17.0 88.2 283.6 ND 95 142.5 20.3 85.8 283.6 ND NF270 (a) 50 112.6 23.7 79.0 436.0 ND 70 112.6 23.4 79.2 436.0 ND 85 112.6 23.9 78.8 436.0 ND 95 112.6 26.6 76.4 436.0 48.4 90.9 NF270 (b) 50 116.3 22.7 80.5 49.8 3.8 92.3 70 116.3 22.9 80.3 49.8 0.7 98.6 85 116.3 23.2 80.1 49.8 5.3 89.4 95 116.3 25.1 78.4 49.8 4.8 90.3 NF270 (c) 50 153.3 30.9 79.8 141.0 21.4 84.8 70 153.3 34.8 77.3 141.0 21.6 84.7 85 153.3 27.3 82.2 141.0 43.0 69.6 95 153.3 28.1 81.7 141.0 53.5 62.1 LFC1 50 143.8 28.7 80.0 131.6 8.5 93.5 70 143.8 25.1 82.5 131.6 26.7 79.7 85 143.8 31.3 78.2 131.6 7.3 94.4 95 143.8 41.7 71.0 131.6 11.6 91.2 NTR 7450 50 145.9 75.1 48.5 220.3 32.7 84.2 70 145.9 87.1 40.3 220.3 52.9 74.5 85 145.9 108.1 25.9 220.3 156.2 24.7 95 145.9 121.3 16.9 220.3 171.0 17.5 1Rej. Percent Rejection 4 THMFP EPA Method 502.2 Detection limit upto 0.1 g/L 2 CF-feed concentration 5 HAAFP EPA Method 552 Detection limit upto 0.01 g/L 3CP-permeate concentration

PAGE 60

45Figure 13. Feed and Permeate THMFP Concentration with Speciation 0 20 40 60 80 100 120 140 160 180F e edw a ter ( N F 90 ) P e r m e a te 50 % Permeate 70 % Permeate 85 % Perme a te 95 % Feed w ater (N F 2 70 a ) Perme a te 50 % Perme a te 70 % Perme a te 85 % Perme a te 95 % F ee d wat er ( NF 2 70 b ) Per m e a te 50 % Perme a te 70 % P e r m e a te 85 % Per m e a te 95 % Feed w ater (N F 2 70 c ) Permeate 50 % Perme a te 70 % Perme a te 85 % Perme a te 9 5 % Feedwater (LFC1 ) Perme a te 5 0 % Perme a te 70 % Perme a te 85 % Perme a te 95 % Feed w ater (N T R 7450 ) P e r m e a te 50 % P e r m e a te 70 % P erm e at e 85 % P e r m e a te 95 % Nanofiltration Membrane and Water TypeTHMFP (ug/L) CHBr3 CHClBr2 CHCl2Br CHCl3 NF 90 19 gfd NF 270 (a) 32 gfd NF 270 (b) 19 gfd NF 270 (c) 12 gfd LFC 1 10 gfd NTR 7450 19 gfd EPA Stage 1 THMFP Regulation EPA Stage 1 THM Regulation

PAGE 61

46 Similar or even higher rejections we re observed for all membranes for the HAA(6)FP. The rejection ranged from 18 to 99 %. Higher HAA(6)FP rejections were observed when compared to THMFP rejections as seen in Table 17. TOC and UV254 rejection were similar to the rejections of DBP precursors monitored and provide a screening tool to assess DBP precursor rejection. The graphical representation of permeate THMFP and SUVA values versus system recovery for NF90 membrane test is shown in Figure 14. Rejection for THMFP di d not correspond to SUVA values as similar recoveries. At higher rec overies, the SUVA values as well as the permeate THMFP increases. The speciation of the disinfection by products in feed water and permeate indicated that chloroform co mprised about 63 % of the total THMFP formed in feed water and permeate water. Currently, MCWT P uses chloramines for disinfection and produces about 60-70% chloroform in finish ed water (Mark Simpson, MCWTP, April 2004). 0 10 20 30 40 50 60 70 80 90 100 405060708090100 Recovery (%)THMFP and SUVA Rejection (%) THMFP Rejection SUVA Rejection Linear (SUVA Rejection) Linear (THMFP Rejection) Parameter Slope Regression (R2) THMFP Rejection -0.0943 0.90 SUVA Rejection -0.7439 0.95 Figure 14. SUVA and THMFP Permeat e Concentration Versus Recoveri es for NF90 Membrane Test

PAGE 62

47Comparison of Membranes for Rejection of Algal Exudates The post filtered Lake Manatee water was spiked with algal exudates for performing the spiked test. These include spiking of microcystin-LR with 10 g/L, geosmin and 2-methylisoborneol (MIB) both 50 ng/ L. The concentrations of feed water, permeate, and concentrate for microcystin-LR ar e presented in Table 18. The feed water concentration of microcys tin-LR varied from 9.7 g/L to 12.4 g/L. The rejection of microcys tin-LR varied from 94% to 99% except for NTR7450 which showed comparably low rejections in the range 80 % to 87 % depending on the recoveries. It should also be noted that even at highe r recoveries of 95% all the membranes, except for NTR 7450 membrane, we re able to achieve, the WHO limit of 1g/L for microcystin-LR at a spiked concentrat ion of 9.7 g/L to 12.4 g/L resulting in a permeate concentrations from 0.12 to 0.96 g /L. The results from the microcystin-LR spiked tests are presented graphically in Figure 17. Microcystin -LR rejection for the three fluxes tested for NF270 memb rane is shown in Figure 15. Higher rejections (greater than 50%) of MIB and geosmi n were also obtained for NF90, NF270 and LFC1 membrane as show n in Table 15 and Figure 18. NF90 and LFC1 membranes performed the best with rejections of greater than 95%. NF270 membrane at flux rates of 12 and 19 gfd perfor med better (> 74% rejection) than at a higher flux rate of 32 gfd (>67% rejection). The feed concentration of geosmin varied from 43 to 215 ng/L and the permeate concentration varied from <1 ng/L to 51 ng/ L. The threshold odor concentration for geosmin is 10 ng/L (McGuire et al, 1981). Geosmin rejection for the three fluxes tested for NF270 membrane is shown in Figure 16.

PAGE 63

48 80 82 84 86 88 90 92 94 96 98 100 101520253035Flux (gfd)Microcystin-LR Rejection (%) 50% Recovery 70% Recovery 85% Recovery 95% Recovery Linear (50% Recovery) Linear (95% Recovery) Linear (85% Recovery) Linear (70% Recovery) Linear (50% Recovery) Parameter Slope Regression (R2) 50% Recovery -0.0796 0.99 70% Recovery -0.0976 0.98 85% Recovery -0.0311 0.96 95% Recovery -0.1471 0.99 Figure 15. NF270 Microcystin-L R Rejection Function of R ecovery and Permeate Flux 20 30 40 50 60 70 80 90 100 101520253035Flux (gfd)Geosmin Rejection (%) 50% Recovery 70% Recovery 85% Recovery 95% Recovery Linear (50% Recovery) Linear (95% Recovery) Linear (85% Recovery) Linear (70% Recovery) Linear(50%Recovery) Parameter Slope Regression (R2) 50% Recovery -1.0413 0.96 70% Recovery -1.7233 0.93 85% Recovery -1.8364 0.99 95% Recovery -1.9422 0.88 Figure 16. NF 270 Geosmin Rejection – Function of Recovery and Permeate Flux

PAGE 64

49 Table 18 Microcystin-LR, Geosmin and 2-Met hylisoborneol Data for NF Membranes Microcystin-LR (g/L) Geosmin (ng/L) 2-Methylisoborneol (ng/L) Membrane Recovery CF 2 CP 3 Rej.1 CF 2 CP 3 Rej.1 CF 2 CP 3 Rej.1 NF90 50 12.4 0.12 99.0 215 6.9 97 210 5.4 97 70 12.4 0.16 99.0 215 7.3 97 210 11.8 95 85 12.4 0.17 99.0 215 7.8 97 210 12.1 94 95 12.4 0.22 98.0 215 8.8 97 210 19.5 92 NF270 (a) 50 9.7 0.28 97.1 115 21.2 81.6 122 12.8 89.5 70 9.7 0.35 96.4 115 29.2 74.6 122 23.0 81.2 85 9.7 0.30 97.0 115 32.8 71.5 122 26.8 78.1 95 9.7 0.59 93.9 115 38.6 66.5 122 44.4 63.7 NF270 (b) 50 10.2 0.20 98.1 52 10.2 82.0 55 5.1 91.6 70 10.2 0.22 97.9 52 10.3 82.0 55 11.6 80.9 85 10.2 0.27 97.5 52 15.4 72.9 55 11.1 81.8 95 10.2 0.42 96.0 52 15.1 73.4 55 17.1 71.9 NF270 (c) 50 10.4 0.14 98.7 40 5.9 85.3 53 5.1 90.8 70 10.4 0.18 98.3 40 6.1 84.9 53 8.0 85.4 85 10.4 0.26 97.6 40 7.1 82.3 53 7.1 87.7 95 10.4 0.34 96.8 40 10.8 73.0 53 12.0 77.9 LFC1 50 9.8 0.10 99.0 52 BDL 43 BDL 70 9.8 0.12 98.8 52 BDL 43 1.97 96.0 85 9.8 0.16 98.5 52 0.45 99.0 43 0.50 99.0 95 9.8 0.21 98.0 52 BDL 43 0.45 99.0 NTR 7450 50 9.4 1.21 87.6 49 45.2 16.9 54 42.4 27.3 70 9.4 1.65 83.1 49 42.1 22.9 54 49.1 15.9 85 9.4 1.85 81.1 49 41.0 24.6 54 46.5 20.4 95 9.4 1.96 79.9 49 37.8 30.6 54 51.5 11.8 1Rej. Percent Rejection 2CF-feed concentration 3 CP-permeate concentration

PAGE 65

50 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0N F 90 Pe r mea t e 50% Permeate 70% Pe r mea t e 85% Permeate 95% NF27 0 (a ) Permeate 50% Permeate 7 0% Permeate 85% P ermeate 9 5% NF270 (b ) Permeate 5 0% P erm e at e 70 % P ermeate 8 5% P erm e at e 95 % N F 270 (c ) Pe r mea t e 50% P er mea t e 70 % Pe r meate 85% Pe r mea t e 95% L F C 1 Permeate 50% Pe r meate 70% Permeate 85% Permeate 95% NTR 7 4 50 P ermeate 5 0% Permeate 70% P ermeate 8 5% Permeate 9 5% Water SampleMicrocystin-LR (ug/L) Microcystin-LR NF 90 19 gfd NF 270 (a) 32 gfd NF 270 (b) 19 gfd NF 270 (c) 12 gfd LFC 1 10 gfd NTR 7450 19 gfd WHO Regulation fo r Microcystin-LR Figure 17. Microcystin-LR Concentration in Pe rmeate Water of Each Nanofiltration Membrane at Recoveries Tested for Water Spiked with 8.5-12.0 g/L

PAGE 66

51 0 10 20 30 40 50 60N F 9 0 Pe r me a te 50 % Permeate 70% Pe r meate 85% Permeate 95% NF270 ( a ) P e r me ate 5 0 % Per me ate 7 0 % P e r me a t e 8 5 % Per me ate 9 5 % N F 27 0 (b ) Pe r me a t e 50 % Pe r meate 70% Pe r me a t e 85 % Permeate 95% N F 270 ( c ) Per me ate 5 0 % Permeate 70% Per me ate 8 5 % Permeate 95% L F C1 P e r me a t e 5 0 % Pe r me a t e 70 % P e r me a t e 8 5 % Pe r meate 95 % N T R 74 5 0 Pe r me a t e 50 % Permeate 70% Perm e ate 8 5 % Permeate 95% Water SampleMIB, Geosmin (ng/L) MIB Geosmin NF 90 19 gfd NF 270 (a) 32 gfd NF 270 (b) 19 gfd NF 270 (c) 12 gfd LFC 1 10 gfd NTR 7450 19 gfd Odor Threshold Limit Geosmin Odor Threshold Limit MIB Figure 18. 2-Methylisoborneol & Geosmin Concentration in Permeate Water of Each Nanofiltration Membrane at Recoveries Tested for Water Spiked with 40-215 ng/L

PAGE 67

52 Membrane tests for NF90, NF270 at 12 gfd, and LFC1 produced permeate that did not exceed the threshold odor concentration. The NT R 7450 and NF 270 at 19 gfd and 32 gfd were among the membrane te sts which did not meet the threshold concentration for geosmin. The feed concen tration of MIB varied from 43 ng/L to 210 ng/L. The NF90 and NF 270 membranes were spiked at higher concentrations to quantify the rejection at hi gher concentrations. The permeate concentration for MIB ranged from < 1ng/L to 51 ng/L. The lower detection limit of MIB was 1 ng/L. NF 270 MIB rejection for recoveries a nd permeate flux tested is presented in Figure 19. All membranes reduced the permeate concen trations of MIB to less than 29 ng/L except for the NTR 7450 and 95% recovery for NF 270 membrane at 32 gfd. It must be noted, however, that the NF 270 membrane at 32 gfd was tested at higher feed concentration of MIB. 0 10 20 30 40 50 60 70 80 90 100 101520253035Flux (gfd)MIB Rejection (%) 50% Recovery 70% Recovery 85% Recovery 95% Recovery Linear (95% Recovery) Linear (85% Recovery) Linear (70% Recovery) Linear (50% Recovery) Parameter Slope Regression (R2) 50% Recovery -0.0772 0.54 70% Recovery -0.1806 0.53 85% Recovery -0.4291 0.93 95% Recovery -0.7002 0.99 Figure 19. NF 270 MIB Re jection Function of Recovery and Permeate Flux

PAGE 68

53 ANOVA analysis indicated that these membranes behaved differently for microcystin-LR, geosmin, and MIB rejection. The NF90 and LFC1 membrane produced lower permeate concentrations of micr ocystin-LR and geosmin whereas NTR7450 membrane produced much higher permeate c oncentrations of microcystin-LR and geosmin when compared to the mean concentrat ions of all membrane tests. For the MIB, LFC1 and NF270 at 12 gfd produced lower pe rmeate concentrations whereas NTR 7450 membrane test produced higher permeat e concentrations. ANOVA analysis was performed on recoveries tested. There was not enough evidence to state that the membrane tests at different recoveries were significantly different for microcystin-LR and MIB. The null hypothesis that the performance was the same at all recoveries was not rejected for the permeate geosmin concentrat ion data. Again, a linear relationship of membrane recovery to permeate c ontaminant concentration exists. Relationship of TOC with Algal Exudates Removal The TOC rejection, if related to the al gal exudates rejection, will provide the water treatment plant a surrogate parameter for estimating algal exudates removal. The TOC rejection is compared to the rejection of microcysti n-LR, geosmin, and MIB and plotted for NF90 membrane test in Figure 20. As seen in Figure 20, a linear relationship exists between TOC rejection and microcystin-LR rejection at di fferent recoveries. Similar relationships exist between TOC rejection and geosmin and MIB rejection. The regression values for these plots are 0.90, 0.91 and 0.90 for microcystin-LR, ge osmin and MIB respectively suggesting a linear relationship with permeate TOC concentr ation. The relations hip could be used to

PAGE 69

54 estimate the rejection of algal exudates as a function of TOC rejection. These correlations are specific for Lake Manatee post filtration water. y = 2.8505x 172.91 R2 = 0.9002 y = 0.377x + 60.80 R2 = 0.9112 y = 0.4713x + 54.23 R2 = 0.9031 90 91 92 93 94 95 96 97 98 99 100 90919293949596TOC Rejection (%)Microcystin-LR, Geosmin and MIB Rejection (%) TOC vs Microcystin-LR Rejection TOC vs Geosmin Rejection TOC vs MIB Rejection Figure 20. TOC Rejection ve rsus Microcystin-LR, Geos min and MIB Rejection for NF90 Membrane at Four Different Recoveries

PAGE 70

55 DISCUSSION In this chapter, the relative performance of the membranes is evaluated in terms of rejection of target compounds. For the f our commercially available NF membranes tested in this study, the RB SMT was effective at identifyi ng a subset of membranes for further testing. Comparison of Membranes Performance To select membranes for water treatment applications, the relative performance of membranes can be compared in terms of rejection of TOC, UV254, DBP precursor, and algal exudates. A comparison of rejection at 70% recovery is given in Table 19. Based on these results, NF90, NF270, and LFC1 memb ranes performed better than NTR7450 in this study. Table 19 Comparison of Membrane Performance base d on Percent Remova l at 70% Recovery Percent Removal Membrane TOC UV254 THMFP HAAFP Microcys tin-LR GeosminMIB NF90 94 98 89 99 99 97 95 NF270 a 93 97 79 99 96 75 81 NF270 b 95 98 80 98 98 82 81 NF270 c 95 98 78 85 99 85 86 LFC1 96 98 83 80 99 99 96 NTR7450 63 58 40 74 83 54 16 a NF270 at 32 gfd b NF270 at 19 gfd c NF270 at 12 gfd

PAGE 71

56 Studies conducted by Falls (2002), suggest ed that the NF90 membrane at 19 gfd displayed a rejection of 95 to 98 % for TOC and 89% for THMFP. Similar results were seen in this research with a 93 to 95 % re jection for TOC and 86 to 90 % rejection for THMFP. Falls (2002) also s uggested that the NF270 membrane s at 32 gfd rejected 94 to 96 % TOC and 88 % THMFP. These results were confirmed in this research by achieving rejections of 91 to 94 % for TOC and 88 to 91 % for THMFP. Taylor (2000) evaluated the NTR7450 membrane at 15 gfd on conventionally treated Floridian surface water and achieved a rejection of 71% for TOC and 68% for THMFP. Lower rejections were obtained from this research by gett ing rejection of 46 – 70% rejection for TOC and 16 to 49 % rejecti on for THMFP. This could be due to the different influent water quality. Reis s (1999) rejected 93% TOC using the LFC1 membrane. TOC rejection was similar in this research for LFC1 membrane with 94 to 96 % rejection. Veerapaneni (2001) used the LFC1 membrane on conventionally treated surface water and achieved 72 % rejection of THMFP. The RBSMT results in this research showed similar rejections of 71 to 80 % THMFP. Although NF90, LFC1, and NF270 membranes displayed good rejection for target contaminants at recovery rates of 85 and 95%, NF270 membrane is recommended for further bench and pilot testing as along with high rejections it was coupled with lower feed pressure. This would increase the MTC and thereby productivity at a lower operational costs. Operational Parameters The advantage of running the membrane at higher recoveries is that more permeate water is produced and less amount of concentrate is generated. It was

PAGE 72

57 important to know the contaminant rejections at higher permeate r ecoveries of 85 % and 95% as the pilot plant testing and full scale pl ant are expected to run at these recoveries rather than recoveries of 50 or 70%. Therefore, the flux rate and the permeate recovery at which the membrane should be operated to optim ize the treatment process is important to know. This study shows that recovery of NF system influences permeate water quality. Other important operational parameter is the feed pressure at which the NF treatment operates. Higher feed pressure fo r the optimized flux rate and recovery will increase the operational costs. Amongs t the membranes tested, NF270 and NTR7450 membranes operated at the lowest feed pressures. However, NF270 membranes displayed higher rejections of target compounds. Wh ile, the NTR7450 had lower rejection of TOC and DBPFP, it has lower opera ting costs (due to lower pressures) and may foul less depending on the treatment goals and the influent water quality. It may meet the design objectives at a lower cost that the other membranes. Concentrate Disposal Concentrate disposal from NF treatment is a significant problem due to increased regulations by state / federal gov ernments on it disposal. This stream could be hazardous in nature due to the high concentration of chemicals and algal toxin contaminants. Research should be performed to determine the treatability of this stream. Pre-treatment may be necessary before the disposal of th is stream. The stream could be diluted by diverting it into the influent of a conventiona l waste water treatment plant. Other options included land application, d eep well injection and evapora tion ponds. The driving force for the decision for waste disposal tec hnology will be the quality and quantity of contaminants.

PAGE 73

58 CONCLUSIONS The objectives of this study were to evaluate four commercially available nanofiltration (NF) membranes at different recoveries and fluxes to determine the feasibility of NF as a polishing treatment fo r removal of disinfection by-products (DBPs) and algal exudates from conventionally treate d surface water. The conclusions derived from this research are: (1) Three of the four NF membranes test ed: NF90, NF270, and LFC1, showed the ability to reject a high per centage of TOCs (greater th an 90%) from conventionally treated surface water at all the recoveries tested. Higher rejections of THM and HAA precursors using NF treatment correlated with higher rejections of TOC and UV254 rejection at the rec overies tested. (2) The rejection of TOC and precursors of THM and HAA(6) decreased with increasing feedwater recovery suggesting th at the transport of these materials is controlled by molecular diffusion across th e membranes than by physical sieving at membrane surface. (3) Three nanofiltration membranes (NF90, NF270, and LFC1) achieved excellent rejection for microcystin-LR, alga l toxin. These NF membranes reduced the permeate concentration of micr ocystin-LR to less than 1 g/L, which is the WHO guideline for all the recoveries tested, from an initial concentration of 10 g/L.

PAGE 74

59 (4) Careful selection of NF membranes is re quired for higher re jections of taste and odor compounds, geosmin and MIB due to th eir low molecular weight (182 and 168). Bench and pilot testing results are important as the recovery and fl ux can be decided to meet the threshold limits for algal compounds as not all NF membranes rejected these compounds to less than thei r threshold limits. (5) The bench testing suggests that NF membranes can be effectively used as a polishing process for the removal of residua l TOC, algal and cyanobacterial exudates from treated surface water.

PAGE 75

60 RECOMMENDATIONS FOR FUTURE RESEARCH To meet the increasing stringe nt regulations from the EPA, alternative approaches are needed. The RBSMT can provide a screenin g tool to investigat e factors influencing removal of organics and trace contaminants. Bench-Scale Testing As a follow-up to this project, the RB SMT can be applied to conduct in-depth bench scale testing. Some potential research projects are listed below. 1. The impacts of seasonal water quality vari ations on the effectiveness of NF for removal of NOM should be investigated. The algal and cyanobacterial blooms are more prevalent during the dry season. 2. Long term analysis (greater than 12 week s) of NF membranes is recommended to assess membrane fouling. It is suggested to test the membrane for a longer period of time (greater than 12 weeks) con tinuously using the RBSMT and monitoring the MTC while keeping the flux rate c onstant. Drop in MTC could indicate membrane fouling. 3. Based on bench scale testing, rejection of TOC and UV254 is correlated to HAA and THM precursor rejections This correlation should be further investigated as these parameters could be used as in direct measures of DBPFP rejection. 4. The feasibility of using NF for removal of other pollutants including other DBPs, other algal by products, pesticides, ph armaceutically active compounds (PHACs)

PAGE 76

61 and endocrine disruptors (EDRs) should be investigated. These contaminants should be spiked individually as well as a group with other contaminants into laboratory Nanopure water as well as c onventionally treated water and then treated using NF and RBSMT to isolate their individual reje ction and/or group rejection. 5. Individual rejection of targ et contaminants rather than combination of various species should be investigated. Microc ystin-LR, geosmin and MIB were spiked in post filtered Lake Manatee water whic h had TOC in the range of 3.5 to 4.6 mg/L in this research. This is impor tant as although geosmin and MIB have molecular weight of 182 and 168, these contaminants are rejected using NF treatment. These contaminants could have been attached to other larger contaminants such as TOC and getting rejected thereby. Spike test using lab Nanopure water and individual contaminant c ould be treated using NF to isolate the rejection of the contaminant. 6. The disposal of concentrate stream is an issue with membrane processes. The concentrate stream will contain high concen trations of chemicals and algal toxins rejected by the NF membrane. Thus, the water quality analysis and disposal of concentrate stream should be investigated. Pre-treatment of this stream prior to its disposal should also be investigated The disposal opti ons could be land application, deep well inj ection, and evaporation ponds Diluting th e stream by mixing it in influent of a waste water trea tment system can also be investigated. 7. A model should be developed and app lied to describe the diffusion and convection transport of target contaminan ts across the membrane. This could be

PAGE 77

62 accomplished by developing equations to address both the diffusion and convection transport and predicting permeate concentration of a given contaminant. Pilot Testing Pilot testing of candidate membranes for a utility should be conducted as an integral component of system design. Impor tant parameters are contaminant rejection, permeate flux, feed water recovery and membra ne fouling. Based on the results from the RBSMT applied in this research, the NF90, NF270, and LFC1 membranes should be considered for pilot testing. Test data from the RBSMT should be us ed in membrane manufacturer’s model, such as Hydranautics “IMS Design Soft ware” model, which would predict design features such as the number of pressure vessels, number of membranes, stages and chemical requirements and their dosages. Me mbrane fouling and cl ean in place (CIP) are important factors which should be investigated during the pi lot testing of selected NF membranes. While running the NF pilot plan t, MTC normalized for temperature values could be calculated. Drop in the MTC valu es generally indicate membrane fouling and time for a CIP. Cleaning of membranes, CIP, should be done with membrane manufacturer recommended chemicals and instructions.

PAGE 78

63 REFERENCES Allgeier, S and Summers, S., (1995). Evalua ting NF for DBP Control with the RBSMT. Jour. AWWA. March, 1995. 87-99. Blau, T., Taylor, J., Morris, K., and Mulford, L., (1992). DBP Control by NF: Cost and Performance. Jour. AWWA, 84, 104-116. Chellam, S. and J. Taylor (2001). Simplifie d Analysis of Contaminant Rejection During Ground and Surface Water Nanofiltration Unde r the Information Collection Rule. Water Research, 35, 10, 2460-2474. Cho, J., Amy, G., and Pellegrino, J., (1999). Membrane Filtration of Natural Organic matter: Intial Comparison of Rejecti on and Flux Decline Characteristics with Ultrafiltration and Na nofiltration Membranes. Wat. Res., 33, 11, 2517-2526. Chorus, I. And Bartram, J., (1999). Toxic Cy anobacteria in Water. E&FN SPON, NY. Chu, F., Huang, X., and Wei, R., (1990). Enzyme-Linked Immunosorbent Assay for Microcystins in Blue-Green Algal Blooms J. Assoc. Off. Anal. Chem., 73, 451-6. Drikas, M., Chow, C., House, J., and Burch, M., (2001). Using Coagulation, Flocculation and Settling to Remove Toxic Cyanobacteria. Jour. AWWA., 93, 100-111. Envirologix Inc., (2001). Microcystin Tube Kit Inst ructions. Catalog No. ET 022. EPA, (April, 1996). ICR Manual for Benchand Pilot-scale Treatment Studies. EPA 814/B-96-003. Falls, V. (2002). Removal of Algal By Pr oducts and Natural Organic Matter from a Florida Surface Water Using Nanofiltration. M.S. Thesis, Department of Civil and Environmental Engineering, University of South Florida. Fu, P., Ruiz, H., Thompson, K., and Spangenbe rg, C., (1994). Selecting Membranes for Removing NOM and DBP Precursors. Membrane Filtration, Dec., 1994, 55-72. Gavrey, E., (2003). Relationships Between Me asures of NOM in Quabbin Watershed. AWWA Jounal, November 2003, 73-84.

PAGE 79

64 Letterman, R., (1999). Water Quality Treatmen t. AWWA, McGraw Hi ll, NY. Chapt. 11. Metcalf, J., Bell, B., and Codd, G., (1999). Production of Novel Po lyclonal Antibodies Against the Cyanobacterial T oxin Microcystin-LR and th eir Application for the Detection and Quantification of Microc ystins and Nodularin. Wat. Res., 34, 2761-2769. Montgomery D.C., Runger G.C., and Hubele N. F (2004). Engineering Statistics. Third Edition. John Wiley & Sons, Inc. Chapter 5. Morris, R.D., Audet, A.M., Angelillo, I.F., Chalmers, T.C. and Mosteller F. (1992). Chlorination, Chlorination by-products and can cer : a meta analysis. Am J Public Health, July 82(7):955-63. Munitsov, M. and Trimboli, P., (1996). Remo val of Algal Toxins Using Membrane Technology. Water, May/June, 34. Roalson, S., Kweon, J., Lawler, D. and Speite l Jr., G, (2003). Enhanced Softening: Effects of Lime Dose and Chemical Additions. Journal AWWA, Novrmber, 2003, 97-109. Reiss, C., Taylor, J., and Robert, C., (1999). Surface Water Treatment Using Nanofiltration-Pilot Testing Results and Design Considerations. Desal., 125, 97112. Smith, D., Falls V., Levine A., Macleod B. and Simpson M. Nanofiltration to Augment Conventional Treatment for Removal of Algal Toxins, Taste and Odor Compounds, and Natural Organic Matt er. Water Quality Technology Conference, American Water Work s Association, Seattle, WA, 2002. Spangenberg, C., Duranceau, S., and Kutile k, J., (2002). Membrane Manufacture and Utility Implement Non-Traditional Memb rane Accepting Testing, Water Quality Technology Conference, American Water Works Association, Seattle, WA 2002. Taylor, J. and Jacobs, E., (1996). Reve rse Osmosis and Nanofiltration. Water Treatment: Membrane Processes, AW WARF, McGraw Hill, HY,Chap 9, 63-64. Taylor, J., Thompson, D., and Carswell, J., ( 1987). Applying Membra ne Processes to Groundwater Sources for Thihalomethane Precursor Control. Jour. AWWA, 79, 72-82. Van der Bruggen, B., Schaep J., Wilms D. a nd Vandecasteele C. (1999). Influence of Molecular Size, Polarity, and Charge on the Retention of Organic Molecules by Nanofiltration. Journal of Membrane Science, 156, 29-41.

PAGE 80

65 Visvanathan, C., Marsono, B., and Bas u, B., (1998). Removal of THMP by Nanofiltration: Effects of Interference Parameters. Wat. Res., 32, 12, 3527-3538. Volk, C., Wood L., Johnson B., Robinson J ., Zhu H.W., and Kaplan L. (2002). Monitoring dissolved organic carbon in su rface and drinking waters. J. Environ. Monit., 2002, 4, 43-47. Waller, K., S. H. Swan, G. DeLorenze and B. Hopkins (1998). Trihalomethanes in drinking water and spontaneous a bortion. Epidemiology 9(2): 134-40. Westrick, J. (2003). Everyt hing a Manager Should Know About Algal Toxins but Was Afraid to Ask. Journal AWWA, September 2003, 26-34.

PAGE 81

66 APPENDICES

PAGE 82

67Appendix A: Equipmen t List & Photographs Equipment List 1. Osmonics Sepa CF System E-316 Stainle ss Steel (SS) Cell Body for Low Foulant Spacers Includes Stainless Steel Cell Body, Alumin um Cell holder, Pressure gauge (system pressure), pressure gauge (concentrate pressure), concentrate control valve, instruction manual, low Foulant feed spacer, permeate carrier and tubing kit) 2. 316 Stainless steel ( SS) tubing 1/4" ODSUPELCO INC Part Nbr: 20527 (Fisher Cat # NC9301000) 3. Various SS compression fitting for 1/4" tubing, 1/8" to 1/4" SS male reducers (for pumps and flow meters), 1/4" SS union tees, 1/2" to 1/4" SS reducing bushings (pressure relief valve) and 1/4" SS couplers 4. 16 Gallon Feed Tank Nalgene Tank 5. Feed Water Pump and Recycle Pump Cole Parmer Variable Speed Gear pump Drive, Model 75211-10 and Micropump head, Model 200.150-000 6. Flowmeter-Scienceware flowmeters--Fi sher Cat # 11-163-75 C, Bel-Art No.: H40407 0035, Size # 5 7. Temperature gauge—Lab Sa fety 3inch Cat# 66479-1A 8. Needle valve -Cole Parmer Cat # EW-68831-00 Needle Valve, Brass, 600 Max psi 9. Pressure gauges (3 nos.) —C at # EW-68110-30 Cole-Parmer Economical Digital Gauges 10. Pressure relief valveCole Parmer Cat #EW-03245-40-316 SS Relief Valve, 1/2" NPT(F) 40-150 psig 11. Check valve for recycle lineCole Parmer Cat # EW-98553-02 12. 16 turn Cole Parmer metering valve

PAGE 83

68Appendix A (Continued) Figure A-1. Osmonics Sepa Test Cell (Top-Compressed Ai r Line (black) and System Pressure for Membrane Cell Holder. Left to Right-Concentrate valve, Concentrate Pressure Gauge, Influent Temperature Gauge and Influent Pressure Gauge) FeedLine Permeate Line Concentrate Line Osmonics Sepa Test Cell

PAGE 84

69Appendix B: Chlorine Demand for Source Water Chlorine demand tests were performed to de termine the amount of chlorine that is consumed or decayed in 7 days at 25C. Chlori ne was added to test water at levels of 6, 10.5 and 13 mg/L (Case 1, 2 and 3 respectiv ely). Lab Nanopure water was spiked with 10 mg/L as a control (Case 4). 1. The chlorine demand of the glass bottles were quenched prior to the chlorine demand tests on test water with soaking th em in water containing 10 mg/L of free chlorine for 4 hours. 2. Sixteen bottles at each level of chlorina ted water were incubated for 7 days at 25C in 300 mL glass bottles without head space in dark conditions. 3. The free chlorine was monitored every 30 mi nutes for the first four hours and then every day and a chlorine demand was de termined at the end of 7 days. Table B-1 Post Filtration Lake Manatee Chlorine Demand Test Results Time pH Free Chlorine (mg/L) (days) Case 1 Case 2 Case 3 Case 4 Case 1 Case 2 Case 3 Case 4 0 8.04 7.94 7.76 5.75 12.6 10 5.5 9.1 0.01 8.02 7.48 7.62 5.43 10.8 9.2 5.1 8.8 0.02 8.01 7.73 7.53 5.39 9.7 8.9 4.9 8.7 0.04 7.91 7.37 7.49 5.48 9.2 8.3 4.4 8.6 0.17 7.67 7.73 7.42 5.51 8.5 8.1 3.1 8.4 1 7.61 7.58 7.21 5.48 6.7 6 2.5 8.2 2 7.45 7.35 7.03 5.44 5.3 4.9 1.8 8.1 3 7.34 7.25 6.98 5.39 5.1 4.5 1.5 8.4 4 7.34 7.23 7.01 5.54 4.8 3.9 1.2 8.6 5 7.32 7.21 6.98 5.51 4.2 3.5 1 8.3 6 7.31 7.2 6.97 5.5 4 3.4 0.9 8.4 7 7.31 7.19 6.96 5.46 4 3.4 0.9 8.4 Chlorine demand trends are shown in Figur e 4. It is seen that the chlorine demand is the highest in the first few hours (4 hou rs) of contact. Further, after four days,

PAGE 85

70Appendix B (Continued) the chlorine demand of the water nearly re duced to zero. The control sample free chlorine remained constant thr oughout the duration of the test. 0 2 4 6 8 10 12 14 01234567Time (days)Free Chlorine (mg/L) Nanopure; 10.1 mg/L Chlorine Dose Manatee Post-Filtration; 13.0 mg/L Chlorine Dose Manatee Post-Filtration; 10.8 mg/L Chlorine Dose Manatee Post-Filtration; 6.2 mg/L Chlorine Dose Figure B-1. Chlorine Dema nd Test Conducted on Lake Manatee Post Filtered Water A ratio was developed using feed TOC a nd chlorine demand. This ratio was calculated to be 2.2 mg of free Cl2 per mg of TOC. The amount of free chlorine to be dosed to NF permeate was estimated on the basi s of TOC concentration at the end of the membrane test and this ratio. For example, if the permeate TOC was measured as 0.5 mg/L then the chlorine demand was estimated as 1.1 mg Cl2. The free residual chlorine at the end of 7 day incubation at 25C chlorine was targeted between 3 mg/L and 5 mg/L. Hence, the amount of chlorine to be sp iked in NF permeate would be 5.2 mg/L approximately for the THMFP and HAAFP tests.

PAGE 86

71 Appendix C: Comparison of Memb ranes for Rejection of Inorganics Inorganics rejection, rejection of total hardness, calcium hardness and conductivity are presented in Table C-1. Th e influent total ha rdness concentration averaged 95 mg/L as CaCO3. The total hardness for Lake Ma natee is in the range of 50 to 70 mg/L as CaCO3. The increase in the total hardness in post filtration water is due to the addition of lime in their c onventional treatment process. Total hardness rejections for the NF270 membrane tests at 12, 19 and 32 gfd at various recoveries are presented in Figure C-1. It is seen that as the recovery and flux rate increases the total hardness rejection decreases. 80 84 88 92 96 100 405060708090100Recovery (%)Total Hardness Rejection (%) 32 gfd 19 gfd 12 gfd Figure C-1. Total Hardness Rejection for NF270 at Different Flux Rates and Recoveries NF membranes tested in this research re jected total hardness in the range from 46 to 99 percent and similarly rejected calciu m hardness in the range of 33 to 99 percent depending on the flux rate and overall system recovery. The permeate total hardness concentration ranged from 1 to 45 mg/L as CaCO3. Similar rejection was observed for conductivity which ranged from 66 to 95 percent.

PAGE 87

72 Appendix C (Continued) Table C-1 Total Hardness, Calcium Hardness and Conduc tivity Data for NF Membranes Tested Total Hardness (mg/L as CaCO3) Calcium Hardness (mg/L as CaCO3) Conductivity (uS/cm) Membrane Recovery CF 2 CP 3 Rej.1 CF 2 CP 3 Rej.1 CF 2 CP 3 Rej.1 NF90 50 99 1 99 69 1 99 233 12 94.9 70 99 3 96 69 2 97 233 17 92.8 85 99 4 96 69 4 95 232 28 88.0 95 99 5 95 69 4 94 230 36 84.4 NF270 (a) 50 92 9 90 67 6 90 228 44 80.8 70 92 11 89 67 7 89 226 49 78.6 85 92 15 84 67 11 83 229 52 78.3 95 92 16 82 67 11 83 231 58 74.9 NF270 (b) 50 91 9 91 64 7 89 225 36 84.0 70 91 10 89 64 7 90 229 48 79.1 85 91 13 86 64 10 84 232 65 72.0 95 91 16 83 64 12 81 229 81 64.7 NF270 (c) 50 97 10 90 66 8 88 236 33 86.1 70 97 11 89 66 7 89 234 44 81.2 85 97 14 86 66 12 82 233 67 71.3 95 97 17 83 66 12 81 235 81 65.6 LFC1 50 94 1 99 64 1 99 224 19 91.6 70 94 3 97 64 2 97 219 26 88.2 85 94 4 96 64 2 97 223 32 85.7 95 94 5 95 64 4 94 227 45 80.2 NTR 7450 50 94 28 70 67 24 63 226 47 79.7 70 94 43 55 67 35 48 229 59 74.3 85 94 49 48 67 43 36 232 68 70.7 95 94 51 46 67 45 33 231 76 67.1 1Rej. Percent Rejection 2CF-feed concentration 3 CP-permeate concentration

PAGE 88

73 Appendix D: Microcystin-LR Analyzed by ELISA Microcystin-LR analyses by Enzyme Li nked Immunosorbent Assay (ELISA) were performed by the Manatee WTP. Microcystin Plate Kit En closure, rev. 04/23/99 This procedure quantitatively analyzes dissolved microcys tin toxins, Microcystin-LR, Microcystin-YR, and Microcys tin-RR, through the use of th e Envirologix Microcystin Plate Kit. Concentrations of the toxins as low as 0.035 ppb of the detoxified microcystin calibrator and 0.108 ppb of microcystin-LR can be detected using the enhanced sensitivity protocol. It shoul d be noted that microcystin-L R, YR and RR when analysed using the kit calibrators freque ntly demonstrate greater than 120% recovery. All results may be confirmed by High Performa nce Liquid Chromatography (HPLC). REAGENTS: Methanol, GC/MS or pesticide grade Microcystin Plate Kit containing: 12 strips of 8 antibody coated wells each 4 calibrators; including negative control, 0.16 ppb, 0.5 ppb, and 1.6 ppb Page Microcystin-enzyme conjugate Wash solution salts Substrate Stop Solution Convenient concentrations of working standard solutions of the above microcystin toxins for preparing check standards in milli-Q water APPARATUS: -Stat Fax 303 microstrip reader designed to read and calculate resu lts of endpoint assays. -Micro syringes, 20 ul, 50 ul, 100 ul, and 200ul -Distilled water for the prep aration of the wash solution -Glassware for the preparation and storage of wash solution -Timer -Parafilm Procedure: 1. Remove samples from refrigerator at least 1 hour before running assay. To prevent condensation onto strips, keep strips tightly sealed in bag until ready to use. 2. Decide prior to starting the assay, wh ether the calibration and samples are to be analyzed in duplicate. Assign each calib rator and sample a location with its duplicate in consecutive order on the strip. 3. Using reagents and samples at room temperature, rapidly add 125 uL of Microcystin Assay Diluent to ea ch well to be used. 4. Immediately add 20 uL of each calibrator and sample to its assigned well. This step and the diluent step should be completed within 10 minutes. 5. Thoroughly mix the contents of the we lls by moving the plates in a circular motion for about 20 to 30 seconds. 6. Cover the wells to prevent evaporat ion and allow them to incubate for 30 minutes.

PAGE 89

74Appendix D (Continued) 7. At the end of 30 minutes add 100 uL of microcystin enzyme conjugate to each well. Mix in the same manner as a bove, cover and incubate for another 30 minutes. 8. At the end of this period, remove th e parafilm and shake the plate contents into the sink. Using a 1000uL micropipette, flood the wells with wash solution, then shake empty again. Repeat this wash step a total of 4 times. After the last washing, tap the plates on a paper towel to remove all traces of wash solution. 9. Turn off the hood lights. Add 100 uL S ubstrate to each well, mix and cover the wells with parafilm and a light impervious covering like a paper towel. Incubate for 30 minutes in the dark. 10. Add 100 uL Stop solution and read imme diately. The plates should be read within 30 minutes of the addi tion of the Stop solution. Increased Sensitivity Assay Procedure: 1. Samples and calibrators should be at room temperature. Remove kit and samples from refrigerator at least 1 hour before running assay. To prevent condensation onto strips, keep strips tightly sealed in ba g until ready to use them. 2. If a calibration is to be performed dilute calibrators 1:3 in Milli-Q water by using100 uL of the Negative Control and each calibrator to 200 uL of water. Mix well. This gives an assigned value of 0. 01 for the negative cont rol dilution and true value concentrations of 0.05 ppb, 0. 15 ppb, and 0.45 ppb for 0.16ppb, 0.5ppb, and 1.6 ppb calibrators, respectively. 3.The calibrators and samples are to be analyzed in duplicate. Assign each calibrator and sample a lo cation with its duplicate in consecutive order on the strip.4. Using reagents and samples at room temperature, ra pidly add 50 uL of Microcystin Assay Diluent to each well to be used. 5. Immediately add 50 uL of each calibrator and sample to its assigned well. This step and the diluent step should be completed within 10 minutes. 6. Thoroughly mix the contents of the we lls by moving the plat es in a circular motion for 20 to 30 seconds. 7.Cover the wells to prevent evaporation and allow them to incubate for 30 minutes. 8. At the end of 30 minutes add 100 uL of microcystin enzyme conjugate to each well. Mix in the same manner as a bove, cover and incubate for another 30 minutes. 9. At the end of this period, remove the parafilm and shake the plate contents into the sink. Using a 1000uL micropipette, flood the wells with wash solution, then shake empty again. Repeat this wash step a total of 4 times. After the last washing, tap the plates on a paper towel to remove all traces of wash solution. 10. Turn off the hood lights. Add 100 uL Substrate to each well, mix and cover the wells with parafilm and a light im pervious covering like a paper towel. Incubate for 30 minutes in the dark. 11. Add 100 uL Stop solution and read imme diately. The plates should be read within 30 minutes of the addi tion of the Stop solution. 12. Consistency and accuracy in measuring reagents and samples is essential. Drips and bubbles at the tip of the mi cropipette should be avoided. 13. Timing is important. Perform each st ep at the same time interval each time the analysis is performed.

PAGE 90

75Appendix D (Continued) Strip Reader Procedure 1. Turn on the Stat Fax 303. When it reads “Ready” and the time, the function should be checked. 2. The check is performed by running a bl ank strip. Set the absorbances to 450 nm and 603nm. Press ABS then select filter by pressing 2 for 450 nm and 4 for 603 nm. Set the carrier, press Blank and then Enter. The reader prints the strip number and the carrier position. It r eads each well position. Each well should read 0.000 + 0.005. Make sure the carrier is cl ean and dust free as this may distort the blank reading. 3. After completing the performance check, return to the main screen by pressing Clear twice 4. Set up the instrument for calibration and analysis. A. Press Alt. The instrument will ask for strip type and number then will go back to ready. B. Press Mult. It will ask if calibration is Regression? Yes. Then select filters, 450 nm and 603 nm again. Then it will ask for the regression type. Use Log Conc with no bla nk and calibrators and samples in duplicate. Use four calibrators, Ne gative Control and three others. Assign the negative control a value of 0.01 as ln 0 is undefined. C. Set carrier to first strip, then press enter. D. When all calibrators are analyz ed, the instrument will calculate the correlation coefficient and the slope of the curve and will ask if the curve should be plotted. Tell it “yes” a nd note the positions of the standards on the curve. Any obvious outliers can be modified by deleting one of the duplicate pair, then reanalyzing the st andards. The program will not let you completely delete a concentrat ion value, only one of the pair. E. When you are satisfied with th e Correlation coefficient, it should be greater than 0.9850, accept the curve a nd analyze the sample s. Set carrier to strip 2 and press enter. F. When analysis is complete, th e curve and “test” can be saved. Essentially, the stored test is a calibra ted template for analysis. To save it, press alt and the question “Save Test?” wi ll appear on the screen. Tell it “yes” and name the test by entering a series of letters from the screen. When the name is complete press enter twice. The test will be saved and the printout will show the name and test number at the end of the analysis printout. If the test is not to be saved, press clear twice. G. To recall a stored/saved test, press menu then enter the test number desired. To recalibrate a particular test, recall that test and plot a new curve. 5. Analytical note: Make sure the bottoms of th e strips are clean and free of fingerprints and dust. Do not read strips that contain bubbles or conden sation. Use the same volume for all samples, check standards, and blanks. Calibration Run a series of calibra tion standards prepared in the sa me manner as the samples will be prepared. The concentrations should brack et the expected sample concentrations.

PAGE 91

76Appendix D (Continued) The calibration curve should ha ve a correlation coefficient of at least 0.9850 and may be saved for future use with the remaining wells of the lot number on which the curve was performed. Do not use ol d calibration curves with new lots of assay chemicals or wells. Quality Control At least one negative control sample and one check stan dard should be performed with each batch of 20 samples or less. A recovery of 65% to 130% is considered acceptable. Safety Perform all tests under a functioning fume hood. Gloves should be worn and the usual personal protective equipment should be used (labcoat, lab glasse s, long pants, and safety shoes). Completed assay may be read on a stri p reader situated on the benchtop.

PAGE 92

77Appendix D (Continued) Microcystins Analyzed by Enzyme Linked Immunosorbent Assay (ELISA) Tube Kit performed at Environmental Engineering Resear ch Lab, University of South Florida. Microcystin Tube Kit En closure, rev. 06/01. The MYCST method has been devel oped by testing and modifying the Envirologix Inc., Microcystin Tube Kit Meth od. The test kit is a competitive enzyme linked immunosorbent assay ( ELISA). Microcystin toxin competes with the enzyme (horseradish peroxidase) labele d Microcystin for a limited number of sites inside the test tube. After the wash step, color visualizati on occurs where the concentration of the toxin is inversely proportional to color development. The tube kit method measures total free Microcystin toxin. The method does not distinguish between the Microcys tin toxins but measures them at varying degrees. The limit of detection (LOD) for this test is 0.3 parts per billi on (ppb). This was established by determining the spread of 2 standard devi ations from the mean population of negative water samples. This is represented as 81. 5% Bo. 100% Bo equals the maximum amount of Microcystin toxin enzyme conjugate th at is bound by the antibody in the absence of any Microcystin in the sample (negative control). %Bo=(Optical Density (OD) of Sample or Calibrator)/(OD of ne gative Control)x100 As reported by Envirologix the following toxi ns are measured in ppb for 50% Bo and 81.5% Bo(LOD). Table D-1 Limit of Detection for Microc ystins using ELISA Tube Kit Compound 50% Bo 81.5 % Bo (LOD) Microcystin LR 0.94 0.30 Microcystin LA 0.78 0.43 Microcystin RR 1.53 0.65 Microcystin YR 2.53 0.69 Nodularin 1.44 0.53 Humic acid was reported not to interf ere at concentrations below 100 ppm. Test Procedure: 1. Allow all tubes and reagents to reach ambient temperature. 2. Add 5 drops of Microcystin assay dilu ent to each tube in the assay. 3. Add two drops of 0.5 ppb Microcystin calibrator to the first tube. 4. Add two drops of 3.0 ppb Microcystin calibrator to the second tube. 5. Add two drops of sample to each of the subsequent tubes, up to a max of 4 samples (6 tubes). 6. Mix tubes for 20-30 seconds. 7. Incubate for 5 minutes at room temperature. 8. Add 5 drops of Microcystin-enzyme conj ugate to each tube and mix for 20-30 seconds.

PAGE 93

78Appendix D (Continued) 9. Incubate for 20 minutes at room temperature. 10. Drain contents into sink and rinse with cool tap water 3 times. Places tubes inverted on towel and tap to remove excess water. 11. Add 10 drops of Substrate to each tube mix for 20-30 seconds and incubate for 10 minutes at room temperature. 12. Pipette 0.7 ml of 1.0N HCl solution to each tube and mix thoroughly. Color should be yellow. Read tubes within 30 minutes on photometer. 13. Using DR4000, set wavele ngth to 450 nanometers (nm) Read absorbance and extrapolate the microcystin concentra tion using the calib ration curve shown below. A calibration curve was made using the microcystin calibrator solution. This solution was diluted and a curve was develope d to determine linearity. This curve was then used to determine microcystin concentrat ions in a sample. This allows for greater quantification as to th e amount of total free Microcystin toxin in a sample. The DR 4000 was programmed to the absorbance method and at a wavelength of 450 nm. Dilutions of the calibration liquid were made to 3, 2, 1, 0.5 and 0.25 ppm using nanopure water and the absorbance were recorded. A plot of the absorbance versus Microcystin concentration was developed. Using the above alternative method, a calib ration curve was completed. A Hach DR 4000 spectrophotometer was us ed for the analysis at a wavelength of 450 nm. Absorbance vs. concentration and the associat ed line fit are shown in the figure below. This calibration curve was used for detection of microcystin-LR in the feed and permeate water. y = -0.237x + 1.3143 R2 = 0.9818 0 0.2 0.4 0.6 0.8 1 1.2 1.40.00 0.25 0.50 0.75 1 00 1 2 5 1 5 0 1.75 2.00 2.25 2.50 2.75 3 00 3 2 5Microcystin-LR (ug/L)Absorbance (@ 450 nm) Figure D-1. Absorbance 450 nm vs Microcystin-LR C oncentration (ppb)

PAGE 94

79Appendix E: Geosmin and MIB: Closed-Loop, GC/FID Protocol ORGANIC: GEOSMIN AND MIB: CLOSED-L00P, GC/FID PROCEDURE Standard Methods 18th Ed, 1992 Method 6040B pg. 6:7-16 This procedure quantit atively analyzes geosmin and 2-me thyl-isoborneol (MIB) in water through the use of closed-loop stripping (CLS)GC/FID analysis. Conc entrations as low as 1.0 ng/L of geosmin and 3.0 ng/L of MIB can successfully be detected. REAGENTS: -Methylene Chloride (CH2Cl2), GC/MS (Fisher Optima) or pesticide grade -Acetone, GC/MS (Fisher Optima) or pesticide grade -Carrier gas, Helium, ultra purifie d grade, moisture and oxygen free -Sodium Sulfate (Na2SO4), granular, anhydrous, pe sticide grade (Bake at 625 C for at least 2 hours before use. Store in oven at 105 C.) -Deionized H2 0, Type I -Milli-Q H2 0 -Internal standards, 1-Chloroocta ne(CI-8), 1-Chlorodecane (Cl-10), 1-Chlorododecane(Cl-12) 1. Internal standard "c ocktail" stock: ~8.300ng/uL a. Fill a 25ml volumetric flask to bottom of neck with acetone. Replace stopper. b. Wait until all parts of flask above so lution are dry, then weigh flask and acetone as tare (analytical balance, 0.1 mg). c. Using 500 ul syringe, inject 250 ul Cl-8 directly into acetone in volumetric without letting it run down the side of flask. Replace stopper. d. Weigh flask to determine actual amount of Cl-8 added. e. Repeat with CI-10 and CI-12, being sure to weigh for tare before each addition. f. Bring to 25.0 ml mark with acetone g. Calculate concentration as ng/ul as follows: 1-Chlorooctane (99%): (grams of Cl-8) x (0.99) x (I09-ng) x ( 1 ml) 25 mL Ig 103ul 1 -Chlorodecane (95%): (grams of CI-10) x (0.95) x (109-nq)x(1mI) 25 mL 1g 103uI

PAGE 95

80Appendix E (Continued) 1-Chlorododecane (98%). (grams of Cl-12) x (0.98)X (109-ng)x( 1ml) 25 mL 1g 103ul 2. Internal standard (I.S.) working solution (WS): ~10ng/uL a. Fill a 25ml volumetric flask to bottom of neck with acetone. b. Wait until all parts above the so lution are dry, then weigh flask and acetone as tare (analytical balance, 0.1 mg). c. With 100ul syringe, in ject 30ul of I.S. "cockta il" stock standard, (Cl-8, CL-1 0, and CI-1 2) directly into acetone without letti ng it run down side of flask, Replace stopper. d. Weigh flask to determine the actual amount added. e. Bring volume up to 25ml mark on flask f. Calculate concentrations of chloroalkanes in ng/ul as follows: Concentration of I.S. "cocktai l" (nq/ul) x (grams of I.S.) x 1.0 mL 25.0 mL 0.79 grams Stock solution of 2-Methyl -Isoborneol (MIB) in acetone Stock solution of Geosmin in acetone 3. Check standard a. Use a I.0ml volumetric flask and add approximately 300 ul of acetone. b. Wait until all parts of the flask above the solution are dry, then weigh flask and acetone as tare (analytical balance, 0.1 mg). c. With 500 ul syringe, acid ~325 ul of I. S. working solution directly into acetone without letting it run down the side of flask. Replace stopper. d. Weigh flask to determine actual amount of I.S. added e. Using a 10 ul syringe, inject 5.0 ul of geosmin stock (205 ng/ul) to volumetric beneath surface of acetone. Rinse syringe 2-3 times with acetone, making sure needle end is below surface of solution. f. Using a 10 ul syringe, add 5.2 ul MIB stock (198 ng/ul) as in step e. g. Bring volume up to 1 0 mark on flask with acetone. h. Calculate concentrations in ng/u l as follows: ~3.3ng each choroalkane

PAGE 96

81Appendix E (Continued) grams I,S. WS x concentration Cl-8 in WS x 1 mL 1.0ml 0.79 grams grams I.S. WS x concentration CI-10 in WS x 1 mL 1.0ml 0.79 grams grams I.S. WS x concentration CI-12 in WS x 1 mL 1.0 mL 0.79 grams 5.2 ul M1B stock x 198 ng/ul = 1.03 ng/uL 1.0 mL 5.0 ul Geosmin stock x 205 ng/ul = 1.02 ng/uL 1.0 mL APPARATUS: -Varian 3300 Flame Ionization Detector Gas Chromatograph, equipped with: 1. Capillary injector, split/splitless 2. Capillary column, 60-m x 0.25mm-ID (SPB-5), fused silica -Tekmar Closed-Loop Stripper, equipped with: 1. Stripping bottle, 1-L 2. Pump, with stainless-steel bellows that provide air flow from 1-1.5L/min 3. Thermostatic water bath, with a thermoregulating system accurate to least +/-0.50 C. 4. Filters, 1.5mg activated carbonGlass co llection vials, 50ul capacity with gastight stoppers -TFE connecting sleeve -TFE stir bar -Micro syringes, 10 ul and 25 ul -Micro pipette, 200-1000ul -1000ml graduated cylinder -Metal clamps -Ice pack -Pipette bulb -Glass sample bottles (1.2L) with TFE lined cars -Peak Simple 32-bit software -UNISYS CMT 510007 with color monitor -Electronic Top loading Balance 0.001g-0.01g -Electronic Analvtical Balance, 0. 1 mg -Stopwatch

PAGE 97

82Appendix E (Continued ) PROCEDURE: Determination of filter flow rate: 1. Find the volume of the trap by filling the longer end of trap above carbon with MeCl2 and measure the volumeadded to the top. 2. After several rinses with solvent to we t filter, measure the time to empty solvent from top of filter to surface of carbon. The ra te should be at least 0.6ml/min. or 10uL/sec. 3. Strip a sample followed by a spiked sa mple to determine Closed-Loop Stripper (CLS)/trap recovery efficiency. CLS/trap effi ciency must be determined for each matrix to be measured. Stripping Procedure: 1. Turn CLSA power and pump on with auxilia ry trap in place, Let warm-up for at least 10 min. 2. The sample temperature should be set at 300 C, the trap temperature set at 400 C, and the line temperature set at 500 C. 3. Pour 1000ml sample into sample bottle. Add stir bar. Add 80g anhydrousNa2SO4, while stirring to dissolve (If sample is from distribution or the blend chamber, add approximately 0.01-002g sodium thiosulfate to neutralize Cl2 prior to Na2SO4 .If the sample is from the reservoir or has hi gh color, use 900ml samp le with 72g of Na2SO4 ). 4. Inject 5ul I.S. working solution (4.5ul for 900ml sample) into sample bottle below surface of sample 5. Reassemble stripper, put in analysis trap, and strip for 1 hour and 30 min. 6. While sample is in the stripping proced ure, set-up G.C. and da ta station,and check response with an injection of check standard. The response sh ould be +/20% of the true value for the standard. 7. When stripping is complete, turn off sti rrer, remove trap, and let trap cool to room temperature. 8. Disassemble and clean CLSA. First remove gas line and rinse consecutively with hot tap water, de ionized water, a nd Milli-Q water. Remove diff user and rinse as above. Pour out sample and clean sample bottle as a bove, however, use tap water overflow and a brush in the initial step. Extraction Procedure: 1. Inject I ul MeCl2 into bottom of collection vial. 2. Place trap in teflon sleeve on vial, short-side down. 3. Inject 5ul MeCl2 into trap, being careful not to touch carbon with needle. 4. Using a pipette bulb and slight pressure, carefully run the MeCl2 over the filter 10 times. 5. Use ice along bottom side of trap to separate MeCl2 from carbon and ta p vial to bring MeCl2 to bottom.

PAGE 98

83Appendix E (Continued) 6. Repeat steps 3-5 with two more 5ul aliquots of MeCl2. 7. Push any remaining MeCl2 through filter with pipette bulb. 8. Remove trap and ad d approximatel y 1 ul MeCl2 to bring volume to marked line on vial. 9. Cap sample and store on ice until injection. GC Program: 1. Carrier gas: He at 1.05ml/minute @ 60C 2. Make-up gas, He at 40mil/minute 3. Detector, 260C 4. Injector a. mode: splitless, vent opened at 1.0 minute b. split flow: about 15ml/minute c. temperature, 200C 5. Column temperature program a. Initial temperature, 60C Hold for 2.0 minutes b Program 1: 60C -124C at 4C /minute Hold for 6.0 minutes c. Program 2: 124C -160C at 7C /minute Hold for 5.0 minutes d. Program 3: 160C -2l0C at 7C /minute Hold for 0.0 minutes e. Program 4: 210C -250C at 20C /minute Hold for 2.0 minutes Computer Set-up: 1. Turn on computer. When main window s screen appears, go to START. Select PROGRAM, then "Peak Simple 32-bit", and finally "Peak Simple''. Enlarge that window to fill the whole screen. 2. To edit Channel 1: go to Edit, select Channels, then Channel 1 a.Activate the DETAILS button. End Time: Should be -1.00 minute after the last analyte of interest 's retention time. Default Display limits should be Max: 250.000mV and Min:-50.000 rmV. Sample Rate should be set at 10 Hz Trigger Group should be "Main" Control by Temperature Unretained Solute Time should be set at 16.000 minutes Close the DETAILS button by telling it OK

PAGE 99

84Appendix E (Continued) b. Activate the POST RUN button. Change the Save File As: Using today's date+000 chr, name the first file to be saved. eg. xxxxx000.chr Do not exceed 8 numbers, The program will assign a new number to each run in ascending order from 0. The first chromatogram will be xxxxx001,chr. Tell it OK to close that button. 3. Just before injecting the first samp le hit the Z button to zero the baseline. 4. Inject the sample into the GC and imme diately hit the space bar. Should you want to stop the acquisition at any time hit the end key. The acquisition will stop and no data will be saved. 5. After acquisition is complete, integr ate the chromatogram manually, The program automatically integrates the run when it is complete, but it usually requires some refinement. Go into Edit, then Manual integration. Choose the "Rubber Band" function which is the next to last button on the far le ft column. To use it, move the mouse pointer to the starting point of the integration. Ho ld down the left mous e button and drag the pointer along the baseline to th e end of the area to be inte grated. Release the button. The area integrated will usually shif t slightly to indicate the new integration. Close Manual integration and save the file by entering File, then Save to make the file permanent. 6. The results can be viewed by entering View, then Results. Injection Procedure: 1. Inject 1.5ul of sample into GC. a. In syringe use 0.2ul MeCl2, 0.3ul air, and 1.5ul extract. b. Inject in one motion and remove after 3 seconds. Clean-Up Procedure: 1. Trap: Clean trap with 2 rinses of Mill-Q water. 2 rinses of acetone, and 2 MeCl2 rinses. Push excess solutions out of top of trap with pipette bulb. 2. Vial: Clean vial with 3 rinses of acetone foll owed by 3 rinses of MeCl2. Use syringe to get solutions out of bottom of vial.

PAGE 100

85 Appendix F: Setting and Recovery Test Graphs The following section presents the graphs for the setting test and the recovery tests for the six membrane tests. These gr aphs follows are in the following order of membranes: NF90 NF270(a) NF270(b) NF270(c) LFC1 NTR7450 The graphs presented here for each me mbrane experiment are as follows: Flow Rates and Permeate Flux vs Time, Setting Test MTC vs Time, Setting Test MTC vs Time, Setting Test MTC vs Time, Recovery Test Flow Rates vs Time, Recovery Test Permeate Flux vs Time, Recovery Test Permeate Conductivity & UV254 vs Time, Recovery Test Pressure vs Time, Recovery Test

PAGE 101

86Appendix F (continued) 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 0.20.40.81.01.31.72.53.75.79.712.021.522.623.724.826.326.8 Time (hrs)Permeate Flux (gfd/psi)0 1 2 3 4 5 6 7 8 9Flow Rate (mL/min) Permeate Flux Permeate Flow Waste Flow Figure F-1. Flow Rates and Permeate Fl ux vs Time, Filmtec NF90, Setting Test 2.5E-04 3.0E-04 3.5E-04 4.0E-04 4.5E-04 5.0E-04 0612182430Time (hours)MTC (gpm/ft2/psi) Uncorrected Data Normalized to 25C Figure F-2. MTC vs. Time, Filmtec NF90, Setting Test

PAGE 102

87Appendix F (continued) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 051015202530Time (hours)MTC (gpd/ft2/psi) Figure F-3. MTC vs. Time, Filmtec NF90, Setting Test 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0102030405060708090100Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-4. MTC vs. Time, F ilmtec NF90, Recovery Test

PAGE 103

88Appendix F (continued) 0 2 4 6 8 10 12 14 16 0102030405060708090100 Time (hours)Flow Rate (ml/min ) Permeate flow Waste Flow 70% 50 % 85% 95% Figure F-5. Flow Rates vs Time Filmtec NF90, Recovery Test 0 5 10 15 20 25 30 35 0102030405060708090100 Time (hours)Flux (gpd/ft2) 70% 50% 8% 95% 85% Figure F-6. Permeate Flux vs Time Filmtec NF90, Recovery Test

PAGE 104

89Appendix F (continued) 0 5 10 15 20 25 30 35 40 011321254148546775828689909293 Time (hours)Conductivity (uS/cm ) 0 0.0005 0.001 0.0015 0.002 0.0025 0.003UV-254 (cm-1) Permeate UV-254 Permeate Conductivity Feed UV-254 0.08 cm-1Feed Conductivity 230 uS/cm 70%50%85% 95% Figure F-7. Permeate Conductivity & UV254 vs. Time, Filmtec NF90, Recovery Test 0 10 20 30 40 50 60 70 80 90 0102030405060708090100 Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-8. Pressure vs. Time, Filmtec NF90, Recovery Test

PAGE 105

90Appendix F (continued) 0 10 20 30 40 50 60 0.20.51.31.72.47.010.318.719.7 Time (hours)Permeate Flux0 5 10 15 20 25Flow Rate (mL/min ) Permeate Flux Permeate Flow Rate Waste Flow Rate Figure F-9. Flow Rates a nd Permeate Flux vs Time, Film tec NF270(a), Setting Test 3.0E-04 3.5E-04 4.0E-04 4.5E-04 5.0E-04 5.5E-04 6.0E-04 036912151821 Time (hours)MTC (gpm/ft2/psi) Uncorrected Data Normalized to 25C Figure F-10. MTC vs. Time, Filmtec NF270 (a), Setting Test

PAGE 106

91Appendix F (continued) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0510152025Time (hours)MTC (gpd/ft2/psi) Figure F-11. MTC vs. Time, Filmtec NF270 (a), Setting Test 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0102030405060708090Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-12. MTC vs. Time, Film tec NF270 (a), Recovery Test

PAGE 107

92Appendix F (continued) 0 5 10 15 20 25 0102030405060708090 Time (hours)Flow Rate (ml/min ) Permeate flow Waste Flow 70% 50% 85% 95% Figure F-13. Flow Rates vs Time, Filmtec NF270(a), Recovery Test 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 0102030405060708090 Time (hours)Flux (gpd/ft2) 70% 50% 85% 95% Figure F-14. Permeate Flux vs Time, Filmtec NF270(a), Recovery Test

PAGE 108

93Appendix F (continued) 0 10 20 30 40 50 60 70 0.34.620.729.853.875.880.882.985.087.989.3Conductivity (uS/cm ) 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004UV-254 (cm-1) Permeate Conductivity Permeate UV-254 Feed Conductivity 230 uS/cm Feed UV-254 0.075 cm-1 70% 95% 85% 50% Figure F-15. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(a), Recovery Test 0 10 20 30 40 50 60 70 80 90 0102030405060708090100 Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-16. Pressure vs. Time, Filmtec NF270(a), Recovery Test

PAGE 109

94Appendix F (continued) 0 5 10 15 20 25 30 35 40 45 50 0.00.51.02.02.33.013.617.019.2 Time (hours)Permeate Flux (gfd/psi)0 5 10 15 20 25Flow Rate (mL/min ) Permeate Flux Permeate Flow Rate Waste Flow Rate Figure F-17. Flow Rates and Permeat e Flux vs Time, Filmtec NF270(b), Setting Test 1.5E-04 2.0E-04 2.5E-04 3.0E-04 3.5E-04 4.0E-04 4.5E-04 5.0E-04 5.5E-04 6.0E-04 0369121518 Time (hours)MTC (gpm/ft2/psi) Uncorrected Data Normalized to 25C Figure F-18. MTC vs. Time, Filmtec NF270 (b), Setting Test

PAGE 110

95Appendix F (continued) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 02468101214161820Time (hours)MTC (gpd/ft2/psi) Figure F-19. MTC vs. Time, Filmtec NF270 (b), Setting Test 0.25 0.27 0.29 0.31 0.33 0.35 0.37 0.39 0.41 0.43 0.45 0102030405060708090100Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-20. MTC vs. Time, Film tec NF270 (b), Recovery Test

PAGE 111

96Appendix F (continued) 0 1 2 3 4 5 6 7 8 9 10 0102030405060708090100 Time (hours)Flow Rate (ml/min ) Permeate flow Waste Flow 70% 50% 85% 95% Figure F-21. Flow Rates vs Time, Filmtec NF270(b), Recovery Test 12.0 14.0 16.0 18.0 20.0 22.0 24.0 0102030405060708090100 Time (hours)Flux (gpd/ft2) 70% 50% 85% 95% Figure F-22. Permeate Flux vs Time, Filmtec NF270(b), Recovery Test

PAGE 112

97Appendix F (continued) 0 10 20 30 40 50 60 70 80 90 100 0.32.510.025.947.066.680.483.685.091.694.2 Time (hours)Conductivity (uS/cm ) 0 0.001 0.002 0.003 0.004 0.005 0.006UV-254 (cm-1) Permeate Conductivity Permeate UV-254 Feed Conductivity 225 uS/cm Feed UV-254 0.074 cm-1 70% 50% 95% 85% Figure F-23. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(b), Recovery Test 0 10 20 30 40 50 60 70 0102030405060708090100 Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-24. Pressure vs. Time, Filmtec NF270(b), Recovery Test

PAGE 113

98Appendix F (continued) 0 1 2 3 4 5 6 0.170.751.251.832.503.2514.1716.0819.58 Time (hrs)Flow-Rate (mL/min)0 2 4 6 8 10 12 14Permeate Flux (gfd/psi) Permeate flow Waste flow Permeate Flux Figure F-25. Flow Rates and Perm eate Flux vs Time, Filmtec NF270(c), Setting Test 1.5E-04 1.7E-04 1.9E-04 2.1E-04 2.3E-04 2.5E-04 2.7E-04 2.9E-04 3.1E-04 3.3E-04 3.5E-04 048121620Time (hours)MTC (gpm/ft2/psi) Uncorrected Data Normalized to 25C Figure F-26. MTC vs. Time, Filmtec NF270 (c), Setting Test

PAGE 114

99Appendix F (continued) 0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 048121620Time (hours)MTC (gpd/ft2/psi) Figure F-27. MTC vs. Time, Filmtec NF270 (c), Setting Test 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.40 0102030405060708090100110Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-28. MTC vs. Time, Film tec NF270 (c), Recovery Test

PAGE 115

100Appendix F (continued) 0 1 2 3 4 5 6 7 8 0102030405060708090100110Time (hours)Flow Rate (ml/min) Permeate flow Waste Flow 70% 50% 85% 95% Figure F-29. Flow Rates vs Time, Filmtec NF270(c), Recovery Test 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 0102030405060708090100110Time (hours)Flux (gpd/ft2) 70% 50% 85% 95% Figure F-30. Permeate Flux vs Time, Filmtec NF270(c), Recovery Test

PAGE 116

101Appendix F (continued) 0 10 20 30 40 50 60 70 80 90 0.12.09.224.342.364.480.484.386.997.7101.6Time (Hours)Permeate Conductivity (uS/cm ) 0 0.001 0.002 0.003 0.004 0.005 0.006UV-254 (cm-1) Permeate Conductivity Permeate UV-254 70% 50% 85% 95% Figure F-31. Permeate Conductivity & UV254 vs. Time, Filmtec NF270(c), Recovery Test 0 5 10 15 20 25 30 35 40 45 50 020406080100120Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-32. Pressure vs. Time, Filmtec NF270(c), Recovery Test

PAGE 117

102Appendix F (continued) 0 1 2 3 4 5 6 0.20.61.01.52.02.53.013.815.318.6Time (Hours)Flow-Rate (mL/min ) 0 2 4 6 8 10 12 14Permeate Flux (gfd ) Permeate flow Waste flow Permeate Flux Figure F-33. Flow Rates and Permeat e Flux vs Time, Hydranautics LFC1, Setting Test 5.0E-05 6.0E-05 7.0E-05 8.0E-05 9.0E-05 1.0E-04 1.1E-04 1.2E-04 1.3E-04 1.4E-04 1.5E-04 036912151821Time (hours)MTC (gpm/ft2/psi ) Uncorrected Data Normalized to 25C Figure F-34. MTC vs. Time, H ydranautics LFC1, Setting Test

PAGE 118

103Appendix F (continued) 0.0 0.1 0.1 0.2 0.2 0.3 0.3 02468101214161820Time (hours)MTC (gpd/ft2/psi) Figure F-35. MTC vs. Time, H ydranautics LFC1, Setting Test 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0102030405060708090100Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-36. MTC vs. Time, Hydr anautics LFC1, Recovery Test

PAGE 119

104Appendix F (continued) 0 1 2 3 4 5 6 7 8 9 10 0102030405060708090100Time (hours)Flow Rate (ml/min) Permeate flow Waste Flow 70% 50% 85% 95% Figure F-37. Flow Rates vs Time, Hy dranautics LFC1, Recovery Test 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 0102030405060708090100Time (hours)Flux (gpd/ft2) 70% 50% 85% 95% Figure F-38. Permeate Flux vs Time, H ydranautics LFC1, Recovery Test

PAGE 120

105Appendix F (continued) 0 5 10 15 20 25 30 35 40 45 50 0.22.210.823.741.561.078.985.390.594.8Time (Hours)Permeate Conductivity0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045Permeate UV-254 Permeate Conductivity Permeate UV-254 70% 85% 95% 85% 50% Figure F-39. Permeate Conductivity & UV254 vs. Time, Hydranautics LFC1, Recovery Test 0 10 20 30 40 50 60 70 80 90 100 0102030405060708090100Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-40. Pressure vs. Time, H ydranautics LFC1, Recovery Test

PAGE 121

106Appendix F (continued) 0 2 4 6 8 10 12 0.10.70.91.41.92.32.912.115.819.8Time (Hours)Flow-Rate (mL/min ) 0 5 10 15 20 25Permeate Flux (gfd ) Permeate flow Waste flow Permeate Flux Figure F-41. Flow Rates and Permeate Flux vs Time, Hydranautics NTR 7450, Setting Test 1.5E-04 2.0E-04 2.5E-04 3.0E-04 3.5E-04 4.0E-04 4.5E-04 5.0E-04 5.5E-04 6.0E-04 036912151821Time (hours)MTC (gpm/ft2/psi ) Uncorrected Data Normalized to 25C Figure F-42. MTC vs. Time, Hydr anautics NTR7450, Setting Test

PAGE 122

107Appendix F (continued) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0510152025Time (hours)MTC (gpd/ft2/psi) Figure F-43. MTC vs. Time, Hydranautics NTR 7450, Setting Test 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0102030405060708090100Time (hours)MTC (gpd/ft2/psi) 70% 50% 85% 95% Figure F-44. MTC vs. Time, Hydran autics NTR 7450, Recovery Test

PAGE 123

108Appendix F (continued) 0 1 2 3 4 5 6 7 8 9 10 0102030405060708090100 Time (hours)Flow Rate (ml/min ) Permeate flow Waste Flow 70% 50% 85% 95% Figure F-45. Flow Rates vs Time, Hy dranautics NTR 7450, Recovery Test 0.0 5.0 10.0 15.0 20.0 25.0 0102030405060708090100Time (hours)Flux (gpd/ft2) 70%50% 8% 95% 85% Figure F-46. Permeate Flux vs Time, H ydranautics NTR7450, Recovery Test

PAGE 124

109Appendix F (continued) 0 10 20 30 40 50 60 70 80 90 02723395777859298Time (Hours)Permeate Conductivity (uS/cm ) 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05Permeate UV-254 (cm-1) Permeate Conductivity Permeate UV-254 70% 50% 85 % 95 % Figure F-47. Permeate Conductivity & UV254 vs. Time, Hydranautics NTR 7450, Recovery Test 0 10 20 30 40 50 60 70 80 0102030405060708090100 Time (hours)Pressure (psi) Influent Pressure Retentate Pressure Figure F-48. Pressure vs. Time, Hydr anautics NTR 7450, Recovery Test

PAGE 125

110Appendix G: NF Membrane Tests Overall Performance The results and graphical presentation of rejection for inorgani cs total hardness and conductivity, organics – UV254, TOC, THMFP and HAAFP, and algal exudates – microcystin-LR, MIB and geosmin are presented in this section. 0 10 20 30 40 50 60 70 80 90 100Total H ardne s s Cond u ctivit y UV -254 T OC T HM FP HA A FP Mic-LR G eos min MIBParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery 0 10 20 30 40 50 60 70 80 90 100Tot a l Hard n ess Co n d u cti vit y UV-254 T O C T H MF P HA A FP Mi c L R G e o sm i n MIBParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery Figure G-1. Summary of NF 90 Membrane Performance at 19 gfd Figure G-2. Summary of NF 270 Membrane Performance at 32 gfd

PAGE 126

111Appendix G (continued) Table G-1 Summary of NF 90 Treatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 18.1 19 19.8 20.3 MTCW,20C gpd/ft2-psi 0.28 0.29 0.30 0.32 Inlet pressure, psi 81.8 82.8 83.2 83.6 Net driving pressure, psi 64.9 65.4 65.6 65.8 Pressure drop, psi 33.8 34.8 35.2 35.6 Microcystin rejection, % 99 99 99 98 Permeate Microcystin, g/L 0.11 0.14 0.17 0.23 1.02 Geosmin rejection, % 97 97 97 97 Permeate geosmin3, ng/L 6.9 7.3 7.8 8.8 10.04 MIB rejection, % 98 96 95 92 Permeate MIB5, ng/L 5.4 11.8 12.1 19.5 29.04 TOC rejection, % 95 95 94 92 Permeate TOC, mg/L 0.237 0.265 0.285 0.367 UV254 rejection, % 99 98 97 96 Permeate UV254, cm-1 0.0008 0.0014 0.0021 0.0026 THMFP rejection, % 90.4 89.1 88.1 85.8 Permeate THMFP, g/L 13.7 15.6 17.0 20.3 806 HAA(6)FP rejection, % Permeate HAA(6)FP, g/L ND7 ND7 ND7 ND7 606 Conductivity rejection, % 96 93 88 85 Permeate Conductivity, Siemens/cm 10 16 28 36 Total Hardness Rejection, % 99 97 96 95 Permeate Total Hardness, mg/L as CaCO3 1.3 3.1 4.1 5.05 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethy l-trans-9-decalol 7 Not Detectable 4 Mallevialle and Suffet, 1987

PAGE 127

112Appendix G (continued) Table G-2 Summary of NF 270 (a) Treatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 32.4 32.8 33.7 33.9 MTCW,20C gpd/ft2-psi 0.50 0.54 0.55 0.57 Inlet pressure, psi 78.3 79.5 80.2 82.4 Net driving pressure, psi 63.1 63.7 64.1 65.2 Pressure drop, psi 30.3 31.5 32.2 34.4 Microcystin rejection, % 97.1 96.4 97.0 93.9 Permeate Microcystin, g/L 0.28 0.35 0.30 0.590 1.02 Geosmin rejection, % 65.2 52.1 46.2 36.7 Permeate geosmin3, ng/L 21.2 29.2 32.8 38.6 10.04 MIB rejection, % 89.5 81.2 78.1 63.7 Permeate MIB5, ng/L 12.8 23.0 26.8 44.4 29.04 TOC rejection, % 96 95 94 93 Permeate TOC, mg/L 0.183 0.215 0.245 0.296 UV254 rejection, % 98 96 94 93 Permeate UV254, cm-1 0.0013 0.0021 0.0029 0.0036 THMFP rejection, % 79.0 79.2 78.8 76.4 Permeate THMFP, g/L 23.7 23.4 23.9 26.6 806 HAA(6)FP rejection, % 90.9 Permeate HAA(6)FP, g/L ND7 ND7 ND7 48.7 606 Conductivity rejection, % 81 78 76 74 Permeate Conductivity, Siemens/cm 44 49 54 58 Total Hardness Rejection, % 89 87 85 84 Permeate Total Hardness, mg/L as CaCO3 10.8 11.9 13.7 14.6 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethy l-trans-9-decalol 7 Not Detectable 4 Mallevialle and Suffet, 1987

PAGE 128

113Appendix G (continued) 0 10 20 30 40 50 60 70 80 90 100T o tal H ardn es s Conductivit y UV-254 TOC THMFP HAA F P M i c-LR Geo s m in MIBParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery Figure G-3. Summary of NF 270 Membrane Performance at 19 gfd 0 10 20 30 40 50 60 70 80 90 100Total Hardne s s Co n duc t i v it y U V-25 4 TOC TH MFP HA A F P Mic -LR Ge os m in MIBParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery Figure G-4. Summary of NF 270 Membrane Performance at 12 gfd

PAGE 129

114Appendix G (continued) Table G-3 Summary of NF 270 (b) Treatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 18.7 18.9 19.2 19.4 MTCW,20C gpd/ft2-psi 0.37 0.36 0.34 0.32 Inlet pressure, psi 59.4 61.3 63.6 64.8 Net driving pressure, psi 54.1 55.2 55.5 56.5 Pressure drop, psi 12.7 14.3 15.9 16.7 Microcystin rejection, % 98.1 97.9 97.5 96.0 Permeate Microcystin, g/L 0.20 0.22 0.26 0.41 1.02 Geosmin rejection, % 82.0 82.0 72.9 73.4 Permeate geosmin3, ng/L 10.2 10.2 15.4 15.1 10.04 MIB rejection, % 91.6 80.9 81.8 71.9 Permeate MIB5, ng/L 5.1 11.6 11.1 17.1 29.04 TOC rejection, % 94.9 94.5 93.2 91.7 Permeate TOC, mg/L 0.183 0.198 0.245 0.302 UV254 rejection, % 98 97 94 92 Permeate UV254, cm-1 0.0013 0.0019 0.0036 0.0052 THMFP rejection, % 80.5 80.3 80.1 78.4 Permeate THMFP, g/L 22.7 22.9 23.2 25.1 806 HAA(6)FP rejection, % 92.3 98.6 89.4 90.3 Permeate HAA(6)FP, g/L 3.81 0.69 5.29 4.84 606 Conductivity rejection, % 83 80 72 67 Permeate Conductivity, Siemens/cm 38 48 66 82 Total Hardness Rejection, % 90.5 88.8 86.0 82.5 Permeate Total Hardness, mg/L as CaCO3 8.6 10.2 12.7 15.9 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethy l-trans-9-decalol 7 Not Detectable 4 Mallevialle and Suffet, 1987

PAGE 130

115Appendix G (continued) Table G-4 Summary of NF 270 (c) Treatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 11.6 11.9 12.3 12.7 MTCW,20C gpd/ft2-psi 0.37 0.36 0.35 0.33 Inlet pressure, psi 39.6 41.4 43.7 44.5 Net driving pressure, psi 33.4 36.1 37.2 38.5 Pressure drop, psi 9.7 12.1 14.2 15.8 Microcystin rejection, % 98.7 98.3 97.6 96.8 Permeate Microcystin, g/L 0.14 0.18 0.26 0.34 1.02 Geosmin rejection, % 85.3 84.9 82.3 73.0 Permeate geosmin3, ng/L 5.9 6.1 7.1 10.8 10.04 MIB rejection, % 90.8 85.4 87.1 77.9 Permeate MIB5, ng/L 5.1 8.0 7.1 12.0 29.04 TOC rejection, % 95.0 94.6 93.4 91.9 Permeate TOC, mg/L 0.18 0.19 0.23 0.29 UV254 rejection, % 99 98 96 95 Permeate UV254, cm-1 0.0011 0.0019 0.0033 0.0046 THMFP rejection, % 79.8 77.3 82.2 81.7 Permeate THMFP, g/L 30.9 34.8 27.3 28.1 806 HAA(6)FP rejection, % 84.8 84.9 41.1 33.7 Permeate HAA(6)FP, g/L 21.4 21.6 83.0 93.5 606 Conductivity rejection, % 84 82 75 71 Permeate Conductivity, Siemens/cm 33 46 65 80 Total Hardness Rejection, % 89.8 89.1 85.5 82.9 Permeate Total Hardness, mg/L as CaCO3 9.9 10.6 14.1 16.6 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethyltrans-9-decalol 4 Mallevialle and Suffet, 1987

PAGE 131

116 Appendix G (continued) 0 10 20 30 40 50 60 70 80 90 100Tota l Hardnes s C onductivit y UV-25 4 TO C THMFP HA A FP M ic -LR G eos m in MIBParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery Figure G-5. Summary of LFC 1 Membrane Performance at 10 gfd 0 10 20 30 40 50 60 70 80 90 100T o tal H a rd n e s s C o ndu c t i vit y U V -254 T OC THM F P HAAF P Mic-LR Ge o smin MI BParameter% Rejection 50% Recovery 70% Recovery 85% Recovery 95 % Recovery Figure G-6. Summary of NTR 7450 Membrane Performance at 19 gfd

PAGE 132

117Appendix G (continued) Table G-5 Summary of LFC 1 Tr eatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 8.8 9.5 10.6 10.9 MTCW,20C gpd/ft2-psi 0.142 0.148 0.153 0.159 Inlet pressure, psi 83.6 84.9 87.4 88.8 Net driving pressure, psi 65.2 66.4 67.7 69.6 Pressure drop, psi 35.7 37.1 38.3 39.6 Microcystin rejection, % 99.0 98.8 98.5 98.0 Permeate Microcystin, g/L 0.10 0.12 0.16 0.21 1.02 Geosmin rejection, % 99 Permeate geosmin3, ng/L ND7 ND7 0.45 ND7 10.04 MIB rejection, % 96 99 99 Permeate MIB5, ng/L ND7 1.97 0.5 0.45 29.04 TOC rejection, % 95.5 95.9 95.1 94.0 Permeate TOC, mg/L 0.17 0.15 0.19 0.23 UV254 rejection, % 99 98 96 94 Permeate UV254, cm-1 0.0009 0.0018 0.0033 0.0039 THMFP rejection, % 80.0 82.5 78.2 71.0 Permeate THMFP, g/L 28.7 25.1 31.3 41.7 806 HAA(6)FP rejection, % 93.5 79.7 94.4 91.2 Permeate HAA(6)FP, g/L 8.5 26.7 7.3 11.6 606 Conductivity rejection, % 93 90 86 84 Permeate Conductivity, Siemens/cm 19 26 35 44 Total Hardness Rejection, % 99.5 97.0 95.8 95.4 Permeate Total Hardness, mg/L as CaCO3 0.5 2.8 3.9 4.4 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethy l-trans-9-decalol 7 Not Detectable 4 Mallevialle and Suffet, 1987

PAGE 133

118Appendix G (continued) Table G-6 Summary of NTR 7450 Tr eatment Performance Recovery 50 70 85 95 Treatment Goal Measured Flux gpd/ft2 18.7 19.4 19.7 20.1 MTCW,20C gpd/ft2-psi 0.31 0.31 0.32 0.32 Inlet pressure, psi 72.5 70.9 73.8 74.3 Net driving pressure, psi 59.5 60.4 60.9 61.4 Pressure drop, psi 24.6 25.3 26.5 26.9 Microcystin rejection, % 87.6 83.1 81.1 79.9 Permeate Microcystin, g/L 1.22 1.65 1.85 1.97 1.02 Geosmin rejection, % 16.9 22.9 24.6 30.6 Permeate geosmin3, ng/L 45.2 42.1 41.0 37.7 10.04 MIB rejection, % 27.3 15.9 20.4 11.8 Permeate MIB5, ng/L 42.5 49.1 46.5 51.5 29.04 TOC rejection, % 69.6 62.4 55.2 46.1 Permeate TOC, mg/L 1.14 1.41 1.68 2.02 UV254 rejection, % 66 59 53 48 Permeate UV254, cm-1 0.025 0.032 0.036 0.042 THMFP rejection, % 48.5 40.3 25.9 16.9 Permeate THMFP, g/L 75.1 87.1 108.1 121.3 806 HAA(6)FP rejection, % 84.2 74.5 24.7 17.5 Permeate HAA(6)FP, g/L 32.7 52.9 156.2 171.0 606 Conductivity rejection, % 81 72 68 65 Permeate Conductivity, Siemens/cm 45 54 65 77 Total Hardness Rejection, % 69.8 54.7 47.4 45.6 Permeate Total Hardness, mg/L as CaCO3 28.4 42.6 49.4 51.1 1 Manufacturer data 5 1,2,7,7-tetramethyl-exo-bicyclo-(2.2.1)heptan-2-ol 2 World Health Organization, 1998 6 USEPA, 1996. HAA(5) Stage I EPA MCL 3 Trans-1,10-dimethyltrans-9-decalol 4 Mallevialle and Suffet, 1987