Removal of Bromate and Perchlorate in Ozone/GAC Systems
Removal of Bromate and Perchlorate in Ozone/GAC Systems
Removal of Bromate and Perchlorate in Ozone/GAC Systems
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
AWWA<br />
Research<br />
Foundation<br />
<strong>Removal</strong> <strong>of</strong> <strong>Bromate</strong><br />
<strong>and</strong> <strong>Perchlorate</strong> <strong>in</strong><br />
Conventional<br />
<strong>Ozone</strong>/<strong>GAC</strong> <strong>Systems</strong><br />
Subject Area:<br />
Water Treatment
<strong>Removal</strong> <strong>of</strong> <strong>Bromate</strong><br />
<strong>and</strong> <strong>Perchlorate</strong> <strong>in</strong><br />
Conventional____<br />
<strong>Ozone</strong>/<strong>GAC</strong> <strong>Systems</strong>
The mission <strong>of</strong> the AWWA Research Foundation is to advance the science <strong>of</strong> water to improve<br />
the quality <strong>of</strong> life. Funded primarily through annual subscription payments from over 1,000 utili<br />
ties, consult<strong>in</strong>g firms, <strong>and</strong> manufacturers <strong>in</strong> North America <strong>and</strong> abroad, AWWARF sponsors<br />
research on all aspects <strong>of</strong> dr<strong>in</strong>k<strong>in</strong>g water, <strong>in</strong>clud<strong>in</strong>g supply <strong>and</strong> resources, treatment, monitor<strong>in</strong>g<br />
<strong>and</strong> analysis, distribution, management, <strong>and</strong> health effects.<br />
From its headquarters <strong>in</strong> Denver, Colorado, the AWWARF staff directs <strong>and</strong> supports the efforts<br />
<strong>of</strong> over 500 volunteers, who are the heart <strong>of</strong> the research program. These volunteers, serv<strong>in</strong>g on<br />
various boards <strong>and</strong> committees, use their expertise to select <strong>and</strong> monitor research studies to ben<br />
efit the entire dr<strong>in</strong>k<strong>in</strong>g water community.<br />
Research f<strong>in</strong>d<strong>in</strong>gs are dissem<strong>in</strong>ated through a number <strong>of</strong> technology transfer activities, <strong>in</strong>clud<br />
<strong>in</strong>g research reports, conferences, videotape summaries, <strong>and</strong> periodicals.
<strong>Removal</strong> <strong>of</strong> <strong>Bromate</strong><br />
<strong>and</strong> Perchlonate <strong>in</strong><br />
Conventional____<br />
<strong>Ozone</strong>/<strong>GAC</strong> <strong>Systems</strong><br />
Prepared by:<br />
Mary Jo Kirisits<br />
Jess C. Brown<br />
Vernon L. Snoey<strong>in</strong>k<br />
Lutgarde M. Rask<strong>in</strong><br />
Joanne C. Chee-Sanford<br />
University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Champaign<br />
3230 Newmark Civil Eng<strong>in</strong>eer<strong>in</strong>g Laboratory<br />
205 North Mathews Avenue<br />
Urbana, Ill<strong>in</strong>ois 61801<br />
<strong>and</strong><br />
Sun Liang<br />
Joon M<strong>in</strong><br />
The Metropolitan Water District <strong>of</strong> Southern California<br />
700 North Moreno Avenue<br />
La Verne, California 91750<br />
Sponsored by:<br />
AWWA Research Foundation<br />
6666 West Qu<strong>in</strong>cy Avenue<br />
Denver, CO 80235-3098<br />
Published by the<br />
AWWA Research Foundation <strong>and</strong> the<br />
American Water Works Association
Disclaimer<br />
This study was funded by the AWWA Research Foundation (AWWARF). AWWARF assumes no responsibility for the<br />
content <strong>of</strong> the research study reported <strong>in</strong> this publication or for the op<strong>in</strong>ions or statements <strong>of</strong> fact expressed <strong>in</strong> the report.<br />
The mention <strong>of</strong> trade names for commercial products does not represent or imply the approval or endorsement <strong>of</strong> AWWARF.<br />
This report is presented solely for <strong>in</strong>formational purposes.<br />
Library <strong>of</strong> Congress Catalog<strong>in</strong>g-<strong>in</strong>-Publication Data has been applied for.<br />
Copyright 2001<br />
by<br />
AWWA Research Foundation<br />
<strong>and</strong><br />
American Water Works Association<br />
Pr<strong>in</strong>ted <strong>in</strong> the U.S.A.<br />
ISBN 1-58321-104-7 Pr<strong>in</strong>ted on recycled paper
CONTENTS<br />
TABLES ................................................................ ix<br />
FIGURES .............................................................. xi<br />
FOREWORD ........................................................... xv<br />
ACKNOWLEDGMENTS .......................;......................... xvii<br />
EXECUTIVE SUMMARY ................................................. xix<br />
CHAPTER 1. INTRODUCTION ............................................. 1<br />
Background - <strong>Bromate</strong> ............................................... 1<br />
Background - <strong>Perchlorate</strong> ............................................. 1<br />
Previous Studies - <strong>Bromate</strong> ............................................ 3<br />
<strong>Bromate</strong> Reduction with Various Influent Dissolved Oxygen<br />
Concentrations .......................................... 5<br />
<strong>Bromate</strong> Reduction to Bromide ................................... 8<br />
<strong>Bromate</strong> Reduction with Various Influent Nitrate Concentrations .......... 9<br />
<strong>Bromate</strong> Reduction with Various EBCTs ........................... 13<br />
Verification <strong>of</strong> Biological <strong>Bromate</strong> <strong>Removal</strong> ......................... 14<br />
Previous Studies - <strong>Perchlorate</strong> ......................................... 16<br />
Objectives ........................................................ 17<br />
CHAPTER 2. MATERIALS AND METHODS ................................. 19<br />
Materials Used for <strong>Bromate</strong> Reduction Experiments ........................ 19<br />
Water ...................................................... 19<br />
Reagents .................................................... 20<br />
Methods Utilized for <strong>Bromate</strong> Reduction Experiments ....................... 20<br />
Biologically Active Carbon- Small Columns ......................... 20<br />
Biologically Active Carbon - Large Filter ........................... 23<br />
Backwash<strong>in</strong>g the BAG Filters .................................... 25<br />
Heterotrophic Plate Counts ..................................... 25<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis ........................... 25<br />
Plat<strong>in</strong>g Experiments ........................................... 26<br />
Dilution-to-Ext<strong>in</strong>ction Experiments ............................... 28
Serum Bottle Experiments ...................................... 30<br />
Materials Used for <strong>Perchlorate</strong> Reduction Experiments ....................... 30<br />
Water ...................................................... 30<br />
Reagents <strong>and</strong> Carbon .......................................... 30<br />
Methods Utilized for <strong>Perchlorate</strong> Reduction Experiments ..................... 31<br />
Biologically Active Carbon Filters ................................. 31<br />
Abiotic Batch Experiments ...................................... 32<br />
Batch Test Wash Experiments ................................... 32<br />
Metal-Catalyzed <strong>GAC</strong> Filter Experiments ........................... 33<br />
Metal-Catalyzed <strong>GAC</strong> Filter Wash Experiments ...................... 33<br />
<strong>Ozone</strong>/Hydrogen Peroxide/<strong>Perchlorate</strong> Batch Tests ................... 33<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis ........................... 34<br />
Plat<strong>in</strong>g Experiments ........................................... 34<br />
Dilution-to-Ext<strong>in</strong>ction Experiments ............................... 35<br />
Serum Bottle Experiments ...................................... 35<br />
Analytical Methods ................................................. 37<br />
CHAPTER 3. RESULTS AND DISCUSSION .................................. 39<br />
<strong>Bromate</strong>.......................................................... 39<br />
Verification <strong>of</strong> Biological <strong>Bromate</strong> <strong>Removal</strong> <strong>in</strong> the BAG Filters .......... 39<br />
Further Test<strong>in</strong>g the Reproducibility <strong>of</strong> BAC Filters .................... 42<br />
Backwash<strong>in</strong>g BAC Filters ....................................... 42<br />
Effect <strong>of</strong> the Initial <strong>Bromate</strong> Concentration .......................... 44<br />
Effect <strong>of</strong> the Influent Dissolved Oxygen Concentration ................. 46<br />
Effect <strong>of</strong> the Influent Nitrate Concentration ......................... 49<br />
Effect <strong>of</strong> the Influent Sulfate Concentration ......................... 51<br />
Effect <strong>of</strong>pH ................................................. 52<br />
Effect <strong>of</strong> Source Water Type .................................... 53<br />
<strong>Perchlorate</strong> Reduction <strong>in</strong> CUW ..........................:........ 60<br />
Microscopic Exam<strong>in</strong>ation <strong>of</strong> Microorganisms <strong>in</strong> the Filters .............. 60<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis Experiment .................. 62<br />
Plat<strong>in</strong>g Experiments <strong>and</strong> Microbial Characterization ................... 64<br />
vi
Dilution-to-Ext<strong>in</strong>ction Experiments ............................... 69<br />
Serum Bottle Experiments ...................................... 72<br />
<strong>Perchlorate</strong> - Abiotic <strong>Removal</strong> ......................................... 73<br />
Calculation <strong>of</strong> Ion Exchange Capacity for Virg<strong>in</strong> Norit <strong>GAC</strong> ............ 73<br />
Abiotic Batch Experiments ...................................... 74<br />
Wash Tests .................................................. 77<br />
Metal-Catalyzed <strong>GAC</strong> Experiments ............................... 77<br />
<strong>Ozone</strong>/Hydrogen Peroxide/<strong>Perchlorate</strong> Batch Tests ................... 80<br />
<strong>Perchlorate</strong> - Biological Reduction ..................................... 80<br />
BAG Filtration Experiments ..................................... 80<br />
Batch Biological Experiments .................................... 89<br />
Microscopic Exam<strong>in</strong>ation <strong>of</strong> Microorganisms <strong>in</strong> the Filters .............. 96<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis Experiment .................. 98<br />
CHAPTER 4. ACTIVITIES AT THE METROPOLITAN WATER DISTRICT<br />
OF SOUTHERN CALIFORNIA ....................................... 99<br />
Introduction ......................................................... 99<br />
Objectives .......................................................... 99<br />
Bench Scale M<strong>in</strong>i-Column Study ....................................... 100<br />
Materials <strong>and</strong> Methods ........................................ 100<br />
Results <strong>and</strong> Discussion ........................................ 104<br />
Pilot-Scale Mobile Pilot Plant Study .................................... Ill<br />
Materials <strong>and</strong> Methods ........................................ Ill<br />
Results .................................................... 115<br />
CHAPTER 5. SUMMARY AND CONCLUSIONS ............................. 119<br />
<strong>Bromate</strong>......................................................... 119<br />
<strong>Perchlorate</strong>....................................................... 122<br />
Metropolitan Water District <strong>of</strong> Southern California: Ozonated Water<br />
(No Added Electron Donor) for <strong>Bromate</strong> <strong>and</strong> <strong>Perchlorate</strong> Reduction ....... 125<br />
CHAPTER 6. RECOMMENDATIONS TO UTILITIES ......................... 127<br />
REFERENCES ......................................................... 129<br />
ABBREVIATIONS ...................................................... 137<br />
vii
TABLES<br />
1.1 <strong>Bromate</strong> removal <strong>in</strong> DFE at pH 6.5 with various DO concentrations .............. 5<br />
1.2 <strong>Bromate</strong> removal <strong>in</strong> CUW at pH 7.5 with different <strong>in</strong>fluent nitrate<br />
concentrations ............................................... 10<br />
1.3 Batch biological bromate reduction experiments .............................. 15<br />
2.1 Average water quality characteristics <strong>of</strong> CUW <strong>and</strong> LMW ..................... 19<br />
2.2 Agar recipe for plat<strong>in</strong>g experiments ..................................... 27<br />
2.3 Experimental conditions: ozone/hydrogen peroxide/perchlorate batch tests ......... 34<br />
2.4 Serum bottle experiment 1 us<strong>in</strong>g perchlorate-reduc<strong>in</strong>g isolate B9<br />
(<strong>in</strong>itial conditions) ............................................. 36<br />
2.5 Serum bottle experiment 2 us<strong>in</strong>g perchlorate-reduc<strong>in</strong>g isolate B9<br />
(<strong>in</strong>itial conditions) ............................................. 37<br />
3.1 Effect <strong>of</strong> backwash<strong>in</strong>g on bromate removal <strong>in</strong> the BAG filter .................. 43<br />
3.2 The effect <strong>of</strong> <strong>in</strong>fluent DO concentration on bromate removal <strong>in</strong> BAG filters ....... 46<br />
3.3 DOC removal <strong>and</strong> DO rema<strong>in</strong><strong>in</strong>g <strong>in</strong> the BAG experiments us<strong>in</strong>g LMW ........... 54<br />
3.4 Test<strong>in</strong>g the isolates for the ability to reduce bromate (12-week <strong>in</strong>cubation;<br />
<strong>in</strong>itial conditions: CUW, 5 mM phosphate buffer, 28.7 ng/L BrO3",<br />
4.82 mg/L NO3-, pH 7.8, 0 mg/L DO) .............................. 65<br />
3.5 Test<strong>in</strong>g the isolates for the ability to reduce bromate (4.5-week <strong>in</strong>cubation;<br />
<strong>in</strong>itial conditions: DDW, 5 mM phosphate buffer, 43.5 ug/L BrO3",<br />
4.92 mg/L NO3', 8.3 mg/L DOC, pH 7.5, 0 mg/L DO) ................. 66<br />
3.6 Challeng<strong>in</strong>g isolates with a high nitrate concentration ........................ 67<br />
3.7 Challeng<strong>in</strong>g isolates with a high bromate concentration ...................... 68<br />
3.8 Challeng<strong>in</strong>g isolates with high concentrations <strong>of</strong> bromate <strong>and</strong> nitrate ............ 68<br />
3.9 Summary <strong>of</strong> experimental evidence for isolates ............................. 70<br />
3.10 Results <strong>of</strong> dilution-to-ext<strong>in</strong>ction experiments .............................. 71<br />
3.11 Abiotic perchlorate batch removal <strong>and</strong> wash tests ........................... 77<br />
3.12 <strong>Perchlorate</strong> mass balance: mass perchlorate removed by <strong>and</strong> recovered from<br />
metal-catalyzed GAG (10 mg/L NO3" r<strong>in</strong>se) ......................... 79<br />
IX
3.13 Extended perchlorate mass balance: total mass removed by <strong>and</strong> recovered from<br />
metal-catalyzed <strong>GAC</strong> (10 + 200 mg/L NO3" r<strong>in</strong>ses) ................... 79<br />
3.14 <strong>Ozone</strong>/hydrogen peroxide/perchlorate batch tests ........................... 81<br />
3.15 Reduction <strong>of</strong> perchlorate <strong>in</strong> aqueous medium us<strong>in</strong>g microbial isolates ............ 90<br />
3.16 Dilution-to-ext<strong>in</strong>ction experiments: perchlorate ............................ 95<br />
4.1 Historical water quality for Metropolitan source water ...................... 100<br />
4.2 Control study for chemical reduction pathway for perchlorate ................ 109<br />
4.3 Characteristics <strong>of</strong> Norit carbon used <strong>in</strong> mobile pilot plant .................... 112<br />
4.4 Mobile pilot plant operat<strong>in</strong>g conditions .................................. 115<br />
4.5 Preload<strong>in</strong>g results from the mobile pilot plant ............................. 116
FIGURES<br />
1.1 Effect <strong>of</strong> <strong>in</strong>fluent dissolved oxygen concentration on bromate breakthrough<br />
from the BAG filter ............................................. 6<br />
1.2 Possible depletion <strong>of</strong> bromate, nitrate, <strong>and</strong> dissolved oxygen <strong>in</strong> the bi<strong>of</strong>ilm ......... 7<br />
1.3 Reduction <strong>of</strong> bromate to bromide <strong>in</strong> a BAG filter ............................ 8<br />
1.4 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate breakthrough from<br />
a BAG filter ................................................. 10<br />
1.5 Nitrogen mass balance <strong>in</strong> experiments 4 <strong>and</strong> 5 (25-m<strong>in</strong>ute BAG filter) - ammonia,<br />
nitrite, <strong>and</strong> nitrate ............................................. 11<br />
1.6 Nitrogen mass balance <strong>in</strong> experiments 6 <strong>and</strong> 7 (49-50-m<strong>in</strong>ute BAG filter) - ammonia,<br />
nitrite, <strong>and</strong> nitrate ............................................. 12<br />
1.7 Nitrogen speciation <strong>in</strong> experiment 4 (25-m<strong>in</strong>ute BAG filter) .................... 12<br />
1.8 Nitrogen speciation <strong>in</strong> experiment 6 (49-m<strong>in</strong>ute BAG filter) ................... 13<br />
1.9 Percent bromate removal versus EBCT with different <strong>in</strong>fluent nitrate<br />
concentrations ............................................... 14<br />
2.1 Schematic <strong>of</strong> BAG filters us<strong>in</strong>g CUW .................................... 22<br />
2.2 Schematic <strong>of</strong> BAG filters us<strong>in</strong>g LMW ................................... 23<br />
2.3 Schematic <strong>of</strong> multi-port BAG filter us<strong>in</strong>g CUW ............................ 24<br />
2.4 Progression <strong>of</strong> plat<strong>in</strong>g <strong>and</strong> aqueous medium experiments with pure cultures ....... 29<br />
3.1 Repeat <strong>of</strong> the 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCT filtration experiment (average <strong>in</strong>fluent<br />
conditions: CUW, pH 7.5, 2 mg/L DO, 5.0 mg/L NO3', 20 ug/L BrO3") .... 40<br />
3.2 Repeat <strong>of</strong> the 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCT filtration experiment (average <strong>in</strong>fluent<br />
conditions: CUW, pH 7.5, 2 mg/L DO, 0.3 mg/L NO3', 20 |ag/L BrO 3") ... 40<br />
3.3 Chang<strong>in</strong>g the <strong>in</strong>fluent nitrate concentration from 0.3 to 5.0 mg/L <strong>in</strong> the<br />
BAG filters .................................................. 41<br />
3.4 <strong>Bromate</strong> removal <strong>in</strong> 1- <strong>and</strong> 2-<strong>in</strong>ch <strong>in</strong>ner-diameter BAG filters ................. 43<br />
3.5 Mass concentration bromate removal with different <strong>in</strong>fluent bromate concentrations<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 2 mg/L DO, 0.3 mg/L NO3") ........ 45<br />
3.6 Mass concentration bromate removal with different <strong>in</strong>fluent bromate concentrations<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 2 mg/L DO, 5.0 mg/L NO3") ........ 45<br />
xi
3.7 Percent bromate removal with various <strong>in</strong>fluent DO concentrations .............. 47<br />
3.8 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate removal <strong>in</strong> BAG filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ug/L BrO3', 2.1 mg/L DO,<br />
26-m<strong>in</strong>ute EBCT)............................................. 50<br />
3.9 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate removal <strong>in</strong> BAG filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ug/L BrO3', 2.1 mg/L DO,<br />
51-m<strong>in</strong>ute EBCT) ............................................. 50<br />
3.10 Effect <strong>of</strong> <strong>in</strong>fluent sulfate concentration on bromate removal <strong>in</strong> BAG filters ........ 52<br />
3.11 Effect <strong>of</strong> <strong>in</strong>fluent pH on bromate removal <strong>in</strong> BAG filters ...................... 53<br />
3.12 <strong>Bromate</strong> rema<strong>in</strong><strong>in</strong>g after first BAG filter us<strong>in</strong>g LMW ........................ 55<br />
3.13 Nitrate rema<strong>in</strong><strong>in</strong>g after first BAG filter us<strong>in</strong>g LMW ......................... 56<br />
3.14 <strong>Bromate</strong> rema<strong>in</strong><strong>in</strong>g after second BAG filter us<strong>in</strong>g LMW ...................... 57<br />
3.15 Nitrate rema<strong>in</strong><strong>in</strong>g after second BAG filter us<strong>in</strong>g LMW ....................... 58<br />
3.16 <strong>Perchlorate</strong> rema<strong>in</strong><strong>in</strong>g after first BAG filter us<strong>in</strong>g LMW ...................... 59<br />
3.17 <strong>Perchlorate</strong> rema<strong>in</strong><strong>in</strong>g after second BAG filter us<strong>in</strong>g LMW .................... 59<br />
3.18 Rods <strong>and</strong> cocci from the BAG filter ..................................... 61<br />
3.19 Cha<strong>in</strong> <strong>of</strong> rods from the BAG filter ...................................... 61<br />
3.20 Aggregated rods from the BAG filter .................................... 61<br />
3.21 Protozoa <strong>in</strong> the BAC filter ............................................ 62<br />
3.22 DGGE pr<strong>of</strong>ile <strong>of</strong> bromate <strong>and</strong> perchlorate samples .......................... 63<br />
3.23 <strong>Bromate</strong> reduction <strong>in</strong> a serum bottle experiment (<strong>in</strong>itial conditions: CUW,<br />
pH 7.7, 19 ug/L BKV, 5.4 mg/L NO3', 25 C) ....................... 72<br />
3.24 <strong>Bromate</strong> reduction <strong>in</strong> a serum bottle experiment (<strong>in</strong>itial conditions: CUW,<br />
pH 7.7, 19 ug/L BrO3-, 5.4 mg/L NO3', 4 C) ........................ 73<br />
3.25 Ion exchange <strong>of</strong> perchlorate to <strong>GAC</strong> .................................... 74<br />
3.26 Batch perchlorate removal (<strong>in</strong>itial conditions: CUW, pH 7.5, 500 mg/L <strong>GAC</strong>) ..... 75<br />
3.27 Batch perchlorate removal (<strong>in</strong>itial conditions: DDW, pH 7.5, 500 mg/L <strong>GAC</strong>) ..... 75<br />
3.28 Batch perchlorate removal (<strong>in</strong>itial conditions: pH 7.5, 500 mg/L <strong>GAC</strong>) .......... 76<br />
3.29 Batch perchlorate removal (<strong>in</strong>itial conditions: DDW, 500 mg/L <strong>GAC</strong>) ........... 76<br />
3.30 <strong>Removal</strong> <strong>of</strong> perchlorate by <strong>GAC</strong> mixed with alum<strong>in</strong>um shot ................... 78<br />
3.31 <strong>Removal</strong> <strong>of</strong> perchlorate by copper- <strong>and</strong> z<strong>in</strong>c oxide-impregnated carbon .......... 78<br />
xii
3.32 <strong>Perchlorate</strong> removal vs. nitrate rema<strong>in</strong><strong>in</strong>g ................................. 82<br />
3.33 Nitrate isolation experiments .......................................... 83<br />
3.34 Comparison <strong>of</strong> perchlorate removal for nitrate preload <strong>and</strong> non-nitrate preload<br />
conditions ................................................... 85<br />
3.35 Comparison <strong>of</strong> perchlorate removal at various EBCTs ....................... 86<br />
3.36 Comparison <strong>of</strong> perchlorate removal for 25-m<strong>in</strong>ute EBCT experiments ........... 87<br />
3.37 BAC filter recovery experiments ....................................... 89<br />
3.38 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, C1O4- 50 ug/L, 0.1<br />
mg/L NCV, DO « 0 mg/L, pH 7.5) ................................ 91<br />
3.39 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50 ng/L C1O4-, 0.4<br />
mg/L NCV, DO « 0 mg/L, pH 7.5) ................................ 92<br />
3.40 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50 |ig/L C1O4-, 1.6<br />
mg/L NCV, DO « 0 mg/L, pH 7.5) ................................ 92<br />
3.41 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50 [ig/L C1O4-, 2.6<br />
mg/L NCV, DO « 0 mg/L, pH 7.5) ................................ 93<br />
3.42 Isolate B9 serum bottle experiment 2: nitrate degradation ..................... 94<br />
3.43 Isolate B9 serum bottle experiment 2: perchlorate degradation ................. 94<br />
3.44 Rods <strong>and</strong> protozoa from the BAC filter .................................. 97<br />
3.45 Bacterial floe from the BAC filter ....................................... 97<br />
3.46 Large protozoan from the BAC filter .................................... 98<br />
4.1 Schematic <strong>of</strong> the bench-scale filter setup at Metropolitan (not to scale) ......... 102<br />
4.2 <strong>Removal</strong> <strong>of</strong> perchlorate by the m<strong>in</strong>i-column under various conditions ........... 105<br />
4.3 <strong>Removal</strong> <strong>of</strong> bromate by the m<strong>in</strong>i-column under various conditions ............. 105<br />
4.4 Mobile pilot plant ....................................................111<br />
4.5 Schematic <strong>of</strong> the mobile pilot plant setup at Metropolitan (not to scale) ......... 112<br />
4.6 Backwash water supply setup ......................................... 114<br />
4.7 Pre-load<strong>in</strong>g <strong>and</strong> biological reduction <strong>of</strong> perchlorate at the pilot-scale ........... 117<br />
4.8 Pre-load<strong>in</strong>g <strong>and</strong> biological reduction <strong>of</strong> bromate at the pilot-scale .............. 117<br />
Xlll
FOREWORD<br />
The AWWA Research Foundation is a nonpr<strong>of</strong>it corporation that is dedicated to the<br />
implementation <strong>of</strong> a research effort to help utilities respond to regulatory requirements <strong>and</strong> traditional<br />
high-priority concerns <strong>of</strong> the <strong>in</strong>dustry. The research agenda is developed through a process <strong>of</strong><br />
consultation with subscribers <strong>and</strong> dr<strong>in</strong>k<strong>in</strong>g water pr<strong>of</strong>essionals. Under the umbrella <strong>of</strong> a Strategic<br />
Research Plan, the Research Advisory Council prioritizes the suggested projects based upon current<br />
<strong>and</strong> future needs, applicability, <strong>and</strong> past work; the recommendations are forwarded to the Board <strong>of</strong><br />
Trustees for f<strong>in</strong>al selection. The foundation also sponsors research projects through the unsolicited<br />
proposal process; the Collaborative Research, Research Applications, <strong>and</strong> Tailored Collaboration<br />
programs; <strong>and</strong> various jo<strong>in</strong>t research efforts with organizations such as the U.S. Environmental<br />
Protection Agency, the U. S. Bureau <strong>of</strong> Reclamation, <strong>and</strong> the Association <strong>of</strong> California Water<br />
Agencies.<br />
This publication is a result <strong>of</strong> one <strong>of</strong> these sponsored studies, <strong>and</strong> it is hoped that its f<strong>in</strong>d<strong>in</strong>gs<br />
will be applied <strong>in</strong> communities throughout the world. The follow<strong>in</strong>g report serves not only as a<br />
means <strong>of</strong> communicat<strong>in</strong>g the results <strong>of</strong> the water <strong>in</strong>dustry's centralized research program but also as<br />
a tool to enlist the further support <strong>of</strong> the nonmember utilities <strong>and</strong> <strong>in</strong>dividuals.<br />
Projects are managed closely from their <strong>in</strong>ception to the f<strong>in</strong>al report by the foundation's staff<br />
<strong>and</strong> large cadre <strong>of</strong> volunteers who will<strong>in</strong>gly contribute their time <strong>and</strong> expertise. The foundation<br />
serves a plann<strong>in</strong>g <strong>and</strong> management function <strong>and</strong> awards contracts to other <strong>in</strong>stitutions such as water<br />
utilities, universities, <strong>and</strong> eng<strong>in</strong>eer<strong>in</strong>g firms. The fund<strong>in</strong>g for this research effort comes primarily from<br />
the Subscription Program, through which water utilities subscribe to the research program <strong>and</strong> make<br />
an annual payment proportionate to the volume <strong>of</strong> water they deliver <strong>and</strong> consultants <strong>and</strong><br />
manufacturers subscribe based on their annual bill<strong>in</strong>gs. The program <strong>of</strong>fers a cost-effective <strong>and</strong> fair<br />
method for fund<strong>in</strong>g research <strong>in</strong> the public <strong>in</strong>terest.<br />
A broad spectrum <strong>of</strong> water supply issues is addressed by the foundation's research agenda:<br />
resources, treatment <strong>and</strong> operations, distribution <strong>and</strong> storage, water quality <strong>and</strong> analysis, toxicology,<br />
economics, <strong>and</strong> management. The ultimate purpose <strong>of</strong> the coord<strong>in</strong>ated effort is to assist water<br />
suppliers to provide the highest possible quality <strong>of</strong> water economically <strong>and</strong> reliably. The true benefits<br />
are realized when the results are implemented at the utility level. The Foundation's trustees are<br />
pleased to <strong>of</strong>fer this publication as a contribution toward that end.<br />
xv
<strong>Perchlorate</strong> <strong>and</strong> chlorate salts are widely used by the chemical, aerospace <strong>and</strong> defense<br />
<strong>in</strong>dustries as oxidizers <strong>in</strong> propellant, explosives, <strong>and</strong> pyrotechnics. In April 1997, the limit <strong>of</strong><br />
detection for perchlorate decreased from 400 ug/L to 4 |ig/L. S<strong>in</strong>ce that time, perchlorate<br />
concentrations rang<strong>in</strong>g from s<strong>in</strong>gle digits to hundreds <strong>of</strong> ug/L have been found <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water<br />
supplies. <strong>Bromate</strong> is a regulated dis<strong>in</strong>fection by-product result<strong>in</strong>g from ozonation <strong>of</strong> water conta<strong>in</strong><strong>in</strong>g<br />
bromide. It is currently regulated at 10 ng/L. This report describes whether conventional ozone<br />
granular/activated carbon systems can be modified to remove perchlorate <strong>and</strong> bromate without<br />
sacrific<strong>in</strong>g system performance.<br />
Julius Ciaccia, Jr. James F. Manwar<strong>in</strong>g, P. E.<br />
Chair, Board <strong>of</strong> Trustees Executive Director<br />
AWWA Research Foundation AWWA Research Foundation<br />
xvi
ACKNOWLEDGMENTS<br />
The authors <strong>of</strong> this report would like to thank the AWWARF project <strong>of</strong>ficer, Kenan Ozek<strong>in</strong>,<br />
<strong>and</strong> the PAC (Project Advisory Committee) - Issam Najm (Water Quality <strong>and</strong> Treatment Solutions,<br />
Inc., Chatsworth, California), Rick Sase (Ma<strong>in</strong> San Gabriel Bas<strong>in</strong> Watermaster, Azusa, California)<br />
<strong>and</strong> Paul Westerh<strong>of</strong>f (Arizona State University, Tempe, Arizona) for all <strong>of</strong> their help <strong>and</strong> advice. The<br />
authors also wish to thank Becky Daugherty, Hatice Inan, <strong>and</strong> Richard L<strong>in</strong> for assistance with data<br />
collection.<br />
xvn
EXECUTIVE SUMMARY<br />
<strong>Bromate</strong> (BrO3~) <strong>and</strong> perchlorate (C1O4~) are <strong>in</strong>organic ions that have been found <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g<br />
water. <strong>Bromate</strong> can be formed dur<strong>in</strong>g ozonation due to the oxidation <strong>of</strong> bromide (Haag <strong>and</strong> Hoigne<br />
1983), <strong>and</strong> it is regulated by the United States Environmental Protection Agency (USEPA) at a<br />
concentration <strong>of</strong> 10 ug/L (Waterweek 1998). Based on the work <strong>of</strong> Kurokawa et al. (1990) <strong>and</strong><br />
DeAngelo et al. (1998), it was determ<strong>in</strong>ed that the <strong>in</strong>gestion <strong>of</strong> aqueous bromate caused renal cell<br />
tumors <strong>in</strong> rats. <strong>Perchlorate</strong> is a component <strong>of</strong> munitions <strong>and</strong> rocket fuel, <strong>and</strong> the contam<strong>in</strong>ation <strong>of</strong><br />
dr<strong>in</strong>k<strong>in</strong>g water may be related to these sources. No state or federal st<strong>and</strong>ard currently exists for<br />
perchlorate, but the USEPA recently set a national action level <strong>of</strong> 32 |ug/L (USEPA 1999).<br />
<strong>Perchlorate</strong> is a health concern due to its ability to disrupt the thyroid gl<strong>and</strong>'s use <strong>of</strong> iod<strong>in</strong>e <strong>in</strong> the<br />
generation <strong>of</strong> metabolic hormones. Thus, the removal <strong>of</strong> bromate <strong>and</strong> perchlorate from dr<strong>in</strong>k<strong>in</strong>g<br />
water or their reduction to <strong>in</strong>nocuous end-products is desirable, <strong>and</strong> requires development <strong>of</strong><br />
appropriate technologies. It was believed that the common ionic structure <strong>of</strong> bromate <strong>and</strong><br />
perchlorate would lend itself to a common treatment method for both anions.<br />
RESEARCH OBJECTIVES<br />
The two ma<strong>in</strong> objectives <strong>of</strong> the project were to <strong>in</strong>vestigate the biological reduction <strong>of</strong><br />
bromate <strong>and</strong> perchlorate <strong>in</strong> biologically active carbon (BAG) filters <strong>and</strong> to <strong>in</strong>vestigate the abiotic<br />
reduction <strong>of</strong> perchlorate <strong>in</strong> advanced oxidation processes <strong>in</strong>volv<strong>in</strong>g granular activated carbon (<strong>GAC</strong>).<br />
Specific project objectives were as follows:<br />
To show that perchlorate can be removed <strong>in</strong> BAG adsorbers under the same conditions that<br />
have proven effective for bromate removal, or to show how these conditions must be<br />
changed to obta<strong>in</strong> biological removal <strong>of</strong> perchlorate.<br />
+ To <strong>in</strong>vestigate the mechanism <strong>of</strong> the biological reduction process, with<br />
special focus on the role <strong>of</strong> nitrate-reduc<strong>in</strong>g organisms, the types <strong>of</strong><br />
microorganisms carry<strong>in</strong>g out the bromate/perchlorate reduction <strong>and</strong> their<br />
growth requirements, the type <strong>and</strong> concentration <strong>of</strong> electron donor, <strong>and</strong> the<br />
xix
effect <strong>of</strong> parameters such as bromate <strong>and</strong> perchlorate concentration, dissolved<br />
oxygen (DO) concentration, nitrate concentration, pH, temperature, <strong>and</strong> the<br />
role <strong>of</strong> ozonation preced<strong>in</strong>g BAG.<br />
To <strong>in</strong>vestigate advanced oxidation processes to determ<strong>in</strong>e whether these reactions can<br />
produce efficient removal <strong>of</strong> perchlorate, <strong>and</strong> if so, to develop the process at the bench-scale.<br />
> Emphasis will be placed on the ozone-hydrogen peroxide-<strong>GAC</strong> process,<br />
virg<strong>in</strong> <strong>GAC</strong>, <strong>GAC</strong> impregnated with metals such as copper <strong>and</strong> z<strong>in</strong>c, <strong>and</strong><br />
<strong>GAC</strong> mixed with solids such as iron, z<strong>in</strong>c, <strong>and</strong> alum<strong>in</strong>um.<br />
To optimize the bromate <strong>and</strong> perchlorate removal processes that show the most encourag<strong>in</strong>g<br />
results <strong>and</strong> to monitor the product water for potability.<br />
To test the most promis<strong>in</strong>g <strong>of</strong> the bromate <strong>and</strong> perchlorate removal processes at the pilot-<br />
scale.<br />
APPROACH<br />
The project <strong>in</strong>cluded bench-scale work conducted at the University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-<br />
Champaign <strong>and</strong> the Metropolitan Water District <strong>of</strong> Southern California (MWDSC). In general, one<br />
water quality or operational parameter was changed at a time, <strong>in</strong> order to observe its effect on<br />
bromate <strong>and</strong> perchlorate removal. Additionally, pilot-scale work for removal <strong>of</strong> bromate <strong>and</strong><br />
perchlorate us<strong>in</strong>g BAG filtration was completed at the La Verne pilot plant at MWDSC; these<br />
experiments were carried out us<strong>in</strong>g filter parameters that were tested at the bench-scale.<br />
CONCLUSIONS<br />
1. Increas<strong>in</strong>g the <strong>in</strong>fluent DO concentration to the BAG filters from 2.1 mg/L to 13.6<br />
mg/L decreased bromate removal from 40 percent to 11 percent at a 20-m<strong>in</strong>ute<br />
empty-bed-contact time (EBCT). The lack <strong>of</strong> bromate reduction observed <strong>in</strong> fullxx
scale ozone-BAC plants is likely due to the high concentrations <strong>of</strong> DO enter<strong>in</strong>g the<br />
filters.<br />
1. Filter history is very important to bromate removal <strong>in</strong> a BAG filter. For <strong>in</strong>stance, the<br />
negative effect <strong>of</strong> high DO concentrations on the bromate-reduc<strong>in</strong>g bacteria was not<br />
immediately reversible when the DO concentration is subsequently reduced.<br />
2. When the <strong>in</strong>fluent nitrate concentration to the BAG filters was <strong>in</strong>creased from 0.3 to<br />
42.3 mg/L, bromate removal decreased from 86 to 49 percent at a 26-m<strong>in</strong>ute emptybed<br />
contact time.<br />
3. The <strong>in</strong>fluent sulfate concentration to the BAG filters did not have a significant effect<br />
on bromate reduction.<br />
4. At a 25-m<strong>in</strong>ute EBCT, bromate removal <strong>in</strong>creased from 20 percent to 40 percent<br />
when the <strong>in</strong>fluent pH to the BAG filters decreased from 8.2 to near neutral pHs (6.8-<br />
7.5). This is beneficial s<strong>in</strong>ce pH control may be used to reduce bromate formation<br />
dur<strong>in</strong>g ozonation <strong>and</strong> <strong>in</strong>crease biological bromate removal s<strong>in</strong>ce bromate formation<br />
decreases as the pH decreases.<br />
5. Under similar operat<strong>in</strong>g conditions, better bromate removal was observed <strong>in</strong> BAG<br />
filters operated with Champaign-Urbana water (CUW, a groundwater source) than<br />
us<strong>in</strong>g Lake Michigan water (LMW). Thus, composition <strong>of</strong> the source water can have<br />
a dramatic impact on bromate reduction <strong>in</strong> a BAG filter.<br />
6. <strong>Bromate</strong> removal results obta<strong>in</strong>ed <strong>in</strong> a one-<strong>in</strong>ch <strong>in</strong>ner-diameter BAG filter were<br />
reproduced <strong>in</strong> a 2-<strong>in</strong>ch <strong>in</strong>ner-diameter BAG filter, suggest<strong>in</strong>g that scale-up <strong>of</strong> the<br />
bench-scale results should not be <strong>of</strong> concern.<br />
7. <strong>Perchlorate</strong> <strong>and</strong> bromate removal were observed <strong>in</strong> pilot-scale BAG filters.<br />
8. Backwash<strong>in</strong>g the BAG filters us<strong>in</strong>g filter effluent with a low DO concentration (2<br />
mg/L) <strong>and</strong> no chlor<strong>in</strong>e did not adversely affect bromate reduction.<br />
9. For lower bromate concentrations (10-20 [ig/L) <strong>and</strong> longer EBCTs (25-50 m<strong>in</strong>utes),<br />
bromate removal <strong>in</strong>creased as the <strong>in</strong>fluent bromate concentration <strong>in</strong>creased.<br />
However, constant mass concentration bromate removal was observed at higher<br />
<strong>in</strong>fluent bromate concentrations (20-50 ug/L) <strong>and</strong> shorter EBCTs (10-20 m<strong>in</strong>utes),<br />
which may <strong>in</strong>dicate that the electron donor concentration was limit<strong>in</strong>g bromate<br />
removal under these conditions.<br />
xxi
10. Eight bromate-reduc<strong>in</strong>g isolates were cultured from the BAG filter, <strong>and</strong> it is unlikely<br />
that any <strong>of</strong> them are denitrifiers.<br />
12. <strong>Perchlorate</strong> was ion exchanged, not reduced, on the surface <strong>of</strong> <strong>GAC</strong>. This <strong>in</strong>cludes<br />
a <strong>GAC</strong> bed as well as <strong>GAC</strong> mixed with alum<strong>in</strong>um shot <strong>and</strong> copper- <strong>and</strong> z<strong>in</strong>cimpregnated<br />
<strong>GAC</strong>.<br />
13. <strong>GAC</strong> used <strong>in</strong> this study had a low ion exchange capacity for perchlorate (0.172 mg<br />
perchlorate per gram <strong>GAC</strong>).<br />
14. <strong>Perchlorate</strong> was not chemically reduced <strong>in</strong> the presence <strong>of</strong> ozone or ozone/hydrogen<br />
peroxide.<br />
15. BAG filters demonstrated efficient biological reduction <strong>of</strong> low ng/L concentrations<br />
<strong>of</strong> perchlorate under anaerobic conditions with the addition <strong>of</strong> an external electron<br />
donor mixture consist<strong>in</strong>g <strong>of</strong> acetate, lactate, <strong>and</strong> pyruvate. The added electron donor<br />
concentration was approximately 1.5-1.8 mg/L as carbon.<br />
16. Biological perchlorate reduction was h<strong>in</strong>dered by the presence <strong>of</strong> nitrate. When<br />
effluent nitrate concentration <strong>in</strong>creased from 0.05 to 0.35 mg/L, perchlorate removal<br />
decreased from >95 to 65 percent.<br />
17. <strong>Perchlorate</strong> removal dur<strong>in</strong>g a zero <strong>in</strong>fluent nitrate experiment <strong>in</strong>creased from 60 to<br />
> 95 percent when the filter was preloaded with nitrate prior to the start <strong>of</strong> the<br />
experiment, as compared to a non-nitrate preloaded condition.<br />
18. The microbial communities present <strong>in</strong> perchlorate-reduc<strong>in</strong>g <strong>and</strong> bromate-reduc<strong>in</strong>g<br />
BAG filters are diverse.<br />
19. Us<strong>in</strong>g bench-scale BAG filters follow<strong>in</strong>g ozonation, complete bromate <strong>and</strong><br />
perchlorate removal was observed with 10-50 ug/L <strong>of</strong> <strong>in</strong>fluent bromate <strong>and</strong><br />
perchlorate, EBCTs <strong>of</strong> 25 <strong>and</strong> 15 m<strong>in</strong>utes, <strong>and</strong> <strong>in</strong>fluent DO concentrations <strong>of</strong> 0-2<br />
mg/L. No external electron donor was required to achieve this removal.<br />
20. Pilot scale treatment, consist<strong>in</strong>g <strong>of</strong> ozonation, conventional treatment, <strong>and</strong> BAG<br />
filtration achieved complete removal <strong>of</strong> 25 <strong>and</strong> 50 ng/L <strong>of</strong> <strong>in</strong>fluent bromate <strong>and</strong><br />
perchlorate, respectively, without the addition <strong>of</strong> an external electron donor.<br />
xxn
RECOMMENDATIONS<br />
1. To utilize BAG filtration for bromate <strong>and</strong> perchlorate reduction, a utility must<br />
determ<strong>in</strong>e if reduction <strong>of</strong> the <strong>in</strong>fluent DO concentration to the filter is feasible. Not<br />
only must it be economically feasible, but the utility should also determ<strong>in</strong>e that no<br />
unwanted effects are caused by DO reduction (i.e. lack <strong>of</strong> taste-<strong>and</strong>-odor removal or<br />
production <strong>of</strong> hydrogen sulfide <strong>in</strong> the filter). This could be accomplished through<br />
bench-scale test<strong>in</strong>g with the water <strong>and</strong> filter medium used by the utility.<br />
2. Bench-scale test<strong>in</strong>g at the utility should also be performed to determ<strong>in</strong>e the required<br />
contact time <strong>in</strong> the BAG filter to obta<strong>in</strong> the desired removal <strong>of</strong> bromate <strong>and</strong><br />
perchlorate. Contact times between 10 <strong>and</strong> 40 m<strong>in</strong>utes should be <strong>in</strong>vestigated.<br />
3 .Depend<strong>in</strong>g on the alkal<strong>in</strong>ity <strong>of</strong> the water, the utility should consider a comb<strong>in</strong>ation<br />
<strong>of</strong> reduction <strong>of</strong> bromate formation dur<strong>in</strong>g ozonation <strong>and</strong> an <strong>in</strong>crease <strong>in</strong> biological<br />
bromate removal <strong>in</strong> the BAC filter by pH control.<br />
4. Utilities that treat water contam<strong>in</strong>ated with both nitrate <strong>and</strong> perchlorate should give<br />
particular consideration to whether or not BAC filtration is a feasible treatment<br />
option. Research has shown that BAC filtration removes nitrate <strong>and</strong> perchlorate very<br />
efficiently. Additionally, nitrate reduction can enhance perchlorate reduction k<strong>in</strong>etics,<br />
mak<strong>in</strong>g BAC filtration particularly attractive for comb<strong>in</strong>ed nitrate-perchlorate<br />
remediation.<br />
FUTURE RESEARCH<br />
The most important modification to the ozone-BAC process for the application <strong>of</strong> bromate<br />
removal is the reduction <strong>of</strong> the <strong>in</strong>fluent DO concentration to the filter. Little or no bromate reduction<br />
has been observed at DO concentrations that are typical <strong>of</strong> the ozone-BAC process. Future work<br />
should address the best way to reduce the <strong>in</strong>fluent DO concentration. Additionally, studies should<br />
be conducted to ascerta<strong>in</strong> the effects <strong>of</strong> a reduced DO concentration on effluent water quality. For<br />
example, a reduced DO concentration to the BAC filter may affect the biodegradation <strong>of</strong> trace<br />
compounds such as 2-methylisoborneol or atraz<strong>in</strong>e.<br />
In order to truly underst<strong>and</strong> how best to optimize the reduction <strong>of</strong> bromate or perchlorate <strong>in</strong><br />
xxiii
a BAG filter, <strong>in</strong>dividual bromate- <strong>and</strong> perchlorate-reduc<strong>in</strong>g cultures must cont<strong>in</strong>ue to be studied. By<br />
underst<strong>and</strong><strong>in</strong>g the physiology <strong>and</strong> nutrient requirements <strong>of</strong> these organisms, the BAG process may<br />
be designed to select for these organisms. Relatedly, the seed<strong>in</strong>g <strong>of</strong> particularly efficient bromate-or<br />
perchlorate-reduc<strong>in</strong>g organisms to the BAG filters could be tested, <strong>and</strong> the retention <strong>of</strong> these<br />
organisms after backwash<strong>in</strong>g should be monitored.<br />
Research to date has <strong>in</strong>dicated that the addition <strong>of</strong> an external, organic electron donor is<br />
important for the stimulation <strong>of</strong> efficient, biological perchlorate reduction <strong>in</strong> a BAG filter. Though<br />
studies done at Metropolitan <strong>in</strong>dicate that ozonated natural organic matter may satisfy electron donor<br />
requirements for biological perchlorate reduction, these results are prelim<strong>in</strong>ary. This pre-filtration<br />
ozonation step needs to be further <strong>in</strong>vestigated, both at the bench- <strong>and</strong> full-scales, to ensure that the<br />
electron donor capability <strong>of</strong> ozonated natural organic matter will suffice to achieve biological<br />
perchlorate reduction under various operat<strong>in</strong>g conditions.<br />
Studies must also be undertaken that focus on treatment requirements after water has been<br />
<strong>in</strong> contact with a BAG filter. The production <strong>of</strong> any undesirable taste <strong>and</strong> odor compounds, the<br />
biological stability <strong>and</strong> chlor<strong>in</strong>e dem<strong>and</strong> <strong>of</strong> the product water, <strong>and</strong> reaeration requirements must each<br />
be addressed to ensure the potability <strong>of</strong> BAC-filtered water.<br />
xxiv
CHAPTER 1. INTRODUCTION<br />
BACKGROUND - BROMATE<br />
<strong>Bromate</strong> is a dis<strong>in</strong>fection by-product result<strong>in</strong>g from the ozonation <strong>of</strong> waters conta<strong>in</strong><strong>in</strong>g<br />
bromide. <strong>Bromate</strong> formation dur<strong>in</strong>g ozonation is affected by various factors <strong>in</strong>clud<strong>in</strong>g pH, <strong>and</strong> the<br />
concentrations <strong>of</strong> bromide, ammonia, natural organic matter (NOM), <strong>and</strong> ozone (von Gunten <strong>and</strong><br />
Hoigne 1996, von Gunten <strong>and</strong> Hoigne 1994, Westerh<strong>of</strong>f et al. 1993, Haag <strong>and</strong> Hoigne 1983).<br />
<strong>Bromate</strong> formation has become an important issue as more utilities are consider<strong>in</strong>g ozone to combat<br />
Cryptosporidium <strong>and</strong> Giardia, color, <strong>and</strong> taste <strong>and</strong> odor problems.<br />
Based on the work <strong>of</strong> Kurokawa et al. (1990) <strong>and</strong> DeAngelo et al. (1998), it was determ<strong>in</strong>ed<br />
that the <strong>in</strong>gestion <strong>of</strong> aqueous bromate caused renal cell tumors <strong>in</strong> rats. For a 70-kg adult dr<strong>in</strong>k<strong>in</strong>g<br />
2 liters <strong>of</strong> water per day with a bromate concentration <strong>of</strong> 5 ng/L, the United States Environmental<br />
Protection Agency (USEP A1998) calculated a lifetime cancer risk <strong>of</strong> 10"4; a 10"4 cancer risk <strong>in</strong>dicates<br />
that 1 person per 10,000 people would develop cancer. The maximum contam<strong>in</strong>ant level (MCL) for<br />
bromate <strong>in</strong> the United States has recently been set at 10 ng/L by the Stage 1<br />
Dis<strong>in</strong>fectants/Dis<strong>in</strong>fection By-Products Rule (Waterweek 1998), based on a practical quantification<br />
limit. Therefore, as analytical capabilities advance, a future, lower MCL may be promulgated.<br />
<strong>Bromate</strong> formation is highly dependent on the bromide concentration. The presence <strong>of</strong><br />
bromide results from natural <strong>and</strong> anthropogenic sources, <strong>in</strong>clud<strong>in</strong>g seawater <strong>in</strong>trusion <strong>and</strong> <strong>in</strong>dustrial<br />
discharges. In a survey <strong>of</strong> 100 U. S. dr<strong>in</strong>k<strong>in</strong>g water utilities, Amy et al. (1993) found bromide<br />
concentrations rang<strong>in</strong>g from 2 - 429 ug/L. Krasner et al. (1993) summarized the f<strong>in</strong>d<strong>in</strong>gs <strong>of</strong> several<br />
pilot- <strong>and</strong> full-scale studies that showed bromate concentrations <strong>in</strong> f<strong>in</strong>ished water rang<strong>in</strong>g from 0-60<br />
Hg/L, depend<strong>in</strong>g on the conditions <strong>and</strong> the amount <strong>of</strong> bromide <strong>in</strong> the source water. Thus, bromate<br />
formation <strong>in</strong> excess <strong>of</strong> the 10 \ig/L st<strong>and</strong>ard is a concern for some utilities.<br />
BACKGROUND-PERCHLORATE<br />
In April 1997, the limit <strong>of</strong> detection (LOD) for perchlorate decreased from 400 ng/L to 4<br />
\igfL (Waterweek 1997). S<strong>in</strong>ce that time, perchlorate concentrations rang<strong>in</strong>g from s<strong>in</strong>gle digits to<br />
hundreds <strong>of</strong> ng/L have been found <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water supplies <strong>in</strong> California, Nevada, <strong>and</strong> Arizona.<br />
1
Ammonium perchlorate is an oxidizer that helps rocket fuel ignite <strong>and</strong> burn, <strong>and</strong> it is <strong>of</strong>ten a<br />
component <strong>of</strong> munitions. S<strong>in</strong>ce ammonium perchlorate has a limited shelf life, it is frequently washed<br />
out <strong>of</strong> the nation's solid rocket supply. Contam<strong>in</strong>ation is likely the result <strong>of</strong> the large volumes <strong>of</strong><br />
perchlorate-conta<strong>in</strong><strong>in</strong>g water generated dur<strong>in</strong>g this process. <strong>Perchlorate</strong> contam<strong>in</strong>ation could also<br />
be due to Wastes produced by the munitions <strong>in</strong>dustry or from areas subjected to substantial use <strong>of</strong><br />
explosives, such as m<strong>in</strong>es. The occurrence <strong>of</strong> perchlorate <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water supplies is now be<strong>in</strong>g<br />
documented. Extensive perchlorate contam<strong>in</strong>ation has been found <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water supplies <strong>in</strong> the<br />
western U.S. For example, perchlorate concentrations <strong>of</strong> greater than 100 ug/L were detected <strong>in</strong> the<br />
San Gabriel Bas<strong>in</strong> <strong>in</strong> Southern California, one <strong>of</strong> the largest groundwater bas<strong>in</strong>s <strong>in</strong> the United States<br />
(Najm et al. 1999). The state <strong>of</strong> California has closed 18 dr<strong>in</strong>k<strong>in</strong>g water wells due to perchlorate<br />
contam<strong>in</strong>ation up to 280 ug/L (Waterweek 1997). <strong>Perchlorate</strong> has also been detected <strong>in</strong> Lake Mead,<br />
one portion <strong>of</strong> the Lower Colorado River system, which supplies dr<strong>in</strong>k<strong>in</strong>g water to 10 million people<br />
<strong>in</strong> Nevada, California, <strong>and</strong> Arizona. It is thought that most <strong>of</strong> the Lake Mead contam<strong>in</strong>ation is from<br />
the Las Vegas Wash, one <strong>of</strong> the lake's source waters, <strong>in</strong> which perchlorate concentrations <strong>of</strong> more<br />
than 1 mg/L have been measured (Rogers 1997). Ten states have been found to have perchloratecontam<strong>in</strong>ated<br />
dr<strong>in</strong>k<strong>in</strong>g water supplies (Betts 1998), <strong>and</strong> this number is likely to <strong>in</strong>crease, as only six<br />
states do not conta<strong>in</strong> users or manufacturers <strong>of</strong> perchlorate (Renner 1999). More complete data will<br />
be collected <strong>in</strong> a survey funded by the American Water Works Association Research Foundation<br />
(AWWARF) that targets both surface <strong>and</strong> groundwater supplies <strong>in</strong> the United States.<br />
<strong>Perchlorate</strong> is <strong>of</strong> concern due to its ability to disrupt the thyroid gl<strong>and</strong>'s use <strong>of</strong> iod<strong>in</strong>e <strong>in</strong> the<br />
generation <strong>of</strong> metabolic hormones. One step <strong>in</strong> the production <strong>of</strong> thyroid hormones is the uptake <strong>and</strong><br />
transport <strong>of</strong> iodide (I") with<strong>in</strong> the thyroid gl<strong>and</strong>. The iodide ion must be present for thyroid hormone<br />
production to occur. The thyroid gl<strong>and</strong> can mistake perchlorate for iodide because perchlorate has<br />
a similar molecular volume <strong>and</strong> charge to iodide. When perchlorate is taken up by the thyroid gl<strong>and</strong><br />
<strong>in</strong>stead <strong>of</strong> iodide, the thyroid gl<strong>and</strong> loses some hormone production capacity (Book 1999).<br />
No state or federal st<strong>and</strong>ard for perchlorate currently exists <strong>in</strong> the U.S. s<strong>in</strong>ce perchlorate<br />
contam<strong>in</strong>ation has only recently come to the forefront. However, based on a 1992 health study<br />
conducted by the USEPA, the California Department <strong>of</strong> Health Services (CDHS) established a<br />
provisional perchlorate action level <strong>of</strong> 18 ng/L (CDHS 1997). However, the national action level is<br />
32 ug/L, as set by the USEPA (USEPA 1998).
PREVIOUS STUDIES - BROMATE 1<br />
Siddiqui et al. (1994) observed bromate reduction by UV light (180 - 254 nm), <strong>and</strong> Mills et<br />
al. (1996a) found that the use <strong>of</strong> a plat<strong>in</strong>ised titanium dioxide catalyst enhanced the rate <strong>of</strong> bromate<br />
reduction by UV light (254 nm). However, UV irradiation for bromate reduction is currently, not<br />
economically viable (Siddiqui et al. 1994).<br />
<strong>Bromate</strong> reduction by granular activated carbon (<strong>GAC</strong>) has been demonstrated (Kirisits et al.<br />
1998, Miller 1996, Mills et al. 1996b, Siddiqui et al. 1994, 1996). <strong>Bromate</strong> removal by <strong>GAC</strong> is<br />
h<strong>in</strong>dered <strong>in</strong> natural water by the presence <strong>of</strong> NOM <strong>and</strong> anions such as chloride, nitrate, <strong>and</strong> sulfate.<br />
Us<strong>in</strong>g a <strong>GAC</strong> filter, Kirisits et al. ( 1 998) observed 50 percent bromate breakthrough after 4 1 ,000 bed<br />
volumes for distilled deionized water (DDW) as compared to less than 1,000 bed volumes for natural<br />
water. It was concluded that abiotic reduction <strong>of</strong> bromate by <strong>GAC</strong> could not be feasibly susta<strong>in</strong>ed<br />
for a significant period <strong>of</strong> time <strong>in</strong> natural water.<br />
Us<strong>in</strong>g half reactions (Equations 1.1-1.3), reduction potentials were calculated for bromate,<br />
nitrate, <strong>and</strong> oxygen. The follow<strong>in</strong>g concentrations, typical for the current study, were used <strong>in</strong> the<br />
calculations: 20 ug/L BrO/, 20 ug/L Br', 5 mg/L NO/, 2 mg/L O2, pH 7.5, <strong>and</strong> the partial pressure<br />
<strong>of</strong> nitrogen (/Vz) was assumed to be 1 s<strong>in</strong>ce the water was sparged with nitrogen gas to remove<br />
oxygen.<br />
BrO~ + 6H+ + 6e~ -» Br' + 3H2O; E = 0.996 V (1.1)<br />
O2 + 4/T + 4e~ -» 2H2 O; E = 0.767 V (1.2)<br />
2NO3 ~ + 12H+ + 10e- -> N2 + 6H2 O; E = 0.665 V (1.3)<br />
<strong>Bromate</strong> is a stronger oxidant than both oxygen <strong>and</strong> nitrate as shown by the calculated<br />
reduction potentials. Although bromate has a higher reduction potential than oxygen <strong>and</strong> nitrate,<br />
reaction k<strong>in</strong>etics must also be considered. Enzymes strongly affect the k<strong>in</strong>etics <strong>of</strong> electron acceptor<br />
utilization. <strong>Bromate</strong> reduction can be catalyzed by sp<strong>in</strong>ach nitrate reductase, an enzyme that catalyzes<br />
the reduction <strong>of</strong> nitrate (Barber <strong>and</strong> Notton 1990). If denitrify<strong>in</strong>g organisms also have the capability<br />
to reduce bromate us<strong>in</strong>g nitrate reductase, the nitrate concentration will affect bromate removal.<br />
1 Adapted from Journal AWWA, 91(8), 74-84.<br />
3
The formation <strong>and</strong> activity <strong>of</strong> nitrate reductase are also affected by oxygen concentration. For<br />
example, approximately 30 uM oxygen (0.96 mg/L) was found to <strong>in</strong>hibit the formation <strong>of</strong> nitrate<br />
reductase <strong>in</strong> Hyphomicrobium (Meiberg et al. 1980) while oxygen concentrations <strong>in</strong> the vic<strong>in</strong>ity <strong>of</strong><br />
10 uM (0.32 mg/L) <strong>in</strong>hibited the activity <strong>of</strong> the enzyme (Tiedje 1988). If the oxygen concentration<br />
is too high, this enzyme may not be available for nitrate or bromate reduction. If conditions permit<br />
nitrate reductase activity, then nitrate <strong>and</strong> bromate could be compet<strong>in</strong>g for use as electron acceptors.<br />
On the other h<strong>and</strong>, more recent work has demonstrated the occurrence <strong>of</strong> denitrification under<br />
aerobic conditions (Robertson <strong>and</strong> Kuenen 1990, Robertson <strong>and</strong> Kuenen 1984). For example,<br />
aerobically grown Thiosphaerapantotropha, when placed <strong>in</strong> an anoxic environment, were able to use<br />
nitrate immediately; the lack <strong>of</strong> a lag time prior to nitrate use <strong>in</strong>dicated that the enzymes <strong>in</strong>volved <strong>in</strong><br />
denitrification were already expressed <strong>and</strong> active (Robertson <strong>and</strong> Kuenen 1984). Furthermore, these<br />
<strong>in</strong>vestigators found that Thiosphaerapantotropha could use oxygen <strong>and</strong> nitrate simultaneously. Thus,<br />
complete removal <strong>of</strong> oxygen may not be necessary for bromate reduction. Moreover, the existence<br />
<strong>of</strong> oxygen <strong>and</strong> nitrate gradients with<strong>in</strong> the bi<strong>of</strong>ilm may permit the use <strong>of</strong> bromate as an electron<br />
acceptor. DO <strong>and</strong> nitrate gradients have been found <strong>in</strong> microbial floes (V<strong>and</strong>enabeele et al. 1995)<br />
For example, these <strong>in</strong>vestigators found that the DO concentration dropped from 0.25 mM (8 mg/L)<br />
<strong>in</strong> the bulk to 0.1 mM (3.2 mg/L) <strong>in</strong> the center <strong>of</strong> the floe (=650 um = 0.03 <strong>in</strong>); the nitrate<br />
concentration dropped from 60.1 mg/L <strong>in</strong> the bulk to 50.8 mg/L approximately 600 um (0.02 <strong>in</strong>)<br />
<strong>in</strong>to the floe. If the DO concentration reaches an appropriate level for the microorganisms <strong>in</strong> the<br />
bi<strong>of</strong>ilm <strong>of</strong> the BAG filters to ma<strong>in</strong>ta<strong>in</strong> active nitrate reductase, then it may be possible that nitrate <strong>and</strong><br />
bromate can be reduced <strong>in</strong> this region.<br />
Hijnen et al. (1995) demonstrated batch microbial bromate reduction <strong>in</strong> the absence <strong>of</strong> oxygen<br />
<strong>and</strong> nitrate. However, a more recent study by Hijnen et al. (1999) showed bromate reduction <strong>in</strong> the<br />
presence <strong>of</strong> nitrate us<strong>in</strong>g a denitrify<strong>in</strong>g bioreactor supplemented with ethanol. The authors suggested<br />
that the bromate reduction was tak<strong>in</strong>g place <strong>in</strong> a portion <strong>of</strong> the bi<strong>of</strong>ilm where nitrate had already been<br />
depleted, although they did not discount the possibility <strong>of</strong> biological bromate reduction below a<br />
threshold nitrate concentration. Hijnen et al. (1999) concluded that the use <strong>of</strong> a denitrify<strong>in</strong>g<br />
bioreactor for bromate reduction would not be practical due to the extended contact times required<br />
for bromate removal to 0.5 ug/L (proposed Dutch st<strong>and</strong>ard) <strong>and</strong> the post-treatment that would be<br />
required to elim<strong>in</strong>ate the excess ethanol <strong>and</strong> biomass. It should be noted that the <strong>in</strong>fluent water used<br />
for this study had a very high nitrate concentration <strong>of</strong> 85 mg/L, <strong>and</strong> it is not surpris<strong>in</strong>g that 50 mg/L
ethanol was required to stimulate enough biological growth to remove the nitrate. However, on the<br />
basis <strong>of</strong> the study by Hijnen et al., one should not conclude that biological bromate reduction is<br />
<strong>in</strong>feasible for a water that has to meet the U.S. st<strong>and</strong>ard for bromate (10 ng/L), especially s<strong>in</strong>ce the<br />
<strong>in</strong>fluent nitrate concentration will likely be significantly less than 85 mg/L.<br />
Experiments concern<strong>in</strong>g the reduction <strong>of</strong> bromate <strong>in</strong> biologically active carbon (BAG) filters<br />
have been completed at the University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Champaign <strong>and</strong> have shown excellent<br />
bromate removal <strong>in</strong> the presence <strong>of</strong> an effluent nitrate concentration between 2 <strong>and</strong> 8 mg/L. These<br />
results were published by Kirisits <strong>and</strong> Snoey<strong>in</strong>k (1999), <strong>and</strong> are discussed <strong>in</strong> the follow<strong>in</strong>g sections.<br />
<strong>Bromate</strong> Reduction with Various Influent Dissolved Oxygen Concentrations<br />
A set <strong>of</strong> BAG filtration experiments were run with different <strong>in</strong>fluent DO concentrations,<br />
rang<strong>in</strong>g from 2.0 to 8.0 mg/L. Danville filter effluent (DFE), obta<strong>in</strong>ed from the Consumer's Ill<strong>in</strong>ois<br />
Treatment Plant (Danville, Ill<strong>in</strong>ois), was used <strong>in</strong> conjunction with an empty bed contact time (EBCT)<br />
<strong>of</strong> approximately 50 m<strong>in</strong>utes <strong>in</strong> the BAG filter. S<strong>in</strong>ce nitrate <strong>and</strong> bromate may compete for use as<br />
electron acceptors, it is noteworthy that the batch <strong>of</strong> DFE used for these experiments conta<strong>in</strong>ed<br />
approximately 25 mg/L nitrate. The pert<strong>in</strong>ent experimental <strong>in</strong>formation is shown <strong>in</strong> Table 1.1.<br />
Table 1.1 <strong>Bromate</strong> removal <strong>in</strong> DFE at pH 6.5 with various DO concentrations<br />
Influent<br />
# EBCT (m<strong>in</strong>) DO<br />
(mg/L) NO3-<br />
(mg/L) BrO3-<br />
("g/L)<br />
1 50 4.5 24.9 10<br />
2 49 2.0 24.9 11<br />
3 50 8.0 27.8 11<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
Effluent<br />
DO<br />
N03"<br />
(mg/L) (mg/L)<br />
_2<br />
_2<br />
24.6<br />
24.2<br />
4.9 26.7<br />
% BrO3-<br />
removal<br />
56<br />
61<br />
42<br />
95%<br />
confidence<br />
<strong>in</strong>terval<br />
50-62<br />
54-68<br />
38-46<br />
Data not available
The bromate breakthrough data for these experiments are shown <strong>in</strong> Figure 1.1. Note that at<br />
the start <strong>of</strong> the experiment, bromate breakthrough was 70 percent (C/C0 = 0.7) but then gradually<br />
decreased. This is considered to be a period <strong>of</strong> acclimation for the microorganisms, <strong>in</strong> which the<br />
microbial communities adjusted to the 50-m<strong>in</strong>ute EBCT <strong>and</strong> the DO concentration <strong>of</strong> 4.5 mg/L. Prior<br />
to this experiment, the BAC filter had been operated with an 8-m<strong>in</strong>ute EBCT <strong>and</strong> an <strong>in</strong>fluent DO<br />
concentration <strong>of</strong> approximately 8.0 mg/L.<br />
The data <strong>in</strong> Table 1.1 show 42 percent bromate removal <strong>in</strong> the presence <strong>of</strong> DO concentrations<br />
<strong>of</strong> 4.9 to 8.0 mg/L (Experiment 3). Increased bromate removal was observed as the <strong>in</strong>fluent DO<br />
concentration decreased from 8.0 mg/L to 4.5 mg/L (Experiments 3 <strong>and</strong> 1); the average percent<br />
bromate removals for these two experiments are different at the 95 percent confidence level s<strong>in</strong>ce<br />
their confidence <strong>in</strong>tervals do not overlap (Table 1.1). The trend <strong>of</strong> improved bromate removal at<br />
lower DO concentrations was expected s<strong>in</strong>ce oxygen may <strong>in</strong>hibit the activity <strong>of</strong> the enzyme (possibly<br />
nitrate reductase) that catalyzes bromate reduction. However, as the <strong>in</strong>fluent DO concentration<br />
decreased from 4.5 to 2.0 mg/L (Experiments 1 <strong>and</strong> 2), the average percent bromate removal did not<br />
<strong>in</strong>crease <strong>in</strong> a statistically significant manner. Table 1.1 shows that the 95 percent confidence <strong>in</strong>tervals<br />
overlap for these experiments. Although the effluent DO concentrations from Experiments 1 <strong>and</strong> 2<br />
GO<br />
c<br />
o<br />
1<br />
U<br />
i.u -<br />
0.8 -<br />
0.6 -<br />
0.4 -<br />
0.2 -<br />
n n -<br />
0<br />
•<br />
• •<br />
4.5 mg/L DO<br />
r •• a .<br />
56% removal<br />
—————— ... i —— _..,„,<br />
2.0 mg/L DO<br />
mm —————•<br />
• • •<br />
• i'. •<br />
61% removal<br />
8.0 mg/L DO<br />
• . ' •<br />
TT. . m . *-<br />
•<br />
, 42% removal<br />
j —.,_. ..... ———————<br />
200 400 600<br />
800 1000<br />
Bed Volumes<br />
Figure 1.1 Effect <strong>of</strong> <strong>in</strong>fluent dissolved oxygen concentration on bromate breakthrough from the<br />
BAC filter (<strong>in</strong>itial conditions: DFE, EBCT «50 m<strong>in</strong>utes, pH 6.5, 10-11 ng/L BrO3")<br />
(Adapted from Journal AWWA, 91(8), 74-84.)
were not measured, they can be presumed to be relatively low s<strong>in</strong>ce 3.1 mg/L DO was consumed <strong>in</strong><br />
Experiment 3. Thus, there may already be a region <strong>of</strong> considerable size <strong>in</strong> the BAG filter <strong>of</strong><br />
Experiment 1 where the DO is quite low, <strong>and</strong> bromate removal can take place. Decreas<strong>in</strong>g the<br />
<strong>in</strong>fluent DO concentration further <strong>in</strong> Experiment 2 does not result <strong>in</strong> additional bromate removal.<br />
The nitrate concentration <strong>in</strong> these experiments was quite high (approximately 25 mg/L) <strong>and</strong> probably<br />
was a major factor affect<strong>in</strong>g bromate removal. Such an effect is expected if nitrate reductase is the<br />
enzyme responsible for bromate reduction. Only low concentrations <strong>of</strong> nitrate (0.3 mg/L -1.1 mg/L)<br />
were removed <strong>in</strong> the BAG filter. However, even with these relatively high concentrations <strong>of</strong> nitrate<br />
<strong>and</strong> a pH <strong>of</strong> 6.5, which was probably less than optimum for denitrification, bromate removals <strong>of</strong> 42 -<br />
61 percent were achieved.<br />
It is likely that microorganisms <strong>in</strong> the early reaches <strong>of</strong> the filter reduce the DO concentration.<br />
Therefore, organisms deeper with<strong>in</strong> the filter are exposed to lower or negligible DO, <strong>and</strong> nitrate<br />
reductase synthesis or activity could occur here unimpeded by the presence <strong>of</strong> DO. Due to<br />
diffusional transport, there will also be a gradient <strong>of</strong> nitrate <strong>and</strong> bromate <strong>in</strong> the bi<strong>of</strong>ilm (Figure 1.2).<br />
As suggested by Hijnen et al. (1999), denitrify<strong>in</strong>g microorganisms may possess the ability to reduce<br />
bromate below a threshold nitrate concentration, <strong>and</strong> the nitrate concentration <strong>in</strong> a portion <strong>of</strong> the<br />
bi<strong>of</strong>ilm may reach a low enough concentration for bromate to be used as an electron acceptor.<br />
Bulk Solution Conta<strong>in</strong><strong>in</strong>g<br />
<strong>Bromate</strong>, Nitrate, <strong>and</strong> DO<br />
<strong>Bromate</strong><br />
<strong>GAC</strong> BIOFILM<br />
Figure 1.2 Possible depletion <strong>of</strong> bromate, nitrate, <strong>and</strong> dissolved oxygen <strong>in</strong> the bi<strong>of</strong>ilm<br />
(Adapted from Journal AWWA, 91(8), 74-84.)
<strong>Bromate</strong> Reduction to Bromide<br />
S<strong>in</strong>ce the previous section demonstrated the removal <strong>of</strong> bromate <strong>in</strong> the BAG filter, it was then<br />
desirable to determ<strong>in</strong>e if bromate was <strong>in</strong>deed reduced to bromide. To this end, a BAG filter with a<br />
20-m<strong>in</strong>ute EBCT was operated with DDW spiked with an electron donor mixture consist<strong>in</strong>g <strong>of</strong><br />
lactate, acetate, <strong>and</strong> pyruvate to ma<strong>in</strong>ta<strong>in</strong> the biological activity. This experiment was performed<br />
<strong>in</strong> DDW so that small changes <strong>in</strong> bromate <strong>and</strong> bromide would be easier to detect analytically. More<br />
than half the data <strong>in</strong> Figure 1.3 show that less bromide was observed than expected on the basis <strong>of</strong><br />
bromate removal. In fact, the data po<strong>in</strong>ts show<strong>in</strong>g bromate removed <strong>and</strong> bromide produced were up<br />
to 3.5 ng/L apart. In k<strong>in</strong>etic studies with GAG <strong>and</strong> bromate, a mass balance on Br showed that the<br />
<strong>in</strong>itial BrO3"-Br <strong>in</strong> the water was equivalent to the BrO3"-Br <strong>and</strong> Br" <strong>in</strong> solution at the end <strong>of</strong> the test,<br />
plus the Br" displaced by a sulfate wash <strong>of</strong> the GAG at the end <strong>of</strong> the experiment (Miller 1996).<br />
However, s<strong>in</strong>ce the BAG filter <strong>in</strong> this study had been cont<strong>in</strong>uously treat<strong>in</strong>g bromate <strong>in</strong> water for over<br />
one year prior to this experiment, the probability <strong>of</strong> bromate or bromide adsorb<strong>in</strong>g to the carbon is<br />
3<br />
1<br />
2<br />
CQ<br />
10<br />
--B-<strong>Bromate</strong> removed<br />
—•-Bromide produced<br />
0<br />
0<br />
—i—<br />
50 100 150<br />
200 250<br />
Bed Volumes<br />
Figure 1.3 Reduction <strong>of</strong> bromate to bromide <strong>in</strong> a BAG filter (<strong>in</strong>itial conditions: DDW, EBCT =<br />
20 m<strong>in</strong>utes, pH 7.5, 1.3 mg/LDO, 2.4 mg/LNO3", 18 ug/LBrO 3")<br />
(Adapted from Journal AWWA, 91(8), 74-84.)
one year prior to this experiment, the probability <strong>of</strong> bromate or bromide adsorb<strong>in</strong>g to the carbon is<br />
quite low. It has been shown that the bromate removal capacity <strong>of</strong> <strong>GAC</strong> <strong>in</strong> the absence <strong>of</strong> biological<br />
activity is exhausted very quickly <strong>in</strong> natural water (Kirisits et al. 1998). Therefore, the lack <strong>of</strong> a<br />
balance on Br <strong>in</strong> the BAG filter may <strong>in</strong>dicate the presence <strong>of</strong> other reduced species that are<br />
<strong>in</strong>termediates <strong>in</strong> the reduction <strong>of</strong> bromate to bromide. Alternatively, Br may be associated with or<br />
adsorbed on the microbes, thereby prevent<strong>in</strong>g a mass balance on Br.<br />
<strong>Bromate</strong> Reduction with Various Influent Nitrate Concentrations<br />
Experiments 4-7 (Table 1.2) employed Champaign-Urbana water (CUW) to show the effect<br />
<strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate removal. S<strong>in</strong>ce the results <strong>of</strong> Experiments 1-3 showed<br />
better bromate removal with lower <strong>in</strong>fluent DO concentrations, Experiments 4-7 were operated with<br />
an <strong>in</strong>fluent DO concentration <strong>of</strong> approximately 2 mg/L. The effluent DO concentration from the<br />
filters was typically 0.1- 0.2 mg/L, although some <strong>of</strong> this oxygen may have been <strong>in</strong>troduced dur<strong>in</strong>g<br />
the measurement. The bromate removal data from Experiments 4 <strong>and</strong> 5 are shown <strong>in</strong> Figure 1.4. The<br />
bromate removal results with the 25-m<strong>in</strong>ute EBCT were particularly encourag<strong>in</strong>g because they<br />
<strong>in</strong>dicated that excellent bromate removal can be obta<strong>in</strong>ed at EBCTs that could be used <strong>in</strong><br />
conventional ozone/<strong>GAC</strong> plants. Typically, BAG filters are operated with EBCTs <strong>in</strong> the range <strong>of</strong> 5<br />
to 20 m<strong>in</strong>utes (Bishop <strong>and</strong> Bourbigot 1996).<br />
As the <strong>in</strong>fluent nitrate concentration <strong>in</strong>creased from 0.2 mg/L to 5.1 mg/L with a 25-m<strong>in</strong>ute<br />
EBCT, the bromate removal decreased from 86 percent to 76 percent (Table 1.2, Figure 1.4).<br />
However, when the EBCT was <strong>in</strong>creased to 50 m<strong>in</strong>utes, the bromate removal decreased<br />
<strong>in</strong>significantly from 95 percent to 93 percent for the same change <strong>in</strong> nitrate concentration (Table 1.2).<br />
Less bromate reduction is generally expected as the nitrate concentration is <strong>in</strong>creased s<strong>in</strong>ce bromate<br />
<strong>and</strong> nitrate may be compet<strong>in</strong>g electron acceptors. However, the <strong>in</strong>creased contact time <strong>in</strong> Experiment<br />
7 was able to counteract the negative effect <strong>of</strong> the elevated nitrate concentration, allow<strong>in</strong>g bromate<br />
to be removed to the detection limit. The importance <strong>of</strong> the effect <strong>of</strong> nitrate concentration on bromate<br />
removal <strong>in</strong>dicated that the nitrogen relationships <strong>in</strong> the filters must be more closely exam<strong>in</strong>ed. The<br />
effluent nitrate concentrations for Experiments 4-7 were always higher than the <strong>in</strong>fluent<br />
concentrations, <strong>in</strong>dicat<strong>in</strong>g that nitrify<strong>in</strong>g organisms were present. CUW typically conta<strong>in</strong>ed 0.3-0.8<br />
mg/L NH3-N, <strong>and</strong> some <strong>of</strong> this was converted to nitrite <strong>and</strong> nitrate, thereby reduc<strong>in</strong>g the amount <strong>of</strong>
Table 1.2 <strong>Bromate</strong> removal <strong>in</strong> CUW at pH 7.5 with different <strong>in</strong>fluent nitrate concentrations<br />
Influent Effluent<br />
# EBCT DO NO3- BrO3" DO NO3' % BrO3'<br />
(m<strong>in</strong>) (mg/L) (mg/L) (ug/L) (mg/L) (mg/L) removal<br />
4 25 2.2 0.2 21 0.1<br />
5 25 2.3 5.1 20 0.1<br />
0.9 86<br />
5.7 76<br />
6 49 2.2 0.2 21 0.2 2.7 95<br />
7 50 2.3 5.1 20 0.2 8.1 93<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
1 0 -|<br />
ao<br />
_c<br />
•a o.s -<br />
1<br />
3 •s uo 0.6 -<br />
-•-5. 1 mg/L <strong>in</strong>fluent nitrate (Expt. 5)<br />
— E3-0.2 mg/L <strong>in</strong>fluent nitrate (Expt. 4)<br />
1 Q 0.4 -<br />
u<br />
m<br />
1<br />
1<br />
—— m ———— 76% removal<br />
-2 0.2 - • B -» Q_ ^-,<br />
8<br />
LJ 0 <br />
u 86% removal<br />
[•y , /•> /\<br />
D D<br />
U.U i i i i i<br />
i i<br />
0 200 400 600 800<br />
Bed Volumes<br />
1000 1200 1400<br />
Figure 1.4 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate breakthrough from a BAG filter<br />
(<strong>in</strong>itial conditions: CUW, EBCT - 25 m<strong>in</strong>utes, pH 7.5, 2.3 mg/L DO, 21 pg/L BrO3")<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
DO <strong>in</strong> the filters. Figures 1.5 <strong>and</strong> 1.6 show nitrogen mass balances, <strong>in</strong>clud<strong>in</strong>g ammonia, nitrite, <strong>and</strong><br />
nitrate, for Experiments 4-7. A few <strong>of</strong> the po<strong>in</strong>ts show that the effluent nitrogen is greater than the<br />
<strong>in</strong>fluent nitrogen by less than 0.1 mg/L as N. This is with<strong>in</strong> the experimental error <strong>of</strong> the analytical<br />
methods used. In general, the <strong>in</strong>fluent nitrogen values agreed quite well with effluent nitrogen values,<br />
except for Experiment 6 (Figure 1.6). S<strong>in</strong>ce effluent nitrogen was less than <strong>in</strong>fluent nitrogen for this<br />
10
2.0 -<br />
e •<br />
• 8<br />
i 8<br />
a
i.s-<br />
2.0 -<br />
S i- D<br />
sg> i.o -<br />
£<br />
0.5 -<br />
n n -<br />
'.8 . 8 «<br />
-<br />
B<br />
• •<br />
^^ n D<br />
H<br />
H u<br />
n n g<br />
• Influent Nitrogen (Expt. 6) D Effluent Nitrogen (Expt. 6)<br />
• Influent Nitrogen (Expt. 7) O Effluent Nitrogen (Expt. 7)<br />
0 100 200 300<br />
Bed Volumes<br />
400 500<br />
Figure 1.6 Nitrogen mass balance <strong>in</strong> experiments 6 <strong>and</strong> 7 (49-50-m<strong>in</strong>ute BAG filter) - ammonia,<br />
nitrite, <strong>and</strong> nitrate<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
00<br />
fi<br />
I.O<br />
1.4 -<br />
1.2 -<br />
1.0 -<br />
0.8 -<br />
0.6 -<br />
0.4 -<br />
0.2 -<br />
nn<br />
• Influent Nitrate<br />
o Effluent Nitrite<br />
X<br />
X<br />
a 7.<br />
X<br />
x.<br />
0<br />
D D<br />
D Effluent Nitrate<br />
- X Influent Ammonia<br />
X<br />
o<br />
8 S<br />
0 100 200 300 400 500<br />
X<br />
o<br />
Bed Volumes<br />
• Influent Nitrite<br />
X Effluent Ammonia<br />
X<br />
X<br />
0 0<br />
° n<br />
o<br />
n<br />
I<br />
x<br />
0<br />
n<br />
i<br />
600 700 800 900<br />
Figure 1.7 Nitrogen speciation <strong>in</strong> experiment 4 (25-m<strong>in</strong>ute BAG filter)<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
12
i S<br />
I<br />
i.o -<br />
1.4 -<br />
1.2 -<br />
1.0 -<br />
0.8 -<br />
0.4 -<br />
0.2 -<br />
no<br />
• Influent Nitrate<br />
• Influent Nitrite<br />
X Influent Ammonia<br />
X X<br />
X<br />
n n n D Q<br />
n D<br />
sk ^t f~\ /^<br />
i i i • Si<br />
n a<br />
0 100 200 300<br />
Bed Volumes<br />
D Effluent Nitrate<br />
o Effluent Nitrite<br />
X Effluent Ammonia<br />
X<br />
i i it i i<br />
400 500<br />
Figure 1.8 Nitrogen speciation <strong>in</strong> experiment 6 (49-m<strong>in</strong>ute BAG filter)<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
<strong>Bromate</strong> Reduction with Various EBCTs<br />
The EBCT was varied by chang<strong>in</strong>g the flowrate through the filters. These data were collected<br />
us<strong>in</strong>g CUW at pH 7.5, spiked with 20 ug/L bromate. The average <strong>in</strong>fluent DO concentration was<br />
approximately 2 mg/L. As before, the data were collected by us<strong>in</strong>g two filters <strong>in</strong> series so that two<br />
EBCTs could be evaluated simultaneously. Two <strong>in</strong>fluent nitrate concentrations were exam<strong>in</strong>ed -<br />
approximately 0.3 <strong>and</strong> 5.0 mg/L nitrate. As shown <strong>in</strong> Figures 1.7 <strong>and</strong> 1.8, nitrification took place<br />
<strong>in</strong> the filters <strong>and</strong> <strong>in</strong>creased the nitrate concentration <strong>and</strong> decreased the oxygen concentration to near<br />
zero. For example, <strong>in</strong> the experiment with the 25- <strong>and</strong> 50-m<strong>in</strong>ute EBCTs, even though the <strong>in</strong>fluent<br />
nitrate concentration was 0.2 mg/L, the effluent nitrate concentration was between 0.7 <strong>and</strong> 1.5 mg/L<br />
from the 25-m<strong>in</strong>ute column <strong>and</strong> between 2.1 <strong>and</strong> 3.5 mg/L from the 50-m<strong>in</strong>ute filter. The trend <strong>of</strong><br />
bromate removal versus EBCT is plotted <strong>in</strong> Figure 1.9. The error bars on each data po<strong>in</strong>t <strong>in</strong>dicate<br />
the 95 percent confidence <strong>in</strong>terval.<br />
Figure 1.9 shows that bromate removal varied widely with contact time. With a 4-m<strong>in</strong>ute<br />
EBCT, bromate removal for both the 0.3 <strong>and</strong> the 5.0 mg/L <strong>in</strong>fluent nitrate experiments was about 10<br />
percent. With a 50-m<strong>in</strong>ute EBCT, the effluent bromate concentration for both the 0.3 <strong>and</strong> the 5.0<br />
13
• 0.3 mg/L <strong>in</strong>fluent nitrate<br />
D 5.0 mg/L <strong>in</strong>fluent nitrate<br />
0<br />
0<br />
10 20 30 40 50 60<br />
EBCT (m<strong>in</strong>)<br />
Figure 1.9 Percent bromate removal versus EBCT with different <strong>in</strong>fluent nitrate concentrations<br />
(<strong>in</strong>itial conditions: CUW, pH 7.5, 2 mg/L DO, 20 ug/L BrO3')<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
mg/L <strong>in</strong>fluent nitrate filtration experiments was approximately at the detection limit. Between these<br />
two extremes, however, better bromate removal was obta<strong>in</strong>ed when less <strong>in</strong>fluent nitrate was present.<br />
The EBCT affects the time available for bromate to diffuse <strong>in</strong>to the bi<strong>of</strong>ilm <strong>and</strong> react. Thus,<br />
for shorter EBCTs, the percent bromate removal is expected to be lower. If EBCT is decreased by<br />
<strong>in</strong>creas<strong>in</strong>g the hydraulic load<strong>in</strong>g to the filter, the thickness <strong>of</strong> the bi<strong>of</strong>ilm should decrease because <strong>of</strong><br />
hydraulic shear forces, <strong>and</strong> bromate removal should accord<strong>in</strong>gly decrease.<br />
Verification <strong>of</strong> Biological <strong>Bromate</strong> <strong>Removal</strong><br />
In order to verify that the observed bromate removal <strong>in</strong> the BAC filters was <strong>in</strong>deed due to<br />
microbial action <strong>and</strong> not merely a catalytic effect <strong>of</strong> the carbon, batch biological experiments were<br />
performed. Microorganisms were harvested from the BAC filters, <strong>and</strong> these microorganisms were<br />
spiked <strong>in</strong>to glass serum bottles conta<strong>in</strong><strong>in</strong>g DDW amended with trace elements or CUW. All bottles<br />
were buffered with 1 mM phosphate buffer, adjusted to pH 7.5, <strong>and</strong> spiked with 20 ug/L bromate,<br />
14
ut the concentrations <strong>of</strong> oxygen <strong>and</strong> nitrate varied <strong>in</strong> the bottles. The bottles were sampled for<br />
bromate <strong>and</strong> nitrate on Days 1 <strong>and</strong> 19 <strong>of</strong> the experiment (Table 1.3). Table 1.3 shows that bromate<br />
removal did not take place <strong>in</strong> the absence <strong>of</strong> microorganisms (Bottle 1). When the same solution<br />
composition was used <strong>in</strong> Bottle 2, <strong>and</strong> <strong>in</strong>oculated with microorganisms from the BAG filters, bromate<br />
reduction did occur. This verifies bromate removal by the microorganisms <strong>in</strong> the BAG filters. In<br />
spite <strong>of</strong> the observation <strong>of</strong> biological bromate removal, this does not rule out the cont<strong>in</strong>ued reduction<br />
<strong>of</strong> bromate by the carbon. However, previous work has demonstrated that abiotic bromate reduction<br />
by GAG is not susta<strong>in</strong>able for long periods <strong>of</strong> time (Kirisits et al. 1998); for example, 85 percent<br />
bromate breakthrough was observed after a n<strong>in</strong>e-day, non-biological GAG filtration experiment <strong>in</strong><br />
natural water with a 2-m<strong>in</strong>ute EBCT. While it is still possible that a small amount <strong>of</strong> abiotic bromate<br />
Table 1.3 Batch biological bromate reduction experiments (<strong>in</strong>itial conditions: pH 7.5, DDW<br />
supplemented with trace salts <strong>and</strong> metals or CUW)<br />
Bottle #<br />
1<br />
2<br />
3<br />
4<br />
5<br />
6<br />
Water<br />
DDW<br />
DDW<br />
CUW<br />
CUW<br />
CUW<br />
CUW<br />
Day<br />
1<br />
19<br />
1<br />
19<br />
1<br />
19<br />
1<br />
19<br />
1<br />
19<br />
1<br />
19<br />
<strong>Bromate</strong><br />
(ug/L)<br />
12.4<br />
11.4<br />
9.4<br />
1.0<br />
11.5<br />
1.0<br />
13.0<br />
2.3<br />
14.1<br />
9.3<br />
13.1<br />
1.0<br />
Nitrate<br />
(mg/L)<br />
0.2<br />
0.2<br />
0.2<br />
0.2<br />
0.3<br />
0.2<br />
6.2<br />
0.2<br />
0.3<br />
4.9<br />
0.3<br />
0.2<br />
DO Inoculum<br />
(mg/L)<br />
0<br />
0 /<br />
0 /<br />
0 /<br />
8.3 /<br />
>0<br />
1.5 /<br />
0<br />
(Adapted from Journal AWWA, 91(8), 74-84.)<br />
15
eduction by <strong>GAC</strong> is occurr<strong>in</strong>g at the extended contact times used <strong>in</strong> the current BAG filtration<br />
experiments, it is more likely that the preponderance <strong>of</strong> the bromate reduction observed is biological.<br />
Bottles 3-6 were set up with various comb<strong>in</strong>ations <strong>of</strong> <strong>in</strong>itial DO concentrations (0 - 8.3 mg/L)<br />
<strong>and</strong> nitrate concentrations (0.2 - 6.2 mg/L), <strong>and</strong> all bottles exhibited microbial bromate removal. By<br />
Day 19, Bottles 3, 4, <strong>and</strong> 6 showed bromate removal to approximately the detection limit while the<br />
microorganisms <strong>in</strong> Bottle 5 had only reduced the bromate concentration from 14.1 ug/L to 9.3 ug/L.<br />
Bottle 5 conta<strong>in</strong>ed higher f<strong>in</strong>al concentrations <strong>of</strong> DO <strong>and</strong> nitrate than did the other bottles. On Day<br />
19, Bottle 5 conta<strong>in</strong>ed 4.88 mg/L nitrate <strong>and</strong> a non-zero DO concentration as compared to 0.2 mg/L<br />
nitrate <strong>and</strong> 0 mg/L DO <strong>in</strong> the other bottles. The elevated concentrations <strong>of</strong> both DO <strong>and</strong> nitrate <strong>in</strong><br />
Bottle 5 probably caused the slower rate <strong>of</strong> bromate reduction <strong>in</strong> this bottle.<br />
PREVIOUS STUDIES - PERCHLORATE<br />
Biological <strong>and</strong> chemical reduction processes, membrane processes, <strong>and</strong> ion exchange are be<strong>in</strong>g<br />
<strong>in</strong>vestigated for perchlorate removal. The latter two are be<strong>in</strong>g <strong>in</strong>vestigated <strong>in</strong> separate research<br />
projects under the supervision <strong>of</strong> AWWARF <strong>and</strong> are not reviewed here. Reduction processes are<br />
based on the fact that perchlorate is thermodynamically unstable. Accord<strong>in</strong>g to free energy<br />
relationships, perchlorate can oxidize water to form oxygen, result<strong>in</strong>g <strong>in</strong> the follow<strong>in</strong>g overall reaction<br />
(Note: H2O appears <strong>in</strong> the oxidation half-reaction but not the overall reaction):<br />
C1O4- - Cl- + 2 O2<br />
This reaction has a large energy barrier that keeps it from tak<strong>in</strong>g place <strong>in</strong> the absence <strong>of</strong> a catalyst.<br />
Similar to free <strong>and</strong> comb<strong>in</strong>ed chlor<strong>in</strong>e, chlorite, chlorate, <strong>and</strong> bromate, the low concentrations <strong>of</strong><br />
perchlorate <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water are k<strong>in</strong>etically stable. An objective <strong>of</strong> this research is to f<strong>in</strong>d a<br />
microorganism with the appropriate enzyme, or a catalytic solid, that can enable this reaction to<br />
proceed rapidly.<br />
In a study <strong>of</strong> biological reduction, Rikken et al. (1996) found that a microorganism <strong>in</strong> the P-<br />
subgroup <strong>of</strong> the Proteobacteria was able to use perchlorate as an electron acceptor. Hurley et al.<br />
(1996) described a pilot system that is <strong>in</strong> place at Tyndall Air Force Base to treat the wash water from<br />
motors conta<strong>in</strong><strong>in</strong>g ammonium perchlorate. In this system, Wol<strong>in</strong>ella succ<strong>in</strong>ogenes is the bacterial<br />
16
species that is able to reduce perchlorate to chloride <strong>and</strong> perchlorate is reduced from 3000 mg/L to<br />
0.5 mg/L. Biological reduction <strong>of</strong> perchlorate <strong>in</strong> the ug/L concentration range has only recently been<br />
achieved (Herman <strong>and</strong> Frankenberger 1999, Logan 1999). Herman <strong>and</strong> Frankenberger (1999) used<br />
a pure culture <strong>of</strong> perclace to reduce perchlorate to below the detection limit <strong>in</strong> s<strong>and</strong> filters. Logan<br />
(1999) isolated perchlorate-reduc<strong>in</strong>g microorganisms from a sample <strong>of</strong> wastewater <strong>and</strong> utilized these<br />
microorganisms <strong>in</strong> various bioreactor configurations to achieve perchlorate reduction. A completely<br />
mixed, autotrophic, hydrogen-oxidiz<strong>in</strong>g bi<strong>of</strong>ilm reactor has been used by Rittmann et al. to exam<strong>in</strong>e<br />
perchlorate, nitrate, <strong>and</strong> hydrogen utilization mechanisms (Rittman et al. 2000).<br />
In each <strong>of</strong> the biological reduction processes found <strong>in</strong> the literature to date, there is a need<br />
to ma<strong>in</strong>ta<strong>in</strong> anaerobic conditions. Oxygen is a major <strong>in</strong>hibitor <strong>of</strong> perchlorate reduction <strong>and</strong> exposure<br />
<strong>of</strong> active perchlorate-reduc<strong>in</strong>g microorganisms to air can immediately term<strong>in</strong>ate perchlorate reduction<br />
(Attaway <strong>and</strong> Smith 1993). Attaway <strong>and</strong> Smith (1993) reported that <strong>in</strong> order for perchlorate<br />
reduction to occur, a redox potential <strong>of</strong> less than -110 mV must be ma<strong>in</strong>ta<strong>in</strong>ed. Other operat<strong>in</strong>g<br />
parameters that have rema<strong>in</strong>ed consistent among biological processes are: 1) neutral pH, 2) moderate<br />
to high temperatures, <strong>and</strong> 3) a carbon source suitable for respiration (Herman <strong>and</strong> Frankenberger<br />
1998).<br />
Chemical reduction by <strong>GAC</strong>, probably with <strong>GAC</strong> serv<strong>in</strong>g as the electron donor, had been<br />
considered to be a promis<strong>in</strong>g method <strong>of</strong> perchlorate reduction. However, studies done by the<br />
California Department <strong>of</strong> Health Services, found that a virg<strong>in</strong> bed <strong>of</strong> <strong>GAC</strong> removed perchlorate for<br />
only two weeks (Forbes 1997).<br />
OBJECTIVES<br />
The two ma<strong>in</strong> objectives <strong>of</strong> the project were to <strong>in</strong>vestigate the biological reduction <strong>of</strong> bromate<br />
<strong>and</strong> perchlorate <strong>in</strong> BAG filters <strong>and</strong> to <strong>in</strong>vestigate the abiotic reduction <strong>of</strong> perchlorate <strong>in</strong> advanced<br />
oxidation processes <strong>in</strong>volv<strong>in</strong>g <strong>GAC</strong>. Specific project objectives were as follows:<br />
• To show that perchlorate can be removed <strong>in</strong> BAC adsorbers under the same conditions that<br />
have proven effective for bromate removal, or to show how these conditions must be changed<br />
to obta<strong>in</strong> biological removal <strong>of</strong> perchlorate.<br />
17
• To <strong>in</strong>vestigate the mechanism <strong>of</strong> the biological reduction process, with<br />
special focus on the role <strong>of</strong> nitrate-reduc<strong>in</strong>g organisms, the types <strong>of</strong><br />
microorganisms carry<strong>in</strong>g out the bromate/perchlorate reduction <strong>and</strong> their<br />
growth requirements, the type <strong>and</strong> concentration <strong>of</strong> electron donor, <strong>and</strong> the<br />
effect <strong>of</strong> parameters such as bromate <strong>and</strong> perchlorate concentration, dissolved<br />
oxygen (DO) concentration, nitrate concentration, pH, temperature, <strong>and</strong> the<br />
role <strong>of</strong> a prefiltration ozonation step.<br />
To <strong>in</strong>vestigate advanced oxidation processes to determ<strong>in</strong>e whether these reactions can<br />
produce efficient removal <strong>of</strong> perchlorate, <strong>and</strong> if so, to develop the process at the bench-scale.<br />
*• Emphasis will be placed on the ozone-hydrogen peroxide-<strong>GAC</strong> process, virg<strong>in</strong><br />
<strong>GAC</strong>, <strong>GAC</strong> impregnated with metals such as copper <strong>and</strong> z<strong>in</strong>c, <strong>and</strong> <strong>GAC</strong><br />
mixed with solids such as iron, z<strong>in</strong>c, <strong>and</strong> alum<strong>in</strong>um.<br />
To optimize the bromate <strong>and</strong> perchlorate removal processes that show the most encourag<strong>in</strong>g<br />
results <strong>and</strong> to monitor the product water for potability.<br />
To test the most promis<strong>in</strong>g <strong>of</strong> the bromate <strong>and</strong> perchlorate removal processes at the pilot-<br />
scale.<br />
18
CHAPTER 2. MATERIALS AND METHODS<br />
MATERIALS USED FOR BROMATE REDUCTION EXPERIMENTS<br />
Water<br />
Deionized, distilled water (DDW) had a total organic carbon concentration less than 0.3<br />
mg/L. Champaign-Urbana, tap water (CUW) <strong>and</strong> Lake Michigan water (LMW) were the two natural<br />
waters used <strong>in</strong> this study. CUW is tapwater (Champaign, Ill<strong>in</strong>ois), <strong>and</strong> it comes from a groundwater<br />
source. Prior to use, CUW was dechlor<strong>in</strong>ated by sodium sulfite addition. Raw LMW was obta<strong>in</strong>ed<br />
from the crib <strong>of</strong> the South Water Production Plant (Chicago, IL); the crib is 3.2 kilometers (2 miles)<br />
from shore at a depth <strong>of</strong> 9.6 to 10.5 meters (32-35 feet). The ma<strong>in</strong> differences between the two<br />
natural waters were the ammonia (NH3) concentrations <strong>and</strong> biodegradability <strong>of</strong> the organic matter.<br />
CUW conta<strong>in</strong>ed less than 1.70 mg/L NH3-N, with an average concentration <strong>of</strong> 0.80 mg/L, <strong>and</strong> the<br />
NH3-N concentration <strong>in</strong> LMW was below the detection limit <strong>of</strong> 0.05 mg/L. The ammonia <strong>in</strong> CUW<br />
led to the production <strong>of</strong> nitrate <strong>in</strong> the BAG filters; little or no nitrification occurred <strong>in</strong> LMW, <strong>and</strong> the<br />
background nitrate concentration was 1.6 mg/L. The organic matter <strong>in</strong> CUW displayed some degree<br />
<strong>of</strong> biodegradability <strong>in</strong> the BAG filter, with an average dissolved organic carbon (DOC) removal <strong>of</strong><br />
25 percent. The organic matter <strong>in</strong> LMW was fairly <strong>in</strong>ert <strong>in</strong> the BAG filter with an average DOC<br />
removal <strong>of</strong> 4 percent. Average water characteristics <strong>of</strong> CUW <strong>and</strong> LMW are listed <strong>in</strong> Table 2.1.<br />
Table 2.1 Average water quality characteristics <strong>of</strong> CUW <strong>and</strong> LMW<br />
Water<br />
Chloride<br />
Nitrate<br />
Sulfate<br />
DOC<br />
Ammonia<br />
(mg/L)<br />
(mg/L)<br />
(mg/L)<br />
(mg/L)<br />
(mg/LN)<br />
CUW<br />
8<br />
0.3<br />
1<br />
1.4<br />
0.80<br />
LMW<br />
10<br />
1.6<br />
23<br />
2.4<br />
Reagents<br />
Reagent grade sodium bromate (Aldrich Chemical Co., Inc., Milwaukee, WI) <strong>and</strong> sodium<br />
perchlorate (Sigma Chemical Co., St. Louis, MO) were used to prepare 1000 mg/L stock solutions.<br />
Other chemicals were <strong>of</strong> reagent grade quality or better. Where appropriate, chemicals were dried<br />
overnight at 105 °C <strong>and</strong> stored <strong>in</strong> a desiccator.<br />
METHODS UTILIZED FOR BROMATE REDUCTION EXPERIMENTS<br />
Biologically Active Carbon - Small Columns<br />
BAC filters were constructed <strong>in</strong> 2.5-cm (1-<strong>in</strong>ch) <strong>in</strong>ner-diameter glass pipes (Ace Glass,<br />
V<strong>in</strong>el<strong>and</strong>, NJ). The pipes were capped with Teflon® <strong>and</strong> steel endcaps. A small layer <strong>of</strong> 3-mm (0.1-<br />
<strong>in</strong>ch) glass beads was placed below the carbon bed <strong>in</strong> order to promote uniform flow characteristics<br />
<strong>and</strong> prevent channel<strong>in</strong>g. Influent was pumped upflow through the columns us<strong>in</strong>g peristaltic<br />
Masterflex pumps (Cole-Parmer, Vernon Hills, EL). Each <strong>in</strong>fluent conta<strong>in</strong>er was fitted with a sta<strong>in</strong>less<br />
steel, Teflon®, or Plexiglas® float<strong>in</strong>g cover, to ma<strong>in</strong>ta<strong>in</strong> the DO concentration. The <strong>in</strong>fluent DO<br />
concentration was adjusted to the desired level by sparg<strong>in</strong>g with compressed nitrogen or oxygen.<br />
Four filters were constructed to evaluate bromate removal from CUW us<strong>in</strong>g Norit RO 0.8<br />
<strong>GAC</strong> (Norit Americas Inc., Atlanta, GA). The average <strong>GAC</strong> pellet diameter <strong>and</strong> length were 0.8 <strong>and</strong><br />
3.4 mm, respectively. Each filter conta<strong>in</strong>ed one-quarter preloaded carbon <strong>and</strong> three-quarters fresh<br />
carbon. The preloaded carbon was taken from a bench-scale filter that had been operated with a 2-<br />
m<strong>in</strong>ute EBCT for 40,000 bed volumes with dechlor<strong>in</strong>ated surface water (Danville filter effluent,<br />
Consumer's Ill<strong>in</strong>ois, Danville, Ill<strong>in</strong>ois). At the end <strong>of</strong> this preload<strong>in</strong>g period, the carbon was used <strong>in</strong><br />
experiments not discussed <strong>in</strong> this report, <strong>and</strong> approximately 30 percent <strong>of</strong> the <strong>in</strong>fluent DOC <strong>and</strong> 10<br />
percent <strong>of</strong> the <strong>in</strong>fluent bromate were consistently be<strong>in</strong>g removed by the microorganisms dur<strong>in</strong>g a 45-<br />
day experiment. At this time, the preloaded carbon was considered to be biologically active <strong>and</strong> at<br />
steady state. Aforementioned, the preloaded carbon was then mixed with fresh carbon <strong>in</strong> the<br />
construction <strong>of</strong> four new filters. This was done <strong>in</strong> order to seed the fresh carbon with<br />
microorganisms, thereby reduc<strong>in</strong>g the preload<strong>in</strong>g time required to stimulate biological activity <strong>in</strong> the<br />
filters.<br />
20
BAG was formed by extensively preload<strong>in</strong>g the carbon filters with CUW that had been<br />
dechlor<strong>in</strong>ated by sulfite addition. The purpose <strong>of</strong> the preload<strong>in</strong>g period was to establish a bi<strong>of</strong>ilm on<br />
the carbon, <strong>and</strong> then the BAG filter could be used for various biological bromate reduction<br />
experiments. The preload<strong>in</strong>g conditions were determ<strong>in</strong>ed <strong>in</strong> accordance with the experiments that<br />
were to immediately follow the preload<strong>in</strong>g time. It was planned to use the BAG filters <strong>in</strong> experiments<br />
with an <strong>in</strong>fluent DO concentration <strong>of</strong> 2 mg/L <strong>and</strong> a variety <strong>of</strong> EBCTs between 4 <strong>and</strong> 50 m<strong>in</strong>utes.<br />
Therefore, the carbon filters were preloaded with 2,700 bed volumes <strong>of</strong> dechlor<strong>in</strong>ated CUW with an<br />
EBCT <strong>of</strong> 10 m<strong>in</strong>utes <strong>and</strong> an <strong>in</strong>fluent DO concentration <strong>of</strong> approximately 2 mg/L. Two parallel<br />
filtration experiments were set up, each with two columns <strong>in</strong> series, for a total <strong>of</strong> four filters. All<br />
filtration experiments with CUW were run with two filters <strong>in</strong> series, so that two EBCTs could be<br />
evaluated simultaneously. For example, <strong>in</strong>fluent water first passed through a 25-m<strong>in</strong>ute filter <strong>and</strong> then<br />
through another 25-m<strong>in</strong>ute filter for a total EBCT <strong>of</strong> 50 m<strong>in</strong>utes. The only difference between the<br />
preload<strong>in</strong>g conditions <strong>of</strong> the parallel filters was the <strong>in</strong>fluent nitrate concentration; one set-up was<br />
operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L <strong>and</strong> the other with an <strong>in</strong>fluent nitrate<br />
concentration <strong>of</strong> 5.0 mg/L. No exogenous electron donor was added to CUW experiments. A<br />
schematic diagram show<strong>in</strong>g the BAG filters us<strong>in</strong>g CUW is shown <strong>in</strong> Figure 2.1.<br />
As the filters were preloaded, the percent bromate removal cont<strong>in</strong>ued to decrease as the<br />
bromate reduction capacity <strong>of</strong> the carbon was approached. However, near the end <strong>of</strong> the preload<strong>in</strong>g<br />
period, the percent bromate removal began to <strong>in</strong>crease. This <strong>in</strong>dicated that while the abiotic reduction<br />
capacity <strong>of</strong> the carbon was approach<strong>in</strong>g exhaustion, microbiological bromate reduction had<br />
commenced. For the filters be<strong>in</strong>g preloaded <strong>in</strong> the presence <strong>of</strong> 5.0 mg/L nitrate, the bromate removal<br />
decreased to reach a m<strong>in</strong>imum <strong>of</strong> 27 percent as the abiotic bromate reduction capacity <strong>of</strong> the carbon<br />
approached exhaustion but then <strong>in</strong>creased to approximately 53 percent removal by the end <strong>of</strong> the<br />
preload<strong>in</strong>g time due to the onset <strong>of</strong> biological bromate reduction. For the filters be<strong>in</strong>g preloaded <strong>in</strong><br />
the presence <strong>of</strong> 0.3 mg/L nitrate, the bromate removal decreased to reach a m<strong>in</strong>imum <strong>of</strong> 37 percent<br />
but then <strong>in</strong>creased to 83 percent removal by the end <strong>of</strong> the preload<strong>in</strong>g period due to biological<br />
bromate reduction. These data show that the BAG filters were biologically active after the preload<strong>in</strong>g<br />
period. To atta<strong>in</strong> steady state bromate removal, the BAG filters were operated with an EBCT <strong>of</strong> 25<br />
m<strong>in</strong>utes (each filter) for 30 days. For the last two weeks <strong>of</strong> this period, the filter be<strong>in</strong>g operated with<br />
an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L <strong>and</strong> a 25-m<strong>in</strong>ute EBCT removed 83 ± 5 percent<br />
bromate.<br />
21
Influent:<br />
CUW<br />
10-50 ug/L <strong>Bromate</strong><br />
pH 6.8-8.2<br />
2-14mg/LDO<br />
0.3 - 42 mg/L Nitrate<br />
t<br />
NOTE: Not to Scale<br />
2.5-cm Inner-Diameter<br />
8.0-cm BAC Bed Height<br />
Support<br />
Material<br />
Figure 2.1 Schematic <strong>of</strong> BAG filters us<strong>in</strong>g CUW<br />
For the filter be<strong>in</strong>g operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L <strong>and</strong> a 25-m<strong>in</strong>ute<br />
EBCT, the last two weeks <strong>of</strong> this experiment showed 81 ± 6 percent bromate removal. The<br />
consistent bromate removals observed <strong>in</strong> these filters demonstrated the atta<strong>in</strong>ment <strong>of</strong> steady state.<br />
New filters were not assembled for each experimental condition (i.e. DO concentration, nitrate<br />
concentration, EBCT). Therefore, zero bed volumes is a term relative to the start <strong>of</strong> the particular<br />
experiment, not necessarily to the overall life <strong>of</strong> the filter. Each experimental condition was run for<br />
approximately 2 weeks, to ensure adequate acclimation to the experimental conditions.<br />
Figure 2.2 shows a schematic diagram <strong>of</strong> another set <strong>of</strong> filters that was constructed to<br />
evaluate bromate reduction us<strong>in</strong>g LMW. Norit 830 <strong>GAC</strong> (Norit Americas Inc., Atlanta, GA) was<br />
the medium <strong>in</strong> these filters, <strong>and</strong> it was taken from the full-scale plant at the Paul M. Neal Water<br />
Treatment Facility (Lake Bluff, IL). This carbon had <strong>and</strong> average particle diameter <strong>of</strong> 0.9 mm <strong>and</strong><br />
had been used for approximately seven years to treat ozonated LMW. Prior to pack<strong>in</strong>g the carbon<br />
<strong>in</strong>to the column, the BAC was r<strong>in</strong>sed with DDW to remove excessive turbidity. S<strong>in</strong>ce this carbon<br />
was already biologically active, no extensive preload<strong>in</strong>g period was required. For some <strong>of</strong> the LMW<br />
experiments, a syr<strong>in</strong>ge pump (74900 series, Cole Farmer Instrument Co., Vernon Hills, IL) was used<br />
to add an electron donor solution conta<strong>in</strong><strong>in</strong>g 1.2 mM<br />
22
Float<strong>in</strong>g Cover<br />
1<br />
Influent:<br />
LMW<br />
20 (ig/L <strong>Bromate</strong><br />
pH7.5<br />
2 mg/L DO<br />
1. 6 mg/L Nitrate<br />
Peristaltic Pump<br />
8.0-cm BAG Bed Height<br />
Support<br />
Material<br />
NOTE: Not to Scale<br />
Syr<strong>in</strong>ge Pump<br />
Lactate +<br />
Pyruvate<br />
2. 5 -cm Inner-Diameter<br />
Figure 2.2 Schematic <strong>of</strong> BAG filters us<strong>in</strong>g LMW<br />
pyruvic acid (10 electron equivalents/mole3) <strong>and</strong> 1.0 mM lactic acid (12 electron equivalents/mole3).<br />
Biologically Active Carbon - Large Filter<br />
A BAG filter was constructed <strong>in</strong> a 5.1-cm (2-<strong>in</strong>ch) <strong>in</strong>ner-diameter glass pipe (Ace Glass,<br />
V<strong>in</strong>el<strong>and</strong>, NJ) <strong>and</strong> capped with rubber stoppers (Figure 2.3). Virg<strong>in</strong> Norit RO 0.8 GAG (Norit<br />
Americas Inc., Atlanta, GA) was washed with DDW to remove carbon f<strong>in</strong>es <strong>and</strong> added to the column.<br />
Three evenly-spaced sampl<strong>in</strong>g ports were added to the column by a glass-blower, <strong>and</strong> four EBCTs<br />
could be monitored <strong>in</strong> every experiment by us<strong>in</strong>g the three sampl<strong>in</strong>g ports <strong>and</strong> the effluent l<strong>in</strong>e <strong>of</strong> the<br />
filter. The filter was run upflow us<strong>in</strong>g a peristaltic pump (Masterflex, Cole-Parmer Instrument Co.,<br />
Vernon Flills, IL). Compressed nitrogen was used to sparge the <strong>in</strong>fluent water to reach the desired<br />
3 The number <strong>of</strong> electron equivalents per mole was calculated assum<strong>in</strong>g m<strong>in</strong>eralization <strong>of</strong><br />
the organic acid to carbon dioxide.<br />
23
DO concentration (2-6 mg/L). The <strong>in</strong>fluent was stored <strong>in</strong> a headspace-free conta<strong>in</strong>er, with a float<strong>in</strong>g<br />
cover made <strong>of</strong> sta<strong>in</strong>less steel, <strong>in</strong> order to ma<strong>in</strong>ta<strong>in</strong> the DO concentration.<br />
The <strong>GAC</strong> was extensively contacted with CUW <strong>in</strong> order to exhaust the adsorptive capacity<br />
<strong>of</strong> the carbon. The <strong>GAC</strong> was preloaded with 7,200 bed volumes <strong>of</strong> chlor<strong>in</strong>ated CUW with an <strong>in</strong>fluent<br />
DO concentration <strong>of</strong> approximately 4 mg/L <strong>and</strong> a total EBCT <strong>of</strong> 5 m<strong>in</strong>utes. CUW was not<br />
dechlor<strong>in</strong>ated dur<strong>in</strong>g this portion <strong>of</strong> the preload<strong>in</strong>g due to the operational difficulty that would have<br />
caused because <strong>of</strong> the high flowrate (-300 mL/m<strong>in</strong>). S<strong>in</strong>ce Kirisits et al. (1998) observed 80 percent<br />
bromate breakthrough from a non-biologically active <strong>GAC</strong> filter after 6,500 bed volumes, it is<br />
unlikely that significant abiotic bromate removal was tak<strong>in</strong>g place <strong>in</strong> the BAG filter after the<br />
preload<strong>in</strong>g period. In order to stimulate biological growth, the filter was subsequently operated for<br />
500 BV with dechlor<strong>in</strong>ated CUW <strong>and</strong> 2 mg/L DO. Dur<strong>in</strong>g this time, the filter had a total EBCT <strong>of</strong><br />
60 m<strong>in</strong>utes, with the sampl<strong>in</strong>g ports allow<strong>in</strong>g the evaluation <strong>of</strong> 15-, 30-, 45-, <strong>and</strong> 60-m<strong>in</strong>ute EBCTs.<br />
. Float<strong>in</strong>g Cover<br />
Influent:<br />
CUW<br />
20 ng/L <strong>Bromate</strong><br />
pH7.5<br />
0.3 mg/L Nitrate<br />
2-6 mg/L DO<br />
Sampl<strong>in</strong>g Ports<br />
30.6-cmBAC Bed Height<br />
3-mm<br />
Gass Beads<br />
NOTE: Not to Scale<br />
Peristaltic Pump<br />
5.1-cm Inner-Diameter<br />
Figure 2.3 Schematic <strong>of</strong> multi-port BAC filter us<strong>in</strong>g CUW<br />
24
Backwash <strong>in</strong>g the BAG filters<br />
Two <strong>of</strong> the small BAG filters us<strong>in</strong>g CUW (Figure 2.1) were backwashed to determ<strong>in</strong>e its<br />
effect on bromate removal. The backwash water consisted <strong>of</strong> filter effluent rather than the usual filter<br />
<strong>in</strong>fluent so that the microorganisms would not be exposed to higher ammonia <strong>and</strong> DOC<br />
concentrations than usual. The backwash<strong>in</strong>g procedure consisted <strong>of</strong> 10 m<strong>in</strong>utes <strong>of</strong> fluidization with<br />
approximately 35 percent bed expansion. The carbon was mechanically agitated from the top <strong>of</strong> the<br />
column to promote mix<strong>in</strong>g <strong>in</strong> the bed.<br />
Heterotrophic Plate Counts<br />
Heterotrophic plate counts (HPCs) were run with water samples collected before <strong>and</strong> after<br />
backwash<strong>in</strong>g. Bacto® R2A agar (Difco Laboratories, Detroit, MI) was the medium used. 9.1 g <strong>of</strong><br />
R2A was dissolved per 500 mL <strong>of</strong> DDW. This was heated to approximately 70 °C <strong>and</strong> stirred for 20<br />
m<strong>in</strong>utes to dissolve the medium. The medium was then sterilized by autoclav<strong>in</strong>g (Eagle/Century<br />
Series, Amsco Scientific) for a 20-m<strong>in</strong>ute liquid cycle. 5-mL portions <strong>of</strong> the medium were poured<br />
<strong>in</strong>to sterile petri dishes <strong>and</strong> allowed to solidify. The Membrane Filter Method was performed<br />
accord<strong>in</strong>g to St<strong>and</strong>ard Method 9215D (APHA, AWWA, <strong>and</strong> WEF, 1995). Various volumes <strong>of</strong> water<br />
were filtered through sterile, gridded filters (0.45 ^im, Millipore, Bedford, MA). Each filter was<br />
placed <strong>in</strong> a petri dish <strong>and</strong> <strong>in</strong>cubated <strong>in</strong> the dark for a m<strong>in</strong>imum <strong>of</strong> 5 days at 25 °C.<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis<br />
Denatur<strong>in</strong>g gradient gel electrophoresis (DGGE) was used to <strong>in</strong>vestigate the diversity <strong>of</strong> the<br />
microbial population <strong>in</strong> the BAG filters. The DGGE technique consists <strong>of</strong> extract<strong>in</strong>g ribosomal DNA<br />
(rDNA) from the microbial community, amplify<strong>in</strong>g it by the polymerase cha<strong>in</strong> reaction (PCR), <strong>and</strong><br />
then apply<strong>in</strong>g it to a polyacrylamide gel that has a gradient <strong>of</strong> DNA denatur<strong>in</strong>g agents. As the rDNA<br />
migrates through the gel, it melts at a particular location <strong>in</strong> the denatur<strong>in</strong>g gradient. The position at<br />
which this melt<strong>in</strong>g occurs is a function <strong>of</strong> the base pairs <strong>in</strong> the DNA fragment. When the fragment<br />
melts, its motility ceases. Thus, one will obta<strong>in</strong> a DGGE pattern conta<strong>in</strong><strong>in</strong>g different b<strong>and</strong>s that have<br />
25
separated accord<strong>in</strong>g to their base pair structure. Each b<strong>and</strong> theoretically represents a unique<br />
microorganism. DGGE gels were analyzed by visual <strong>in</strong>spection.<br />
Plat<strong>in</strong>g Experiments<br />
The purpose <strong>of</strong> the plat<strong>in</strong>g experiments was to isolate <strong>in</strong>dividual active bromate-reducers from<br />
the BAC filters operated with CUW. One straightforward <strong>and</strong> common approach to do this is<br />
through the use <strong>of</strong> solid medium, which facilitates the isolation <strong>of</strong> a pure culture. Agar medium<br />
(MR2A), conta<strong>in</strong><strong>in</strong>g essential nutrients, bromate, <strong>and</strong> nitrate, was prepared for the plat<strong>in</strong>g<br />
experiments (Table 2.2). MR2A was selected as an appropriate heterotrophic medium for<br />
environmental isolates <strong>and</strong> is particularly well-suited for denitrifiers, which have been suggested as<br />
an important group <strong>of</strong> microorganisms <strong>in</strong> biological bromate reduction. The MR2A was adjusted to<br />
pH 7.0 <strong>and</strong> autoclaved. The vitam<strong>in</strong> 2 solution was filter sterilized, stored at 4 °C, <strong>and</strong> added to the<br />
sterile agar medium before the plates were poured.<br />
The sterile agar plates were degassed <strong>in</strong> an anaerobic chamber for a period <strong>of</strong> one week before<br />
an <strong>in</strong>oculum was applied. Biomass was obta<strong>in</strong>ed from the BAC filters operated with CUW <strong>and</strong> was<br />
diluted <strong>in</strong> degassed <strong>in</strong>fluent medium <strong>in</strong> the follow<strong>in</strong>g ratios: 1:10,000, 1:1,000, <strong>and</strong> 1:100. 0.1 mL<br />
from each dilution was taken <strong>and</strong> added to a plate us<strong>in</strong>g a sterile syr<strong>in</strong>ge. The biomass was spread<br />
over the surface <strong>of</strong> the agar. The plates were sealed with parafilm <strong>and</strong> placed <strong>in</strong>to anaerobic jars that<br />
had been purged with argon. These jars were then <strong>in</strong>cubated at room temperature. After 3 weeks,<br />
the plates were observed for growth. There were a number <strong>of</strong> fast-grow<strong>in</strong>g microorganisms which<br />
caused colonies to <strong>in</strong>trude upon each other. For this reason, some biomass was taken from each plate<br />
<strong>and</strong> restreaked onto a fresh plate. These plates were <strong>in</strong>cubated for only 5 days before they were<br />
checked for growth. Different colony morphologies were observed, <strong>and</strong> each different morphology<br />
was streaked onto a fresh plate <strong>in</strong> order to obta<strong>in</strong> a pure culture. These plates were <strong>in</strong>cubated <strong>in</strong> an<br />
argon atmosphere for 41 days.<br />
At this time, some <strong>of</strong> the pure cultures were harvested. Us<strong>in</strong>g a sterile syr<strong>in</strong>ge, 0.1 mL <strong>of</strong><br />
sterile medium was used to resuspend colonies <strong>of</strong> pure cultures to use as <strong>in</strong>ocula for further test<strong>in</strong>g<br />
<strong>in</strong> aqueous medium conta<strong>in</strong><strong>in</strong>g bromate.<br />
26
Table 2.2 Agar recipe for plat<strong>in</strong>g experiments<br />
MR2A Constituent<br />
KH2PO4<br />
K2HP04<br />
CaCl2 * 2H2O<br />
MgCl2 * 6H2O<br />
FeSO4 * 7H20<br />
NH4C1<br />
Trace metals 1<br />
Vitam<strong>in</strong> 2<br />
Yeast extract<br />
Peptone<br />
Casam<strong>in</strong>o acids<br />
Sodium pyruvate<br />
Bacto agar<br />
Br03'(ng/L)<br />
NO3- (mM)<br />
DDW<br />
1 Trace metals (500X)<br />
MnCl2 * 4H2O<br />
H3B03<br />
ZnCl2<br />
CoCl2 * 6H2O<br />
NiSO4 * 6H2O<br />
CuCl2 * 2H2O<br />
NaMoO2 * 2H2O<br />
DDW<br />
2 Vitam<strong>in</strong> solution (1000X)<br />
B 12<br />
Biot<strong>in</strong><br />
Calcium pantothenate<br />
Folic acid<br />
Nicot<strong>in</strong>amide<br />
Pyridox<strong>in</strong>e HC1<br />
Rib<strong>of</strong>lav<strong>in</strong><br />
Thiam<strong>in</strong>e HC1<br />
DDW<br />
(g/L)<br />
0.25<br />
0.40<br />
0.015<br />
0.02<br />
0.007<br />
0.005<br />
0.8<br />
2mL<br />
ImL<br />
0.5<br />
0.5<br />
0.5<br />
0.3<br />
15.0<br />
100<br />
1000<br />
1.0<br />
1000 mL<br />
g/lOOmL<br />
0.25<br />
0.025<br />
0.025<br />
0.025<br />
0.025<br />
0.015<br />
0.0005<br />
100 mL<br />
mg/L<br />
1.0<br />
20.0<br />
50.0<br />
20.0<br />
50.0<br />
100.0<br />
50.0<br />
50.0<br />
1000 mL<br />
27
Three different media were used. The first aqueous medium consisted <strong>of</strong> CUW, 28.7 ug/L<br />
bromate, 4.8 mg/L nitrate, <strong>and</strong> 0 mg/L DO at pH 7.8 with 2.5 mM phosphate buffer. The second<br />
aqueous medium was DDW, 43.5 ug/L bromate, 4.9 mg/L nitrate, 8.3 mg/L DOC (0.10 mM acetate,<br />
0.08 mM pyruvate, <strong>and</strong> 0.07 mM lactate), <strong>and</strong> 0 mg/L DO at pH 7.5 with 5.0 mM phosphate buffer.<br />
The third aqueous medium consisted <strong>of</strong> MR2A broth, supplemented with higher-than-usual<br />
concentrations <strong>of</strong> bromate <strong>and</strong>/or nitrate. This medium was amended with one <strong>of</strong> the follow<strong>in</strong>g: 1<br />
mM nitrate, 45 uM nitrate plus 70 uM bromate, or 70 uM bromate; the cultures <strong>in</strong> these media were<br />
monitored for visible growth <strong>and</strong> gas production (N2O). Follow<strong>in</strong>g <strong>in</strong>cubation <strong>of</strong> the aqueous<br />
medium experiments, samples were taken for bromate, bromide, nitrite, <strong>and</strong> nitrate analyses. Figure<br />
2.4 schematically shows the progression <strong>of</strong> the plat<strong>in</strong>g <strong>and</strong> aqueous medium experiments.<br />
The pure cultures were also tested for facultative growth on MR2 A agar by <strong>in</strong>cubation under<br />
aerobic conditions.<br />
Dilution-to-Ext<strong>in</strong>ction Experiments<br />
In addition to the traditional plat<strong>in</strong>g technique, another way to isolate microorganisms is<br />
through dilution-to-ext<strong>in</strong>ction experiments (progressive dilutions). Two <strong>in</strong>ocula were used for these<br />
experiments. The first <strong>in</strong>oculum consisted <strong>of</strong> effluent water collected from the second BAG filter <strong>in</strong><br />
series operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5 mg/L <strong>in</strong> CUW. The second <strong>in</strong>oculum<br />
consisted <strong>of</strong> carbon <strong>and</strong> water taken from the effluent side <strong>of</strong> the second BAG filter <strong>in</strong> series operated<br />
with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L <strong>in</strong> CUW. Medium for the experiments consisted<br />
<strong>of</strong> the follow<strong>in</strong>g: dechlor<strong>in</strong>ated CUW, 5 mg/L nitrate, 30 pgfL bromate, 2.5 mM phosphate buffer.<br />
15 mL <strong>of</strong> medium were added to each Balch tube. Butyl rubber stoppers were <strong>in</strong>serted <strong>and</strong> alum<strong>in</strong>um<br />
caps were crimped <strong>in</strong> place. The tubes were autoclaved, cooled, <strong>and</strong> then <strong>in</strong>oculated with the<br />
appropriate biomass. Each experimental condition was set up <strong>in</strong> triplicate. The tubes were then<br />
monitored for bromate, bromide, nitrate, <strong>and</strong> nitrite. Active dilution samples can be <strong>in</strong>oculated onto<br />
agar medium for isolation <strong>and</strong> test<strong>in</strong>g. However, s<strong>in</strong>ce 8 bromate-reduc<strong>in</strong>g isolates have been<br />
cultured us<strong>in</strong>g the traditional plate technique, the dilution-to-ext<strong>in</strong>ction samples have not yet been<br />
used to isolate additional bromate-reducers.<br />
28
Streaked <strong>in</strong>dividual colony morphologies<br />
to get pure isolate (Incubated 41 days)<br />
Pure isolates <strong>in</strong>oculated to aqueous media<br />
Biomass taken from<br />
BAG filter<br />
A\<br />
3 Dilutions prepared<br />
1:10,000 1:1,000 1:100<br />
J T \ Plates streaked (Incubated 3 weeks)<br />
Confluent growth; plates restreaked<br />
(Incubated 5 days)<br />
Aqueous media: CUW, DDW, M2RA<br />
Bottles sampled to monitor bromate/nitrate reduction<br />
I<br />
Active bromate-reducers will be characterized<br />
by 16S rDNA sequenc<strong>in</strong>g <strong>and</strong> nutritional requirements<br />
Figure 2.4 Progression <strong>of</strong> plat<strong>in</strong>g <strong>and</strong> aqueous medium experiments with pure cultures
Serum Bottle Experiments<br />
S<strong>in</strong>ce the column filtration experiments are labor <strong>and</strong> time <strong>in</strong>tensive, it is useful to set up batch<br />
biological experiments to test the dependence <strong>of</strong> bromate reduction on various parameters. 160 mL<br />
serum bottles, equipped with gas-tight butyl rubber stoppers, were used as the batch reactors. Each<br />
bottle conta<strong>in</strong>ed 140 mL <strong>of</strong> aqueous medium, which was autoclaved after the bottles were capped.<br />
After the bottles were cooled, biomass from the BAG filters was added.<br />
The medium consisted <strong>of</strong> the follow<strong>in</strong>g: CUW, 5 mg/L nitrate, 5 mg/L sulfate, 19 ug/L<br />
bromate, <strong>and</strong> 1 mM phosphate buffer. Biomass was taken from the BAG filter operated with an<br />
<strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L <strong>in</strong> CUW, <strong>and</strong> 1 mL <strong>of</strong> this <strong>in</strong>oculum was spiked to each<br />
serum bottle. The serum bottle experiments were used to evaluate the effect <strong>of</strong> temperature (4 <strong>and</strong><br />
25 °C) on bromate reduction.<br />
MATERIALS USED FOR PERCHLORATE REDUCTION EXPERIMENTS<br />
Water<br />
Two types <strong>of</strong> water were used <strong>in</strong> this study - DDW <strong>and</strong> CUW. Typical properties <strong>of</strong> CUW<br />
are listed <strong>in</strong> Table 2.1.<br />
Reagents <strong>and</strong> Carbon<br />
The chemicals used <strong>in</strong>cluded reagent grade sodium perchlorate (Sigma Chemical Co., St.<br />
Louis, MO) <strong>and</strong> sodium bromate (Aldrich Chemical Co., Inc., Milwaukee, WI). Other chemicals<br />
were <strong>of</strong> reagent grade quality or better. Where appropriate, the chemicals were dried overnight at<br />
105 °C <strong>and</strong> stored <strong>in</strong> a desiccator. Norit RO 0.8 GAG, lot number 72303-5 (Norit Americas Inc.,<br />
Atlanta, GA), was used <strong>in</strong> the BAG filters. This carbon is an extruded peat activated by steam. The<br />
30x40 fraction <strong>of</strong> Calgon F-400 (Calgon Inc., Pittsburgh, PA), lot number 5925-LG, was used for<br />
some k<strong>in</strong>etic studies. The F-400 is made from bitum<strong>in</strong>ous coal.<br />
30
METHODS UTILIZED FOR PERCHLORATE REDUCTION EXPERIMENTS<br />
Biologically Active Carbon Filters<br />
Four filters were constructed to evaluate perchlorate removal. Each filter conta<strong>in</strong>ed fresh<br />
carbon. BAG was formed by extensively preload<strong>in</strong>g the filters with effluent from the previously<br />
described bromate reduc<strong>in</strong>g filters. The <strong>in</strong>fluent was spiked with 50 ng/L perchlorate <strong>and</strong> 20 ug/L<br />
bromate. The columns were operated with a 25-m<strong>in</strong>ute EBCT for 600 bed volumes. At the end <strong>of</strong><br />
the preload<strong>in</strong>g period, approximately 60 percent <strong>of</strong> the <strong>in</strong>fluent DOC <strong>and</strong> 90 percent <strong>of</strong> the <strong>in</strong>fluent<br />
bromate were consistently be<strong>in</strong>g removed. Therefore, by the end <strong>of</strong> the preload<strong>in</strong>g period, the filters<br />
were considered biologically active. Two parallel filtration experiments were set up, each with two<br />
columns <strong>in</strong> series, for a total <strong>of</strong> four filters. The filtration experiments were <strong>of</strong>ten run with two filters<br />
<strong>in</strong> series, so that two EBCTs could be evaluated simultaneously. Once the filters were rendered<br />
biologically active, the <strong>in</strong>fluent base water for one set <strong>of</strong> filters was switched from bromate filter<br />
effluent to DDW. This allowed for better control over <strong>in</strong>fluent composition. The other set <strong>of</strong> filters<br />
utilized CUW as the base water for approximately one year. However, s<strong>in</strong>ce perchlorate removal was<br />
not consistent <strong>in</strong> these filters due to the occurrence <strong>of</strong> nitrification, the data from these filters are not<br />
reported here. The schematic for the biological bromate filters shown <strong>in</strong> Figure 2.3 also serves as the<br />
schematic for the biological perchlorate filters.<br />
Syr<strong>in</strong>ge pumps (kd Scientific, Boston, Massachusetts) add an electron donor mixture to the<br />
<strong>in</strong>fluent at a po<strong>in</strong>t shortly before the <strong>in</strong>fluent enters the first filter. The electron donor solution was<br />
either a mixture <strong>of</strong> acetate, lactate, <strong>and</strong> pyruvate or a mixture <strong>of</strong> acetate, benzoate, <strong>and</strong> pyruvate.<br />
Benzoate was sometimes substituted for lactate due to the fact that carboxylic acids are common<br />
ozonation by-products (Camel <strong>and</strong> Bermond 1998). S<strong>in</strong>ce the focus <strong>of</strong> this project is conventional<br />
ozone/<strong>GAC</strong> systems, the use <strong>of</strong> benzoate was very practical. The electron donor concentration was<br />
always 1.5 mg/L <strong>of</strong> equivalent nitrate dem<strong>and</strong>4 <strong>in</strong> excess <strong>of</strong> that required to stoichiometrically remove<br />
the <strong>in</strong>fluent concentrations <strong>of</strong> DO <strong>and</strong> nitrate. Thus, electron donor concentration was never the<br />
limit<strong>in</strong>g variable.<br />
4The equivalent nitrate dem<strong>and</strong> was determ<strong>in</strong>ed by calculat<strong>in</strong>g the concentration <strong>of</strong><br />
electron donor required to stoichiometrically reduce NO3" completely to N2 .<br />
31
Influent was prepared by add<strong>in</strong>g 50 ug/L <strong>of</strong> perchlorate, the appropriate nitrate concentration,<br />
<strong>and</strong> 1 mM phosphate buffer to DDW. Flowrates through the BAG filters were adjusted to provide<br />
the desired EBCT. Samples were taken at the <strong>in</strong>fluent, at the effluent <strong>of</strong> the first filter, <strong>and</strong> at the<br />
effluent <strong>of</strong> the second filter. Samples were stored at 4 °C until they were analyzed.<br />
Abiotic Batch Experiments<br />
Us<strong>in</strong>g ajar test apparatus (B-KER2, Model 7790-400, Phipps <strong>and</strong> Bird, Richmond, Virg<strong>in</strong>ia),<br />
a series <strong>of</strong> batch tests were run to determ<strong>in</strong>e if chemical reduction <strong>of</strong> perchlorate could be achieved<br />
us<strong>in</strong>g <strong>GAC</strong> under various conditions. Four different <strong>GAC</strong>s were used: 1) virg<strong>in</strong> Norit, 2) virg<strong>in</strong><br />
Calgon F-400,3) acid-washed outgassed (AWOG) Norit, <strong>and</strong> 4) AWOGCalgon F-400. The AWOG<br />
carbon was prepared by Kirisits (2000). AWOG carbon typically conta<strong>in</strong>s lower concentrations <strong>of</strong><br />
surface oxygen groups <strong>and</strong> has a higher pH (Miller 1996). The mean pellet diameter <strong>and</strong> length for<br />
the Norit carbon were 0.8 <strong>and</strong> 3.4 mm, respectively. The mean particle diameter for the Calgon F-<br />
400 carbon was approximately 0.7 mm. The carbon to be used <strong>in</strong> the batch studies was weighed <strong>and</strong><br />
stored <strong>in</strong> DDW overnight to wet it. Two liters <strong>of</strong> DDW or CUW was placed <strong>in</strong> each jar for the batch<br />
test studies. The pH <strong>of</strong> each solution was adjusted, <strong>and</strong> 1 mM phosphate buffer was added to prevent<br />
pH fluctuation. 50 ug/L <strong>of</strong> perchlorate was spiked <strong>in</strong>to the batch test jars. A total <strong>of</strong> 9 different<br />
experiments were run. The behavior <strong>of</strong> each <strong>of</strong> the four carbons was exam<strong>in</strong>ed <strong>in</strong> both CUW <strong>and</strong><br />
DDW at pH 7.5. The n<strong>in</strong>th experiment used DDW, virg<strong>in</strong> Norit carbon <strong>and</strong> pH 2. Once the n<strong>in</strong>e<br />
batch test solutions were appropriately prepared, 500 mg/L <strong>of</strong> <strong>GAC</strong> was added to each jar, <strong>and</strong> the<br />
stirrers were turned on to suspend the <strong>GAC</strong>. The experiments were run for 5 hours <strong>and</strong> were<br />
sampled regularly dur<strong>in</strong>g this time.<br />
Batch Test Wash Experiments<br />
Upon completion <strong>of</strong> the batch test experiments, the <strong>GAC</strong> from the four CUW experiments<br />
was removed <strong>and</strong> placed <strong>in</strong> separate glass columns equipped with stopcocks <strong>and</strong> glass wool to<br />
support the carbon. Solutions conta<strong>in</strong><strong>in</strong>g a nitrate concentration <strong>of</strong> 200 mg/L were used to wash the<br />
<strong>GAC</strong>. The effluent from the columns were collected <strong>in</strong> known volumes <strong>and</strong> analyzed for perchlorate.<br />
The <strong>GAC</strong> was washed until no perchlorate was detected <strong>in</strong> the effluent.<br />
32
Metal-Catalyzed GAG Filter Experiments<br />
The two filters constructed for the abiotic, metal-catalyzed experiments were similar to those<br />
used for the BAG filter experiments. One filter was composed <strong>of</strong> Calgon copper- <strong>and</strong> z<strong>in</strong>c oxideimpregnated<br />
carbon. The other filter conta<strong>in</strong>ed virg<strong>in</strong> Norit RO 0.8 carbon, to which alum<strong>in</strong>um shot<br />
was added (10% w/w). The alum<strong>in</strong>um was mixed thoroughly with the <strong>GAC</strong>, <strong>and</strong> the mixture was<br />
added to the column. Influent was made with the follow<strong>in</strong>g parameters: 1) DDW, 2) DO = ambient<br />
(-8.6 mg/L), 3) pH 2.5, 4) ClO/^, = 3 mg/L, NO3-<strong>in</strong>fluent = 0.1 mg/L (ambient). A 3 mg/L <strong>in</strong>fluent<br />
perchlorate concentration was used so that reduction products could be monitored. A 25-m<strong>in</strong>ute<br />
EBCT was used for each filter, <strong>and</strong> samples were taken at the <strong>in</strong>fluent <strong>and</strong> effluent <strong>of</strong> each column.<br />
Metal-Catalyzed <strong>GAC</strong> Filter Wash Experiments<br />
At the end <strong>of</strong> the aforementioned experiment, the <strong>GAC</strong> was r<strong>in</strong>sed with a 10 mg/L nitrate<br />
solution to wash <strong>of</strong>f any perchlorate or chloride that was ion-exchanged at the carbon surface. This<br />
procedure would help determ<strong>in</strong>e if ion exchange or chemical reduction was the mechanism <strong>of</strong><br />
perchlorate removal. The effluent from these tests was collected <strong>and</strong> analyzed for perchlorate.<br />
R<strong>in</strong>s<strong>in</strong>g cont<strong>in</strong>ued until no perchlorate was detected <strong>in</strong> the effluent. The mass <strong>of</strong> perchlorate desorbed<br />
was calculated <strong>and</strong> compared to the mass <strong>of</strong> perchlorate removed by each filter. Once it was<br />
determ<strong>in</strong>ed that not all the perchlorate had been recovered after the 10 mg/L nitrate r<strong>in</strong>se, a 200 mg/L<br />
nitrate r<strong>in</strong>se was run until no perchlorate was detected <strong>in</strong> the effluent.<br />
<strong>Ozone</strong>/Hydrogen Peroxide/<strong>Perchlorate</strong> Batch Tests<br />
Six batch experiments were set up to exam<strong>in</strong>e the effect <strong>of</strong> ozone <strong>and</strong> hydrogen peroxide on<br />
perchlorate reduction. The same jar test apparatus that was used <strong>in</strong> other batch tests was utilized for<br />
these experiments. Table 2.3 describes the experimental conditions for these experiments.<br />
Gaseous ozone was generated by a model GL-1 ozone generator (PCI <strong>Ozone</strong> Corporation,<br />
West Caldwell, New Jersey) from pure oxygen. <strong>Ozone</strong> was applied to the batch solutions as gas <strong>and</strong><br />
concentrations were determ<strong>in</strong>ed accord<strong>in</strong>g to the Indigo Colorimetric Method (4500-O3) described<br />
<strong>in</strong> St<strong>and</strong>ard Methods for the Exam<strong>in</strong>ation <strong>of</strong> Water <strong>and</strong> Wastewater, 19* Edition (1995).<br />
33
Table 2.3 Experimental conditions: ozone/hydrogen peroxide/perchlorate batch tests<br />
Experiment #<br />
1<br />
2<br />
3<br />
4<br />
5<br />
6<br />
Initial<br />
<strong>Perchlorate</strong><br />
46.6<br />
46.5<br />
46.2<br />
43.2<br />
43.5<br />
*<br />
42.8<br />
pH<br />
8.8<br />
8.8<br />
8.8<br />
8.8<br />
8.8<br />
8 .8<br />
<strong>Ozone</strong>:DOC<br />
(mg/mg)<br />
2<br />
2<br />
2<br />
3<br />
3<br />
3<br />
<strong>Ozone</strong> (mg/L)<br />
2.94<br />
2.94<br />
2.94<br />
4.41<br />
4.41<br />
4.41<br />
H2O2 : <strong>Ozone</strong><br />
(mg/mg)<br />
0.0<br />
0.5<br />
1.0<br />
0.0<br />
0.5<br />
1 .0<br />
Each jar was prepared by add<strong>in</strong>g appropriate concentrations <strong>of</strong> perchlorate <strong>and</strong> hydrogen<br />
peroxide to 2 liters <strong>of</strong> CUW. The appropriate amount <strong>of</strong> ozone from the batch solution was then<br />
added to the CUW, <strong>and</strong> samples were taken at time = 0, 5,10, <strong>and</strong> 15 m<strong>in</strong>utes. Samples were filtered<br />
(0.45 |am) prior to storage at 4 °C <strong>and</strong> analysis. No solutions or samples were quenched.<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis<br />
DGGE was used to <strong>in</strong>vestigate the diversity <strong>of</strong> the microbial population <strong>in</strong> the BAC filters.<br />
The same DGGE technique used for the bromate samples (described earlier) was used for the<br />
perchlorate samples.<br />
Plat<strong>in</strong>g Experiments<br />
The solid medium plat<strong>in</strong>g technique used to isolate perchlorate-reduc<strong>in</strong>g bacteria was the<br />
same as that used to isolate bromate-reduc<strong>in</strong>g bacteria (described earlier). To check for perchloratereduc<strong>in</strong>g<br />
ability, microorganisms isolated from the BAC filters were <strong>in</strong>oculated <strong>in</strong>to an aqueous<br />
medium conta<strong>in</strong><strong>in</strong>g 50 |ig/L perchlorate, 20 |ig/L bromate, 1.0 mg/L nitrate, 5 mM phosphate buffer,<br />
<strong>and</strong> 0 mg/L DO at pH 7.5. Electron donor was supplied at 1.5 mg/L <strong>of</strong> equivalent nitrate dem<strong>and</strong><br />
<strong>in</strong> excess <strong>of</strong> that required to stoichiometrically remove 1.0 mg/L nitrate. Figure 2.4 schematically<br />
shows the progression <strong>of</strong> these plat<strong>in</strong>g experiments. Follow<strong>in</strong>g <strong>in</strong>cubation, samples were taken for<br />
34
perchlorate, nitrite, nitrate, <strong>and</strong> bromate analyses. To verify the ability <strong>of</strong> pure cultures to reduce<br />
perchlorate, transfers were made to fresh medium <strong>and</strong> monitored. Individual isolates will be further<br />
characterized once they are confirmed active for bromate <strong>and</strong> perchlorate reduction.<br />
Dilution-to-Ext<strong>in</strong>ction Experiments<br />
The technique used for the perchlorate dilution-to-ext<strong>in</strong>ction experiments was the same as that<br />
used for the bromate dilution-to-ext<strong>in</strong>ction experiments (described earlier). The <strong>in</strong>oculum used for<br />
these experiments was effluent water from the BAG filter. The medium consisted <strong>of</strong> DDW, 50 ug/L<br />
perchlorate, 20 ng/L bromate, 1.0 mg/L nitrate, <strong>and</strong> 5 mM phosphate buffer. The electron donor<br />
concentration, comprised <strong>of</strong> acetate, lactate, <strong>and</strong> pyruvate, was 1.5 mg/L <strong>of</strong> equivalent nitrate dem<strong>and</strong><br />
<strong>in</strong> excess <strong>of</strong> that required to stoichiometrically reduce 1.0 mg/L nitrate. Each experimental condition<br />
was run <strong>in</strong> triplicate.<br />
Serum Bottle Experiments<br />
Serum bottle experiments were run as screen<strong>in</strong>g tests for the effect <strong>of</strong> various parameters on<br />
perchlorate reduction. 160 mL serum bottles, equipped with gas-tight, butyl rubber stoppers, were<br />
used as batch reactors. All serum bottle experiments were performed us<strong>in</strong>g a s<strong>in</strong>gle microbial species<br />
isolated dur<strong>in</strong>g the plat<strong>in</strong>g experiments, described <strong>in</strong> the previous section. This isolate, B9,<br />
demonstrated the ability to reduce 50 (ig/L perchlorate to below the detection limit after be<strong>in</strong>g<br />
<strong>in</strong>oculated to aqueous medium.<br />
Two different serum bottle experiments were run us<strong>in</strong>g isolate B9. Medium for the first<br />
experiment was prepared accord<strong>in</strong>g to Table 2.4. This experiment was designed to compare several<br />
different operat<strong>in</strong>g conditions: 1) nitrate concentration (0.0, 0.3, <strong>and</strong> 1.5 mg/L), 2) DDW amended<br />
with external electron donor added versus CUW with no external electron donor, 3) bromate<br />
concentration (0, 20 ug/L), 4) lactate versus benzoate as an electron donor. The results <strong>of</strong> this<br />
experiment led to the design <strong>of</strong> the second serum bottle experiment, as described <strong>in</strong> Table 2.5. This<br />
experiment focused on the difference <strong>in</strong> perchlorate reduction k<strong>in</strong>etics when the <strong>in</strong>itial nitrate<br />
concentration was varied.<br />
Us<strong>in</strong>g 5 mM phosphate buffer, each medium was ma<strong>in</strong>ta<strong>in</strong>ed at pH 7.5. 150 mL <strong>of</strong> each<br />
35
medium was placed <strong>in</strong>to a serum bottle <strong>and</strong> sparged with argon gas for 30 m<strong>in</strong>utes to remove the DO.<br />
Butyl stoppers were <strong>in</strong>serted <strong>and</strong> alum<strong>in</strong>um caps were crimped <strong>in</strong> place. The bottles were autoclaved<br />
<strong>and</strong> then cooled. To produce a uniform <strong>in</strong>oculum among serum bottles, a stock solution <strong>of</strong> biomass<br />
was first prepared. For this solution, 5 mL <strong>of</strong> DDW was placed <strong>in</strong>to a serum bottle, stoppered,<br />
capped, <strong>and</strong> then autoclaved. 0.5 mL <strong>of</strong> sterile DDW was aspirated on the plate conta<strong>in</strong><strong>in</strong>g the B9<br />
isolate to suspend the colonies. This 0.5 mL volume, conta<strong>in</strong><strong>in</strong>g the microbial suspension, was added<br />
to the other 4.5 mL sterile DDW. The bottle was then vortexed for several m<strong>in</strong>utes to create a<br />
uniform microbial suspension. 0.1 mL <strong>of</strong> this suspension was spiked to each serum bottle. Serum<br />
bottles with no <strong>in</strong>oculum were also prepared as controls for each medium. Bottles were manually<br />
shaken once per day, <strong>and</strong> samples were taken once, twice, or three times daily. Each serum bottle<br />
was run <strong>in</strong> duplicate.<br />
Table 2.4 Serum bottle experiment 1 us<strong>in</strong>g perchlorate-reduc<strong>in</strong>g isolate B9 (<strong>in</strong>itial conditions)<br />
Serum<br />
bottle<br />
#<br />
1<br />
2<br />
3<br />
4<br />
5<br />
6<br />
7<br />
8<br />
9<br />
10<br />
Water<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
CUW<br />
(dechlor<strong>in</strong>ated)<br />
50<br />
50<br />
50<br />
50<br />
50<br />
50<br />
50<br />
50<br />
50<br />
50<br />
cio4-<br />
NO3-<br />
(mg/L)<br />
0.0<br />
0.3<br />
1.5<br />
0.0<br />
0.3<br />
1.5<br />
0.0<br />
0.3<br />
1.5<br />
0.0<br />
Calculated concentration<br />
(mg/L as C)<br />
<strong>and</strong> type <strong>of</strong> electron donor<br />
2.17, acetate/lactate/pyruvate<br />
2.17, acetate/lactate/pyruvate<br />
2.17, acetate/lactate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
2.17, acetate/benzoate/pyruvate<br />
No External Donor Added<br />
^<br />
20<br />
20<br />
20<br />
20<br />
20<br />
20<br />
0<br />
0<br />
0<br />
20<br />
11 CUW 50 0.3<br />
(dechlor<strong>in</strong>ated)<br />
12 CUW 50 1.5<br />
(dechlor<strong>in</strong>ated)<br />
No External Donor Added<br />
No External Donor Added<br />
20<br />
20<br />
36
Table 2.5 Serum bottle experiment 2 us<strong>in</strong>g perchlorate-reduc<strong>in</strong>g isolate B9 (<strong>in</strong>itial conditions)<br />
Serum<br />
bottle #<br />
Water<br />
cio4-<br />
(Hg/L)<br />
N03-<br />
(mg/L)<br />
Calculated concentration (mg/L as C)<br />
<strong>and</strong> type <strong>of</strong> electron donor<br />
1<br />
DDW<br />
50<br />
0.1<br />
2.17, acetate/benzoate/pyruvate<br />
2<br />
DDW<br />
50<br />
0.4<br />
2.17, acetate/benzoate/pyruvate<br />
3<br />
DDW<br />
50<br />
1.6<br />
2.17, acetate/benzoate/pyruvate<br />
4<br />
DDW<br />
50<br />
.2.6<br />
2.17, acetate/benzoate/pyruvate<br />
ANALYTICAL METHODS<br />
<strong>Bromate</strong>, nitrite, bromide, chloride, nitrate, <strong>and</strong> sulfate were analyzed by ion chromatography.<br />
The ion chromatography system consisted <strong>of</strong> a Dionex Series 300 ion chromatograph (Sunnyvale,<br />
CA) with an anion self-regenerat<strong>in</strong>g suppressor (ASRS-1) <strong>and</strong> conductivity meter. Three columns<br />
were used <strong>in</strong> series - Dionex lonpac NG1, AG9-HC <strong>and</strong> AS9-HC. The purpose <strong>of</strong> the NGl was to<br />
adsorb organic material that could foul the analytical column. The analytical method <strong>in</strong>cluded a 9.0<br />
mM sodium carbonate eluant, 1.0 mL/m<strong>in</strong> flowrate, <strong>and</strong> 250 uL <strong>in</strong>jection loop. <strong>Perchlorate</strong> was also<br />
analyzed by ion chromatography. Three columns were used <strong>in</strong> series for perchlorate analysis -Dionex<br />
lonpac NGl, AG11, <strong>and</strong> AS 11. The analytical method for perchlorate <strong>in</strong>cluded a 100 mM sodium<br />
hydroxide eluant, 1.0 mL/m<strong>in</strong> flowrate, <strong>and</strong> a 250 uL <strong>in</strong>jection loop. All anions were detected by<br />
suppressed conductivity.<br />
The LOD for bromate <strong>and</strong> perchlorate were calculated by the method outl<strong>in</strong>ed by Skoog <strong>and</strong><br />
Leary (1992) to be approximately 2 ug/L. Any bromate or perchlorate concentration that appeared<br />
to be lower than 2 ug/L was recorded as be<strong>in</strong>g at one half <strong>of</strong> the LOD. Other studies have shown<br />
that this simple method <strong>of</strong> h<strong>and</strong>l<strong>in</strong>g concentrations below the LOD is as good or better than more<br />
complicated mathematical methods (Clarke 1998).<br />
DO was measured us<strong>in</strong>g a YSI Model 58 DO meter with a 5905 probe (Yellow Spr<strong>in</strong>gs, OH).<br />
An Orion 720A pH/Ion Selective Electrode Meter with an Orion ammonia probe 95-12 or pH probe<br />
(Beverly, MA) was used to measure ammonia <strong>and</strong> pH, respectively. DOC was analyzed with a<br />
Phoenix 8000 TOC Analyzer (Tekmar-Dohrmann, C<strong>in</strong>c<strong>in</strong>nati, OH) by UV-persulfate oxidation.<br />
The method detection limits were as follows: bromate = 2 ug/L, bromide
perchlorate = 2 ug/L, chloride
CHAPTER 3. RESULTS AND DISCUSSION<br />
BROMATE<br />
Verification <strong>of</strong> Biological <strong>Bromate</strong> <strong>Removal</strong> <strong>in</strong> the BAC Filters<br />
S<strong>in</strong>ce <strong>GAC</strong> can abiotically reduce bromate <strong>and</strong> can also serve as the attachment medium for<br />
microorganisms that use bromate as an electron acceptor, there exists the question <strong>of</strong> whether both<br />
processes are occurr<strong>in</strong>g simultaneously <strong>in</strong> the BAC filters. Due to the fact that the abiotic bromate<br />
reduction capacity <strong>of</strong> <strong>GAC</strong> is quickly exhausted <strong>in</strong> natural water (Kirisits et al. 1998), it seemed<br />
highly unlikely that the BAC filters were simultaneously support<strong>in</strong>g abiotic <strong>and</strong> biotic bromate<br />
reduction. However, to gather further evidence for the hypothesis that most <strong>of</strong> the observed bromate<br />
reduction <strong>in</strong> the BAC filters was <strong>in</strong>deed biological, repeat bromate removal experiments were run to<br />
verify the percent bromate removals obta<strong>in</strong>ed with the same BAC filters approximately one year ago.<br />
The 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCT BAC experiments first run <strong>in</strong> March 1998 were repeated <strong>in</strong> February<br />
1999 with the same BAC filters; if the percent bromate removals rema<strong>in</strong>ed the same, this would be<br />
an <strong>in</strong>dication that bromate removal <strong>in</strong> the BAC filters was biological s<strong>in</strong>ce it is doubtful that abiotic<br />
bromate reduction was still occurr<strong>in</strong>g 1.5 years after the filters were first constructed. The 10- <strong>and</strong><br />
20-m<strong>in</strong>ute EBCT BAC filtration experiments were repeated for <strong>in</strong>fluent nitrate concentrations <strong>of</strong> 5.0<br />
<strong>and</strong> 0.3 mg/L.<br />
Figure 3.1 shows the results for the BAC experiment with an <strong>in</strong>fluent nitrate concentration<br />
<strong>of</strong> 5.0 mg/L, <strong>and</strong> it demonstrates that the percent bromate removal did not change after an additional<br />
year <strong>of</strong> operation. Thus, only biological removal was occurr<strong>in</strong>g <strong>in</strong> both March 1998 <strong>and</strong> February<br />
1999. However, when the same comparison was performed for the BAC experiment with an <strong>in</strong>fluent<br />
nitrate concentration <strong>of</strong> 0.3 mg/L, the percent removal <strong>in</strong> February 1999 did not correspond to the<br />
percent removal <strong>in</strong> March 1998 (Figure 3.2). This is not <strong>in</strong>terpreted to mean that abiotic bromate<br />
reduction was occurr<strong>in</strong>g <strong>in</strong> the experiments with a 0.3 mg/L <strong>in</strong>fluent nitrate concentration <strong>in</strong> March<br />
1998, which would have artificially <strong>in</strong>flated the percent bromate removals. Rather, these results are<br />
evidence <strong>of</strong> a microbial disturbance that occurred <strong>in</strong> these filters <strong>in</strong> December 1998. In December<br />
1998, the EBCT <strong>in</strong> the BAC filters with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L was <strong>in</strong>creased<br />
to approximately 115 m<strong>in</strong>utes so that they could run with a s<strong>in</strong>gle batch <strong>of</strong> <strong>in</strong>fluent for<br />
39
100<br />
80 -<br />
• March 1998<br />
D February 1999<br />
5 60 H<br />
S<br />
40 -<br />
20 -<br />
0<br />
0 10 15<br />
EBCT (m<strong>in</strong>)<br />
20 25<br />
Figure 3.1 Repeat <strong>of</strong> the 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCT filtration experiment (average <strong>in</strong>fluent<br />
conditions: CUW, pH 7.5, 2 mg/L DO, 5.0 mg/L NO3', 20 ug/L BrO3')<br />
100<br />
80<br />
• March 1998<br />
D February 1999<br />
£ 60<br />
a two-week period. Over the two-week period, the flowrate discharged by the peristaltic pump<br />
decreased so that the EBCT was well over 200 m<strong>in</strong>utes. It is possible that the extended contact time<br />
<strong>in</strong> the filter resulted <strong>in</strong> much lower DOC concentrations reach<strong>in</strong>g the lower portions <strong>of</strong> the filter,<br />
caus<strong>in</strong>g that microbial community to die or to change composition. This microbial disturbance had<br />
apparently not been rectified by the time the verification study was run <strong>in</strong> February 1999. Even<br />
though these filters have been operated for an additional year s<strong>in</strong>ce the disturbance, they have been<br />
unable to reproduce the bromate removals orig<strong>in</strong>ally observed <strong>in</strong> March 1998. This <strong>in</strong>dicates that<br />
filter history has a very important effect on biological bromate removal. The microbial community<br />
that is developed under a particular water quality condition might affect the bromate removal obta<strong>in</strong>ed<br />
even after the water quality conditions become more favorable for bromate removal.<br />
To further <strong>in</strong>vestigate the performance <strong>of</strong> the BAG filters that had undergone a disturbance<br />
(Figure 3.2), the <strong>in</strong>fluent nitrate concentration to these filters was switched to 5.0 mg/L (Figure 3.3).<br />
In Figure 3.3, the series entitled "Orig<strong>in</strong>al 5.0 mg/L <strong>in</strong>fluent nitrate filters" refers to the BAC filters<br />
that have been operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L for two years; the series<br />
entitled "Filters switched from 0.3 to 5.0 mg/L <strong>in</strong>fluent nitrate" refers to the BAC filters that were<br />
operated for 2 years with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L but were recently switched<br />
to an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L. There is no statistically significant difference 1 <strong>in</strong><br />
bromate removal between the two sets <strong>of</strong> filters (Figure 3.3). These data could <strong>in</strong>dicate a difference<br />
<strong>in</strong> the microbial communities that were first developed <strong>in</strong> the two sets <strong>of</strong> filters. That is, when the<br />
BAG filters operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L were first run, a nitratesensitive<br />
microbial community may have developed. This community may have washed out dur<strong>in</strong>g<br />
the two-week disturbance <strong>in</strong> December 1998 <strong>and</strong> has not been able to re-establish itself s<strong>in</strong>ce that<br />
time.<br />
Further Test<strong>in</strong>g the Reproducibility <strong>of</strong> BAG Filters<br />
To test the reproducibility <strong>of</strong> the data obta<strong>in</strong>ed with the BAG filter operated with a 0.3 mg/L<br />
<strong>in</strong>fluent nitrate concentration, a new BAG filter was developed. This BAG filter was constructed <strong>in</strong><br />
a larger column (2-<strong>in</strong>ch <strong>in</strong>ner diameter, Figure 2.3) as compared to the smaller columns used for other<br />
BAG filtration experiments (1-<strong>in</strong>ch <strong>in</strong>ner diameter, Figure 2.1). Due to the multi-port configuration<br />
<strong>of</strong> this filter (Figure 2.3), EBCTs <strong>of</strong> 15, 30, 45, <strong>and</strong> 60 m<strong>in</strong>utes were simultaneously tested. Figure<br />
3.4 compares the bromate removal results from the 1 -<strong>in</strong>ch <strong>in</strong>ner diameter filter (prior to the December<br />
1998 disturbance) to the results from the 2-<strong>in</strong>ch <strong>in</strong>ner diameter filter <strong>and</strong> demonstrates the<br />
reproducibility <strong>of</strong> the data.<br />
S<strong>in</strong>ce Figure 3.4 demonstrates that a one-<strong>in</strong>ch <strong>in</strong>ner diameter BAG filter produces the same<br />
results as a 2-<strong>in</strong>ch <strong>in</strong>ner diameter BAG filter, this suggests that wall effects are m<strong>in</strong>imal. This is a<br />
useful result from an operational st<strong>and</strong>po<strong>in</strong>t because the smaller filter requires preparation <strong>of</strong> smaller<br />
volumes <strong>of</strong> <strong>in</strong>fluent water.<br />
Backwash<strong>in</strong>g BAG Filters<br />
S<strong>in</strong>ce the filters utilized for bromate removal did not accumulate appreciable head loss (as<br />
evidenced by a constant flow rate across the filter for a given pump<strong>in</strong>g rate), they were not rout<strong>in</strong>ely<br />
backwashed. However, the question arose as to whether bromate removal would decrease due to<br />
backwash<strong>in</strong>g.<br />
'Statistical significance was based on overlap <strong>of</strong> 95 percent confidence <strong>in</strong>tervals.<br />
42
To establish the effect <strong>of</strong> backwash<strong>in</strong>g on filter performance, two filters operat<strong>in</strong>g with<br />
CUW, a 25-m<strong>in</strong>ute EBCT, <strong>and</strong> an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 or 5.0 mg/L were chosen for<br />
backwash<strong>in</strong>g. Prior to backwash<strong>in</strong>g, the filters were run to obta<strong>in</strong> the pre-backwash<strong>in</strong>g bromate<br />
removal percentages. Then the filters were backwashed <strong>and</strong> run aga<strong>in</strong> to obta<strong>in</strong> the postbackwash<strong>in</strong>g<br />
bromate removal percentages.<br />
As shown <strong>in</strong> Table 3.1, backwash<strong>in</strong>g did not negatively impact the bromate removal rate.<br />
It should be noted that the backwash<strong>in</strong>g procedure removed a substantial amount <strong>of</strong> microbial<br />
material from the filters; the backwash water was visually, turbid <strong>and</strong> conta<strong>in</strong>ed floes.<br />
100<br />
^ 80<br />
o 60<br />
Samples were collected for heterotrophic plate counts for the follow<strong>in</strong>g waters: <strong>in</strong>fluent<br />
water, pre-backwash filter effluent, post-backwash filter effluent, <strong>and</strong> the backwash water. Plate<br />
counts were performed accord<strong>in</strong>g to the membrane filter method. However, the volume <strong>of</strong> water<br />
filtered through the membranes resulted <strong>in</strong> extensive microbial growth, <strong>and</strong> <strong>in</strong>dividual colonies were<br />
too numerous to count (TNTC).<br />
Effect <strong>of</strong> the Initial <strong>Bromate</strong> Concentration<br />
The mass concentration <strong>of</strong> bromate removed by the BAG filters was monitored as the <strong>in</strong>fluent<br />
bromate concentrations <strong>in</strong>creased from 10 to 40 ug/L with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3<br />
mg/L (Figure 3.5) <strong>and</strong> from 10 to 50 ug/L with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L (Figure<br />
3.6). Figure 3.5 shows constant mass concentration bromate removal as the bromate concentration<br />
was <strong>in</strong>creased from 20 to 40 ug/L <strong>in</strong> the 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCT filters which <strong>in</strong>dicates that the<br />
microorganisms <strong>in</strong> the BAG filters were not able to reduce additional bromate at the 10- <strong>and</strong> 20-<br />
m<strong>in</strong>ute EBCTs as the <strong>in</strong>fluent bromate concentration <strong>in</strong>creased. One may have expected that an<br />
<strong>in</strong>crease <strong>in</strong> the <strong>in</strong>fluent bromate concentration would have <strong>in</strong>creased the driv<strong>in</strong>g force for bromate <strong>in</strong><br />
the bulk solution to diffuse deeper with<strong>in</strong> the bi<strong>of</strong>ilm, thereby <strong>in</strong>creas<strong>in</strong>g the mass <strong>of</strong> bromate<br />
removed. However, it seems that bromate reduction under these conditions (10- <strong>and</strong> 20- m<strong>in</strong>ute<br />
EBCTs <strong>and</strong> an <strong>in</strong>fluent bromate concentration between 20-40 ug/L) is limited by the reaction rate,<br />
which may <strong>in</strong>dicate that the electron donor concentration is the limit<strong>in</strong>g factor. Figure 3.5 shows that<br />
the mass concentration <strong>of</strong> bromate removed <strong>in</strong>creased as the <strong>in</strong>fluent bromate concentration <strong>in</strong>creased<br />
from 10 to 20 ug/L <strong>in</strong> the 25- <strong>and</strong> 50-m<strong>in</strong>ute EBCT filters, <strong>in</strong>dicat<strong>in</strong>g that the electron donor<br />
concentration was not limit<strong>in</strong>g under these conditions.<br />
Similar experiments were carried out <strong>in</strong> the BAG filters with an <strong>in</strong>fluent nitrate concentration<br />
<strong>of</strong> 5.0 mg/L. Figure 3.6 shows a general <strong>in</strong>crease <strong>in</strong> mass concentration bromate removal as the<br />
<strong>in</strong>fluent bromate concentration was <strong>in</strong>creased from 10 to 50 ug/L for 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCTs.<br />
However, as the <strong>in</strong>fluent bromate concentration was <strong>in</strong>creased from 40 to 50 ug/L, the data po<strong>in</strong>ts<br />
show that constant mass concentration bromate removal was observed <strong>in</strong> the 10- <strong>and</strong> 20-m<strong>in</strong>ute<br />
EBCT filters, at the 95 percent confidence level. This <strong>in</strong>dicates that the microorganisms <strong>in</strong> the BAG<br />
filters were not able to reduce additional bromate at the 10- <strong>and</strong> 20-m<strong>in</strong>ute EBCTs when the <strong>in</strong>fluent<br />
44
o<br />
S<br />
25<br />
20 -<br />
15 -<br />
10 -<br />
O 50-m<strong>in</strong> EBCT<br />
• 25-m<strong>in</strong> EBCT<br />
D 20-rrdn EBCT<br />
• 10-m<strong>in</strong>EBCT<br />
I 2<br />
ffl<br />
5 -<br />
0<br />
0 10 20 30 40 50<br />
Influent <strong>Bromate</strong> Concentration (u,g/L)<br />
Figure 3.5 Mass concentration bromate removal with different <strong>in</strong>fluent bromate<br />
concentrations (average <strong>in</strong>fluent conditions: CUW, pH 7.5, 2 mg/L DO, 0.3 mg/L NO3")<br />
20 -T<br />
15 -<br />
10 -<br />
o 50-m<strong>in</strong> EBCT<br />
• 25-m<strong>in</strong> EBCT<br />
D 20-m<strong>in</strong> EBCT<br />
• 10-m<strong>in</strong>EBCT<br />
2<br />
CQ<br />
5 -<br />
10 20 30 40<br />
Influent <strong>Bromate</strong> Concentration (|j,g/L)<br />
50 60<br />
Figure 3.6 Mass concentration bromate removal with different <strong>in</strong>fluent bromate concentrations<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 2 mg/L DO, 5.0 mg/L NO3")<br />
bromate concentration was above 40 ug/L. As discussed previously, this may <strong>in</strong>dicate that the<br />
electron donor concentration is limit<strong>in</strong>g bromate removal.<br />
Figure 3.6 also shows that the mass concentration <strong>of</strong> bromate removed <strong>in</strong>creased as the<br />
<strong>in</strong>fluent bromate concentration <strong>in</strong>creased from 10 to 20 ug/L <strong>in</strong> the 25- <strong>and</strong> 50-m<strong>in</strong>ute EBCT filters,<br />
45
<strong>in</strong>dicat<strong>in</strong>g that the electron donor concentration was not limit<strong>in</strong>g under these conditions. Hijnen et<br />
al. (1999) also observed an <strong>in</strong>crease <strong>in</strong> bromate removal <strong>in</strong> a denitrify<strong>in</strong>g bioreactor as the <strong>in</strong>fluent<br />
bromate concentration <strong>in</strong>creased. S<strong>in</strong>ce ethanol was fed to the bioreactor, the concentration <strong>of</strong><br />
electron donor did not limit bromate removal <strong>in</strong> that study.<br />
Effect <strong>of</strong> the Influent Dissolved Oxygen Concentration<br />
To evaluate the effect <strong>of</strong> DO concentration on bromate removal, the BAG filters were<br />
•operated with CUW <strong>and</strong> vary<strong>in</strong>g DO concentrations (Table 3.2). The <strong>in</strong>fluent nitrate concentration<br />
was relatively constant for all experiments (4.6 -5.1 mg/L), although the nitrate concentration <strong>in</strong> the<br />
filters <strong>in</strong>creased due to nitrification <strong>of</strong> the <strong>in</strong>fluent ammonia (Table 3.2). The data <strong>in</strong> Table 3.2 show<br />
that the effluent DO concentration was below the detection limit (0.1 mg/L) when the <strong>in</strong>fluent DO<br />
concentration was less than or equal to 5.8 mg/L. When the <strong>in</strong>fluent DO concentration was raised<br />
to 8.0 mg/L <strong>and</strong> higher, the effluent DO concentration was above the detection limit.<br />
Table 3.2 The effect <strong>of</strong> <strong>in</strong>fluent DO concentration on bromate removal <strong>in</strong> BAC filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ug/L BrO3")<br />
Average DO Average nitrate<br />
(mg/L)<br />
Expt. EBCT Influent Effluent<br />
Number (m<strong>in</strong>)<br />
Influent Effluent Average bromate removal<br />
with 95% confidence limits<br />
1 10<br />
20<br />
2.1
In general, the average percent bromate removal decreased as the <strong>in</strong>fluent DO concentration<br />
<strong>in</strong>creased (Table 3.2 <strong>and</strong> Figure 3.7). This was expected s<strong>in</strong>ce oxygen, a common electron acceptor,<br />
may <strong>in</strong>hibit the expression or activity <strong>of</strong> the enzymes required for bromate reduction. Additionally,<br />
the <strong>in</strong>creased DO concentration could cause more competition for electron donor. Note that the<br />
average percent bromate removals cannot be considered statistically different at the 95 percent<br />
confidence level for experiments with <strong>in</strong>fluent DO concentrations between 3.8 <strong>and</strong> 13.6 mg/L.<br />
However, there is a statistically significant decrease <strong>in</strong> bromate removal, at the 95 percent confidence<br />
level, as the <strong>in</strong>fluent DO concentration was <strong>in</strong>creased from 2.1 to 3.8 mg/L.<br />
The average percent bromate removal observed <strong>in</strong> the experiment with an <strong>in</strong>fluent DO<br />
concentration <strong>of</strong> 13.6 mg/L was only 11 percent after a 21-m<strong>in</strong>ute EBCT (Table 3.2). Even though<br />
nitrification <strong>and</strong> aerobic respiration consumed 8.5 mg/L DO dur<strong>in</strong>g the 21-m<strong>in</strong>ute EBCT, the<br />
rema<strong>in</strong><strong>in</strong>g DO (> 5.1 mg/L throughout the filter) prevented significant bromate removal. <strong>Bromate</strong><br />
removal may also have been prevented by electron donor limitation due to electron donor<br />
consumption by aerobic respiration. Therefore, it is likely that full-scale ozone-B AC plants are not<br />
observ<strong>in</strong>g biological bromate reduction <strong>in</strong> their BAG filters because the DO concentration is too high.<br />
Treatment systems that generate ozone from pure oxygen may have <strong>in</strong>fluent DO concentrations<br />
o<br />
50<br />
40 -<br />
30 -<br />
20 -<br />
• 2.0 mg/L <strong>in</strong>fluent DO<br />
D 3.9 mg/L <strong>in</strong>fluent DO<br />
• 5.9 mg/L <strong>in</strong>fluent DO<br />
• 8.2 mg/L <strong>in</strong>fluent DO<br />
X 13.8 mg/L <strong>in</strong>fluent DO<br />
2 10 H<br />
CQ<br />
0<br />
0 10 15<br />
20 25<br />
EBCT (m<strong>in</strong>)<br />
Figure 3.7 Percent bromate removal with various <strong>in</strong>fluent DO concentrations (average<br />
<strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ug/L Br03", 5 mg/L NO3')<br />
47
greater than 10 mg/L. For example, <strong>in</strong> the full-scale, post-ozonation BAG filters at the Paul M. Neal<br />
Water Treatment Facility (Lake Bluff, IL), bromate is not reduced from the <strong>in</strong>fluent concentration<br />
(less than 6 ug/L) after a 15-m<strong>in</strong>ute EBCT (Soucie 1999); this is likely due to the fact that the<br />
average <strong>in</strong>fluent <strong>and</strong> effluent DO concentrations are 11.5 <strong>and</strong> 10.7 mg/L, respectively.<br />
Tiedje (1988) noted that the expression <strong>and</strong> activity <strong>of</strong> the enzymes required for<br />
denitrification are <strong>of</strong>ten affected by the DO concentration. While denitrification is primarily thought<br />
<strong>of</strong> as an anoxic process, denitrification has also been observed under aerobic conditions (Robertson<br />
<strong>and</strong> Kuenen 1984). In Thiosphaera pantotropha (now Paracoccus denitrificans [Ludwig et al.<br />
1993]), Bell et al. (1990) found that the nitrate reductases <strong>in</strong>volved under aerobic conditions are<br />
different from those used under anaerobic conditions. A membrane bound nitrate reductase is<br />
primarily expressed under anaerobic conditions; a periplasmic nitrate reductase is primarily<br />
expressed under aerobic conditions but is active under aerobic <strong>and</strong> anaerobic conditions. Ye et al.<br />
(1994) noted that aerobic denitrifiers, such as Paracoccus denitrificans, are rare organisms.<br />
However Lloyd et al. (1987) demonstrated aerobic denitrification with eight bacterial isolates <strong>and</strong><br />
concluded that aerobic denitrification may be common <strong>in</strong> the environment. Bell et al. (1990)<br />
observed that the aerobic rate <strong>of</strong> nitrate reduction was approximately one tenth <strong>of</strong> the anaerobic rate<br />
<strong>of</strong> nitrate reduction; thus, the slower rate <strong>of</strong> aerobic denitrification may make aerobic denitrifiers<br />
more difficult to identify. It is possible that aerobic denitrifiers are present <strong>in</strong> the BAG filters used<br />
<strong>in</strong> the current study s<strong>in</strong>ce bromate reduction was observed <strong>in</strong> the presence <strong>of</strong> DO (Experiments 4 <strong>and</strong><br />
5 <strong>in</strong> Table 3.2) for 36 days. The low percent bromate removal observed under aerobic conditions<br />
would be consistent with the slow rate <strong>of</strong> aerobic denitrification observed by Bell et al. (1990).<br />
The effluent nitrate concentration <strong>in</strong>creased as the <strong>in</strong>fluent DO concentration <strong>in</strong>creased due<br />
to biological nitrification <strong>of</strong> ammonia (Table 3.2). However, note that <strong>in</strong> Experiments 4 <strong>and</strong> 5, the<br />
average effluent nitrate from the 21-m<strong>in</strong>ute EBCT filter was the same as or close to the effluent<br />
concentration from the 10-m<strong>in</strong>ute EBCT filter. This <strong>in</strong>dicates that all <strong>of</strong> the nitrification took place<br />
<strong>in</strong> the first (10-m<strong>in</strong>ute EBCT) filter for these experiments.<br />
In general, the measured DO consumption (Table 3.2) was greater than the calculated<br />
consumption <strong>of</strong> DO required for the observed nitrate production. This <strong>in</strong>dicates that some <strong>of</strong> the DO<br />
was also used <strong>in</strong> aerobic respiration. Also, the ammonia consumption can account for the observed<br />
<strong>in</strong>crease <strong>in</strong> nitrate concentration. On average, there was only a 6 percent difference between the<br />
48
measured nitrate concentration <strong>and</strong> the calculated nitrate concentration, based on ammonia<br />
consumption (data not shown).<br />
There is no evidence that significant denitrification took place <strong>in</strong> the filters, s<strong>in</strong>ce the <strong>in</strong>fluent<br />
nitrogen (ammonia, nitrite, <strong>and</strong> nitrate) was equal to the effluent nitrogen (data not shown). Based<br />
on the results <strong>of</strong> Hijnen et al. (1999), one might expect a faster rate <strong>of</strong> nitrate reduction as compared<br />
to bromate reduction s<strong>in</strong>ce the nitrate concentration was two orders <strong>of</strong> magnitude greater than the<br />
bromate concentration <strong>in</strong> the BAG filters. Due to the low concentration <strong>of</strong> biodegradable organic<br />
matter <strong>in</strong> dr<strong>in</strong>k<strong>in</strong>g water, it is likely that microbial activity - <strong>in</strong>clud<strong>in</strong>g denitrification - was limited<br />
by the electron donor concentration. Additionally, a low concentration <strong>of</strong> electron donor can result<br />
<strong>in</strong> measurable effluent DO (Experiments 4 <strong>and</strong> 5, Table 3.2) which can also <strong>in</strong>hibit the enzymes<br />
required for denitrification. These po<strong>in</strong>ts are illustrated <strong>in</strong> a later section (see - Effect <strong>of</strong> Source<br />
Water Type) where significant nitrate removal was observed upon addition <strong>of</strong> an external electron<br />
donor to a dr<strong>in</strong>k<strong>in</strong>g water source.<br />
Effect <strong>of</strong> the Influent Nitrate Concentration<br />
Kirisits <strong>and</strong> Snoey<strong>in</strong>k (1999) showed a decrease <strong>in</strong> bromate removal as the <strong>in</strong>fluent nitrate<br />
concentration <strong>in</strong>creased over a relatively narrow concentration range (0.3 to 5.0 mg/L). Figures 3.8<br />
<strong>and</strong> 3,9 exp<strong>and</strong> this result to show how bromate removal decreased as the <strong>in</strong>fluent nitrate<br />
concentration <strong>in</strong>creased from 0.3 to 42.3 mg/L. With a 26-m<strong>in</strong>ute EBCT, bromate removal dropped<br />
from 86 to 49 percent as the <strong>in</strong>fluent nitrate concentration <strong>in</strong>creased from 0.3 to 42.3 mg/L (Figure<br />
3.8). It should be noted that the 86 percent removal performance was susta<strong>in</strong>able (data po<strong>in</strong>t covers<br />
45 days <strong>of</strong> flow). With a 51-m<strong>in</strong>ute EBCT, bromate removal decreased from 95 to 79 percent for<br />
the same <strong>in</strong>crease <strong>in</strong> nitrate concentration (Figure 3.9). The change <strong>in</strong> bromate removal was<br />
seem<strong>in</strong>gly smaller for the 51-m<strong>in</strong>ute EBCT than for the 26-m<strong>in</strong>ute EBCT (16 percent decrease<br />
versus 37 percent decrease) because <strong>of</strong> the first two po<strong>in</strong>ts <strong>in</strong> Figure 3.9. These po<strong>in</strong>ts show 95<br />
percent bromate removal, which is removal to the detection limit. In these experiments, it is likely<br />
that the BAG filters could have removed more than the 20 |ag/L bromate present <strong>in</strong> the <strong>in</strong>fluent.<br />
Therefore, the 51-m<strong>in</strong>ute EBCT seemed to dampen the negative effect <strong>of</strong> an <strong>in</strong>creased nitrate<br />
concentration due to bromate limitation for the experiments with <strong>in</strong>fluent nitrate concentrations <strong>of</strong><br />
0.3 <strong>and</strong> 5.0 mg/L.<br />
49
ffl<br />
10 - Immediately preced<strong>in</strong>g these experiments, the filters were<br />
exposed to an <strong>in</strong>fluent DO concentration <strong>of</strong> 13.6 mg/L DO for 2 weeks<br />
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0<br />
40.0 45.0<br />
Influent Nitrate Concentration (mg/L)<br />
Figure 3.8 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate removal <strong>in</strong> BAG filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ng/L BrO3', 2.1 mg/L DO,<br />
26-m<strong>in</strong>ute EBCT)<br />
£<br />
13<br />
o<br />
s<br />
O<br />
90<br />
80 H<br />
70<br />
60<br />
50 H<br />
40<br />
30 -<br />
20 -<br />
10 -<br />
0<br />
0.0<br />
Immediately preced<strong>in</strong>g these experiments, the filters were<br />
exposed to an <strong>in</strong>fluent DO concentration <strong>of</strong> 13.6 mg/L for 2 weeks<br />
—I————————,———————!———————!————————!————————!————————,————————!——————<br />
5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0<br />
Influent Nitrate Concentration (mg/L)<br />
Figure 3.9 Effect <strong>of</strong> <strong>in</strong>fluent nitrate concentration on bromate removal <strong>in</strong> BAG filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 20 ug/L BrtV, 2.1 mg/L DO,<br />
51 -m<strong>in</strong>ute EBCT)<br />
50
Two data po<strong>in</strong>ts <strong>in</strong> each <strong>of</strong> Figures 3.8 <strong>and</strong> 3.9 are denoted as outliers (<strong>in</strong>fluent nitrate<br />
concentrations <strong>of</strong> 10.9 <strong>and</strong> 32.1 mg/L). These experiments were run with a different set <strong>of</strong> BAG<br />
filters than the rest <strong>of</strong> the experiments shown <strong>in</strong> Figures 3.8 <strong>and</strong> 3.9. Prior to the experiments that<br />
generated these data po<strong>in</strong>ts, the filters were exposed to an <strong>in</strong>fluent DO concentration <strong>of</strong> 13.6 mg/L<br />
for two weeks (Experiment 5, Table 3.2), whereas the other data po<strong>in</strong>ts were generated from filters<br />
that had always been exposed to an <strong>in</strong>fluent DO concentration <strong>of</strong> 2 mg/L. It is likely that filter history<br />
had an effect on subsequent experiments, s<strong>in</strong>ce these data po<strong>in</strong>ts are significantly below the curve<br />
fitted through the other data <strong>in</strong> Figures 3.8 <strong>and</strong> 3.9. The negative effect <strong>of</strong> higher DO concentrations<br />
on the bromate-reduc<strong>in</strong>g bacteria was not immediately reversible. These data suggest that the DO<br />
concentration <strong>of</strong> the backwash water must be carefully considered. Optimally, the backwash water<br />
should have a low DO concentration <strong>in</strong> order to m<strong>in</strong>imize its deleterious effect on bromate-reduc<strong>in</strong>g<br />
bacteria. However, the effects <strong>of</strong> short-term, high DO concentrations on biological bromate removal<br />
have not yet been assessed; the effect on bromate removal would likely depend on the frequency <strong>and</strong><br />
duration <strong>of</strong> backwash. Simpk<strong>in</strong> <strong>and</strong> Boyle (1988) studied the effect <strong>of</strong> aerobic/anoxic cycles on the<br />
synthesis <strong>of</strong> nitrate <strong>and</strong> nitrite reductase <strong>in</strong> activated sludge <strong>and</strong> found that these enzymes were<br />
synthesized to at least 50 percent <strong>of</strong> their maximum levels; they concluded that the repression <strong>of</strong><br />
nitrate reductase <strong>and</strong> nitrite reductase synthesis by oxygen was not as important as the effect <strong>of</strong><br />
oxygen on the enzyme activities. Thus, if nitrate reductase is <strong>in</strong>volved <strong>in</strong> bromate reduction, it is<br />
possible that a short aerated backwash would not be detrimental to bromate removal. However, the<br />
current study showed that the operation <strong>of</strong> the BAG filters with an <strong>in</strong>fluent DO concentration <strong>of</strong> 13.6<br />
mg/L for two weeks was detrimental <strong>in</strong> subsequent bromate removal experiments.<br />
Effect <strong>of</strong> the Influent Sulfate Concentration<br />
In contrast to bromate, nitrate, <strong>and</strong> DO - which are high potential term<strong>in</strong>al electron acceptors<br />
(£ 0/gr0 -/Br -=+1.03F; £°'0///0=+0.82F; E°'NO -/NO -=+OA3V)- sulfate is a low potential term<strong>in</strong>al<br />
electron acceptor E° so ^-/HS -=-0.22V. Thus, sulfate reduction would tend to take place after<br />
bromate, nitrate, <strong>and</strong> DO have been depleted. Based on this, the sulfate concentration was not<br />
expected to have a considerable impact on bromate reduction <strong>in</strong> the BAG filter. To test this<br />
hypothesis, a BAG filter was operated with <strong>in</strong>fluent sulfate concentrations <strong>of</strong> 11.1 <strong>and</strong> 102.7 mg/L.<br />
Sulfate reduction did not take place <strong>in</strong> any <strong>of</strong> the experiments, s<strong>in</strong>ce the <strong>in</strong>fluent <strong>and</strong> effluent sulfate<br />
51
concentrations were equal (Figure 3.10). As the sulfate concentration <strong>in</strong>creased, the average percent<br />
bromate removal decl<strong>in</strong>ed just slightly <strong>in</strong> the BAC with the 25- <strong>and</strong> 50-m<strong>in</strong>ute EBCTs (Figure 3.10).<br />
This is similar to the f<strong>in</strong>d<strong>in</strong>gs <strong>of</strong> Chen <strong>and</strong> Hao (1996), who observed that 120 mg/L sulfate slightly<br />
<strong>in</strong>hibited microbial chromate reduction even though no sulfate reduction was observed.<br />
Effect <strong>of</strong> pH<br />
It is useful to evaluate the sensitivity <strong>of</strong> bromate-reduc<strong>in</strong>g microorganisms to pH changes, so<br />
that conditions for biological bromate reduction can be optimized. To assess the effect <strong>of</strong> pH on<br />
bromate reduction, four BAC filtration experiments were run with pH values <strong>of</strong> 6.8,7.2,7.5, <strong>and</strong> 8.2.<br />
The system was buffered with 1-2 mM phosphate buffer, <strong>and</strong> the effluent pH values deviated from<br />
the <strong>in</strong>fluent values by less than 0.1 pH units. <strong>Bromate</strong> removal generally decreased as the <strong>in</strong>fluent pH<br />
<strong>in</strong>creased from 6.8 to 8.2 (Figure 3.11). The percent bromate removals observed at pH values <strong>of</strong> 6.8<br />
<strong>and</strong> 7.2 are not different at the 95 percent confidence level, which might <strong>in</strong>dicate that the optimum<br />
pH for bromate removal is <strong>in</strong> this region.<br />
1 o1<br />
4><br />
1 O<br />
ffl<br />
90 -<br />
80 -<br />
70 -<br />
60 -<br />
50 -<br />
40 -<br />
D 50-m<strong>in</strong>, bromate • 25-rrr<strong>in</strong>, bromate X <strong>in</strong>fluent sulfate<br />
O 50-m<strong>in</strong>, sulfate — 25-rrr<strong>in</strong>, sulfate<br />
f--------.j----..........^ ;<br />
^ —— ~~~— — --*—— _ *<br />
T~~— —— — — — r<br />
- 180.0 c<br />
r 160.0 -B<br />
- 140.0 |<br />
- 120.0 g u<br />
- 100.0 i<br />
- 80.0 3<br />
X<br />
en<br />
30 -<br />
a<br />
- 60.0 |<br />
20 -<br />
- 40.0 $|<br />
10 -<br />
20.0<br />
*<br />
0 - 11111<br />
- 0.0<br />
0.0 20.0 40.0 60.0 80.0 100.0 120.0<br />
Influent Sulfate Concentration (mg/L)<br />
Figure 3.10 Effect <strong>of</strong> <strong>in</strong>fluent sulfate concentration on bromate removal <strong>in</strong> BAC filters<br />
(average <strong>in</strong>fluent conditions: CUW, pH 7.5, 21 ug/L BrO3", 2.1 mg/L DO, 0.3 mg/L NO/;<br />
52
90 -<br />
^ 80 -<br />
70 -<br />
J 60 - o<br />
11<br />
1<br />
50 -<br />
40 -<br />
30 -<br />
20 -<br />
10 -<br />
n -<br />
• 50-m<strong>in</strong>EBCT<br />
D25-m] mEBCT<br />
4 ' 5<br />
f ' ——<br />
i * i<br />
6.5 7.5<br />
pH<br />
i<br />
8.5<br />
Figure 3.11 Effect <strong>of</strong> <strong>in</strong>fluent pH on bromate removal <strong>in</strong> BAG filters<br />
(average <strong>in</strong>fluent conditions: CUW, 20 \ig/L BrO3', 2.1 mg/L DO, 5.0 mg/L NO3')<br />
S<strong>in</strong>ce bromate formation dur<strong>in</strong>g ozonation is dependent on the pH, bromate formation could<br />
be reduced <strong>and</strong> biological bromate reduction could be <strong>in</strong>creased by pH control. For example, Song<br />
et al. (1996) observed a 50 percent reduction <strong>in</strong> bromate formation by reduc<strong>in</strong>g the pH from 8.0 to<br />
7.0. Thus, if the pH for ozonation were decreased to the near-neutral range (6.8 to 7.2) to reduce<br />
bromate formation, the data presented <strong>in</strong> Figure 3.11 show that biological bromate removal would<br />
be improved. Although pH 6.8 <strong>and</strong> 7.2 yielded the highest percent bromate removal, the optimum<br />
pH can only be found by evaluat<strong>in</strong>g a narrower range <strong>of</strong> pH values s<strong>in</strong>ce biological reduction rates<br />
can be very sensitive to small changes <strong>in</strong> pH. For <strong>in</strong>stance, Chen <strong>and</strong> Hao (1996) observed a doubl<strong>in</strong>g<br />
<strong>in</strong> the <strong>in</strong>itial specific microbial chromate reduction rate as the pH <strong>in</strong>creased from 7.0 to 7.3.<br />
Effect <strong>of</strong> Source Water Type<br />
To evaluate if the type <strong>of</strong> source water would have an effect on bromate reduction,<br />
experiments were run with raw water from Lake Michigan us<strong>in</strong>g seven-year-old BAG from the Paul<br />
M. Neal Water Treatment Facility (Lake Bluff, IL). S<strong>in</strong>ce this plant treats water from Lake Michigan,<br />
it was appropriate to use this BAG for the LMW experiments. The <strong>in</strong>fluent DO concentration was<br />
53
approximately 2 mg/L <strong>and</strong> the pH was 7.5. Table3.3 shows pert<strong>in</strong>ent experimental <strong>in</strong>formation. The<br />
LMW experiment was separated <strong>in</strong>to five phases - A through E. Each phase had a different electron<br />
donor concentration <strong>and</strong>/or EBCT. Table 3.3 shows the concentration <strong>of</strong> exogenous electron donor<br />
<strong>and</strong> the total DOC. The total DOC is the sum <strong>of</strong> the organic matter present <strong>in</strong> LMW (2.4 mg C/L)<br />
<strong>and</strong> the concentration <strong>of</strong> exogenous electron donor.<br />
Two BAG filters were run <strong>in</strong> series for these experiments so that two EBCTs could be<br />
evaluated simultaneously (Figure 2.2). The results from the lead BAG filter will be discussed first.<br />
Figures 3.12 <strong>and</strong> 3.13 show the bromate <strong>and</strong> nitrate rema<strong>in</strong><strong>in</strong>g <strong>in</strong> the effluent <strong>of</strong> the lead BAG filter<br />
for each phase <strong>of</strong> the experiment. Note that each phase <strong>of</strong> the experiment <strong>in</strong> Figure 3.12 is labeled<br />
as "no electron donor added", "low electron donor", or "high electron donor" <strong>in</strong> order to rem<strong>in</strong>d the<br />
reader about the concentration <strong>of</strong> external electron donor added <strong>in</strong> each phase.<br />
At the beg<strong>in</strong>n<strong>in</strong>g <strong>of</strong> the experiment (Phase A), no external electron donor was added to the<br />
<strong>in</strong>fluent. Table 3.3 shows that little DOC removal was observed <strong>in</strong> Phase A <strong>in</strong> the 15-m<strong>in</strong>ute EBCT<br />
BAG filter, <strong>in</strong>dicat<strong>in</strong>g the lack <strong>of</strong> biodegradable organic matter <strong>in</strong> LMW. In the full-scale plant from<br />
which the BAG was obta<strong>in</strong>ed, an average <strong>of</strong> 20 percent DOC removal from LMW is observed <strong>in</strong> a<br />
15-m<strong>in</strong>ute EBCT filter (Soucie 1999). The difference <strong>in</strong> DOC removal between the<br />
Table 3.3 DOC removal <strong>and</strong> DO rema<strong>in</strong><strong>in</strong>g <strong>in</strong> the BAG experiments us<strong>in</strong>g LMW (average<br />
<strong>in</strong>fluent conditions: LMW, 2.0 mg/L DO, pH 7.5, 21 ug/L BrO3", 50 ug/L C1O4", 1.6 mg/L NO3')<br />
Phase<br />
EBCT<br />
(m<strong>in</strong>)<br />
Exogenous electron donor<br />
lactate + pyruvate (mg C/L)<br />
Total DOC<br />
(mg/L)<br />
DOC<br />
removal (%)<br />
Effluent DO<br />
(mg/L)<br />
A<br />
15<br />
30<br />
0<br />
2.4<br />
4<br />
2<br />
0.8<br />
0.3<br />
B<br />
15<br />
30<br />
1.0<br />
3.4<br />
34<br />
42<br />
0.2<br />
0.2<br />
C<br />
30<br />
60<br />
3.1<br />
5.5<br />
62<br />
67<br />
0.2<br />
0.2<br />
D<br />
15<br />
30<br />
3.6<br />
6.0<br />
56<br />
73<br />
0.2<br />
0.2<br />
E<br />
15<br />
30<br />
0.6<br />
3.0<br />
36<br />
40<br />
0.2<br />
0.2<br />
54
ench-scale <strong>and</strong> full-scale B AC filters was likely due to the fact that the bench-scale <strong>in</strong>fluent was not<br />
ozonated while the full-scale filter receives ozonated <strong>in</strong>fluent (0.1 mg ozone/mg DOC) which can<br />
<strong>in</strong>crease the concentration <strong>of</strong> biodegradable organic matter. Very little bromate removal was<br />
observed <strong>in</strong> Phase A (Figure 3.12). Figure 3.13 shows that only a small amount <strong>of</strong> nitrate was<br />
produced <strong>in</strong> the BAG (C/C0 > 1) because the <strong>in</strong>fluent NH3-N concentration was rout<strong>in</strong>ely below the<br />
detection limit <strong>of</strong> 0.05 mg/L. The effluent DO concentration was 0.8 mg/L (Table 3.3), which<br />
<strong>in</strong>dicates the presence <strong>of</strong> biological activity <strong>in</strong> the filter. Kirisits <strong>and</strong> Snoey<strong>in</strong>k (1999) showed that<br />
an 18-m<strong>in</strong>ute EBCT BAG filter us<strong>in</strong>g CUW under similar conditions (pH 7.5, <strong>in</strong>fluent DO = 2.0<br />
mg/L, effluent DO = 0.2 mg/L, <strong>in</strong>fluent nitrate = 0.3 mg/L, effluent nitrate = 2.1 mg/L, 20 ng/L<br />
bromate) removed 60 percent <strong>of</strong> the bromate. The reduction <strong>of</strong> bromate <strong>in</strong> LMW was likely h<strong>in</strong>dered<br />
by a comb<strong>in</strong>ation <strong>of</strong> factors which are expla<strong>in</strong>ed as follows. Only a small amount <strong>of</strong> DO was<br />
consumed <strong>in</strong> the biological nitrification reaction because <strong>of</strong> the low ammonia concentration. This DO<br />
could then compete with bromate for the biodegradable organic matter. Alternatively, if nitrate<br />
reductase catalyzes bromate reduction, the DO could also have <strong>in</strong>hibited the synthesis or activity <strong>of</strong><br />
the enzyme.<br />
0-m<strong>in</strong>EB<br />
e don<br />
15-m<strong>in</strong> EBCT<br />
Low e" donor<br />
15-m<strong>in</strong> EBCT<br />
High e" donor<br />
15-mmEBCT<br />
No e" donor added<br />
15-m<strong>in</strong> EBCT<br />
Low e" donor<br />
0 1000 2000 3000 4000 5000 6000<br />
Bed Volumes<br />
7000 8000 9000 10000<br />
Figure 3.12 <strong>Bromate</strong> rema<strong>in</strong><strong>in</strong>g after first BAG filter us<strong>in</strong>g LMW (average <strong>in</strong>fluent conditions: Lake<br />
County BAG, LMW, 2.0 mg/L DO, pH 7.5, 21 ug/L BrO3', 50 ng/L C1O4', 1.6 mg/L NO3')<br />
55
fl<br />
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000<br />
Bed Volumes<br />
Figure 3.13 Nitrate rema<strong>in</strong><strong>in</strong>g after first BAG filter us<strong>in</strong>g LMW (average <strong>in</strong>fluent conditions: Lake<br />
County BAG, LMW, 2.0 mg/L DO, pH 7.5, 21 ug/L BrO3', 50 ug/L CIO/, 1.6 mg/L NO3')<br />
Lactate <strong>and</strong> pyruvate (1 mg C/ L) were added to the <strong>in</strong>fluent <strong>in</strong> Phase B, <strong>and</strong> the EBCT was<br />
ma<strong>in</strong>ta<strong>in</strong>ed at 15 m<strong>in</strong>utes. There was little difference <strong>in</strong> the amount <strong>of</strong> bromate removed between<br />
Phases A <strong>and</strong> B (Figure 3.12), despite the fact that significant DOC removal was observed <strong>in</strong> Phase<br />
B (Table 3.3). The low effluent DO concentration (0.2 mg/L) suggests that much <strong>of</strong> the DOC <strong>in</strong><br />
Phase B was be<strong>in</strong>g used by aerobic microorganisms. S<strong>in</strong>ce nitrate removal became significant <strong>in</strong><br />
Phase B (Figure 3.13), DOC was also be<strong>in</strong>g consumed by nitrate reducers. This may <strong>in</strong>dicate that<br />
a nitrate-reduc<strong>in</strong>g population was develop<strong>in</strong>g on the carbon but was not able to reduce bromate due<br />
to an <strong>in</strong>sufficient concentration <strong>of</strong> electron donor.<br />
In Phase C, the EBCT <strong>in</strong> the BAC was doubled from 15 to 30 m<strong>in</strong>utes, <strong>and</strong> the external<br />
electron donor concentration was <strong>in</strong>creased to 3.1 mg C/L (Table 3.3) to provide a favorable<br />
environment for bromate reduction. Figure 3.13 shows that the effluent nitrate concentrations were<br />
very low. Figure 3.12 shows that bromate removal <strong>in</strong>creased until the effluent bromate concentration<br />
reached the detection limit.<br />
When the EBCT was reduced from 30 to 15 m<strong>in</strong>utes (Phase D) while the external electron<br />
donor concentration was ma<strong>in</strong>ta<strong>in</strong>ed close to the concentration <strong>in</strong> Phase C (Table 3.3), complete<br />
56
omate removal was susta<strong>in</strong>ed (Figure 3.12). The effluent nitrate concentration also rema<strong>in</strong>ed very<br />
low (Figure 3.13). This shows that the electron donor concentration, <strong>and</strong> not the EBCT, was limit<strong>in</strong>g<br />
bromate removal <strong>in</strong> these experiments. This is different from the results observed <strong>in</strong> CUW because<br />
bromate removal <strong>in</strong>creased as the EBCT <strong>in</strong>creased, <strong>in</strong>dicat<strong>in</strong>g that EBCT was limit<strong>in</strong>g bromate<br />
removal (Kirisits <strong>and</strong> Snoey<strong>in</strong>k 1999).<br />
F<strong>in</strong>ally, the external electron donor concentration was decreased to 0.6 nig C/L while<br />
ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g an EBCT <strong>of</strong> 15 m<strong>in</strong>utes <strong>in</strong> Phase E (Table 3.3). Figures 3.12 <strong>and</strong> 3.13 show that the<br />
effluent bromate <strong>and</strong> nitrate concentrations began to <strong>in</strong>crease. It is likely that the electron donor<br />
concentration was limit<strong>in</strong>g the growth <strong>of</strong> the nitrate-reduc<strong>in</strong>g <strong>and</strong> bromate-reduc<strong>in</strong>g organisms. With<br />
less electron donor present, there was a decreased dem<strong>and</strong> for nitrate <strong>and</strong> bromate as electron<br />
acceptors.<br />
Table 3.3 <strong>and</strong> Figures 3.14 <strong>and</strong> 3.15 show the results from the two BAG filters <strong>in</strong> series. The<br />
trend <strong>of</strong> the data <strong>in</strong> each phase <strong>of</strong> Figures 3.14 <strong>and</strong> 3.15 is the same as that <strong>in</strong> Figures 3.12 <strong>and</strong> 3.13.<br />
Compar<strong>in</strong>g phase A between Figures 3.12 <strong>and</strong> 3.14, when no exogenous electron donor was added<br />
to LMW, bromate removal did not improve even though the contact time was <strong>in</strong>creased from 15 to<br />
30 m<strong>in</strong>utes. This is another <strong>in</strong>dication that the electron donor concentration was limit<strong>in</strong>g bromate<br />
removal from LMW.<br />
1.0<br />
•1 °- 8 H<br />
•a<br />
i<br />
ft 0.6 H<br />
fi<br />
0 1000 2000 3000<br />
4000 5000<br />
Bed Volumes<br />
Figure 3.15 Nitrate rema<strong>in</strong><strong>in</strong>g after second BAG filter us<strong>in</strong>g LMW (average <strong>in</strong>fluent conditions:<br />
Lake County BAG, LMW, 2.0 mg/L DO, pH 7.5, 21 ng/L BrCV, 50 ug/L CIO;, 1.6 mg/L NO3')<br />
Based on the results <strong>of</strong> the LMW experiments, it can be concluded that the biodegradability<br />
<strong>of</strong> the NOM <strong>in</strong> the source water can significantly affect the biological bromate removal. Ozonat<strong>in</strong>g<br />
the LMW to <strong>in</strong>crease the biodegradable fraction <strong>of</strong> NOM may improve bromate removal.<br />
Although bromate removal was the focus <strong>of</strong> the work us<strong>in</strong>g Lake County BAG, perchlorate<br />
was also added to the column <strong>in</strong>fluent beg<strong>in</strong>n<strong>in</strong>g <strong>in</strong> Phase B (Figures 3.16 <strong>and</strong> 3.17). At this time,<br />
the concentration <strong>of</strong> perchlorate <strong>in</strong> the effluent was <strong>in</strong>creas<strong>in</strong>g, as the ion exchange capacity for<br />
perchlorate on the <strong>GAC</strong> was approached. <strong>Perchlorate</strong> reduction <strong>in</strong>creased as the concentration <strong>of</strong><br />
exogenous electron donor <strong>in</strong>creased <strong>and</strong> the effluent nitrate concentration was very low (Phases C<br />
<strong>and</strong> D). As the effluent nitrate concentration began to <strong>in</strong>crease <strong>in</strong> Phase E, due to a lower<br />
concentration <strong>of</strong> external electron donor, perchlorate removal rapidly decl<strong>in</strong>ed. The dependence <strong>of</strong><br />
perchlorate removal on nitrate concentration is further explored later <strong>in</strong> this chapter.<br />
58
Fraction <strong>Perchlorate</strong> Rema<strong>in</strong><strong>in</strong>g, Fraction <strong>Perchlorate</strong> Rema<strong>in</strong><strong>in</strong>g,<br />
C/Co
<strong>Perchlorate</strong> Reduction <strong>in</strong> CITW<br />
The <strong>in</strong>fluent <strong>and</strong> effluent concentrations <strong>of</strong> perchlorate to the BAG filters operated with CUW<br />
were also monitored. In all experiments, no evidence <strong>of</strong> reduction <strong>of</strong> the <strong>in</strong>fluent perchlorate<br />
concentration (50 ng/L) was observed. As discussed later <strong>in</strong> this chapter, perchlorate reduction is<br />
extremely sensitive to nitrate concentration. The threshold nitrate concentration, at which 95 percent<br />
perchlorate reduction could be obta<strong>in</strong>ed, was between 0.35 <strong>and</strong> 0.05 mg/L. S<strong>in</strong>ce CUW conta<strong>in</strong>s<br />
ammonia, nitrification took place <strong>in</strong> the BAG filters. Depend<strong>in</strong>g on the experimental conditions,<br />
nitrification caused the nitrate concentration to <strong>in</strong>crease by 1.8-4.6 mg/L. Significant nitrate<br />
reduction did not take place s<strong>in</strong>ce no external electron donor was added to CUW. Therefore, the<br />
0.35/0.05 mg/L nitrate threshold for perchlorate reduction could not be obta<strong>in</strong>ed <strong>in</strong> these<br />
experiments.<br />
Microscopic Exam<strong>in</strong>ation <strong>of</strong> Microorganisms <strong>in</strong> the Filters<br />
A microscopic exam<strong>in</strong>ation <strong>of</strong> biomass taken from the BAG filters operated with CUW was<br />
completed. Four representative pictures are shown <strong>in</strong> Figures 3.18-3.21, which were obta<strong>in</strong>ed us<strong>in</strong>g<br />
phase contrast microscopy with 1 OOOx magnification. The dark objects are microorganisms while the<br />
lighter, refractile images are likely debris.<br />
The bacterial morphologies were generally observed to be cocci, rods, cha<strong>in</strong>-form<strong>in</strong>g rods,<br />
<strong>and</strong> aggregated rods. All <strong>of</strong> these morphologies were found <strong>in</strong> both sets <strong>of</strong> BAC filters operated with<br />
CUW.<br />
Rods <strong>and</strong> cocci can be observed <strong>in</strong> Figure 3.18. Large aggregates <strong>of</strong> bacteria associated with<br />
non-biological particulates were commonly observed. The figure also shows some flocculated<br />
material that has bacteria <strong>in</strong>corporated <strong>in</strong>to the floe structure. Despite the fact that the carbon filters<br />
looked very "clean" <strong>and</strong> did not experience significant headless, the microscope images confirm that<br />
there are a lot <strong>of</strong> microorganisms <strong>in</strong> the filters.<br />
A cha<strong>in</strong> <strong>of</strong> rods, illustrated <strong>in</strong> Figure 3.19, was a common morphology observed <strong>in</strong> BAC filter<br />
samples. It is <strong>in</strong>terest<strong>in</strong>g to note that Hijnen et al. (1995) isolated some cha<strong>in</strong>-form<strong>in</strong>g rods that were<br />
able to reduce bromate. Figure 3.20 shows a loose aggregation <strong>of</strong> rods.<br />
60
Figure 3.18 Rods <strong>and</strong> cocci from the BAG filter<br />
Figure 3.19 Cha<strong>in</strong> <strong>of</strong> rods from the BAG filter<br />
Figure 3.20 Aggregated rods from the BAG filter<br />
61
Figure 3.21 Protozoa <strong>in</strong> the BAG filter<br />
Protozoa were also found <strong>in</strong> the filters. The common morphology seemed to be the shape <strong>of</strong><br />
a round-bottom flask. Protozoa can be seen <strong>in</strong> Figure 3.21. It is not uncommon to f<strong>in</strong>d protozoa <strong>in</strong><br />
BAG filters, especially filters that have only been backwashed once <strong>in</strong> two years <strong>of</strong> operation. Higher<br />
organisms, such as protozoa, graze on the microbial population.<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis Experiment<br />
DGGE is a genetic f<strong>in</strong>gerpr<strong>in</strong>t<strong>in</strong>g technique used to pr<strong>of</strong>ile the bacterial diversity present <strong>in</strong><br />
a sample. The method is based on the separation <strong>of</strong> PCR-amplified ribosomal DNA (rDNA)<br />
fragments <strong>of</strong> the same length but with different sequences. Each b<strong>and</strong> theoretically represents a<br />
unique organism. Figure 3.22 shows a DGGE pr<strong>of</strong>ile that was obta<strong>in</strong>ed with DNA extracted from<br />
the BAC filters operated with CUW. DNA from each sample was extracted <strong>and</strong> analyzed <strong>in</strong><br />
duplicate. The first lane is a st<strong>and</strong>ard marker. The next four lanes are from the BAC filters operated<br />
for perchlorate removal <strong>and</strong> will be discussed <strong>in</strong> a later section. Lanes 6 <strong>and</strong> 7 were run with a<br />
microbial sample from the effluent side <strong>of</strong> the BAC operated with CUW <strong>and</strong> an <strong>in</strong>fluent nitrate<br />
concentration <strong>of</strong> 5.0 mg/L. Lanes 8 <strong>and</strong> 9 were run with material isolated from the backwash water<br />
<strong>of</strong> the BAC filter operated with CUW <strong>and</strong> an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L. Lanes 10<br />
<strong>and</strong> 11 were run with material isolated from the backwash water <strong>of</strong> the BAC filter operated with<br />
CUW <strong>and</strong> an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L.<br />
S<strong>in</strong>ce the number <strong>of</strong> b<strong>and</strong>s is a measure <strong>of</strong> the diversity <strong>of</strong> the microbial community, it can be<br />
62
seen that the BAG samples conta<strong>in</strong> a diverse array <strong>of</strong> microorganisms (Figure 3.22). It is possible<br />
that the genetic diversity is even greater <strong>in</strong> the samples than <strong>in</strong>dicated by the b<strong>and</strong><strong>in</strong>g pr<strong>of</strong>iles. By<br />
analyz<strong>in</strong>g the DNA fragments us<strong>in</strong>g a narrow denatur<strong>in</strong>g gradient, it is possible to further resolve<br />
s<strong>in</strong>gle b<strong>and</strong>s <strong>in</strong>to multiple b<strong>and</strong>s represent<strong>in</strong>g dist<strong>in</strong>ct sequences. Individual b<strong>and</strong>s, excised from<br />
DGGE gels, can also be further analyzed for DNA sequences, <strong>and</strong> thus the identity <strong>of</strong> specific<br />
members <strong>of</strong> the microbial community can be determ<strong>in</strong>ed.<br />
There are conserved b<strong>and</strong>s across the samples taken from the BAG filters operated to remove<br />
bromate, <strong>and</strong> there are some b<strong>and</strong>s that are unique to a particular sample. There are conserved b<strong>and</strong>s<br />
between the samples taken from the BAG filters operated to remove bromate <strong>and</strong> the BAG filters<br />
operated to remove perchlorate; this was expected s<strong>in</strong>ce the BAG filters operated to remove<br />
perchlorate were first run with effluent from the BAG filters that were remov<strong>in</strong>g bromate.<br />
Lanes 123456789 10 11<br />
Figure 3.22 DGGE pr<strong>of</strong>ile <strong>of</strong> bromate <strong>and</strong> perchlorate samples<br />
63
Future work may use DGGE pr<strong>of</strong>il<strong>in</strong>g as a technique for observ<strong>in</strong>g significant perturbations <strong>in</strong> the<br />
population that may relate to the bromate reduction performance <strong>of</strong> the BAG filter. DGGE could be<br />
run on BAG filter biomass samples after chang<strong>in</strong>g water quality parameters or operational conditions<br />
<strong>in</strong> the BAG filters. For <strong>in</strong>stance, DGGE could be run on biomass samples from the BAG filters that<br />
have been exposed to different pH values, to see if the bromate-reduc<strong>in</strong>g isolates are able to persist<br />
under different pH conditions.<br />
Plat<strong>in</strong>g Experiments <strong>and</strong> Microbial Characterization<br />
Biomass was taken from the BAG filters operated with CUW <strong>and</strong> an <strong>in</strong>fluent nitrate<br />
concentration <strong>of</strong> 5 mg/L, <strong>and</strong> it was diluted <strong>and</strong> streaked on agar plates. Fourteen isolates were<br />
obta<strong>in</strong>ed from the dilutions <strong>of</strong> the mixed culture.<br />
<strong>Bromate</strong> Reduction <strong>in</strong> CUW<br />
To test some <strong>of</strong> the isolates for bromate reduction ability, biomass was taken from the plates<br />
<strong>and</strong> <strong>in</strong>oculated to Balch tubes conta<strong>in</strong><strong>in</strong>g CUW spiked with bromate <strong>and</strong> nitrate. The results <strong>of</strong> the<br />
experiment are summarized <strong>in</strong> Table 3.4. Little or no bromate removal was demonstrated by the<br />
isolates under the tested conditions. The most bromate reduction was demonstrated by isolate B15,<br />
which reduced bromate from 28 to 2 \igfL after 12 weeks. Nitrite <strong>and</strong> nitrate data were also obta<strong>in</strong>ed<br />
<strong>in</strong> this experiment. Isolate Bl showed nitrate removal with no nitrite accumulation. Isolates B2,B3,<br />
B4, B9, B11, B12, <strong>and</strong> B13 showed nitrate removal with nitrite accumulation. Isolates B6, B8, B14,<br />
<strong>and</strong> B15 did not show substantial nitrate removal. A mass balance on nitrite + nitrate showed that<br />
isolates Bl <strong>and</strong> Bl 1 may have converted some <strong>of</strong> the nitrate to unmonitored species (such as NH3<br />
or N2). The nitrogen balance showed that some DO may have been present <strong>in</strong> the tubes conta<strong>in</strong><strong>in</strong>g<br />
B3 <strong>and</strong> B13, lead<strong>in</strong>g to nitrification. The nitrite + nitrate balance for the rest <strong>of</strong> the isolates was close<br />
to the control value, <strong>in</strong>dicat<strong>in</strong>g that nitrite <strong>and</strong> nitrate were not converted to unmonitored nitrogen<br />
species.<br />
64
<strong>Bromate</strong> Reduction <strong>in</strong> DDW with Exogenous Electron Donor<br />
Based on the poor bromate removal exhibited <strong>in</strong> the previous experiment, the experiment was<br />
modified <strong>and</strong> repeated. The aqueous medium was DDW spiked with bromate, nitrate, <strong>and</strong> exogenous<br />
electron donor. The concentration <strong>of</strong> electron donor <strong>in</strong> the medium (0.07 mM lactate, 0.10 mM<br />
acetate, <strong>and</strong> 0.08 mM pyruvate) was such that 10 mg/L nitrate could be reduced to N2 . Not all <strong>of</strong><br />
the same isolates used <strong>in</strong> the previous experiment (Table 3.4) were used <strong>in</strong> this experiment. B3, B4,<br />
B12, B13, <strong>and</strong> B14 could not be sub-cultured after the first culture was grown.<br />
Table 3.4 Test<strong>in</strong>g the isolates for the ability to reduce bromate (12-week <strong>in</strong>cubation; <strong>in</strong>itial<br />
conditions: CUW, 5 mM phosphate buffer, 28.7 \igfL BrO3', 4.82 mg/L NO 3',<br />
pH 7.8, 0 mg/L DO)<br />
Isolate<br />
Control -<br />
no <strong>in</strong>oculum<br />
Bl<br />
B2<br />
B3<br />
B4<br />
B6<br />
B8<br />
B9<br />
Bll<br />
B12<br />
B13<br />
B14<br />
B15<br />
<strong>Bromate</strong><br />
(Hg/L)<br />
28.4<br />
26.6<br />
31.0<br />
26.0<br />
26.0<br />
27.5<br />
28.8<br />
25.7<br />
28.6<br />
26.7<br />
28.9<br />
27.7<br />
24.6<br />
Nitrite<br />
(mg/L)<br />
0.05<br />
0.03<br />
3.51<br />
5.04<br />
3.78<br />
0.20<br />
0.31<br />
3.44<br />
1.50<br />
3.42<br />
4.68<br />
0.04<br />
0.05<br />
Nitrate<br />
(mg/L)<br />
4.82<br />
0.39<br />
0.20<br />
0.21<br />
0.17<br />
4.74<br />
4.48<br />
0.38<br />
0.13<br />
0.13<br />
0.12<br />
4.68<br />
4.84<br />
Nitrite + nitrate<br />
(mg/L as N)<br />
1.10<br />
0.09<br />
1.11<br />
1.58<br />
1.19<br />
1.13<br />
1.11<br />
1.13<br />
0.49<br />
1.07<br />
1.45<br />
1.07<br />
1.11<br />
65
Table 3.5 shows that all <strong>of</strong> the isolates tested (Bl, B2, B6, B7, B8, B9, BIO, Bl 1, <strong>and</strong> B15)<br />
were able to reduce bromate to some degree. Isolate B7 showed the best bromate removal after the<br />
4.5-week <strong>in</strong>cubation period, s<strong>in</strong>ce it was able to reduce bromate to the detection limit. Isolates B2<br />
<strong>and</strong> B8 showed the least bromate removal. Even with the external electron donor addition, isolates<br />
B6, B8, <strong>and</strong> B15 were not able to reduce nitrate (Table 3.5) as was observed earlier <strong>in</strong> the absence<br />
<strong>of</strong> an external electron donor (Table 3.4). Isolate B2 was unable to remove a substantial amount <strong>of</strong><br />
nitrate <strong>in</strong> the presence <strong>of</strong> external electron donor (Table 3.5), <strong>in</strong> contrast to the nitrate removal <strong>and</strong><br />
nitrite accumulation observed <strong>in</strong> the absence <strong>of</strong> external electron donor (Table 3.4). Isolates Bl, B9,<br />
BIO, <strong>and</strong> Bl 1 were able to reduce nitrate with nitrite accumulation. Isolate B7 was able to remove<br />
nitrate with no nitrite accumulation.<br />
Table 3.5 Test<strong>in</strong>g the isolates for the ability to reduce bromate (4.5-week <strong>in</strong>cubation; <strong>in</strong>itial<br />
conditions: DDW, 5 mM phosphate buffer, 43.5 ug/L BrO3', 4.92 mg/L NO3', 8.3 mg/L DOC,<br />
pH 7.5, 0 mg/L DO)<br />
Isolate<br />
Control -<br />
no <strong>in</strong>oculum<br />
<strong>Bromate</strong><br />
(Hg/L)<br />
43.5<br />
Nitrite<br />
(mg/L)<br />
0.00<br />
Nitrate<br />
(mg/L)<br />
4.92<br />
Nitrite + nitrate<br />
(mg/L as N)<br />
1.11<br />
Bl<br />
B2<br />
B6<br />
B7<br />
B8<br />
B9<br />
BIO<br />
Bll<br />
B15<br />
36.6<br />
40.6<br />
36.3<br />
1.0<br />
39.5<br />
34.1<br />
32.5<br />
37.6<br />
37.3<br />
0.42<br />
0.16<br />
0.09<br />
0.00<br />
0.14<br />
3.64<br />
3.18<br />
2.49<br />
0.06<br />
2.99<br />
4.35<br />
4.85<br />
0.41<br />
4.85<br />
0.41<br />
0.42<br />
0.42<br />
4.77<br />
0.68<br />
1.03<br />
1.12<br />
0.09<br />
1.14<br />
1.20<br />
1.06<br />
0.85<br />
1.10<br />
66
The mass balance on nitrate + nitrite (Table 3.5) showed that isolates B1, B7, <strong>and</strong> Bll may<br />
have converted some <strong>of</strong> the nitrate to unmonitored species (such as NH3 or N2). Isolates B1 <strong>and</strong> Bll<br />
also demonstrated this behavior <strong>in</strong> the previous experiment (Table 3.4). The mass balance on nitrate<br />
+ nitrite for the rest <strong>of</strong> the isolates was close to the control value, <strong>in</strong>dicat<strong>in</strong>g that these<br />
microorganisms did not convert nitrate to unmonitored nitrogen species.<br />
Challeng<strong>in</strong>g the Isolates<br />
The isolates were also challenged with higher than usual concentrations <strong>of</strong> bromate <strong>and</strong><br />
nitrate. Each isolate was <strong>in</strong>oculated <strong>in</strong>to aqueous medium conta<strong>in</strong><strong>in</strong>g 14.0 mg/LNO 3 " -N (Table 3.6),<br />
5.6 mg/L BrO 3--Br (Table 3.7), or 0.6 mg/L NO 3'-N / 5.6 mg/L BrO 3'-Br (Table 3.8).<br />
Table 3.6 shows that all <strong>of</strong> the isolates, except for B15, were able to reduce nitrate. Table<br />
3.7 shows that isolates B2, B6, BIO, <strong>and</strong> B15 were able to substantially reduce bromate when no<br />
nitrate was added to the medium. Table 3.8 shows that isolates B6, B7, BIO, Bll, <strong>and</strong> B15 were<br />
able to substantially reduce bromate when nitrate was added to the medium.<br />
Table 3.6 Challeng<strong>in</strong>g isolates with a high nitrate concentration (<strong>in</strong>itial conditions: modified<br />
MR2A broth, 14.0 mg/L NO3 -N, pH 7.0, 0 mg/L DO)<br />
Isolate<br />
Bl<br />
B2<br />
B6<br />
B7<br />
B8<br />
B9<br />
BIO<br />
Bll<br />
B15<br />
Nitrate removed<br />
(mg/L as N)<br />
1.1<br />
13.3<br />
13.6<br />
13.6<br />
4.3<br />
13.6<br />
13.6<br />
13.5<br />
-0.6<br />
Nitrite produced<br />
(mg/L as N)<br />
0<br />
12.7<br />
0<br />
0<br />
0<br />
0<br />
0<br />
0<br />
0<br />
67
Table 3.7 Challeng<strong>in</strong>g isolates with a high bromate concentration (<strong>in</strong>itial conditions: modified<br />
MR2A broth, 5.6 mg/L BrO 3' -Br, pH 7.0, 0 mg/L DO)<br />
Isolate<br />
Bl<br />
B2<br />
B6<br />
B7<br />
B8<br />
B9<br />
BIO<br />
Bll<br />
B15<br />
<strong>Bromate</strong> removed<br />
(mg/L as Br)<br />
0.4<br />
1.2<br />
2.4<br />
0.2<br />
0.2<br />
0.2<br />
1.5<br />
0.2<br />
1.2<br />
Bromide produced<br />
(mg/L as Br)<br />
0.4<br />
1.5<br />
1.5<br />
0.4<br />
0.1<br />
0.1<br />
1.0<br />
0.2<br />
1.1<br />
Table 3.8 Challeng<strong>in</strong>g isolates with high concentrations <strong>of</strong> bromate <strong>and</strong> nitrate (<strong>in</strong>itial conditions:<br />
MR2A broth, 5.6 mg/L BrO 3' -Br, 0.6 mg/L NO3" -N, pH 7.0, 0 mg/L DO)<br />
Isolate<br />
<strong>Bromate</strong> removed<br />
(mg/L as Br)<br />
Bromide produced<br />
(mg/L as Br)<br />
Nitrate removed<br />
(mg/L as N)<br />
Nitrite produced<br />
(mg/L as N)<br />
Bl<br />
0.2<br />
0.3<br />
0.4<br />
0<br />
B2<br />
0.2<br />
0.2<br />
0.5<br />
0.6<br />
B6<br />
2.1<br />
1.6<br />
0.5<br />
0<br />
B7<br />
0.7<br />
0.5<br />
0.5<br />
0<br />
B8<br />
0.3<br />
-0.2<br />
0.5<br />
0<br />
B9<br />
0.4<br />
0.2<br />
0.5<br />
0<br />
BIO<br />
2.3<br />
2.3<br />
0.6<br />
0<br />
Bll<br />
1.1<br />
0.9<br />
0.6<br />
0<br />
B15<br />
0.8<br />
1.1<br />
0.2<br />
-0.2<br />
68
N2O was not detected <strong>and</strong> no evidence <strong>of</strong> trapped gases were present <strong>in</strong> Durham tubes <strong>in</strong> any<br />
cultures where nitrate consumption had occurred, <strong>in</strong>dicat<strong>in</strong>g that the primary products <strong>of</strong> nitrate<br />
reduction were not likely to be through denitrification pathways (i.e. N2O or N2 gases). It is likely<br />
that nitrate is be<strong>in</strong>g reduced to ammonia, <strong>and</strong> this will be tested <strong>in</strong> future experiments.<br />
Additional Isolate Information<br />
All cultures are facultative, s<strong>in</strong>ce they were able to grow on agar plates that were <strong>in</strong>cubated<br />
anaerobically <strong>and</strong> those that were <strong>in</strong>cubated aerobically. Isolate B6 was further tested <strong>in</strong> MR2A<br />
supplemented with yeast extract, casam<strong>in</strong>o acids, acetate, <strong>and</strong> pyruvate, without additional electron<br />
acceptors present (i.e. no bromate, nitrate, or DO). Isolate B6 grew but growth was significantly less<br />
than when additional electron acceptor was present. This <strong>in</strong>dicates that isolate B6 can grow by<br />
fermentation.<br />
A summary <strong>of</strong> all the <strong>in</strong>formation concern<strong>in</strong>g the isolates is shown <strong>in</strong> Table 3.9. Isolate B8<br />
is the pure culture that has not shown much promise regard<strong>in</strong>g bromate reduction. Isolates Bl <strong>and</strong><br />
B9 have shown promise for bromate reduction at low bromate concentrations (~ ug/L) but not at high<br />
bromate concentrations (- mg/L). Isolates B6, B10, <strong>and</strong> B15 seem quite robust, s<strong>in</strong>ce they were able<br />
to reduce bromate under all conditions tested. All <strong>of</strong> the isolates shown <strong>in</strong> Table 3.9 will be<br />
sequenced <strong>and</strong> phylogenetically classified.<br />
Dilution-to-Ext<strong>in</strong>ction Experiments<br />
Us<strong>in</strong>g a non-traditional enrichment technique, consist<strong>in</strong>g <strong>of</strong> progressive dilutions, Jackson<br />
et al. (1998) was able to enrich for different microorganisms at different dilutions. For example,<br />
microorganisms that grow very well <strong>in</strong> the culture medium will be selected for at lower dilutions;<br />
they will be able to outcompete organisms that are orig<strong>in</strong>ally present <strong>in</strong> higher numbers but are not<br />
well-suited to the chosen culture medium. On the other h<strong>and</strong>, microorganisms that were abundant<br />
<strong>in</strong> number <strong>in</strong> the orig<strong>in</strong>al culture will be selected for at higher dilutions s<strong>in</strong>ce the less abundant<br />
organisms will essentially be diluted out <strong>of</strong> the culture. In contrast to the plat<strong>in</strong>g experiments<br />
described earlier, which targeted pure cultures us<strong>in</strong>g an artificial nutrient-rich condition, the dilutionto-ext<strong>in</strong>ction<br />
method considers microorganisms which may be "specialists" or whose activity may<br />
69
Table 3.9 Summary <strong>of</strong> experimental evidence for isolates<br />
High BrO3" concentration Low BrO 3 " concentration<br />
Isolate<br />
Bl<br />
B2<br />
B6<br />
B7<br />
B8<br />
B9<br />
BIO<br />
Bll<br />
B15<br />
<strong>Bromate</strong>- <strong>Bromate</strong>- <strong>Bromate</strong>-reducer, Nitrate- Facultative<br />
reducer, reducer, with nitrate reducer<br />
with nitrate without nitrate<br />
X XX<br />
X X X<br />
X X X XX<br />
X XXX<br />
X X<br />
X XX<br />
XX X XX<br />
X XXX<br />
XXX X<br />
be dependent upon a consortium. This is an alternative strategy that targets the isolation <strong>of</strong> pert<strong>in</strong>ent<br />
microorganisms under conditions that more closely parallel those under which the BAG filters are<br />
operated (i.e. low nutrient conditions).<br />
The results <strong>of</strong> this experiment are tabulated <strong>in</strong> Table 3.10. Us<strong>in</strong>g biomass from the BAG<br />
operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5 mg/L, bromate reduction was observed for all<br />
dilutions <strong>and</strong> decreased as the dilution <strong>in</strong>creased. Close attention should be paid to the highest<br />
dilution, 1:10,000, s<strong>in</strong>ce bromate reduction was still observed here. At this high dilution, organisms<br />
not well-suited to the medium were diluted out <strong>of</strong> the mixture, thereby lessen<strong>in</strong>g the complexity <strong>of</strong><br />
the microbial population. For all tubes, significant nitrate removal was observed. It is not known<br />
if bromate removal took place before, dur<strong>in</strong>g or after the nitrate removal.<br />
70
Table 3.10 Results <strong>of</strong> dilution-to-ext<strong>in</strong>ction experiments (3.5-month <strong>in</strong>cubation; <strong>in</strong>itial conditions:<br />
CUW, 28 ug/L BrCV, 7.25 mg/L N(V, pH 7.5, 0 mg/L DO)<br />
Biomass Source - B AC filters operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 5.0 mg/L<br />
Dilution Undiluted<br />
1:10<br />
1:100<br />
1:1,000<br />
1:10,000<br />
<strong>Bromate</strong> (ng/L) 2.5<br />
9.3<br />
19.1<br />
21.0<br />
23.8<br />
Nitrite (mg/L) 0.07<br />
ND<br />
ND<br />
ND<br />
ND<br />
Nitrate (mg/L)<br />
0.13<br />
0.10<br />
0.10<br />
0.09<br />
0.09<br />
Biomass Source - B AC filters operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3 mg/L<br />
Dilution Undiluted<br />
1:10<br />
1:100<br />
1:1,000<br />
1:10,000<br />
<strong>Bromate</strong> (ug/L) 1.0<br />
2.0<br />
1.0<br />
1.0<br />
2.6<br />
Nitrite (mg/L) ND<br />
ND<br />
ND<br />
ND<br />
ND<br />
Nitrate (mg/L)<br />
0.09<br />
0.10<br />
0.09<br />
0.09<br />
0.09<br />
Us<strong>in</strong>g biomass from the BAG filters operated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.3<br />
mg/L, bromate removal was observed to levels at or near the detection limit for all dilutions. This<br />
suggests that the bromate-reduc<strong>in</strong>g bacteria were numerically dom<strong>in</strong>ant <strong>in</strong> this particular <strong>in</strong>oculum.<br />
The organisms from the dilution-to-ext<strong>in</strong>ction experiments were not plated onto agar medium.<br />
Characterization <strong>of</strong> the eight bromate-reduc<strong>in</strong>g cultures that have been isolated by traditional<br />
techniques (Table 3.9) will be completed before attempt<strong>in</strong>g to isolate additional bromate-reducers.<br />
71
Serum Bottle Experiments<br />
To evaluate the effect <strong>of</strong> temperature on biological bromate reduction, batch experiments<br />
were set up <strong>in</strong> serum bottles with aqueous medium <strong>and</strong> <strong>in</strong>ocula from the BAG filters <strong>and</strong> <strong>in</strong>cubated<br />
at 25 <strong>and</strong> 4 °C. External electron donor (lactate, acetate, <strong>and</strong> pyruvate) was added to the bottles after<br />
33 days. As expected, bromate was reduced <strong>in</strong> the 25 °C experiment (Figure 3.23) while bromate was<br />
not reduced <strong>in</strong> the 4 °C experiment (Figure 3.24). Figure 3.23 shows that nitrate was also removed<br />
at 25 °C, <strong>and</strong> bromate reduction took place <strong>in</strong> the presence <strong>of</strong> nitrate.<br />
c<br />
_o<br />
1
0<br />
so<br />
o<br />
U<br />
).0 -<br />
I 5.0<br />
U<br />
ffl<br />
Addition <strong>of</strong> external electron donor<br />
6.0<br />
-- 5.0<br />
-- 4.0<br />
u/ u ' v<br />
3.0 ai o --<br />
2.0<br />
1.0<br />
o •-S<br />
U S<br />
0.0<br />
0 20 40 60 80<br />
Time (days)<br />
100 120<br />
0.0<br />
Figure 3.24 <strong>Bromate</strong> reduction <strong>in</strong> a serum bottle experiment (<strong>in</strong>itial conditions: CUW, pH<br />
7.7, 19 ug/L BrO3-, 5.4 mg/L NO3', 4 °C)<br />
PERCHLORATE - ABIOTIC REMOVAL<br />
Calculation <strong>of</strong> Ion Exchange Capacity for Virg<strong>in</strong> Norit <strong>GAC</strong><br />
CUW spiked with 50 ug/L <strong>of</strong> perchlorate was filtered through a bed <strong>of</strong> virg<strong>in</strong> Norit <strong>GAC</strong> with<br />
a 25-m<strong>in</strong>ute EBCT. Figure 3.25 shows the perchlorate breakthrough curve for the first 5,800 bed<br />
volumes (BV). A mass balance calculation showed that the total mass <strong>of</strong> perchlorate applied to the<br />
carbon through 5,800 BV equaled the total mass <strong>of</strong> perchlorate <strong>in</strong> the effluent through 5,800 BV.<br />
Therefore, any observed perchlorate removal up to 5,800 BV was the result <strong>of</strong> adsorption, probably<br />
ion exchange. The first 1,600 B V correspond to the phase when perchlorate was undergo<strong>in</strong>g primary<br />
ion exchange. To calculate the ion exchange capacity <strong>of</strong> the <strong>GAC</strong>, the area above the curve <strong>and</strong><br />
below C/C0= 1 from O^BV< 1,600 (shaded area, Figure 3.25) was <strong>in</strong>tegrated. This <strong>in</strong>tegration<br />
showed an ion exchange capacity <strong>of</strong> 0.172 mg perchlorate (1.7 x 10"6 mole) per gram <strong>GAC</strong>.<br />
73
1000 2000 3000 4000<br />
Bed Volumes<br />
5000 6000<br />
Figure 3.25 Ion exchange <strong>of</strong> perchlorate to <strong>GAC</strong> (<strong>in</strong>itial conditions: CUW, 2.5 mg/L DO, 5.0<br />
mg/L NO3-, pH 7.5, 50 ug/L C1O4')<br />
Abiotic Batch Experiments<br />
Figures 3.26 <strong>and</strong> 3.27 show the batch perchlorate removal results for four <strong>GAC</strong>s <strong>in</strong> CUW<br />
<strong>and</strong> DDW, respectively. <strong>Perchlorate</strong> was abiotically removed by each carbon. In both DDW <strong>and</strong><br />
CUW, the virg<strong>in</strong> Calgon F-400 carbon performed better than the virg<strong>in</strong> Norit carbon. This is most<br />
likely due to the fact that the Norit carbon particles are much larger than the Calgon F-400 carbon<br />
particles. A larger particle size requires a perchlorate molecule to travel a greater diffusional distance<br />
before reach<strong>in</strong>g an active site on the carbon. The two figures also show that the AWOG treated<br />
carbons perform better than the virg<strong>in</strong> carbons. This is <strong>in</strong> agreement with the earlier f<strong>in</strong>d<strong>in</strong>gs <strong>of</strong> Miller<br />
(1996) <strong>and</strong> Kirisits et al. (1998), both <strong>of</strong> whom attributed the differences <strong>in</strong> virg<strong>in</strong> <strong>and</strong> AWOG carbon<br />
reactivity to changes <strong>in</strong> surface chemistry generated dur<strong>in</strong>g the process <strong>of</strong> acid wash<strong>in</strong>g <strong>and</strong><br />
outgass<strong>in</strong>g.<br />
Figure 3.28 compares perchlorate removal by virg<strong>in</strong> Calgon F-400 <strong>in</strong> CUW <strong>and</strong> DDW.<br />
<strong>Perchlorate</strong> removal performance was better <strong>in</strong> DDW than <strong>in</strong> CUW. These results were comparable<br />
for each <strong>of</strong> the 4 carbons. NOM or other <strong>in</strong>organic anions <strong>in</strong> CUW are likely compet<strong>in</strong>g with<br />
perchlorate for active sites on the carbon, thus decreas<strong>in</strong>g perchlorate removal.<br />
74
o<br />
I<br />
Io<br />
U<br />
D<br />
o<br />
ou -<br />
50 1<br />
40 -<br />
30 -<br />
20 -<br />
10 -<br />
!•••<br />
"*£ 2 o<br />
A S 5 D D<br />
A A<br />
• Virg<strong>in</strong> Norit<br />
D Virg<strong>in</strong> F-400<br />
AAWOGNorit<br />
AAWOGF-400<br />
*<br />
A<br />
n<br />
0 50 100 150 200<br />
Time (m<strong>in</strong>)<br />
250 300 350<br />
Figure 3.26 Batch perchlorate removal (<strong>in</strong>itial conditions: CUW, pH 7.5, 500 mg/L <strong>GAC</strong>)<br />
••s 0<br />
3 co<br />
o tH<br />
O 5b<br />
J5<br />
o<br />
hd 0 >t> "0<br />
o<br />
— . g «><br />
1 <strong>Perchlorate</strong> Concentration f <strong>Perchlorate</strong> Concentration f f'<br />
S 0*/L) I WV 1 1<br />
« o 5 £ S £ 3 § £ o 5 £ £ £ 3 § £ f<br />
P<br />
1 ° -<br />
w ° 1<br />
—— i —— i —— i —— i —— an —<br />
^ OQ 3r*<br />
S<br />
•n<br />
S cr<br />
0s-<br />
O. 0)<br />
•o<br />
TJ 0<br />
• D<br />
(6<br />
S 1-<br />
O <br />
I °<br />
»<br />
i -<br />
< -<br />
SL 0<br />
2.'<br />
/*N<br />
«-»-<br />
£<br />
§ si § -<br />
B: §<br />
1' i"<br />
a B .<br />
to S tO<br />
^ 0 -<br />
0 °<br />
D<br />
$<br />
o to<br />
§ 3 -<br />
1O u><br />
> 0 -<br />
h °<br />
u><br />
Wash Tests<br />
Table 3.11 lists the results <strong>of</strong> the high nitrate concentration wash tests performed on four <strong>of</strong><br />
the carbons used <strong>in</strong> the batch abiotic perchlorate tests. It can be noted that, with<strong>in</strong> the limits <strong>of</strong><br />
detection for the ion chromatographic analysis, the entire mass <strong>of</strong> perchlorate that was removed by<br />
the <strong>GAC</strong> was washed <strong>of</strong>f the <strong>GAC</strong>. Therefore, no chemical reduction <strong>of</strong> perchlorate took place on<br />
the <strong>GAC</strong>.<br />
Metal-Catalyzed <strong>GAC</strong> Experiments<br />
Figures 3.30 <strong>and</strong> 3.31 show that both the filter conta<strong>in</strong><strong>in</strong>g <strong>GAC</strong> mixed with alum<strong>in</strong>um shot<br />
<strong>and</strong> the filter conta<strong>in</strong><strong>in</strong>g copper- <strong>and</strong> z<strong>in</strong>c oxide-impregnated <strong>GAC</strong> were efficient at remov<strong>in</strong>g<br />
perchlorate from <strong>in</strong>fluent water.<br />
Both <strong>GAC</strong> filters were stopped after approximately 500 BV. Up to that po<strong>in</strong>t the<br />
<strong>GAC</strong>/alum<strong>in</strong>um filter had removed a total <strong>of</strong> 52.2 mg <strong>of</strong> perchlorate <strong>and</strong> the copper- <strong>and</strong> z<strong>in</strong>c oxideimpregnated<br />
<strong>GAC</strong> filter had removed a total <strong>of</strong> 51.8 mg <strong>of</strong> perchlorate. At 500 BV, there was no<br />
<strong>in</strong>dication that perchlorate removal efficiencies were decreas<strong>in</strong>g for either filter.<br />
<strong>GAC</strong><br />
Virg<strong>in</strong> Norit<br />
Virg<strong>in</strong> F-400<br />
AWOG Norit<br />
AWOG F-400<br />
Virg<strong>in</strong> Norit<br />
Virg<strong>in</strong> F-400<br />
AWOG Norit<br />
AWOG F-400<br />
Virg<strong>in</strong> Norit<br />
Table 3.11 Abiotic perchlorate batch removal <strong>and</strong> wash tests<br />
Water<br />
CUW<br />
CUW<br />
CUW<br />
CUW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
DDW<br />
PH<br />
7.5<br />
7.5<br />
7.5<br />
7.5<br />
7.5<br />
7.5<br />
7.5<br />
7.5<br />
2.0<br />
Mass <strong>of</strong> perchlorate<br />
removed<br />
G*g)<br />
42.0<br />
71.8<br />
80.5<br />
72.3<br />
60.4<br />
90.7<br />
95.5<br />
95.7<br />
85.3<br />
Mass <strong>of</strong> perchlorate<br />
washed <strong>of</strong>f carbon<br />
fog)<br />
Not measured<br />
70.2<br />
79.0<br />
72.1<br />
Not measured<br />
Not measured<br />
Not measured<br />
Not measured<br />
83.1<br />
77
00<br />
•1<br />
1<br />
& a o<br />
2 u<br />
0<br />
fc<br />
OH<br />
§<br />
* i-H<br />
-4-*<br />
2<br />
i.u -<br />
0.9 -<br />
0.8 -<br />
0.7 -<br />
0.6 -<br />
0.5 -<br />
0.4 -<br />
0.3 -<br />
0.2 o.i -<br />
0.0 Jr ¥ i • i<br />
• • A A • ***<br />
0 200 400<br />
Bed Volumes<br />
600<br />
Figure 3.30 <strong>Removal</strong> <strong>of</strong> perchlorate by <strong>GAC</strong> mixed with alum<strong>in</strong>um shot (<strong>in</strong>itial<br />
conditions: DDW, EBCT = 25 m<strong>in</strong>utes, pH 2.0, C1O4" = 2.7 mg/L)<br />
n<br />
c3<br />
'5<br />
'51<br />
The next step was to determ<strong>in</strong>e if perchlorate was be<strong>in</strong>g removed by adsorption or chemical<br />
reduction. If perchlorate was be<strong>in</strong>g reduced completely to chloride, then it should have been possible<br />
to observe a stoichiometric <strong>in</strong>crease <strong>in</strong> the effluent chloride concentration, based on the amount <strong>of</strong><br />
perchlorate be<strong>in</strong>g removed <strong>in</strong> the filters. However, no <strong>in</strong>crease <strong>in</strong> effluent chloride concentration was<br />
observed. If adsorption was the dom<strong>in</strong>ant removal mechanism, it should have been possible to wash<br />
the entire mass <strong>of</strong> adsorbed perchlorate <strong>of</strong>f the carbon us<strong>in</strong>g a concentrated nitrate solution. The<br />
<strong>GAC</strong> was washed until no perchlorate could be detected <strong>in</strong> the filter effluent. A 10 mg/L nitrate<br />
solution was used to desorb perchlorate from the metal-catalyzed <strong>GAC</strong> filters (Table 3.12) followed<br />
by a 200 mg/L nitrate solution (Table 3.13).<br />
When the 10 mg/L nitrate solution was applied, 79 percent <strong>of</strong> the perchlorate removed by the<br />
<strong>GAC</strong>/alum<strong>in</strong>um filter was desorbed, but almost no perchlorate was released from the copper- <strong>and</strong> z<strong>in</strong>c<br />
oxide-impregnated <strong>GAC</strong> filter. When the 200 mg/L nitrate solution was applied, little perchlorate<br />
came <strong>of</strong>f <strong>of</strong> the <strong>GAC</strong>/alum<strong>in</strong>um filter, while essentially all <strong>of</strong> the perchlorate was released from the<br />
copper- <strong>and</strong> z<strong>in</strong>c oxide-impregnated <strong>GAC</strong> filter.<br />
Table 3.12 <strong>Perchlorate</strong> mass balance: mass perchlorate removed by <strong>and</strong> recovered from metal-<br />
catalyzed <strong>GAC</strong> (10 mg/L NO3' r<strong>in</strong>se)<br />
Filter<br />
<strong>GAC</strong>/alum<strong>in</strong>um<br />
Cu- <strong>and</strong> ZnOimpregnated<br />
<strong>GAC</strong><br />
Mass perchlorate removed<br />
by carbon (mg)<br />
52.10<br />
51.80<br />
Mass perchlorate<br />
recovered (mg)<br />
41.10<br />
0.97<br />
% recovery<br />
78.9<br />
1.9<br />
Table 3.13 Extended perchlorate mass balance: total mass removed by <strong>and</strong> recovered from metal-<br />
catalyzed <strong>GAC</strong> (10 + 200 mg/L NO3 " r<strong>in</strong>ses)<br />
Filter Mass perchlorate Mass perchlorate % recovery<br />
removed by carbon washed <strong>of</strong>f carbon<br />
__________________(mg)__________(mg)__________________<br />
<strong>GAC</strong>/alum<strong>in</strong>um 52.10 42.73 82.0<br />
Cu- <strong>and</strong> ZnOimpregnated<br />
<strong>GAC</strong>_______51.80__________51.39__________99.2_____<br />
79
Thus, perchlorate was removed by the copper- <strong>and</strong> z<strong>in</strong>c oxide-impregnated <strong>GAC</strong> filter by ion<br />
exchange. Most <strong>of</strong> the perchlorate that was removed by the <strong>GAC</strong>/alum<strong>in</strong>um filter was also removed<br />
by ion exchange, although 18 percent (9.37 mg) <strong>of</strong> the perchlorate removed by the filter was not<br />
recovered. A fraction <strong>of</strong> the <strong>in</strong>fluent perchlorate may have been chemically reduced or may still be<br />
on the carbon, but there is no evidence to support either possibility. Even under the low pH <strong>and</strong> high<br />
perchlorate concentration conditions <strong>in</strong> these experiments, ion exchange is the dom<strong>in</strong>ant mechanism<br />
<strong>of</strong> perchlorate removal by <strong>GAC</strong>.<br />
<strong>Ozone</strong>/Hydrogen Peroxide/<strong>Perchlorate</strong> Batch Tests<br />
Table 3.14 shows the results <strong>of</strong> the ozone/hydrogen peroxide/perchlorate batch tests, <strong>and</strong> no<br />
perchlorate reduction occurred <strong>in</strong> the presence <strong>of</strong> ozone or ozone <strong>and</strong> hydrogen peroxide.<br />
PERCHLORATE - BIOLOGICAL REDUCTION<br />
BAC Filtration Experiments<br />
Electron Donor Limited Experiments<br />
Figure 3.32 illustrates the results <strong>of</strong> the BAC filter/syr<strong>in</strong>ge pump experiments. The figure<br />
exam<strong>in</strong>es the fraction <strong>of</strong> perchlorate rema<strong>in</strong><strong>in</strong>g versus the concentration <strong>of</strong> nitrate <strong>in</strong> the effluent <strong>of</strong><br />
the 25-m<strong>in</strong>ute EBCT filter. It is important to note that biological removal <strong>of</strong> perchlorate has been<br />
achieved. The ion exchange capacity <strong>of</strong> the <strong>GAC</strong> had been exhausted prior to the start <strong>of</strong> this<br />
experiment. There has been no evidence that perchlorate can be abiotically reduced by <strong>GAC</strong>.<br />
Therefore, any removal demonstrated dur<strong>in</strong>g this experiment is the result <strong>of</strong> biological reduction.<br />
The second <strong>in</strong>terest<strong>in</strong>g result illustrated by Figure 3.32 is that with low <strong>in</strong>fluent DO<br />
concentrations, the biological reduction <strong>of</strong> perchlorate is highly sensitive to the concentration <strong>of</strong><br />
nitrate present <strong>in</strong> the water. As nitrate concentration <strong>in</strong>creases, perchlorate removal decreases.<br />
Figure 3.32 also <strong>in</strong>dicates that even though the <strong>in</strong>fluent nitrate concentrations were constant,<br />
nitrate removed by the BAC filters was variable. This was caused by the fact that the float<strong>in</strong>g,<br />
80
Jar<br />
cio4-<br />
42.8<br />
40.4<br />
40.5<br />
41.1<br />
6<br />
C/C0<br />
1.00<br />
0.95<br />
0.95<br />
0.96<br />
Table 3.14 <strong>Ozone</strong>/hydrogen peroxide/perchlorate batch tests<br />
(<strong>in</strong>itial conditions: DDW, pH 8.8)<br />
Time<br />
(m<strong>in</strong>)<br />
Jar<br />
cio4-<br />
1<br />
C/C0<br />
Jar<br />
cio4-<br />
2<br />
C/C0<br />
Jar<br />
cio;<br />
3<br />
C/C0<br />
Jar<br />
cio 4-<br />
4<br />
C/C0<br />
Jar<br />
cio 4-<br />
5<br />
C/C0<br />
oo<br />
0<br />
46.6<br />
1.00<br />
46.5<br />
1.00<br />
46.2<br />
1.00<br />
43.2<br />
1.00<br />
43.5<br />
1.00<br />
5<br />
45.9<br />
0.98<br />
45.6<br />
0.98<br />
43.1<br />
0.93<br />
44.3<br />
1.03<br />
41.8<br />
0.96<br />
10<br />
47.1<br />
1.01<br />
45.9<br />
0.99<br />
44.7<br />
0.97<br />
43.6<br />
1.01<br />
41.0<br />
0.94<br />
15<br />
45.9<br />
0.98<br />
45.3<br />
0.98<br />
43.9<br />
0.95<br />
42.4<br />
0.98<br />
41.4<br />
0.95
2.5mg/L<br />
Influent NOs- =<br />
<strong>of</strong> nitrate on perchlorate reduction. Three different <strong>in</strong>fluent nitrate concentrations were used: 0.1<br />
(concentration <strong>of</strong> nitrate present <strong>in</strong> DDW with no external spike), 1.0, <strong>and</strong> 1.5 mg/L. Figure 3.33<br />
shows the results <strong>of</strong> the experiments with <strong>in</strong>fluent nitrate concentrations <strong>of</strong> 1.0 <strong>and</strong> 1.5 mg/L.<br />
Several important results were obta<strong>in</strong>ed from these experiments. First, due to the excess<br />
electron donor, the effluent nitrate concentration was constant for a given <strong>in</strong>fluent nitrate<br />
concentration. That is, EBCT was controll<strong>in</strong>g nitrate removal <strong>in</strong> the filters. This is <strong>in</strong> contrast to the<br />
previous BAG filter experiments which showed fluctuat<strong>in</strong>g effluent nitrate concentrations for a given<br />
<strong>in</strong>fluent nitrate concentration. Second, for a given nitrate concentration, the degree <strong>of</strong> perchlorate<br />
removal was essentially constant. Therefore, these experiments have been successful at elim<strong>in</strong>at<strong>in</strong>g<br />
the condition where electron donor concentration limits the removal <strong>of</strong> perchlorate. The<br />
concentration <strong>of</strong> nitrate present <strong>in</strong> the filter has become the variable controll<strong>in</strong>g the removal <strong>of</strong><br />
perchlorate. Third, it was observed that the effluent nitrate concentration needed to be ma<strong>in</strong>ta<strong>in</strong>ed<br />
between 0.05 <strong>and</strong> 0.35 mg/L to obta<strong>in</strong> >95 percent perchlorate removal. When the <strong>in</strong>fluent nitrate<br />
concentration was 1.5 mg/L, the average effluent nitrate concentration was 0.35 mg/L. Under these<br />
1.0<br />
§ 0.9<br />
oB 0.8<br />
1 0-5 H<br />
I 0.4 J<br />
o<br />
f£ c<br />
1 0.2 H<br />
ts<br />
0.0<br />
B B B<br />
v/<br />
/<br />
Influent NO3"= 1.5 mg/L<br />
Effluent NCh" = 0.35 mg/L<br />
B<br />
Influent NO3~= 1.0 mg/L<br />
Effluent NCb" = 0.05 mg/L<br />
0 200 400 600 800 1000 1200<br />
Bed Volumes<br />
Figure 3.33 Nitrate isolation experiments (<strong>in</strong>itial conditions: 2.5 mg/L DO, 50 ng/L C104",<br />
-1.7 mg/L as C added electron donor, pH 7.5)<br />
83
conditions, 65 percent perchlorate removal was observed (Figure 3.33). When the <strong>in</strong>fluent nitrate<br />
concentration was 1.0 mg/L, the average effluent nitrate concentration was 0.05 mg/L, <strong>and</strong><br />
perchlorate removal was >95 percent (Figure 3.33). S<strong>in</strong>ce it has been noted that there is an <strong>in</strong>verse<br />
relationship between the effluent nitrate concentration <strong>and</strong> perchlorate removal, the 0.05-0.35 mg/L<br />
range <strong>of</strong> nitrate concentrations was taken directly from Figure 3.33. This means that these particular<br />
BAG filters can remove perchlorate <strong>in</strong> the presence <strong>of</strong> nitrate, but only a low concentration <strong>of</strong> nitrate<br />
can be present if >95 percent perchlorate removal is expected.<br />
The experiment with no nitrate added was run for 950 bed volumes. For the duration <strong>of</strong> the<br />
experiment, no perchlorate was measured <strong>in</strong> the effluent. Thus, <strong>in</strong> presence <strong>of</strong> very low nitrate<br />
concentrations (
Effluent <strong>Perchlorate</strong>, C/C<br />
l.U<br />
0.8 -<br />
0.6-<br />
0.4 -<br />
* * * *<br />
• Non-Nitrate pre-loaded<br />
• Nitrate pre-loaded<br />
0.2 -<br />
B<br />
sa a u • H B B<br />
o.o - ———————,———————,———————————<br />
0 200 400 600 800 1000 1200 1400<br />
Bed Volumes<br />
Figure 3.34 Comparison <strong>of</strong> perchlorate removal for nitrate preload <strong>and</strong> non-nitrate<br />
preload conditions (<strong>in</strong>itial conditions: DDW, DO = 2.5 mg/L, C1O4' = 50 ug/L,<br />
NCV = 0 mg/L, 1.5 mg/L as C added electron donor, pH 7.5, EBCT = 15 m<strong>in</strong>utes)<br />
perchlorate-reducers could be reduc<strong>in</strong>g nitrate to ammonia for assimilatory purposes. Other than<br />
endogenous respiration, nitrate is the only nitrogen source for the microorganisms <strong>in</strong> the BAC filter<br />
s<strong>in</strong>ce there is no ammonia <strong>in</strong> the <strong>in</strong>fluent. If endogenous respiration was not provid<strong>in</strong>g sufficient<br />
nitrogen to satisfy microbial nutrient requirements, a portion <strong>of</strong> the biomass would die, thus<br />
decreas<strong>in</strong>g the quantity <strong>of</strong> perchlorate-reduc<strong>in</strong>g biomass. The more nitrate present, the more nitrogen<br />
available for assimilation <strong>and</strong> growth. Increased growth means that there would be more cells present<br />
to degrade perchlorate, thus caus<strong>in</strong>g an <strong>in</strong>creased rate <strong>of</strong> perchlorate reduction. This scenario<br />
suggests that ammonia is the key component <strong>of</strong> the system for the microbial community. Nitrate is<br />
not required per se, but is only used to generate ammonia. A second possible explanation for the<br />
improved perchlorate removal performance when the filter was preloaded with nitrate is that<br />
microorganisms <strong>in</strong> the filter dissimilatively reduce nitrate to ammonia, nitrous oxide, or d<strong>in</strong>itrogen<br />
gas to ga<strong>in</strong> energy for growth. The electron transport <strong>in</strong>volved <strong>in</strong> these reduction pathways directly<br />
leads to the production <strong>of</strong> energy. Microorganisms can then use this energy to grow, <strong>and</strong> the<br />
generation <strong>of</strong> more cells would <strong>in</strong>crease the rate <strong>of</strong> perchlorate reduction. This scenario suggests that<br />
nitrate is more important for microbial growth than ammonia.<br />
It is also possible that the decrease <strong>in</strong> perchlorate reduction ability was <strong>in</strong>duced by the<br />
85
suppression <strong>of</strong> the active perchlorate-reduc<strong>in</strong>g enzyme due to a lack <strong>of</strong> nitrate reduction activity. The<br />
enzymes <strong>in</strong>volved <strong>in</strong> nitrate reduction may also be responsible for perchlorate reduction, <strong>and</strong> nitrate<br />
reduction may be required to ma<strong>in</strong>ta<strong>in</strong> the activity <strong>of</strong> these enzymes Without nitrate reduction, these<br />
enzymes may have become <strong>in</strong>active.<br />
<strong>Perchlorate</strong> <strong>Removal</strong> Versus EBCT with No Influent Nitrate<br />
As an extension <strong>of</strong> the nitrate preload experiment, perchlorate removal versus EBCT without<br />
an <strong>in</strong>fluent nitrate spike was <strong>in</strong>vestigated. EBCTs <strong>of</strong> 25-, 15-, 10-, <strong>and</strong> 5-m<strong>in</strong>utes were used,<br />
respectively. Each EBCT was first run without any nitrate <strong>in</strong> the <strong>in</strong>fluent. The <strong>in</strong>fluent parameters<br />
were as follows: DDW, C1O4" = 50 ug/L, NO3" = 0.1 mg/L (ambient), DO = 2.5 mg/L, phosphate<br />
buffer = 1 mM, <strong>and</strong> pH 7.5. Electron donor (acetate, benzoate, <strong>and</strong> pyruvate) was supplied at 1.5<br />
mg/L <strong>of</strong> equivalent nitrate dem<strong>and</strong> <strong>in</strong> excess <strong>of</strong> that required to remove all <strong>in</strong>fluent DO. Figure 3.35<br />
illustrates the results <strong>of</strong> this experiment.<br />
Each EBCT was run for two weeks, which allowed the microbial community time to adjust<br />
to changes <strong>in</strong> flowrate. Figure 3.35 shows that, on average, perchlorate removal decl<strong>in</strong>ed as EBCT<br />
1.2<br />
«j uH<br />
|<br />
H a
was decreased. It is <strong>in</strong>terest<strong>in</strong>g to note, however, that dur<strong>in</strong>g the 10-m<strong>in</strong>ute EBCT, perchlorate<br />
removal steadily decl<strong>in</strong>ed. For the first few days <strong>of</strong> the 10-m<strong>in</strong>ute experiment, approximately 70<br />
percent <strong>of</strong> the <strong>in</strong>fluent perchlorate was be<strong>in</strong>g removed. By the end <strong>of</strong> the two-week period,<br />
perchlorate removal had dropped to about 10 percent. Then, the 5-m<strong>in</strong>ute EBCT showed essentially<br />
no perchlorate removal. To determ<strong>in</strong>e if this loss <strong>in</strong> perchlorate reduction capacity was solely a<br />
function <strong>of</strong> EBCT, the 25-m<strong>in</strong>ute EBCT experiment was rerun after the 5-m<strong>in</strong>ute experiment was<br />
complete. Figure 3.36 shows the results <strong>of</strong> this experiment (25-m<strong>in</strong>ute EBCT 2) <strong>and</strong> compares it to<br />
the 25-m<strong>in</strong>ute experiment run previously (25-m<strong>in</strong>ute EBCT 1). The second 25-m<strong>in</strong>ute EBCT was<br />
run approximately six weeks after the end <strong>of</strong> the first 25-m<strong>in</strong>ute EBCT experiment.<br />
The data <strong>in</strong> Figure 3.36 clearly <strong>in</strong>dicate that the filter had lost most <strong>of</strong> its ability to reduce<br />
perchlorate. Average perchlorate removal for the first 25-m<strong>in</strong>ute experiment was approximately 80<br />
percent. Six weeks later, the second 25-m<strong>in</strong>ute experiment showed average perchlorate removal<br />
levels <strong>of</strong> less than 5 percent. There are three possible explanations for this phenomenon <strong>and</strong> are the<br />
same as those described for the nitrate preload experiments. The first relates to the nitrogen<br />
requirements <strong>of</strong> microbial communities. It is possible that the microbial community <strong>in</strong> the filter relied,<br />
<strong>in</strong> part, on assimilative nitrate reduction to produce the nitrogen they require for cell growth.<br />
1 .Z<br />
Effluent <strong>Perchlorate</strong>,<br />
1.0 -<br />
0.8 -<br />
I 0 - 6 -<br />
0.4 -<br />
0.2 -<br />
n n<br />
*— ——— ^— --— -"""<br />
-«— 25 -M<strong>in</strong>ute EBCT 2<br />
-H&- 25 -M<strong>in</strong>ute EBCT 1<br />
^^a — — 0 —— n ——— B ——— B^^<br />
100 200 300 400 500 600<br />
Bed Volumes<br />
700<br />
Figure 3.36 Comparison <strong>of</strong> perchlorate removal for 25-m<strong>in</strong>ute EBCT experiments<br />
(<strong>in</strong>itial conditions: DDW, 2.5 mg/L DO, 50 ug/L C1O4", 0 mg/L NO3", 1.5 mg/L as<br />
C added electron donor, pH 7.5)<br />
87
Without this nitrogen, the community may have died. There is some evidence to suggest, though,<br />
that this may not entirely expla<strong>in</strong> the loss <strong>of</strong> perchlorate reduction ability. Throughout the<br />
experiment, bromate was added to the <strong>in</strong>fluent at a concentration <strong>of</strong> 20 ug/L. Dur<strong>in</strong>g the first 25-<br />
m<strong>in</strong>ute EBCT experiment, <strong>in</strong>fluent bromate was reduced to below detection <strong>in</strong> the filter. Dur<strong>in</strong>g the<br />
second 25-m<strong>in</strong>ute experiment, an average <strong>of</strong> 63 percent <strong>of</strong> the <strong>in</strong>fluent bromate was removed. While<br />
this clearly <strong>in</strong>dicates a decrease <strong>in</strong> the filter's ability to reduce bromate, significant bromate reduction<br />
ability still existed even after operat<strong>in</strong>g without an exogenous source <strong>of</strong> nitrogen for over two months.<br />
This suggests that endogenous respiration was provid<strong>in</strong>g some <strong>of</strong> the required nutritional nitrogen.<br />
It is possible that the perchlorate-reduc<strong>in</strong>g organisms are different than the bromate-reduc<strong>in</strong>g<br />
organisms <strong>and</strong> that the perchlorate-degraders dim<strong>in</strong>ished because they were outcompeted for<br />
nitrogen.<br />
A second possible explanation is that there was plenty <strong>of</strong> nitrogen available to the perchloratedegraders<br />
<strong>and</strong> that the loss <strong>of</strong> perchlorate reduction ability was caused by suppression <strong>of</strong> the<br />
perchlorate-reduc<strong>in</strong>g enzyme due to a lack <strong>of</strong> nitrate reduction. The enzymes responsible for nitrate<br />
reduction may also be <strong>in</strong>volved <strong>in</strong> perchlorate reduction, <strong>and</strong> nitrate reduction may be required to<br />
ma<strong>in</strong>ta<strong>in</strong> the activity <strong>of</strong> the perchlorate-reduc<strong>in</strong>g enzymes.<br />
A third possible explanation for the loss <strong>of</strong> perchlorate reduction ability is that the perchloratereduc<strong>in</strong>g<br />
microorganisms grow via dissimilative nitrate reduction, <strong>and</strong> a lack <strong>of</strong> nitrate reduction<br />
could reduce the population <strong>of</strong> perchlorate-reducers.<br />
Filter Recovery<br />
In an attempt to recover biological perchlorate reduction <strong>in</strong> the filter, nitrate was aga<strong>in</strong> added<br />
to the <strong>in</strong>fluent (Figure 3.37). Dur<strong>in</strong>g the first 1,700 BV, the C/C0 was greater than 1 as perchlorate<br />
desorbed from the <strong>GAC</strong>. Dur<strong>in</strong>g the previous period when no nitrate was added to the <strong>in</strong>fluent, the<br />
<strong>GAC</strong> ion exchange sites equilibrated with an <strong>in</strong>fluent nitrate concentration <strong>of</strong> 0.1 mg/L, thereby<br />
releas<strong>in</strong>g adsorbed nitrate from the <strong>GAC</strong>. <strong>Perchlorate</strong> could then adsorb to these sites. When nitrate<br />
was re<strong>in</strong>troduced to the <strong>in</strong>fluent dur<strong>in</strong>g the filter recovery experiments, nitrate replaced perchlorate<br />
on the ion exchange sites caus<strong>in</strong>g C/C0 for perchlorate to be greater than 1 from 0-1,700 BV.<br />
Figure 3.37 clearly <strong>in</strong>dicates that the filter is rega<strong>in</strong><strong>in</strong>g its ability to reduce perchlorate. The<br />
actual amount <strong>of</strong> biological perchlorate reduction that was occurr<strong>in</strong>g <strong>in</strong> the filter throughout the<br />
88
1<br />
1.6<br />
1.4 -<br />
1.2 -<br />
.1.0-<br />
1.5mg/L<br />
Influent NO3" = 1.0 mg/L 0.5 ng/L<br />
0.8 -<br />
0.6 -<br />
w<br />
0.4<br />
0.2 H<br />
0.0<br />
0 1000 2000<br />
Bed Volumes<br />
3000 4000<br />
Figure 3.37 BAG filter recovery experiments (<strong>in</strong>itial conditions: DDW, 2.5 mg/L<br />
DO, 50 ug/L C1CV, 1.5-1.8 mg/L as C added electron donor pH 7.5, EBCT = 25<br />
m<strong>in</strong>utes)<br />
recovery experiments was not detectable due to the fact that biological reduction may have been<br />
masked by desorption. That is, biological reduction may have been occurr<strong>in</strong>g throughout the<br />
recovery experiments but could not be measured, as perchlorate was simultaneously desorb<strong>in</strong>g from<br />
the <strong>GAC</strong>. Experiments will be run to further <strong>in</strong>vestigate why the BAG filter's perchlorate<br />
reduction ability deteriorated when nitrate was removed from the <strong>in</strong>fluent. These experiments will<br />
isolate whether the perchlorate-reduc<strong>in</strong>g microorganisms <strong>in</strong> the filter simply died due to nitrogen<br />
deprivation, lack <strong>of</strong> energy for growth, or loss <strong>of</strong> enzymatic activity caused by a lack <strong>of</strong> nitrate<br />
reduction.<br />
Batch Biological Experiments<br />
Plat<strong>in</strong>g Experiments <strong>and</strong> Serum Bottle Tests<br />
Twelve isolates which had been grow<strong>in</strong>g on agar plates were <strong>in</strong>oculated <strong>in</strong>to aqueous medium,<br />
as described <strong>in</strong> Chapter 2. Table 3.15 shows the results <strong>of</strong> samples taken 3 weeks after the aqueous<br />
medium was <strong>in</strong>oculated.<br />
89
While some <strong>of</strong> the isolates reduced small amounts <strong>of</strong> perchlorate, isolate B9 reduced<br />
perchlorate to below the detection limit. B9 was then used <strong>in</strong> two sets <strong>of</strong> serum bottle experiments.<br />
At the <strong>in</strong>ception <strong>of</strong> the first serum bottle experiment, it was not known how fast isolate B9 would<br />
reduce perchlorate. Therefore, as sample volume was limited, samples were taken on the third, sixth,<br />
<strong>and</strong> eighth day after <strong>in</strong>oculation. Even after the eighth day, the bottles conta<strong>in</strong><strong>in</strong>g CUW <strong>and</strong> no<br />
external electron donor showed no perchlorate removal. The biodegradable fraction <strong>of</strong> NOM does<br />
not suffice as an electron donor for perchlorate reduction. The DDW serum bottles did demonstrate<br />
perchlorate removal. By the third day, each bottle conta<strong>in</strong><strong>in</strong>g an <strong>in</strong>itial nitrate concentration <strong>of</strong> 0.3<br />
Table 3.15 Reduction <strong>of</strong> perchlorate <strong>in</strong> aqueous medium us<strong>in</strong>g microbial isolates<br />
(<strong>in</strong>itial conditions: DDW, 46.3 ug/L CIO/, 1.0 mg/L NCV, 0 mg/L DO, pH 7.5)<br />
Isolate<br />
Control<br />
(no <strong>in</strong>oculum)<br />
Bl<br />
B2<br />
B3<br />
B4<br />
B5<br />
B6<br />
B7<br />
B8<br />
B9<br />
BIO<br />
Bll<br />
B12<br />
cio4-<br />
(Hg/L)<br />
46.5<br />
45.2<br />
47.3<br />
42.5<br />
46.4<br />
45.3<br />
45.7<br />
42.5<br />
45.1<br />
or 1.5 mg/L nitrate showed no perchlorate rema<strong>in</strong><strong>in</strong>g. The bottles us<strong>in</strong>g benzoate with 0 mg/L nitrate<br />
removed 50 percent <strong>of</strong> the <strong>in</strong>fluent perchlorate by the third day while the bottles us<strong>in</strong>g lactate with<br />
0 mg/L nitrate removed 17 percent <strong>of</strong> the <strong>in</strong>fluent perchlorate dur<strong>in</strong>g the same time period. These<br />
data suggest that pechlorate removal k<strong>in</strong>etics can be enhanced by the presence <strong>of</strong> nitrate <strong>and</strong> that B9<br />
might prefer benzoate slightly over lactate as an electron donor. There were no differences <strong>in</strong><br />
perchlorate removal performance between the bottles conta<strong>in</strong><strong>in</strong>g bromate <strong>and</strong> those without bromate.<br />
Thus, it can be concluded that perchlorate <strong>and</strong> bromate reduction are <strong>in</strong>dependent <strong>of</strong> each other.<br />
The second serum bottle experiment was designed to further test the differences <strong>in</strong> perchlorate<br />
reduction k<strong>in</strong>etics caused by differences <strong>in</strong> <strong>in</strong>itial nitrate concentration. Figures 3.38-3.41 show<br />
perchlorate <strong>and</strong> nitrate removal versus <strong>in</strong>cubation time.<br />
0123456<br />
Incubation Time (days)<br />
0.0<br />
Figure 3.38 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50<br />
ug/L C1CV, 0.1 mg/L NO3-, DO ^0 mg/L, pH 7.5)<br />
91
~\<br />
I<br />
0<br />
012345<br />
Incubation Time (days)<br />
Figure 3.39 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50<br />
Hg/L CIO/, 0.4 mg/L NO3', DO ^0 mg/L, pH 7.5)<br />
2.5<br />
0 1<br />
23456<br />
Incubation Time (days)<br />
Figure 3.40 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50<br />
Hg/L C1CV, 1.6 mg/L NO3', DO *0 mg/L, pH 7.5)<br />
0<br />
92
^——s<br />
0 4<br />
0 234<br />
Incubation Time (days)<br />
Figure 3.41 Isolate B9 serum bottle experiment 2 (<strong>in</strong>itial conditions: DDW, 50<br />
\igfL C1O4', 2.6 mg/L NO3', DO ^0 mg/L, pH 7.5)<br />
The first observation that can be made from Figure 3.42 is that isolate B9 does have the ability to<br />
reduce nitrate. Consequently, there is at least one microbial species <strong>in</strong> the BAG filters that has the<br />
ability to reduce both nitrate <strong>and</strong> perchlorate. Another <strong>in</strong>terest<strong>in</strong>g observation from these plots is that<br />
significant perchlorate removal did not occur <strong>in</strong> any <strong>of</strong> the bottles until the nitrate concentration was<br />
^0.1 mg/L. This falls <strong>in</strong> the range <strong>of</strong> nitrate concentrations required for greater than 95 percent<br />
perchlorate removal <strong>in</strong> the BAG filters, as determ<strong>in</strong>ed dur<strong>in</strong>g the nitrate isolation experiments.<br />
Nitrate <strong>and</strong> perchlorate degradation k<strong>in</strong>etics are compared <strong>in</strong> Figures 3.42 <strong>and</strong> 3.43,<br />
respectively. Figure 3,42 also shows that dur<strong>in</strong>g the first 1.5 days <strong>of</strong> <strong>in</strong>cubation, the nitrate was not<br />
reduced by isolate B9. It is possible that there was some residual DO <strong>in</strong> the bottles that may have<br />
<strong>in</strong>ibited the reduction <strong>of</strong> nitrate. Once nitrate degradation began, however, B9 required less than one<br />
day to reduce the nitrate concentration to 0.1 mg/L or lower, regardless <strong>of</strong> the <strong>in</strong>itial nitrate<br />
concentration.<br />
93
j.u -<br />
f-x — x —— x x-x<br />
2.5 -<br />
ll 2.0-<br />
S<br />
u 1.5 -<br />
1<br />
g i.o -<br />
"A — A ——— &-A-A \<br />
0-5 - İ<br />
•• • ^ '*^*, • •.-• H ' i • ——<br />
() 1 2 3 4 5 f<br />
Incubation Time (days)<br />
Figure 3.42 Isolate B9 serum bottle experiment 2: nitrate degradation (<strong>in</strong>itial<br />
conditions: DDW, 50 ng/L CIO/, DO ^0 mg/L, pH 7.5)<br />
0.1mg/LNO3-<br />
0.4 mg/L N03<br />
1.6 mg/L NO3.<br />
2.6 mg/L N03.<br />
0<br />
0 1 23456<br />
Incubation Time (days)<br />
Figure 3.43 Isolate B9 serum bottle experiment 2: perchlorate degradation (<strong>in</strong>itial<br />
conditions: DDW, 50 ng/L C1O4', DO ^0 mg/L, pH 7.5)<br />
94
Figure 3.43 illustrates the differences <strong>in</strong> perchlorate reduction rates when the <strong>in</strong>itial nitrate<br />
concentration was varied. As the <strong>in</strong>itial nitrate concentration <strong>in</strong>creased, the rate <strong>of</strong> perchlorate<br />
reduction <strong>in</strong>creased. B9 required approximately 3 days to reduce perchlorate to below the LOD<br />
when the <strong>in</strong>itial nitrate concentration was 1.6 or 2.6 mg/L, while it took over 5 days when the <strong>in</strong>itial<br />
nitrate concentration was 0.1 or 0.4 mg/L. Similar to the BAG filter experiments, there are three<br />
possible explanations for this: 1) <strong>in</strong>creased growth due to dissimilative nitrate reduction, 2) <strong>in</strong>creased<br />
growth due to assimilative nitrate reduction to ammonia (nutrient requirements), <strong>and</strong> 3) <strong>in</strong>creased<br />
perchlorate-reduc<strong>in</strong>g enzymatic activity due to <strong>in</strong>creased nitrate reduction.<br />
Dilution-to-Ext<strong>in</strong>ction Experiments<br />
Table 3.16 shows the results <strong>of</strong> the perchlorate dilution-to-ext<strong>in</strong>ction experiments. After six<br />
weeks, the microorganisms reduced perchlorate to below detection <strong>in</strong> every dilution except 1:100.<br />
<strong>Perchlorate</strong> reduction may occur at a high dilution <strong>and</strong> not at a low dilution because at the low<br />
dilution there may be non-perchlorate reducers present that outcompete the numerically dom<strong>in</strong>ant<br />
perchlorate-reducers for substrate. This can occur if the non-perchlorate reducers are faster growers<br />
than the perchlorate reducers. At the higher dilutions, these non-pechlorate reducers are diluted out,<br />
leav<strong>in</strong>g the numerically dom<strong>in</strong>ant perchlorate-reducers alone <strong>in</strong> the sample.<br />
Table 3.16 Dilution-to-ext<strong>in</strong>ction experiments: perchlorate (<strong>in</strong>itial conditions: DDW, 50 |u.g/L<br />
C1CV, 20 ng/L BrCV, 1.0 mg/L NO3", <strong>and</strong> 5 mM phosphate buffer<br />
Dilution<br />
C104-(ug/L)<br />
t = 6 weeks<br />
C104- (ng/L)<br />
t = 8 weeks<br />
N03- (mg/L)<br />
t = 8 weeks<br />
N02- (mg/L)<br />
t = 8 weeks<br />
Br0 3-(ug/L)<br />
t = 8 weeks<br />
Control -<br />
no <strong>in</strong>oculum<br />
53.1<br />
Not<br />
measured<br />
1.05<br />
After eight weeks, the microorganisms reduced all <strong>of</strong> the bromate <strong>and</strong> almost all <strong>of</strong> the nitrate,<br />
<strong>and</strong> significant nitrite accumulation did not occur except <strong>in</strong> the zero dilution Balch tubes. The nitrite<br />
degraders may have been outcompeted by other microorganisms under the zero dilution condition.<br />
Microscopic Exam<strong>in</strong>ation <strong>of</strong> Microorganisms <strong>in</strong> the Filters<br />
Biomass from the perchlorate-reduc<strong>in</strong>g BAG filters has been exam<strong>in</strong>ed microscopically.<br />
Three representative pictures are shown <strong>in</strong> Figures 3.44-3.46. The microscopy was done us<strong>in</strong>g phase<br />
contrast with lOOOx magnification. Therefore, the dark objects <strong>in</strong> the pictures are microorganisms<br />
while the lighter, retractile images are likely debris.<br />
Rods were the most common bacterial morphology observed. The dark specks <strong>in</strong> Figures<br />
3.44 <strong>and</strong> 3.45 are these rod-shaped bacteria. They were observed to exist most <strong>of</strong>ten as s<strong>in</strong>gle<br />
organisms (Figure 3.44) but were also found <strong>in</strong> floe form (Figure 3.45).<br />
Some common morphologies were observed between the biomass from the bromate <strong>and</strong><br />
perchlorate BAG filters. Protozoa shaped like round-bottom flasks are one example <strong>of</strong> this (Figures<br />
3.21 <strong>and</strong> 3.44). Other morphologies found <strong>in</strong> the perchlorate samples were clearly dist<strong>in</strong>ctive. The<br />
best example <strong>of</strong> this is the large protozoan that was observed <strong>in</strong> biomass taken from the the filter <strong>of</strong><br />
the DDW apparatus (Figure 3.46). The micrograph shows about one-half <strong>of</strong> the protozoan. The<br />
small black dots <strong>in</strong>side this protozoan are bacteria which it has consumed - evidence <strong>of</strong> microbial<br />
graz<strong>in</strong>g. From this micrograph, it can be noted that this particular protozoan is extremely large -<br />
thous<strong>and</strong>s <strong>of</strong> times larger than the typical bacteria found <strong>in</strong> these filters.<br />
96
Figure 3.44 Rods <strong>and</strong> protozoa from the BAG filter<br />
Figure 3.45 Bacterial floe from the BAG filter<br />
97
Figure 3.46 Large protozoan from the BAG filter<br />
Denatur<strong>in</strong>g Gradient Gel Electrophoresis Experiment<br />
Figure 3.22 shows the DGGE pr<strong>of</strong>ile obta<strong>in</strong>ed with material harvested from various BAC<br />
filters. Each sample was run <strong>in</strong> duplicate. The first lane is a st<strong>and</strong>ard sample. Lanes 2 <strong>and</strong> 3 were run<br />
with a fresh microbial sample from the effluent side <strong>of</strong> the CUW perchlorate filter that used CUW for<br />
one year. The microbial sample used <strong>in</strong> lanes 4 <strong>and</strong> 5 came from the effluent side <strong>of</strong> the DDW filter.<br />
Similar to the DGGE pr<strong>of</strong>iles generated us<strong>in</strong>g samples from the BAC filters operated for bromate<br />
removal, the pr<strong>of</strong>iles for the microbial community from the perchlorate-reduc<strong>in</strong>g filters reflect a<br />
diverse array <strong>of</strong> microorganisms. There are conserved b<strong>and</strong>s across the two different samples, while<br />
there are some b<strong>and</strong>s that are unique to each sample.<br />
98
CHAPTER 4. ACTIVITIES AT THE METROPOLITAN WATER DISTRICT<br />
OF SOUTHERN CALIFORNIA<br />
INTRODUCTION<br />
Bench- <strong>and</strong> pilot-scale studies were conducted at the Metropolitan Water District <strong>of</strong> Southern<br />
California (MWDSC) to <strong>in</strong>vestigate the reduction <strong>of</strong> perchlorate <strong>and</strong> bromate <strong>in</strong> biologically active<br />
carbon (BAG) filters. The foundation for these studies was developed at the University <strong>of</strong> Ill<strong>in</strong>ois<br />
at Urbana-Champaign (UIUC). Work done at UIUC demonstrated that efficient biological<br />
perchlorate <strong>and</strong> bromate reduction could be achieved <strong>and</strong> susta<strong>in</strong>ed us<strong>in</strong>g BAG filtration. Natural<br />
organic matter (NOM) <strong>and</strong> exogenous organic acids served as the electron donors for bromate <strong>and</strong><br />
perchlorate reduction for the bench-scale UIUC experiments, respectively. One unique aspect <strong>of</strong> the<br />
work done at the MWDSC is the fact that <strong>in</strong>fluent water was ozonated to provide the electron donor<br />
necessary for biological reduction. Ozonation <strong>of</strong> natural water can break large NOM molecules down<br />
<strong>in</strong>to more readily assimilable organic compounds such as aldehydes <strong>and</strong> carboxylic acids. Other<br />
biological perchlorate treatment processes reported <strong>in</strong> the literature require the addition <strong>of</strong> an<br />
exogenous electron donor to support biological reduction (Logan et al. 1999).<br />
Tests were performed at the MWDSC us<strong>in</strong>g blends <strong>of</strong> Colorado River Water (CRW) <strong>and</strong><br />
State Project Water (SPW). Results at the bench-scale were used to generate the design <strong>of</strong> pilotscale<br />
BAG filters, which were utilized to exam<strong>in</strong>e the issue <strong>of</strong> scale-up for biological perchlorate <strong>and</strong><br />
bromate reduction.<br />
OBJECTIVES<br />
The objective <strong>of</strong> the work done at the MWDSC was to <strong>in</strong>vestigate the reduction <strong>of</strong><br />
perchlorate <strong>and</strong> bromate <strong>in</strong> B AC filters after ozonation. Specific goals <strong>of</strong> the project were as follows:<br />
To demonstrate that low ug/L concentrations <strong>of</strong> perchlorate <strong>and</strong> bromate can be<br />
biologically reduced to below detection us<strong>in</strong>g the ozonation/activated carbon process<br />
for CRW/SPW blends at the bench-scale.<br />
• To <strong>in</strong>vestigate the mechanism <strong>of</strong> biological reduction, with special focus on the effects<br />
<strong>of</strong> perchlorate, bromate, nitrate, <strong>and</strong> dissolved oxygen (DO) concentrations.<br />
• To determ<strong>in</strong>e the optimal operat<strong>in</strong>g conditions to biologically reduce perchlorate <strong>and</strong><br />
bromate us<strong>in</strong>g BAG filtration.<br />
• To exam<strong>in</strong>e perchlorate <strong>and</strong> bromate removal us<strong>in</strong>g BAG filtration at the pilot-scale.<br />
99
BENCH-SCALE MINI-COLUMN STUDY<br />
Materials <strong>and</strong> Methods<br />
Water<br />
Currently the La Verne Pilot Plant at the Metropolitan Water District <strong>of</strong> Southern California<br />
(MWDSC) receives a blend <strong>of</strong> 25 percent State Project Water (SPW) <strong>and</strong> 75 percent Colorado River<br />
Water (CRW). Typical water characteristics are given <strong>in</strong> Table 4.1 based on historical data. More<br />
detailed water quality parameters are presented <strong>in</strong> a later section.<br />
The raw water blend was ozonated with approximately 2 mg/L <strong>of</strong> applied ozone us<strong>in</strong>g 100-<br />
gallon, 6-<strong>in</strong>ch ozone contactors located at the La Verne Pilot Plant. Compressed air was used as a<br />
feed gas for the ozone generator. The <strong>in</strong>itial DO concentration us<strong>in</strong>g compressed air as a feed gas<br />
yielded approximately 8-9 mg/L <strong>of</strong> DO after the ozonation process. The batch <strong>of</strong> ozonated water was<br />
spiked with sodium thiosulfate (Certified ACS, Fisher, Pittsburgh, PA) to reduce the DO level, <strong>and</strong><br />
the amount <strong>of</strong> sodium thiosulfate varied depend<strong>in</strong>g on the desired DO level (23-160 mg/L sodium<br />
thiosulfate to ma<strong>in</strong>ta<strong>in</strong> 0-5 mg/L DO).<br />
Table 4.1 Historical water quality for Metropolitan source water<br />
Parameter Typical Level at Metropolitan<br />
Alkal<strong>in</strong>ity 13 0 mg/L as CaCO3<br />
Bromide between 0.07 <strong>and</strong> 0.20 mg/L<br />
<strong>Bromate</strong> ND (< 3 (ig/L) - 55 ug/L (2 mg/L ozone dose)<br />
<strong>Perchlorate</strong> ND (< 4 ug/L) - 9 ug/L for CRW<br />
ND for SPW<br />
Nitrate 0.25 to 0.40 mg/L as N<br />
Color 3 cu<br />
Iron 0.020 mg/L<br />
Manganese 0.005 mg/L<br />
pH ~8<br />
UVA 0.1 cm- 1<br />
TOC 2.5 mg/L ____<br />
100
<strong>Perchlorate</strong> is only found <strong>in</strong> CRW, <strong>and</strong> the background level concentration ranged from < 4<br />
Mg/L to 9 ug/L (Table 4.1). Thus, perchlorate was spiked (10 to 50 ng/L) to the ozonated water for<br />
the bench-scale study. Flowrates <strong>of</strong> 1.6 <strong>and</strong> 2.7 mL/m<strong>in</strong>ute were used to achieve 25- <strong>and</strong> 15-m<strong>in</strong>ute<br />
empty-bed contact times (EBCTs), respectively, for a s<strong>in</strong>gle filter <strong>and</strong> 50- <strong>and</strong> 30-m<strong>in</strong>ute EBCTs,<br />
respectively, for two filters <strong>in</strong> series.<br />
Column Setup<br />
The bench-scale columns were constructed <strong>in</strong> May 1999 at the La Verne Pilot Plant, <strong>and</strong> have<br />
been <strong>in</strong> operation s<strong>in</strong>ce the beg<strong>in</strong>n<strong>in</strong>g <strong>of</strong> June 1999. The columns were modeled after those already<br />
<strong>in</strong> use at the University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Champaign <strong>in</strong> terms <strong>of</strong> physical setup <strong>and</strong> flow<br />
characteristics.<br />
Two 60-cm glass columns with 2.5-cm <strong>in</strong>ner diameter (Sigma, St. Louis, MO) were used as<br />
filter columns. The columns have V4-28 PTFE fitt<strong>in</strong>gs, <strong>and</strong> Tygon® tub<strong>in</strong>g was used to connect the<br />
columns <strong>and</strong> the reservoir. The reservoir was constructed from a 15-gallon Low Density<br />
Polyethylene (LDPE) tank with a spigot, <strong>and</strong> a commercially available plastic device was used as a<br />
float<strong>in</strong>g cover with 1/16" clearance with the reservoir wall. The schematic <strong>of</strong> the bench-scale setup<br />
is shown <strong>in</strong> Figure 4.1.<br />
Spent Norit <strong>GAC</strong> 8x12 (Norit America Inc., Atlanta, GA) was taken from the La Verne Pilot<br />
Plant filters, which had been <strong>in</strong> service for 18 months, <strong>and</strong> was used to pack the m<strong>in</strong>i-columns. Each<br />
column was packed with 1 cm <strong>of</strong> 3-mm glass beads at the top <strong>and</strong> bottom <strong>of</strong> an 8-cm bed <strong>of</strong> <strong>GAC</strong>.<br />
After spik<strong>in</strong>g the water with 10 to 50 ng/L <strong>of</strong> perchlorate (sodium perchlorate, Sigma<br />
Chemical Co. St. Louis, MO), a float<strong>in</strong>g cover was placed <strong>in</strong>to the reservoir. The reservoir water was<br />
changed once a week with freshly ozonated water.<br />
101
Effluent<br />
Sample<br />
Bottle<br />
M<strong>in</strong>i<br />
Column<br />
Float<strong>in</strong>g<br />
Cover<br />
Peristaltic<br />
Pump<br />
Reservoir<br />
Mid Sampl<strong>in</strong>g Po<strong>in</strong>t<br />
Figure 4.1. Schematic <strong>of</strong> the bench-scale filter setup at Metropolitan (not to scale)<br />
Sampl<strong>in</strong>g<br />
The ozonated water from the bench-scale set-up was analyzed for perchlorate, bromate, <strong>and</strong><br />
ultraviolet (UV) absorbance, total organic carbon (TOC), conductivity, <strong>and</strong> pH. Aldehydes <strong>and</strong><br />
carboxylic acids were also measured once bromate <strong>and</strong> perchlorate reduction were established <strong>in</strong> the<br />
columns. When appropriate, the concentrations <strong>of</strong> DO <strong>and</strong> nitrate were also analyzed. Samples were<br />
collected at the <strong>in</strong>fluent, after the first column (midpo<strong>in</strong>t), <strong>and</strong> after the second column (effluent), as<br />
shown <strong>in</strong> Figure 4.1.<br />
Analytical Methods<br />
All <strong>in</strong>organic <strong>and</strong> organic analyses were performed at Metropolitan's Water Quality<br />
Laboratory <strong>in</strong> accordance with the procedures described <strong>in</strong> St<strong>and</strong>ard Methods (1995), except as<br />
noted <strong>in</strong> the follow<strong>in</strong>g sections.<br />
102
Residual ozone. To reduce sample overbleach<strong>in</strong>g <strong>and</strong> dilution problems dur<strong>in</strong>g residual<br />
ozone measurements, the gravimetric method was used, which <strong>in</strong>corporates a slight modification <strong>of</strong><br />
the <strong>in</strong>digo colorimetric method (St<strong>and</strong>ard Methods, 1995). With this method, the sample was<br />
collected <strong>in</strong> a 125-mL Erlenmeyer flask conta<strong>in</strong><strong>in</strong>g 10 mL <strong>of</strong> <strong>in</strong>digo trisulfonate reagent to decolorize<br />
the solution to a light blue. The exact sample volume was gravimetrically determ<strong>in</strong>ed by compar<strong>in</strong>g<br />
the total weight <strong>of</strong> the <strong>in</strong>digo-sample mixture <strong>and</strong> the flask with the <strong>in</strong>itial weight <strong>of</strong> the <strong>in</strong>digo <strong>and</strong><br />
flask. This procedure reduced the time needed to collect <strong>and</strong> analyze the samples. The sample<br />
absorbance was measured by a spectrophotometer (Lambda-SB model; Perk<strong>in</strong>-Elmer Corp.,<br />
Norwalk, Conn.) with a 1-cm light path length, <strong>and</strong> the measurement was corrected for the<br />
absorbance <strong>of</strong> background organics at 800 nm.<br />
<strong>Perchlorate</strong>. <strong>Perchlorate</strong> samples were analyzed us<strong>in</strong>g an ion chromatograph (model DX300,<br />
Dionex Corp., Sunnyvale, Calif.) modified with a 200-|oL sample loop. An ion-chromatography (1C)<br />
analytical column (model AG11, Dionex Corp.), an anion micromembrane suppressor, <strong>and</strong><br />
a conductivity detector were used. The method detection limit was 1.4 ng/L, <strong>and</strong> the report<strong>in</strong>g limit<br />
was 4.0 ug/L.<br />
Bromide. Bromide analyses were conducted on an ion chromatograph (model 2010, Dionex<br />
Corp., Sunnyvale, Calif.) with a 20- or 50-uL sample loop. An 1C analytical column (model AS4A/<br />
Dionex), an anion micromembrane suppresser, <strong>and</strong> a conductivity detector were used. The method<br />
detection limit was 0.01 mg/L, <strong>and</strong> the report<strong>in</strong>g limit was 0.02 mg/L.<br />
<strong>Bromate</strong>. <strong>Bromate</strong> analyses were performed us<strong>in</strong>g a modification <strong>of</strong> an 1C method described<br />
by Pfaff <strong>and</strong> Brockh<strong>of</strong>f (1990). The modified method is described by Kuo et al. (1990). An 1C<br />
analytical column (AS4A, Dionex), an anion micromembrane suppressor, <strong>and</strong> a conductivity detector<br />
were used. The method detection limit was 1.0 ng/L, <strong>and</strong> the report<strong>in</strong>g limit was 3.0 ug/L.<br />
Aldehydes. Formaldehyde <strong>and</strong> acetaldehyde were analyzed by a derivatization-extraction gaschromatographic<br />
(GC) electron-capture detection method described by Glaze <strong>and</strong> colleagues (1989).<br />
Other aldehydes, such as glyoxal <strong>and</strong> methyl glyoxal, were analyzed by a modification (heated<br />
derivatization) <strong>of</strong> this method, as described by Sclimenti <strong>and</strong> colleagues (1990).<br />
103
Carboxylic Acids. Carboxylic acids were measured us<strong>in</strong>g ion chromatography as described<br />
by C<strong>of</strong>fey et al. 1997. Organic acids were separated on a high-capacity ion exchange Dionex ASH<br />
column, us<strong>in</strong>g a sodium hydroxide solution eluent <strong>in</strong> the gradient mode, followed by conductivity<br />
detection. The <strong>in</strong>dividual method detection limits for each <strong>of</strong> the three carboxylic acid species<br />
analyzed (acetate, formate, <strong>and</strong> oxalate) was 15 ug/L.<br />
Dissolved Oxygen. The modified W<strong>in</strong>kler or iodometric method was used to measure the DO<br />
concentration <strong>in</strong> water. The samples were analyzed accord<strong>in</strong>g to the St<strong>and</strong>ard Method 4500-O F<br />
(Sulfamic acid flocculation modification). A DO probe was not used s<strong>in</strong>ce high backgrounds <strong>of</strong><br />
sodium thiosulfate confounded the measurement.<br />
pH. pH was measured us<strong>in</strong>g a pH meter with an accuracy <strong>of</strong> ±0.002 pH units (model 920A;<br />
Orion Research, Inc., Boston, Massachusetts).<br />
Results <strong>and</strong> Discussion<br />
<strong>Perchlorate</strong> <strong>and</strong> <strong>Bromate</strong> Reduction<br />
The removal <strong>of</strong> perchlorate <strong>and</strong> bromate has been observed s<strong>in</strong>ce the third week<br />
(approximately 660 B Vs) <strong>of</strong> the m<strong>in</strong>i-column experiment. The columns cont<strong>in</strong>ued to show the<br />
reduction <strong>of</strong> perchlorate <strong>and</strong> bromate after 10 months <strong>of</strong> operation under various conditions. Figures<br />
4.2 <strong>and</strong> 4.3 show the removal <strong>of</strong> perchlorate <strong>and</strong> bromate by the m<strong>in</strong>i-columns. These figures are<br />
divided up <strong>in</strong>to regions, <strong>and</strong> different regions are compared <strong>in</strong> the follow<strong>in</strong>g sections. Dur<strong>in</strong>g these<br />
tests, there were few <strong>in</strong>stances when the <strong>in</strong>fluent perchlorate concentration decreased <strong>in</strong> the reservoir<br />
(<strong>in</strong>fluent sample). This occurred when a freshly ozonated water supply was not available due to pilot<br />
plant shut down. In this case, the <strong>in</strong>fluent water was <strong>in</strong> the reservoir for more than a week (up to 10<br />
days), <strong>and</strong> the <strong>in</strong>fluent concentration is plotted separately (denoted as -7th day) from the regular<br />
<strong>in</strong>fluent sample. The <strong>in</strong>fluent perchlorate concentrations for these cases were about half the target<br />
value. Dur<strong>in</strong>g the clean<strong>in</strong>g <strong>of</strong> the reservoir after these periods when the <strong>in</strong>fluent was unchanged for<br />
over a week, a thick, yellowish bi<strong>of</strong>ilm was observed at the bottom <strong>of</strong> the reservoir. This bi<strong>of</strong>ilm may<br />
have been responsible for the premature degradation <strong>of</strong> perchlorate <strong>and</strong> bromate.<br />
As shown <strong>in</strong> Figure 4.2, after three months (-2,800 BVs) <strong>of</strong> m<strong>in</strong>i-column operation, the<br />
<strong>in</strong>fluent concentration <strong>of</strong> perchlorate was reduced from 50 to 10 ng/L (regions 1 <strong>and</strong> 2). Complete<br />
perchlorate reduction cont<strong>in</strong>ued to take place. From 3,500 to 5,000 BV, the effect <strong>of</strong> EBCT was<br />
104
^ <strong>in</strong>fluent 0 effluent A mid • <strong>in</strong>f (~7th day)<br />
Region<br />
1<br />
2<br />
3<br />
4<br />
5<br />
6<br />
7<br />
8<br />
Initial Conditions<br />
C1O4 EBCT DO NO,<br />
studied with an <strong>in</strong>fluent perchlorate concentration <strong>of</strong> 50 ug/L (Figure 4.2, region 3). The <strong>in</strong>fluent<br />
flow rate was <strong>in</strong>creased to 2.7 mL/m<strong>in</strong> to achieve a 15-m<strong>in</strong> EBCT after the first column (mid-po<strong>in</strong>t).<br />
This condition was tested for approximately one month to confirm the f<strong>in</strong>d<strong>in</strong>gs, <strong>and</strong> the condition was<br />
repeated with a lower <strong>in</strong>fluent perchlorate concentration <strong>of</strong> 10 ug/L (region 4 <strong>in</strong> Figure 4.2). Under<br />
each <strong>of</strong> these conditions, complete removal <strong>of</strong> perchlorate was achieved.<br />
<strong>Bromate</strong> is formed dur<strong>in</strong>g ozonation at the La Verne pilot plant, <strong>and</strong> therefore bromate is<br />
present <strong>in</strong> the <strong>in</strong>fluent to the BAG filters. Average bromate formation has been 17 ug/L with a<br />
maximum concentration <strong>of</strong> 55 ug/L. As found for perchlorate, bromate removal by the m<strong>in</strong>i-columns<br />
has also been observed. Complete bromate removal has been observed for 15- <strong>and</strong> 25-m<strong>in</strong>ute EBCTs<br />
(Figure 4.3).<br />
Effect <strong>of</strong> Dissolved Oxygen<br />
The effect <strong>of</strong> DO concentration was also studied <strong>in</strong> the bench-scale setup us<strong>in</strong>g a 15-m<strong>in</strong>ute<br />
EBCT. When a low <strong>in</strong>fluent DO concentration (0.6-2 mg/L) was used (Figure 4.2, region 5 <strong>and</strong><br />
Figure 4.3, region 3), complete perchlorate <strong>and</strong> bromate removal was observed except for one<br />
perchlorate sample. When the <strong>in</strong>fluent DO concentration was <strong>in</strong>creased to 5 mg/L (Figure 4.2, region<br />
6), mid-po<strong>in</strong>t samples showed effluent perchlorate concentrations between 8 <strong>and</strong> 13 ug/L, <strong>in</strong>dicat<strong>in</strong>g<br />
that perchlorate removal decl<strong>in</strong>ed as the <strong>in</strong>fluent DO concentration was <strong>in</strong>creased. By contrast, when<br />
the <strong>in</strong>fluent DO concentration was <strong>in</strong>creased to 5 mg/L (Figure 4.3, region 4), bromate was removed<br />
to the detection limit.<br />
After the tests at 5 mg/L <strong>in</strong>fluent DO were complete, the columns were equilibrated with a<br />
zero <strong>in</strong>fluent DO condition to allow for the exam<strong>in</strong>ation <strong>of</strong> nitrate effects without the occurrence <strong>of</strong><br />
any residual DO effects. Influent perchlorate was also <strong>in</strong>creased from 10 to 50 |jg/L dur<strong>in</strong>g this<br />
equilibration period. Four days after the new condition was imposed, less than 10 percent perchlorate<br />
reduction was observed <strong>in</strong> the 15-m<strong>in</strong>ute EBCT filter (Figure 4.2, region 7). However, by the second<br />
week <strong>of</strong> equilibration, complete removal <strong>of</strong> <strong>in</strong>fluent perchlorate was aga<strong>in</strong> observed <strong>in</strong> the 15-m<strong>in</strong>ute<br />
EBCT filter.<br />
106
Effect <strong>of</strong> Nitrate<br />
The background nitrate concentration <strong>in</strong> the source water at Metropolitan is typically very<br />
low. The <strong>in</strong>fluent nitrate concentration was below the detection limit (0.05 mg/L) <strong>in</strong> the bench-scale<br />
setup except for a few samples dur<strong>in</strong>g the experiments. There were three <strong>in</strong>stances where nitrate was<br />
detected <strong>in</strong> the <strong>in</strong>fluent, <strong>and</strong> the values ranged from 0.2 to 1.0 mg/L. Consequently, the filters were<br />
exposed to only low nitrate concentrations for the first 8,000 B Vs. Nitrate was spiked to the <strong>in</strong>fluent<br />
at concentrations <strong>of</strong> 4.0 <strong>and</strong> 8.3 mg/L for a two-week period with zero <strong>in</strong>fluent DO (Figure 4.2,<br />
region 8 <strong>and</strong> Figure 4.3, region 5). Dur<strong>in</strong>g these experiments, complete removal <strong>of</strong> nitrate,<br />
perchlorate, <strong>and</strong> bromate was observed. With complete nitrate removal <strong>in</strong> the BAG filters,<br />
perchlorate <strong>and</strong> bromate removal cont<strong>in</strong>ued unperturbed by the elevated <strong>in</strong>fluent nitrate<br />
concentration.<br />
DO Control with Sodium Thiosulfate<br />
In order to facilitate biological utilization <strong>of</strong> perchlorate <strong>and</strong> bromate, a zero or low <strong>in</strong>fluent<br />
DO concentration was <strong>of</strong>ten ma<strong>in</strong>ta<strong>in</strong>ed. Purg<strong>in</strong>g <strong>of</strong> the <strong>in</strong>fluent with gaseous nitrogen to remove<br />
DO was tested but not used because <strong>of</strong> the possibility <strong>of</strong> also purg<strong>in</strong>g ozonated by-products needed<br />
as electron donors. Thus, the DO concentration was chemically controlled with sodium thiosulfate,<br />
but there have been some complications <strong>in</strong> its use. The fast reaction <strong>of</strong> thiosulfate with DO yields<br />
tetrathionate (S 4O62~), <strong>and</strong> this stoichiometry was used to calculate the mass <strong>of</strong> thiosulfate needed for<br />
DO reduction. Based on this stoichiometry (equation 4.1), the sodium thiosulfate dosage required<br />
to reduce the DO concentration to 0 mg/L was approximately 160 mg/L.<br />
4S2O32" + O2(aq) + 4H+ - 2S 4O62' + 2H2O (4.1)<br />
However, tetrathionate can slowly be oxidized to sulfate as it reacts with DO. Assum<strong>in</strong>g<br />
oxidation <strong>of</strong> thiosulfate to sulfate, the dose <strong>of</strong> thiosulfate needed for DO reduction is only one eighth<br />
<strong>of</strong> the amount calculated by equation 4.1 (20 mg/L). However, reduction <strong>of</strong> the DO concentration<br />
to 0 mg/L could not be achieved for the latter case with<strong>in</strong> a reasonable time period. After 5 days,<br />
with a 20 mg/L dose <strong>of</strong> sodium thiosulfate, the DO concentration was 5 mg/L <strong>in</strong> a study done us<strong>in</strong>g<br />
107
the W<strong>in</strong>kler method (St<strong>and</strong>ard Method 4500-O F) to measure DO. The fast k<strong>in</strong>etics reaction model<br />
was used <strong>and</strong> tested to determ<strong>in</strong>e the amount <strong>of</strong> sodium thiosulfate spike to prepare a batch <strong>of</strong> water.<br />
However, tetrathionate can be oxidized further to sulfate, as it reacts with dissolved oxygen,<br />
<strong>and</strong> the amount <strong>of</strong> sodium thiosulfate spike based on this stoichiometry is only 1/8 <strong>of</strong> the amount<br />
compared with the reaction shown above. However, the reduction to zero DO level <strong>in</strong> the <strong>in</strong>fluent<br />
could not be achieved for the latter case with<strong>in</strong> the reasonable time frame <strong>of</strong> the experiments. After<br />
5 days, with 20 mg/L sodium thiosulfate dose, the DO concentration was 5 mg/L. Therefore, the<br />
k<strong>in</strong>etically fast reaction (equation 4.1) was used to determ<strong>in</strong>e the sodium thiosulfate spike for DO<br />
control.<br />
One concern with high concentrations <strong>of</strong> sodium thiosulfate <strong>in</strong> the reservoir is the possibility<br />
<strong>of</strong> chemical reduction <strong>of</strong> perchlorate by sodium thiosulfate. Three control studies were done to verify<br />
that thiosulfate did not reduce perchlorate, <strong>and</strong> a summary <strong>of</strong> these results is given <strong>in</strong> Table 4.3. In<br />
the deionized (Super-Q) water, no perchlorate reduction was found over a two-week period.<br />
Similarly, no perchlorate reduction was observed when this experiment was repeated us<strong>in</strong>g an<br />
ozonated blend <strong>of</strong> CRW/SPW. In the third control study, the same spent <strong>GAC</strong> used for the m<strong>in</strong>icolumns<br />
was used to determ<strong>in</strong>e if thiosulfate can reduce perchlorate <strong>in</strong> the presence <strong>of</strong> <strong>GAC</strong>. The<br />
<strong>GAC</strong> used for this study was probably not biologically active s<strong>in</strong>ce it had not been used for dr<strong>in</strong>k<strong>in</strong>g<br />
water filtration for about 6 months prior to this experiment. It was found that perchlorate reduction<br />
did not take place <strong>in</strong> the presence <strong>of</strong> thiosulfate <strong>and</strong> <strong>GAC</strong>.<br />
108
Table 4.2 Control study for chemical reduction pathway for perchlorate (<strong>in</strong>itial values N/A)<br />
Reaction time Water Sodium<br />
(week) thiosulfate used<br />
1<br />
1<br />
Super-Q<br />
Super-Q<br />
no<br />
yes<br />
<strong>GAC</strong> used<br />
no<br />
no<br />
CKV(ng/L)<br />
173<br />
174<br />
2<br />
Super-Q<br />
yes<br />
no<br />
208<br />
1<br />
Ozonated CRW<br />
no<br />
no<br />
180<br />
1<br />
Ozonated CRW<br />
yes<br />
no<br />
182<br />
2<br />
Ozonated CRW<br />
yes<br />
no<br />
221<br />
1<br />
Super-Q<br />
yes<br />
yes<br />
126<br />
1<br />
Super-Q<br />
no<br />
yes<br />
142<br />
1<br />
Ozonated CRW<br />
yes<br />
yes<br />
100<br />
Ozonation By-products<br />
Carboxylic acids <strong>and</strong> aldehydes were analyzed from the m<strong>in</strong>i-column experiments s<strong>in</strong>ce the<br />
third month <strong>of</strong> filter operation. There has been no clear decrease <strong>in</strong> carboxylic acids across the filters.<br />
Some <strong>of</strong> the aldehydes decl<strong>in</strong>e <strong>in</strong> concentration across the filters, which may be <strong>in</strong>dicative <strong>of</strong> aldehyde<br />
use as an microbial electron donor. However, <strong>in</strong>terpretation <strong>of</strong> the data was potentially confounded<br />
by 1) difficulty <strong>in</strong> ma<strong>in</strong>ta<strong>in</strong><strong>in</strong>g a stable concentration <strong>of</strong> ozone by-products <strong>in</strong> the water, 2)<br />
<strong>in</strong>terference <strong>of</strong> sodium thiosulfate with aldehydes dur<strong>in</strong>g GC analysis, <strong>and</strong> 3) changes <strong>in</strong> sodium<br />
thiosulfate concentration as the reaction occurs between the reservoir <strong>and</strong> the columns. A more<br />
detailed <strong>and</strong> controlled study is required to isolate <strong>and</strong> identify the by-products used <strong>in</strong> the BAC<br />
filters.<br />
TOC <strong>and</strong> 254<br />
Dur<strong>in</strong>g the first three months <strong>of</strong> the experiment, the effluent total organic carbon (TOC)<br />
concentration from the BAC filters was approximately 30 percent higher than the <strong>in</strong>fluent<br />
concentration. Typical <strong>in</strong>fluent <strong>and</strong> effluent TOC concentrations were 1.8 <strong>and</strong> 2.4 mg/L, respectively.<br />
109
This <strong>in</strong>crease <strong>in</strong> TOC concentration across the filters may have been due to bi<strong>of</strong>ilm slough<strong>in</strong>g from<br />
the spent <strong>GAC</strong>. S<strong>in</strong>ce the third month <strong>of</strong> operation, the average <strong>in</strong>fluent <strong>and</strong> effluent TOC<br />
concentrations have been approximately 2.11 <strong>and</strong> 2.21 mg/L respectively, which is only a 5 percent<br />
<strong>in</strong>crease.<br />
The UV absorption at 254 nm (filters were not filtered) <strong>of</strong> the filter effluent, 0.11 cm" 1 , was<br />
approximately 40 percent less than the average <strong>in</strong>fluent value (0.18 cm" 1) for the first three months<br />
<strong>of</strong> the experiment. After this time, the average UV254 values have been 0.13 cm" 1 at the filter <strong>in</strong>fluent<br />
<strong>and</strong> 0.11 cm' 1 at the effluent, which is a decrease <strong>of</strong> about 15 percent across the filter.<br />
The historical sulfate concentration found <strong>in</strong> Metropolitan's source water is approximately<br />
250 mg/L, which is also the average value measured <strong>in</strong> the <strong>in</strong>fluent <strong>of</strong> the m<strong>in</strong>i-column four days after<br />
dos<strong>in</strong>g the reservoir with sodium thiosulfate. This <strong>in</strong>dicates that sodium thiosulfate is not oxidized<br />
to sulfate <strong>in</strong> the reservoir with<strong>in</strong> four days. However, the sulfate concentration is higher at the mid<br />
po<strong>in</strong>t (13 percent higher than the <strong>in</strong>fluent concentration) <strong>and</strong> <strong>in</strong> the effluent from the second filter (20<br />
percent higher than the <strong>in</strong>fluent concentration). The average effluent sulfate concentration from the<br />
second filter is about 300 mg/L. Tetrathionate formed <strong>in</strong> the reservoir could be oxidized to sulfate<br />
<strong>in</strong> the presence <strong>of</strong> <strong>GAC</strong>, thus elevat<strong>in</strong>g the effluent sulfate concentrations.<br />
Sulfur-based reduc<strong>in</strong>g chemicals can be effective <strong>in</strong> reduc<strong>in</strong>g the <strong>in</strong>fluent DO concentration,<br />
but taste-<strong>and</strong>-odor problems associated with sulfide may be a concern. For <strong>in</strong>stance, hydrogen sulfide<br />
has been observed effluent samples from the m<strong>in</strong>i-columns. Post-treatment could be employed to<br />
remove hydrogen sulfide, but it may be possible to use other reduc<strong>in</strong>g agents that will not cause taste<strong>and</strong>-odor<br />
problems. The use <strong>of</strong> ascorbic acid would facilitate faster reduction <strong>of</strong> DO; this could<br />
potentially decrease the amount <strong>of</strong> chemical required for DO reduction <strong>and</strong> reduce the overall<br />
chemical costs. However, ascorbic acid was not used <strong>in</strong> this study s<strong>in</strong>ce ascorbic acid could break<br />
down <strong>in</strong> the reservoir or filters <strong>and</strong> provide an additional electron donor source.<br />
110
PILOT-SCALE MOBILE PILOT PLANT STUDY<br />
Materials <strong>and</strong> Methods<br />
Water<br />
The typical blend <strong>of</strong> water to the mobile pilot plant is 25 percent SPW <strong>and</strong> 75 percent CRW.<br />
This is the same blend <strong>of</strong> water used for the m<strong>in</strong>i-column study. The mobile pilot plant has the<br />
capability to supply 100 percent CRW <strong>in</strong> case the SPW supply to the mobile pilot plant is shut down<br />
dur<strong>in</strong>g regular ma<strong>in</strong>tenance or for special test<strong>in</strong>g.<br />
Mobile Pilot Plant Setup<br />
The mobile pilot plant, located at the MWDSC's La Verne facility (Figure 4.4), consists <strong>of</strong><br />
two parallel tra<strong>in</strong>s, which <strong>in</strong>clude ozone contactors <strong>and</strong> conventional treatment. The raw CRW or<br />
CRW/SPW blend is treated by 1) pre-ozonation, 2) rapid mix (high-<strong>in</strong>tensity mix<strong>in</strong>g with addition <strong>of</strong><br />
coagulant <strong>and</strong> cationic polymer), 3) flocculation (slow mix<strong>in</strong>g <strong>of</strong> water <strong>in</strong> comb<strong>in</strong>ation with coagulant<br />
<strong>and</strong> polymer <strong>in</strong> order to form floes), 4) sedimentation (settl<strong>in</strong>g <strong>of</strong> floe), <strong>and</strong> 5) biological filtration (no<br />
Figure 4.4 Mobile pilot plant<br />
111
DO residual across the filters, allow<strong>in</strong>g for biodegradation <strong>of</strong> perchlorate <strong>and</strong> bromate).<br />
A schematic <strong>of</strong> the mobile pilot plant is shown <strong>in</strong> Figure 4.5, <strong>and</strong> the arrow <strong>in</strong>dicates<br />
direction <strong>of</strong> water flow. The flow to the ozone contactor is equal to the flow rate to the head <strong>of</strong> the<br />
tra<strong>in</strong>, which is set at 1.4 gallons per m<strong>in</strong>ute (gpm). When small shells or debris clog the pre-filter<br />
outside <strong>of</strong> the mobile pilot plant, the flow rate decreases; thus the filter is checked <strong>and</strong> cleaned<br />
regularly.<br />
Mobile Pilot Plant Filter<br />
A 10-ft column with an <strong>in</strong>ner-diameter <strong>of</strong> 4 <strong>in</strong>ches was filled with 4 feet <strong>of</strong> virg<strong>in</strong> Norit<br />
carbon (Hydrodarco GCW 10 X 20). The flow rate to the column is set at 0.1 gpm to atta<strong>in</strong> an<br />
EBCT <strong>of</strong> 25 m<strong>in</strong>utes. The characteristics <strong>of</strong> the virg<strong>in</strong> carbon are given <strong>in</strong> Table 4.3.<br />
Influent<br />
Alum,<br />
Polymer<br />
<strong>Perchlorate</strong><br />
<strong>Bromate</strong><br />
Spike<br />
Biological<br />
Filter<br />
Sodium<br />
Tbiosulfate<br />
Spike<br />
CD CD<br />
Air O 3<br />
Compressor Generator O 3 Contactor<br />
Rapid Flocculators Sedimentation<br />
Mix Bas<strong>in</strong><br />
Effluent<br />
Figure 4.5 Schematic <strong>of</strong> the mobile pilot plant setup at Metropolitan (not to scale)<br />
Table 4.3 Characteristics <strong>of</strong> Norit carbon used <strong>in</strong> mobile pilot plant<br />
______(Norit Americas, Inc., GA)______<br />
Characteristics<br />
Value<br />
BET surface area 1025 m2/g<br />
Average particle diameter<br />
1 mm<br />
Uniformity coefficient<br />
1.2<br />
Total pore volume 0.8 mL/g<br />
Apparent density 0.48 g/mL<br />
112
Mobile Pilot Plant <strong>Ozone</strong><br />
The ozonation process consisted <strong>of</strong> an ozone contactor column <strong>and</strong> an ozone reactor column.<br />
The raw water was pumped to the top <strong>of</strong> the ozone contactor at 1.4 gpm, <strong>and</strong> ozone was applied <strong>in</strong><br />
countercurrent fashion. <strong>Ozone</strong> was bubbled through 0.5-<strong>in</strong> diameter (1-<strong>in</strong> long) sta<strong>in</strong>less steel static<br />
diffiisers to generate f<strong>in</strong>e bubbles. The ozone generator used was a tube-type, water-cooled unit<br />
operat<strong>in</strong>g on a low-frequency (60 Hz), variable-voltage power supply (model Labo-76,<br />
Praxair/Trailigaz <strong>Ozone</strong> Co., C<strong>in</strong>c<strong>in</strong>nati, Ohio). The ozone contactor column was operated with a<br />
1.2-m<strong>in</strong>ute residence time, <strong>and</strong> the ozone reactor had a 4.7- m<strong>in</strong>ute residence time. The water level<br />
was ma<strong>in</strong>ta<strong>in</strong>ed at 10 feet <strong>in</strong> the 2-<strong>in</strong>ch diameter contact column <strong>and</strong> <strong>in</strong> the 4-<strong>in</strong>ch diameter reactor<br />
column.<br />
The ozone generator capacity is 1 Ib/day, with compressed air as the feed gas. The target gasphase<br />
concentration <strong>of</strong> ozone produced is approximately 0.95 percent by weight. The ozone<br />
concentration <strong>in</strong> the feed gas was measured by a gaseous ozone monitor (model HC-12, PCI <strong>Ozone</strong><br />
Corp., West Caldwell, NJ). Off-gas from the contactor flows through a thermocatalytic ozone<br />
destruction unit prior to discharge to the atmosphere.<br />
Backwash Setup<br />
The backwash water supply setup has been modified to dechlor<strong>in</strong>ate <strong>and</strong> deoxygenate the<br />
water, so that the BAG filters are protected dur<strong>in</strong>g the backwash cycle. The backwash water supplied<br />
to the mobile pilot plant is treated Weymouth Plant utility water with approximately 2 mg/L <strong>of</strong> free<br />
chlor<strong>in</strong>e residual <strong>and</strong> an ambient DO concentration. The modification <strong>in</strong>cluded a 55-gallon sta<strong>in</strong>less<br />
steel tank packed with virg<strong>in</strong> Norit carbon (HYDRODARCO GCW 10 X 20) for dechlor<strong>in</strong>ation, an<br />
HOPE tank to spike sodium thiosulfate to deoxygenate the backwash water, <strong>and</strong> a pump to fluidize<br />
the beds (Figure 4.6). The automatic backwash cycle was disabled to allow a plant operator to<br />
manually backwash the filters at least once every 48 hours.<br />
113
Sodium<br />
Thiosulfate<br />
to reduce<br />
Treated<br />
Backwas<br />
Water<br />
Weymouth<br />
Plant<br />
Utility<br />
Water<br />
GAG tank to<br />
dechlor<strong>in</strong>ate<br />
Figure 4.6 Backwash water supply setup<br />
Operat<strong>in</strong>g Conditions<br />
The typical operat<strong>in</strong>g conditions for the mobile pilot plant are listed <strong>in</strong> Table 4.5. A<br />
perchlorate spike <strong>of</strong> 500 u,g/L was used for the first few weeks to pre-load the virg<strong>in</strong> carbons. The<br />
spike was subsequently reduced to 50 ug/L. Similarly, 250 (o,g/L <strong>of</strong> bromate was used for preload<strong>in</strong>g,<br />
<strong>and</strong> 25 |u,g/L was used dur<strong>in</strong>g the experimental phase. Chemicals were cont<strong>in</strong>uously <strong>in</strong>jected<br />
us<strong>in</strong>g <strong>in</strong>dividual peristaltic pumps from the 5-gallon carboys. The secondary feed stocks for alum,<br />
polymer (poly-DADMAC, NeoSolutions, Beaver, Pa.), perchlorate, <strong>and</strong> bromate were prepared<br />
weekly. Sodium thiosulfate <strong>in</strong>jection started after the exchange capacity <strong>of</strong> the carbon for perchlorate<br />
<strong>and</strong> bromate was exhausted (~ 6000 BVs). When used, sodium thiosulfate was spiked after the<br />
sediment bas<strong>in</strong>, <strong>and</strong> three <strong>in</strong>-l<strong>in</strong>e static mixers (six mix<strong>in</strong>g elements per mixer; Cole-Parmer, Vernon<br />
Hills, 111.) were <strong>in</strong>stalled to provide adequate mix<strong>in</strong>g so that zero DO was achieved before <strong>in</strong>fluent<br />
water entered the filters.<br />
114
Table 4.4 Mobile pilot plant operat<strong>in</strong>g conditions<br />
Operat<strong>in</strong>g Conditions Current Sett<strong>in</strong>g<br />
Influent flow rate ~1.4 gpm<br />
Applied ozone 2 mg/L<br />
Rapid mix <strong>in</strong>fluent flow rate 1 gpm<br />
<strong>Perchlorate</strong> spike 50 jag/L<br />
<strong>Bromate</strong> spike 0 to 25 (ig/L<br />
Alum dosage 4.5 mg/L<br />
Polymer dosage 1.5 mg/L<br />
Filter <strong>in</strong>fluent flow rate 0.1 gpm<br />
Sodium thiosulfate spike -160 mg/L<br />
Influent DO concentration 0 mg/L<br />
Sampl<strong>in</strong>g<br />
For the pilot-scale study, perchlorate <strong>and</strong> bromate concentrations were analyzed at the <strong>in</strong>fluent<br />
<strong>and</strong> effluent <strong>of</strong> the BAG filter. For selected tests, perchlorate <strong>and</strong> bromate were also analyzed after<br />
the rapid mix step. Once complete perchlorate <strong>and</strong> bromate reduction was observed, samples were<br />
also taken for UV, TOC, aldehydes, <strong>and</strong> carboxylic acids.<br />
Results<br />
<strong>Perchlorate</strong> <strong>and</strong> <strong>Bromate</strong> Preload<strong>in</strong>g<br />
S<strong>in</strong>ce the pilot plant filters were comprised <strong>of</strong> virg<strong>in</strong> carbon, the filters were subjected to<br />
perchlorate <strong>and</strong> bromate preload<strong>in</strong>g periods to exhaust the ion exchange capacity <strong>of</strong> the carbon. This<br />
would ensure that subsequent removal was via biological reduction <strong>and</strong> not ion exchange. Table 4.5<br />
shows the results <strong>of</strong> preload<strong>in</strong>g the <strong>GAC</strong> <strong>in</strong> the pilot plant experiments. The virg<strong>in</strong> carbon effectively<br />
removed perchlorate dur<strong>in</strong>g the <strong>in</strong>itial phase <strong>of</strong> operation <strong>and</strong> removed bromate to a lesser degree.<br />
After a few weeks <strong>of</strong> preload<strong>in</strong>g the carbon with 500 ng/L perchlorate, the effluent perchlorate<br />
concentration from the <strong>GAC</strong> filter steadily <strong>in</strong>creased.<br />
115
Table 4.5 Preload<strong>in</strong>g results from the mobile pilot plant<br />
Day<br />
1<br />
2<br />
3<br />
4<br />
8<br />
9<br />
Sampl<strong>in</strong>g Po<strong>in</strong>t<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
rapid mix effluent<br />
filter <strong>in</strong>fluent<br />
filter effluent<br />
C1CV (ug/L)<br />
632<br />
667<br />
ND<br />
532<br />
577<br />
ND<br />
592<br />
572<br />
ND<br />
640<br />
641<br />
ND<br />
580<br />
580<br />
250<br />
590<br />
570<br />
230<br />
Br03-(ug/L)<br />
257<br />
300<br />
169<br />
248<br />
270<br />
183<br />
288<br />
289<br />
171<br />
238<br />
249<br />
203<br />
262<br />
259<br />
160<br />
265<br />
241<br />
185<br />
S<strong>in</strong>ce samples were diluted for ion chromatographic analysis, the ND <strong>in</strong>dicated <strong>in</strong> Table 4.5 could be<br />
as high as 40 u,g/L. After approximately one week <strong>of</strong> <strong>GAC</strong> preload<strong>in</strong>g, perchlorate removal <strong>in</strong> the<br />
filter decreased to about 60 percent. <strong>Bromate</strong> was also spiked to the water at 250 |ig/L. <strong>Bromate</strong><br />
removal decl<strong>in</strong>ed steadily for the first few days, as <strong>in</strong>dicated <strong>in</strong> Table 4.5. <strong>Perchlorate</strong> <strong>and</strong> bromate<br />
concentrations were nearly constant from the rapid mix effluent to the filter <strong>in</strong>fluent, <strong>in</strong>dicat<strong>in</strong>g that<br />
virtually no perchlorate or bromate reduction was tak<strong>in</strong>g place dur<strong>in</strong>g coagulation or sedimentation.<br />
Figures 4.7 <strong>and</strong> 4.8 show the exhaustion <strong>of</strong> the <strong>GAC</strong>'s ion exchange capacity for perchlorate<br />
<strong>and</strong> bromate, respectively. <strong>Perchlorate</strong> <strong>and</strong> bromate concentrations <strong>of</strong> 500 u,g/L <strong>and</strong> 250 u,g/L were<br />
used <strong>in</strong>itially to m<strong>in</strong>imize the time required to preload the <strong>GAC</strong>. When the perchlorate spike was<br />
dereased to approximately 60 |ig/L (at 800 BV), the aqueous perchlorate concentration was lower<br />
than the sorbed perchlorate concentration. This negative concentration gradient caused perchlorate<br />
to desorb from the carbon, yield<strong>in</strong>g C/C0 values greater than one. Though the <strong>in</strong>fluent bromate spike<br />
was also decreased from its <strong>in</strong>itial concentration, similar desorption behavior was not observed. Both<br />
116
p o<br />
O. i<br />
Fraction <strong>Bromate</strong> Rema<strong>in</strong><strong>in</strong>g<br />
C/Co<br />
n><br />
Fraction <strong>Perchlorate</strong> Rema<strong>in</strong><strong>in</strong>g<br />
C/Co<br />
if<br />
-•— ' O<br />
CE. o'<br />
§1<br />
Ks* ^^<br />
CZ • \J<br />
ET o'<br />
3 5T<br />
I o'<br />
2,<br />
cr<br />
3<br />
aO)<br />
O<br />
O "-*><br />
•o<br />
n><br />
3<br />
o<br />
09<br />
a><br />
"2.<br />
o"<br />
r*<br />
CO<br />
O<br />
a<br />
iCO<br />
I<br />
CD
perchlorate <strong>and</strong> bromate reached equilibrium (C/C0 * 1) before the filters were switched to biological<br />
mode (ozonation <strong>and</strong> deaeration). No desorption <strong>of</strong> perchlorate or bromate was observed when the<br />
filters were operat<strong>in</strong>g <strong>in</strong> the biological mode.<br />
<strong>Perchlorate</strong> <strong>and</strong> <strong>Bromate</strong> Reduction <strong>in</strong> BAC Filters<br />
Approximately 3,000 BVs after the filter was switched to biological mode, greater than 87<br />
percent perchlorate removal was observed (Figure 4.7). Figure 4.8 shows that biological bromate<br />
reduction was also occurr<strong>in</strong>g <strong>in</strong> the filter. Dur<strong>in</strong>g the third week <strong>of</strong> operation, however, bromate was<br />
not be<strong>in</strong>g formed dur<strong>in</strong>g ozonation, necessitat<strong>in</strong>g the addition <strong>of</strong> a 25 u.g/L bromate spike. <strong>Bromate</strong><br />
removal cont<strong>in</strong>ued to be very efficient.<br />
At about 10,000 BV, the DO control system failed, subject<strong>in</strong>g the filter to ambient DO<br />
concentrations (8-9 mg/L) for approximately 48 hours. Repairs were made <strong>and</strong> the thiosulfate spike<br />
reduced the <strong>in</strong>fluent DO concentration back down to zero. Dur<strong>in</strong>g the two weeks follow<strong>in</strong>g the<br />
repair, perchlorate removal was less than 15 percent <strong>and</strong> bromate removal decreased to 75 percent.<br />
These data are shown as open circles <strong>in</strong> Figures 4.7 <strong>and</strong> 4.8, respectively. Subsequently, complete<br />
perchlorate <strong>and</strong> bromate removal was achieved <strong>in</strong> the filter. This lagged recovery period was<br />
previously observed for perchlorate when the m<strong>in</strong>i-columns were subjected to 5 mg/L DO dur<strong>in</strong>g a<br />
controlled study.<br />
Both bench- <strong>and</strong> pilot-scale experimentation demonstrated efficient biological perchlorate <strong>and</strong><br />
bromate reduction <strong>in</strong> a BAC filter. These results suggest that ozonation followed by deaeration <strong>and</strong><br />
biological filtration could comprise an effective strategy for remov<strong>in</strong>g low u,g/L concentrations <strong>of</strong><br />
perchlorate <strong>and</strong> bromate from dr<strong>in</strong>k<strong>in</strong>g water.<br />
118
CHAPTER 5. SUMMARY AND CONCLUSIONS<br />
BROMATE<br />
A ma<strong>in</strong> purpose <strong>of</strong> this project was to <strong>in</strong>vestigate bromate reduction <strong>in</strong> BAG filters. BAG<br />
filters were constructed at the University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Champaign to evaluate various<br />
parameters affect<strong>in</strong>g bromate reduction.<br />
To rule out cont<strong>in</strong>ued abiotic reduction <strong>of</strong> bromate <strong>in</strong> the BAG filters, BAG filtration<br />
experiments were repeated one year apart. The same percent bromate removal was observed <strong>in</strong><br />
March 1998 as <strong>in</strong> February 1999 for a set <strong>of</strong> BAG filters. These results <strong>in</strong>dicated that abiotic bromate<br />
reduction was not significant <strong>in</strong> these experiments, s<strong>in</strong>ce abiotic reduction could not likely be<br />
susta<strong>in</strong>ed for 11 months between the repeat experiments. A similar experiment was done with<br />
another set <strong>of</strong> BAG filters, but the percent bromate removal obta<strong>in</strong>ed <strong>in</strong> February 1999 was<br />
significantly less than the percent bromate removal obta<strong>in</strong>ed <strong>in</strong> March 1998. A microbial disturbance,<br />
probably caused by a two-week period <strong>in</strong> December 1998 with an EBCT greater than 200 m<strong>in</strong>utes,<br />
is likely responsible for the decl<strong>in</strong>e <strong>in</strong> bromate removal <strong>in</strong> these filters. This <strong>in</strong>dicates that filter history<br />
has a very important effect on biological bromate removal. The microbial community that develops<br />
under a particular water quality condition might affect the bromate removal obta<strong>in</strong>ed even after the<br />
water quality conditions become more favorable for bromate removal. A study <strong>of</strong> the change <strong>in</strong> the<br />
microbial population result<strong>in</strong>g from such events will be very useful <strong>in</strong> underst<strong>and</strong><strong>in</strong>g <strong>and</strong> optimiz<strong>in</strong>g<br />
BAG filter performance.<br />
The bromate removal results obta<strong>in</strong>ed <strong>in</strong> a one-<strong>in</strong>ch <strong>in</strong>ner diameter BAG filter were able to<br />
be reproduced <strong>in</strong> a 2-<strong>in</strong>ch <strong>in</strong>ner diameter BAG filter, suggest<strong>in</strong>g that wall-effects are m<strong>in</strong>imal. This<br />
<strong>in</strong>dicates that scale-up <strong>of</strong> the bench-scale results should not be a concern.<br />
Backwash<strong>in</strong>g the BAG filters us<strong>in</strong>g filter effluent with a low DO concentration (2 mg/L) did<br />
not adversely affect bromate reduction. <strong>Bromate</strong> removal before <strong>and</strong> after backwash<strong>in</strong>g rema<strong>in</strong>ed<br />
essentially the same, despite the fact that significant microbial material was removed from the filters<br />
dur<strong>in</strong>g backwash<strong>in</strong>g.<br />
The mass concentration <strong>of</strong> bromate removed <strong>in</strong> the BAG filters was <strong>of</strong>ten dependent on the<br />
<strong>in</strong>fluent bromate concentration. For lower bromate concentrations (10-20 ng/L) <strong>and</strong> longer EBCTs<br />
(25-50 m<strong>in</strong>utes), bromate removal <strong>in</strong>creased as the <strong>in</strong>fluent bromate concentration <strong>in</strong>creased. This<br />
119
<strong>in</strong>dicated that bromate removal was not limited by the electron donor concentration under these<br />
conditions. However, constant mass concentration bromate removal was observed at higher <strong>in</strong>fluent<br />
bromate concentrations (20-50 ng/L) <strong>and</strong> shorter EBCTs (10-20 m<strong>in</strong>utes) which may <strong>in</strong>dicate that<br />
the electron donor concentration was limit<strong>in</strong>g bromate removal under these conditions. Based on this<br />
<strong>in</strong>formation, percent bromate removal would not necessarily rema<strong>in</strong> constant as the <strong>in</strong>fluent bromate<br />
concentration changed. The percent bromate removal obta<strong>in</strong>ed for a given water will depend on a<br />
comb<strong>in</strong>ation <strong>of</strong> factors <strong>in</strong>clud<strong>in</strong>g the <strong>in</strong>fluent bromate concentration, EBCT, <strong>and</strong> concentration <strong>of</strong><br />
electron donor.<br />
Increas<strong>in</strong>g the concentration <strong>of</strong> <strong>in</strong>fluent DO to the BAG filter caused a decrease <strong>in</strong> bromate<br />
removal. S<strong>in</strong>ce it is possible that denitrify<strong>in</strong>g organisms are us<strong>in</strong>g nitrate reductase to reduce<br />
bromate, the presence <strong>of</strong> DO may be <strong>in</strong>hibit<strong>in</strong>g the synthesis or activity <strong>of</strong> this enzyme. An <strong>in</strong>creased<br />
DO concentration may also cause the use <strong>of</strong> more electron donor for aerobic respiration, leav<strong>in</strong>g less<br />
electron donor available for bromate reduction. The lack <strong>of</strong> bromate reduction observed <strong>in</strong> full-scale<br />
ozone-BAC plants is likely due to the high concentrations <strong>of</strong> DO enter<strong>in</strong>g the filters. Therefore, it<br />
would be advantageous for treatment plants to reduce the DO concentration enter<strong>in</strong>g the BAG filters<br />
as much as possible.<br />
The negative effect <strong>of</strong> higher concentrations <strong>of</strong> DO on the bromate-reduc<strong>in</strong>g bacteria was not<br />
immediately reversible when the DO concentration was reduced. A BAG filter operated with an<br />
<strong>in</strong>fluent DO concentration <strong>of</strong> 2.0 mg/L, which had previously been exposed to an <strong>in</strong>fluent DO<br />
concentration <strong>of</strong> 13.6 mg/L for two weeks, yielded significantly lower bromate removal efficiency<br />
than expected. This result suggests that the DO concentration <strong>of</strong> the water used to backwash the<br />
BAG filters should be carefully considered s<strong>in</strong>ce it may affect subsequent bromate removal.<br />
Increas<strong>in</strong>g the <strong>in</strong>fluent nitrate concentration to the BAG filter caused a decrease <strong>in</strong> bromate<br />
removal. If denitrifiers are reduc<strong>in</strong>g bromate <strong>in</strong> the BAG filter, nitrate <strong>and</strong> bromate may be compet<strong>in</strong>g<br />
for use as term<strong>in</strong>al electron acceptors by nitrate reductase. S<strong>in</strong>ce it has been shown that a number<br />
<strong>of</strong> non-denitrify<strong>in</strong>g microbial isolates from the BAG can reduce bromate <strong>and</strong> nitrate, it also is possible<br />
that nitrate <strong>and</strong> bromate are simply compet<strong>in</strong>g for the available electron donor.<br />
Increas<strong>in</strong>g the concentration <strong>of</strong> sulfate from 11.1 to 102.7 mg/L to the BAG filters only<br />
caused a slight decl<strong>in</strong>e <strong>in</strong> bromate removal for the BAG filters. No significant sulfate reduction was<br />
detected <strong>in</strong> the filters.<br />
120
<strong>Bromate</strong> removal <strong>in</strong>creased as the <strong>in</strong>fluent pH to the BAG filters decreased from 8.2 to the<br />
near-neutral range <strong>of</strong> 6.8 to 7.2. It may be possible to reduce bromate formation dur<strong>in</strong>g ozonation<br />
<strong>and</strong> <strong>in</strong>crease biological bromate removal through pH control s<strong>in</strong>ce bromate formation decreases as<br />
the pH decreases.<br />
Under similar operat<strong>in</strong>g conditions, better bromate removal was observed <strong>in</strong> the BAG filters<br />
us<strong>in</strong>g CUW (a groundwater source) than us<strong>in</strong>g LMW. The reduction <strong>of</strong> bromate <strong>in</strong> LMW may have<br />
been h<strong>in</strong>dered directly <strong>and</strong>/or <strong>in</strong>directly by the DO. Only a small amount <strong>of</strong> DO was consumed <strong>in</strong> the<br />
biological nitrification reaction because <strong>of</strong> the low ammonia concentration. This DO could then<br />
compete with bromate for biodegradable organic matter to use <strong>in</strong> aerobic respiration. Alternatively,<br />
if nitrate reductase catalyzes bromate reduction, the DO could also have <strong>in</strong>hibited the synthesis or<br />
activity <strong>of</strong> the enzyme. It is possible that bromate reduction <strong>in</strong> LMW could have been improved if<br />
the water had been ozonated, s<strong>in</strong>ce bromate removal did improve when enough exogenous electron<br />
donor was added to remove the DO <strong>and</strong> nitrate present <strong>in</strong> LMW. Thus, the biodegradability <strong>of</strong> the<br />
source water NOM is a key aspect for efficient bromate reduction.<br />
<strong>Perchlorate</strong> removal was observed <strong>in</strong> a BAG filter treat<strong>in</strong>g LMW spiked with electron donor.<br />
<strong>Perchlorate</strong> was removed when the nitrate concentration <strong>in</strong> the filter was very low. No perchlorate<br />
removal was observed <strong>in</strong> a BAG filter treat<strong>in</strong>g CUW because no external electron donor was added<br />
to reduce the nitrate concentration.<br />
Microscopic exam<strong>in</strong>ation <strong>of</strong> the biomass present <strong>in</strong> the BAG filters showed that the<br />
morphology ma<strong>in</strong>ly consists <strong>of</strong> rods <strong>and</strong> cocci. Some <strong>of</strong> the rods formed cha<strong>in</strong>s <strong>and</strong> a number <strong>of</strong><br />
aggregated rods were observed. Round-bottom protozoa were also seen.<br />
DGGE showed that the microbial population present <strong>in</strong> the BAG was quite diverse. DGGE<br />
can be run on the 16S rDNA <strong>of</strong> the bromate-reduc<strong>in</strong>g isolates, <strong>and</strong> these pr<strong>of</strong>iles can be compared<br />
to the DGGE pr<strong>of</strong>ile <strong>of</strong> the mixed biomass sample from the BAG filter. In this way, the presence <strong>of</strong><br />
<strong>in</strong>dividual bromate-reduc<strong>in</strong>g isolates <strong>in</strong> the BAG could be confirmed. Future work may use DGGE<br />
pr<strong>of</strong>il<strong>in</strong>g as a technique for observ<strong>in</strong>g significant perturbations <strong>in</strong> the population that may relate to<br />
the bromate reduction performance <strong>of</strong> the BAG filter. DGGE could be run on BAG filter biomass<br />
samples after chang<strong>in</strong>g water quality parameters or operational conditions <strong>in</strong> the BAG filters <strong>in</strong> order<br />
to observe changes <strong>in</strong> the bromate-reduc<strong>in</strong>g community.<br />
Eight bromate-reduc<strong>in</strong>g isolates were cultured. All cultures are facultative, s<strong>in</strong>ce they were<br />
able to grow on agar plates that were <strong>in</strong>cubated anaerobically <strong>and</strong> those that were <strong>in</strong>cubated<br />
121
aerobically. All but one <strong>of</strong> the isolates can reduce nitrate. However, due to the lack <strong>of</strong> gas<br />
production by the isolates, it is unlikely that any <strong>of</strong> these bromate-reduc<strong>in</strong>g organisms are denitrifiers.<br />
Identification <strong>and</strong> further characterization <strong>of</strong> these bromate-reduc<strong>in</strong>g microorganisms may aid <strong>in</strong><br />
optimiz<strong>in</strong>g conditions for bromate reduction <strong>in</strong> the BAG filters.<br />
Dilution-to-ext<strong>in</strong>ction experiments were also performed so that additional bromate-reduc<strong>in</strong>g<br />
organisms may be isolated. S<strong>in</strong>ce eight bromate-reduc<strong>in</strong>g organisms have already been isolated by<br />
traditional plat<strong>in</strong>g techniques, the organisms obta<strong>in</strong>ed <strong>in</strong> the dilution-to-ext<strong>in</strong>ction experiments will<br />
be reta<strong>in</strong>ed for future analyses.<br />
<strong>Bromate</strong> reduction was sensitive to temperature. In batch experiments, <strong>in</strong>oculated with<br />
biomass from the BAG filters, bromate was reduced at 25 °C but not at 4 °C. Therefore, as the water<br />
temperature changes at the water treatment plant dur<strong>in</strong>g the year, bromate removal will not likely<br />
rema<strong>in</strong> constant.<br />
The most important modification to the ozone-BAC process for the application <strong>of</strong> bromate<br />
removal is the reduction <strong>of</strong> the <strong>in</strong>fluent DO concentration to the filter. Little or no bromate reduction<br />
has been observed at DO concentrations that are typical <strong>of</strong> the ozone-BAC process. Future work<br />
should address the best way to reduce the <strong>in</strong>fluent DO concentration. Additionally, studies should<br />
be conducted to ascerta<strong>in</strong> the effects <strong>of</strong> a reduced DO concentration on effluent water quality. For<br />
example, a reduced DO concentration to the BAG filter may affect the biodegradation <strong>of</strong> trace<br />
compounds such as 2-methylisoborneol or geosm<strong>in</strong>.<br />
In conjunction with a reduced DO concentration, other modifications (such as pH adjustment)<br />
may be considered as a means to improve biological bromate removal. However, <strong>in</strong> order to truly<br />
underst<strong>and</strong> how best to optimize the reduction <strong>of</strong> bromate <strong>in</strong> a BAG filter, <strong>in</strong>dividual bromatereduc<strong>in</strong>g<br />
cultures must cont<strong>in</strong>ue to be studied. By underst<strong>and</strong><strong>in</strong>g the physiology <strong>and</strong> nutrient<br />
requirements <strong>of</strong> these organisms, the BAG process may be designed to select for these organisms.<br />
PERCHLORATE<br />
Abiotic <strong>and</strong> biological removal <strong>of</strong> perchlorate at low ng/L concentrations was <strong>in</strong>vestigated.<br />
GAG filter experiments demonstrated that perchlorate is not chemically reduced on the surface <strong>of</strong><br />
GAG. Rather, ion exchange is the mechanism by which perchlorate is removed by GAG. An ion<br />
exchange capacity <strong>of</strong> virg<strong>in</strong> Norit GAG for perchlorate was calculated. For the <strong>in</strong>fluent matrix used,<br />
122
virg<strong>in</strong> Norit <strong>GAC</strong> showed an ion exchange capacity <strong>of</strong> 0.172 mg perchlorate per gram <strong>GAC</strong>. Batch<br />
experiments us<strong>in</strong>g four different carbons, CUW <strong>and</strong> DDW, <strong>and</strong> two different pHs confirmed the filter<br />
results that showed ion exchange not chemical reduction as the mechanism <strong>of</strong> perchlorate removal<br />
by <strong>GAC</strong>.<br />
Efficient abiotic removal <strong>of</strong> perchlorate was demonstrated by metal-catalyzed <strong>GAC</strong> filters.<br />
Norit carbon mixed with alum<strong>in</strong>um shot <strong>and</strong> copper- <strong>and</strong> z<strong>in</strong>c oxide-impregnated carbon filters were<br />
used for these experiments. Subsequent wash tests demonstrated that no perchlorate was chemically<br />
reduced <strong>in</strong> the Cu- <strong>and</strong> ZnO-impregnated <strong>GAC</strong> filter. Though a small percentage <strong>of</strong> the perchlorate<br />
removed by the alum<strong>in</strong>um/<strong>GAC</strong> filter was not accounted for, the dom<strong>in</strong>ant perchlorate removal<br />
mechanism <strong>in</strong> both metal-catalyzed filters was ion exchange.<br />
Experiments <strong>in</strong>vestigat<strong>in</strong>g the removal <strong>of</strong> perchlorate <strong>in</strong> ozone <strong>and</strong> ozone plus hydrogen<br />
peroxide <strong>in</strong>dicated that no perchlorate reduction occurred.<br />
The <strong>GAC</strong> filters were rendered biologically active by extensively contact<strong>in</strong>g them with effluent<br />
from a set <strong>of</strong> filters that had demonstrated the ability to biologically reduce bromate. Once the <strong>GAC</strong><br />
filters were biologically active, the <strong>in</strong>fluent matrix for one B AC filter was switched to DDW to allow<br />
for better control over <strong>in</strong>fluent conditions. Influent dissolved oxygen concentrations were kept low.<br />
An electron donor solution was added to the <strong>in</strong>fluent tank when it was noted that no reduction <strong>of</strong><br />
perchlorate was occurr<strong>in</strong>g. Due to premature degradation <strong>of</strong> the electron donor solution, syr<strong>in</strong>ge<br />
pumps were <strong>in</strong>cluded <strong>in</strong> the treatment scheme to add electron donor at the po<strong>in</strong>t <strong>of</strong> entry to the BAG<br />
bed. Biological reduction <strong>of</strong> perchlorate was immediately observed. The DDW filter demonstrated<br />
efficient perchlorate removal <strong>and</strong> results from this filter revealed that perchlorate reduction is highly<br />
sensitive to the concentration <strong>of</strong> nitrate <strong>in</strong> the filter. As nitrate concentrations <strong>in</strong>crease, perchlorate<br />
removal decreases.<br />
In order to isolate the effect <strong>of</strong> nitrate concentration on perchlorate reduction <strong>in</strong> the BAG<br />
filters, electron donor concentration was <strong>in</strong>creased <strong>and</strong> EBCT was reduced. This ensured that<br />
electron donor availability was not a constra<strong>in</strong>t. Under these conditions, effluent nitrate<br />
concentrations were constant for a given <strong>in</strong>fluent nitrate concentration. A threshold nitrate<br />
concentration between 0.05 mg/L <strong>and</strong> 0.35 mg/L was established for the atta<strong>in</strong>ment <strong>of</strong> >95 percent<br />
perchlorate removal <strong>in</strong> the BAG filter.<br />
To <strong>in</strong>vestigate the effect <strong>of</strong> nitrate preload<strong>in</strong>g on the removal <strong>of</strong> perchlorate, an experiment<br />
was set up that compared perchlorate removal after the filter had been loaded with nitrate for several<br />
123
months to perchlorate removal after the filter did not see nitrate for 2 !/2 weeks prior to the start <strong>of</strong><br />
the experiment. The data showed that when compared to the nitrate preload condition, perchlorate<br />
removal for the non-nitrate preload condition decreased approximately 3 5 percent. It is not currently<br />
known if this phenomenon is caused by the deprivation <strong>of</strong> nutritional nitrogen, a loss <strong>in</strong> energy<br />
required for growth, or a suppression <strong>of</strong> perchlorate-reduc<strong>in</strong>g enzymes.<br />
<strong>Perchlorate</strong> removal versus EBCT without <strong>in</strong>fluent nitrate was also <strong>in</strong>vestigated. EBCTs <strong>of</strong><br />
25, 15,10, <strong>and</strong> 5 m<strong>in</strong>utes were utilized. Dur<strong>in</strong>g the course <strong>of</strong> this experiment, the BAG filter lost the<br />
ability to reduce perchlorate. When a second 25 m<strong>in</strong>ute EBCT experiment was run two months after<br />
the first 25 m<strong>in</strong>ute EBCT experiment, perchlorate removal decreased to less than 5 percent.<br />
<strong>Perchlorate</strong> removals dur<strong>in</strong>g the first 25 m<strong>in</strong>ute EBCT experiment were approximately 80 percent.<br />
Aga<strong>in</strong>, a loss <strong>in</strong> microbial growth, a suppression <strong>of</strong> perchlorate-reduc<strong>in</strong>g enzymes, or the deprivation<br />
<strong>of</strong> nutritional nitrogen may expla<strong>in</strong> this phenomenon. An experiment is currently be<strong>in</strong>g run to recover<br />
the perchlorate-reduction ability <strong>in</strong> the filter. The filter has been operat<strong>in</strong>g for 2 '/2 months with<br />
nitrate <strong>in</strong> the <strong>in</strong>fluent. Desorption <strong>of</strong> perchlorate from the filter may have masked the degree to which<br />
perchlorate was be<strong>in</strong>g biologically reduced dur<strong>in</strong>g the early portions <strong>of</strong> the filter recovery experiment.<br />
However, biological perchlorate reduction to below detection is now occurr<strong>in</strong>g <strong>in</strong> the filter.<br />
Batch biological perchlorate experiments have also been performed. Microorganisms that<br />
were isolated from the DDW BAG filter us<strong>in</strong>g agar plat<strong>in</strong>g techniques were <strong>in</strong>oculated <strong>in</strong>to aqueous<br />
media. One <strong>of</strong> the 12 isolates reduced perchlorate to below the detection limit-- isolate B9. Serum<br />
bottle tests us<strong>in</strong>g this isolate were run to exam<strong>in</strong>e the k<strong>in</strong>etic effects <strong>of</strong> different <strong>in</strong>itial nitrate<br />
concentrations on perchlorate removal. Several <strong>in</strong>terest<strong>in</strong>g observations were made. First, isolate<br />
B9 showed the ability to reduce both nitrate <strong>and</strong> perchlorate. Second, perchlorate removal did not<br />
occur <strong>in</strong> any serum bottle until the nitrate concentration was ^ 0.1 mg/L. This falls <strong>in</strong> the range <strong>of</strong><br />
nitrate concentrations required for 2t95 percent perchlorate removal <strong>in</strong> the BAG filters, as determ<strong>in</strong>ed<br />
dur<strong>in</strong>g the nitrate isolation experiments. Third, it was found that the greater the <strong>in</strong>itial nitrate<br />
concentration <strong>in</strong> the serum bottle, the greater the rate <strong>of</strong> perchlorate reduction. Aga<strong>in</strong>, the exact<br />
casue <strong>of</strong> this has not yet been determ<strong>in</strong>ed.<br />
Dilution to ext<strong>in</strong>ction experiments showed that perchlorate-reduc<strong>in</strong>g microorganisms are<br />
numerically dom<strong>in</strong>ant <strong>in</strong> the BAG filter. <strong>Bromate</strong> was reduced to below the detection limit at every<br />
dilution. Nitrate was reduced from 1.0 mg/L to
Microscopic exam<strong>in</strong>ation <strong>of</strong> the biomass present <strong>in</strong> the BAG filters revealed the most common<br />
morphology to be that <strong>of</strong> s<strong>in</strong>gle rod bacteria. These bacteria were found both as isolates <strong>and</strong> <strong>in</strong> floes.<br />
Large protozoa were also found <strong>in</strong> the perchlorate filters. Some morphologies were common<br />
between the bromate <strong>and</strong> the perchlorate filter samples while others were unique to each sample. A<br />
DGGE analysis run with samples from the perchlorate filters confirmed the microscopic analysis: the<br />
microbial population present <strong>in</strong> the BAG is diverse, shares some common species with the bromate<br />
filters, <strong>and</strong> conta<strong>in</strong>s some unique microbial species.<br />
METROPOLITAN WATER DISTRICT OF SOUTHERN CALIFORNIA: OZONATED<br />
WATER (NO ADDED ELECTRON DONOR) FOR BROMATE AND PERCHLORATE<br />
REDUCTION<br />
Bench-scale BAG filters follow<strong>in</strong>g ozonation demonstrated efficient biological removal <strong>of</strong><br />
both bromate <strong>and</strong> perchlorate for approximately 10,000 BV under various operat<strong>in</strong>g conditions.<br />
Complete bromate <strong>and</strong> perchlorate removal was observed with 10-50 ug/L <strong>of</strong> <strong>in</strong>fluent bromate <strong>and</strong><br />
perchlorate, EBCTs <strong>of</strong> 25 <strong>and</strong> 15 m<strong>in</strong>utes, <strong>and</strong> DO concentrations <strong>of</strong> 0-2 mg/L. When the <strong>in</strong>fluent<br />
DO was <strong>in</strong>creased to 5mg/L, perchlorate removal decreased by 50 percent while complete reduction<br />
<strong>of</strong> bromate cont<strong>in</strong>ued to be observed.<br />
Ozonation by-products are believed to serve as the electron donors needed by the perchlorate<strong>and</strong><br />
bromate- reduc<strong>in</strong>g microorganisms <strong>in</strong> the BAG filter. Further study is recommended to isolate<br />
<strong>and</strong> identify the specific by-products utilized by microorganisms as electron donors.<br />
Given that the filter operates anaerobically, the choice <strong>of</strong> a sulfur-based reduc<strong>in</strong>g chemical<br />
poses a taste-<strong>and</strong>-odor problem. Post-treatment may be required to meet aesthetic st<strong>and</strong>ards if the<br />
current DO reduc<strong>in</strong>g strategy is adopted for future studies. However, the use <strong>of</strong> an alternative<br />
reduc<strong>in</strong>g agent should be <strong>in</strong>vestigated to prevent any potential taste-<strong>and</strong>-odor problems.<br />
Pilot-scale experiments were conducted at the Metropolitan's Mobile Pilot Plant facility,<br />
which consists <strong>of</strong> ozone columns, conventional treatment, <strong>and</strong> BAG filters. The virg<strong>in</strong> carbon used<br />
for the pilot-scale work showed saturation <strong>of</strong> the reduction capacity for bromate <strong>and</strong> the ion exchange<br />
capacity for perchlorate at approximately 6,000 BV. When the system reached equilibrium (C/C0<br />
si), ozonation <strong>and</strong> deaeration were started to provide favorable conditions for biological reduction<br />
<strong>of</strong> bromate <strong>and</strong> perchlorate. 4,000 BV after switch<strong>in</strong>g the filter to this biological mode, complete<br />
125
omate <strong>and</strong> perchlorate reduction was observed. These results suggest that ozonation followed by<br />
deaeration <strong>and</strong> biological filtration could comprise an effective strategy for remov<strong>in</strong>g low |ig/L<br />
concentrations <strong>of</strong> bromate <strong>and</strong> perchlorate from dr<strong>in</strong>k<strong>in</strong>g water.<br />
126
CHAPTER 6. RECOMMENDATIONS TO UTILITIES<br />
1. To utilize BAG filtration for bromate <strong>and</strong> perchlorate reduction, a utility must<br />
determ<strong>in</strong>e if reduction <strong>of</strong> the <strong>in</strong>fluent DO concentration to the filter is feasible. Not<br />
only must it be economically feasible, but the utility should also determ<strong>in</strong>e that no<br />
unwanted effects are caused by DO reduction (i.e. lack <strong>of</strong> taste-<strong>and</strong>-odor removal or<br />
production <strong>of</strong> hydrogen sulfide <strong>in</strong> the filter). This could be accomplished through<br />
bench-scale test<strong>in</strong>g with the water <strong>and</strong> filter medium used by the utility.<br />
2. Bench-scale test<strong>in</strong>g at the utility should also be performed to determ<strong>in</strong>e the required<br />
contact time <strong>in</strong> the BAG filter to obta<strong>in</strong> the desired removal <strong>of</strong> bromate <strong>and</strong><br />
perchlorate.<br />
3. Depend<strong>in</strong>g on the alkal<strong>in</strong>ity <strong>of</strong> the water, the utility should consider a comb<strong>in</strong>ation <strong>of</strong><br />
reduction <strong>of</strong> bromate formation dur<strong>in</strong>g ozonation <strong>and</strong> an <strong>in</strong>crease <strong>in</strong> biological<br />
bromate removal <strong>in</strong> the BAG filter by pH control.<br />
4. Utilities that treat water contam<strong>in</strong>ated with both nitrate <strong>and</strong> perchlorate should give<br />
particular consideration to whether or not BAG filtration is a feasible treatment<br />
option. Research has shown that BAG filtration removes nitrate <strong>and</strong> perchlorate very<br />
efficiently. Additionally, nitrate reduction can enhance perchlorate reduction k<strong>in</strong>etics,<br />
mak<strong>in</strong>g BAG filtration particularly attractive for comb<strong>in</strong>ed removal <strong>of</strong> nitrate <strong>and</strong><br />
perchlorate.<br />
127
REFERENCES<br />
Amy, G., M. Siddiqui, W. Zhai, J. DeBroux, <strong>and</strong> W. Odem. 1993. Nation-Wide Survey <strong>of</strong> Bromide<br />
Ion Concentrations <strong>in</strong> Dr<strong>in</strong>k<strong>in</strong>g Water Sources. In Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works<br />
Association. Denver, Colorado: AWWA.<br />
APHA, AWWA, <strong>and</strong> WEF. (American Public Health Association, American Water Works<br />
Association, <strong>and</strong> Water Environment Federation). 1995. St<strong>and</strong>ard Methods for the Exam<strong>in</strong>ation<br />
<strong>of</strong> Water <strong>and</strong> Wastewater. 19* edition. Wash<strong>in</strong>gton, D. C.: APHA.<br />
Attaway, H. <strong>and</strong> M. Smith. 1993. Reduction <strong>of</strong> <strong>Perchlorate</strong> by an Anaerobic Enrichment Culture.<br />
Journal <strong>of</strong> Industrial Microbiology. 12: 408-412.<br />
Barber, M. J. <strong>and</strong> B. A. Norton. 1990. Sp<strong>in</strong>ach Nitrate Reductase - Effects <strong>of</strong> Ionic Strength <strong>and</strong> pH<br />
on the Full <strong>and</strong> Partial Enzyme Activities. Plant Physiology, 93:537-540.<br />
Bell L. C., D. J. Richardson, <strong>and</strong> S. J. Ferguson. 1990. Periplasmic <strong>and</strong> Membrane-Bound<br />
Respiratory Nitrate Reductases <strong>in</strong> Thiosphaera pantotropha. FEBS Letters, 265(l-2):85-87.<br />
Betts, K. S. 1998. Rotat<strong>in</strong>g Ion Exchange System Removes <strong>Perchlorate</strong>. Environmental Science<br />
<strong>and</strong> Technology. 32(19):454A-455A.<br />
Bishop, S. L., <strong>and</strong> M. Bourbigot. 1996. Meet<strong>in</strong>g the Proposed Regulations Us<strong>in</strong>g <strong>Ozone</strong> <strong>and</strong><br />
Biologically Active Carbon. Journal <strong>of</strong> the New Engl<strong>and</strong> Water Works Association, C V( 1): 1 -12.<br />
Book, S. 1999. Health Effects <strong>of</strong> <strong>Perchlorate</strong>. Presented at the 1999 <strong>Perchlorate</strong> Conference,<br />
Ontario, California, March 18-19,1999.<br />
129
Brock, T. D., M. T. Madigan, J. M. Mart<strong>in</strong>ko, <strong>and</strong> J. Parker. 1994. Biology <strong>of</strong> Microorganisms, 7th<br />
ed. Englewood Cliffs, New Jersey: Prentice Hall.<br />
California Department <strong>of</strong> Health Services (CDHS) Division <strong>of</strong> Dr<strong>in</strong>k<strong>in</strong>g Water <strong>and</strong> Environmental<br />
Management Dr<strong>in</strong>k<strong>in</strong>g Water Program. 1997. <strong>Perchlorate</strong> <strong>in</strong> Dr<strong>in</strong>k<strong>in</strong>g Water Basis <strong>of</strong> DHS=<br />
Provisional Action Level. In Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works Association Research<br />
Foundation <strong>Perchlorate</strong> Research Issue Group Conference. Ontario, California.<br />
Camel, V. <strong>and</strong> A. Bermond. 1998. The Use <strong>of</strong> <strong>Ozone</strong> <strong>and</strong> Associated Oxidation Processes <strong>in</strong><br />
Dr<strong>in</strong>k<strong>in</strong>g Water Treatment. Water Research. 32(11): 3208-3222<br />
Chen, J. M. <strong>and</strong> O. J. Hao. 1996. Environmental Factors <strong>and</strong> Model<strong>in</strong>g <strong>in</strong> Microbial Chromium<br />
(VI) Reduction. Water Environment Research, 68 (7): 1156-1164.<br />
Clarke, J. U. 1998. Evaluation <strong>of</strong> Censored Data Methods to Allow Statistical Comparisons Among<br />
Very Small Samples with Below Detection Limit Observations. Environmental Science <strong>and</strong><br />
Technology, 32 (1): 177-183.<br />
Coates, J.D. U. Michaelidou, R.A. Bruce, S.M. O' Conner, J.N. Crespi, <strong>and</strong> L.A. Auchenbach.<br />
1999. Ubiquity <strong>and</strong> Diversity <strong>of</strong> Dissimilatory (Per)chlorate-Reduc<strong>in</strong>g Bacteria. Applied <strong>and</strong><br />
Environmental Microbiology, 65(12): 5234-5241.<br />
DeAngelo, A. B., M. H. George., S. R. Kilburn, T. M. Moore, <strong>and</strong> D. C. Wolf. 1998.<br />
Carc<strong>in</strong>ogenicity <strong>of</strong> Potassium <strong>Bromate</strong> Adm<strong>in</strong>istered <strong>in</strong> the Dr<strong>in</strong>k<strong>in</strong>g Water to Male B6C3Fj Mice<br />
<strong>and</strong> F344/N Rats. Toxicologic Pathology, 26 (5): 587-594.<br />
Glaze, W. H., M. Koga, <strong>and</strong> D. Cancilla. 1989. Ozonation By Products 2: Improvement <strong>of</strong> Aqueous-<br />
Phase Derivatization Method for the Detection <strong>of</strong> Formaldehyde <strong>and</strong> Other Carbonyl Compounds<br />
Formed by the Ozonation <strong>of</strong> Dr<strong>in</strong>k<strong>in</strong>g Water. Environmental Science <strong>and</strong> Technology, 23: 838.<br />
130
Forbes, C. 1997. Fax sent from Calif. Dept. <strong>of</strong> Health Services (Dr<strong>in</strong>k<strong>in</strong>g Water Field Operations<br />
Branch) to Robert Mart<strong>in</strong> (General Manager <strong>of</strong> East Valley Water District, CA).<br />
Frankenberger, William T., Jr., et al. 1999. Bacterial Reduction <strong>of</strong> <strong>Perchlorate</strong> <strong>and</strong> Nitrate <strong>in</strong><br />
Groundwater by Perclace. Presented at the <strong>Perchlorate</strong> Conference, Ontario, California, March 18-<br />
19,1999.<br />
H aag, W. R. <strong>and</strong> J. Hoigne. 1983. Ozonation <strong>of</strong> Bromide-Conta<strong>in</strong><strong>in</strong>g Waters: K<strong>in</strong>etics <strong>of</strong> Formation<br />
<strong>of</strong> Hypobromous Acid <strong>and</strong> <strong>Bromate</strong>. Environmental Science <strong>and</strong> Technology, 17: 261-267.<br />
Herman, D. C., <strong>and</strong> W. T. Frankenberger. 1998. Microbial-Mediated Reduction <strong>of</strong> <strong>Perchlorate</strong> <strong>in</strong><br />
Groundwater. Journal <strong>of</strong> Environmental Quality 27:750-754.<br />
Herman, D.C. <strong>and</strong> W.T. Frankenberger. 1999. Bacterial Reduction <strong>of</strong> <strong>Perchlorate</strong> <strong>and</strong> Nitrate <strong>in</strong><br />
Water. Journal <strong>of</strong> Environmental Quality. 28: 1018-1024.<br />
Hijnen, W. A. M., R. long, <strong>and</strong> D. van der Kooij. 1999. <strong>Bromate</strong> <strong>Removal</strong> <strong>in</strong> a Denitrify<strong>in</strong>g<br />
Bioreactor used <strong>in</strong> Water Treatment. Water Research, 33 (4): 1049 - 1053.<br />
Hijnen, W. A. M., R. Voogt, H. R. Veenendaal, H. van der Jagt, <strong>and</strong> D. van der Kooij. 1995. <strong>Bromate</strong><br />
Reduction by Denitrify<strong>in</strong>g Bacteria. Applied <strong>and</strong> Environmental Microbiology, 61 (1): 239 - 244.<br />
Hurley, J.M., W. Wallace, <strong>and</strong> E. Coppola. 1996. The Military Eng<strong>in</strong>eer. August-September, 56.<br />
Jackson, C. R., E. E. Roden, <strong>and</strong> P. F. Churchill. 1998. Changes <strong>in</strong> Bacterial Species Composition<br />
<strong>in</strong> Enrichment Cultures with Various Dilutions <strong>of</strong> Inoculum as Monitored by Denatur<strong>in</strong>g Gradient<br />
Gel Electrophoresis. Applied <strong>and</strong> Environmental Microbiology, 64 (12): 5046-5048.<br />
Kirisits, M. J. <strong>and</strong> V. L. Snoey<strong>in</strong>k. 1999. Reduction <strong>of</strong> <strong>Bromate</strong> <strong>in</strong> a BAG Filter. Journal AWWA,<br />
91 (8): 74-84.<br />
131
Kirisits, M. J., V. L. Snoey<strong>in</strong>k, <strong>and</strong> J. C. Kruith<strong>of</strong>. 1998. Reduction <strong>of</strong> <strong>Bromate</strong> by Granular<br />
Activated Carbon. In Proceed<strong>in</strong>gs <strong>of</strong> the 1998 National Conference on Environmental Eng<strong>in</strong>eer<strong>in</strong>g.<br />
Chicago, Ill<strong>in</strong>ois.<br />
Krasner, S. W., W. H. Glaze, H. S. We<strong>in</strong>berg, <strong>and</strong> P. A. Daniel. 1993. <strong>Bromate</strong> Occurrence <strong>and</strong><br />
Control: Pilot- <strong>and</strong> Full-Scale Studies. In Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works Association.<br />
Denver, Colorado: AWWA.<br />
Kuo, C. Y., et al. 1990 Analysis <strong>of</strong> Inorganic Dis<strong>in</strong>fection By-Products <strong>in</strong> Ozonated Dr<strong>in</strong>k<strong>in</strong>g Water<br />
by Ion Chromatography. In Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works Association. San Diego,<br />
California: AWWA.<br />
Kurokawa, Y., A. Maekawa, M. Takahashi, <strong>and</strong> Y. Hayashi. 1990. Toxicity <strong>and</strong> Carc<strong>in</strong>ogenicity<br />
<strong>of</strong> Potassium <strong>Bromate</strong> - A New Renal Carc<strong>in</strong>ogen. Environmental Health Perspectives, 87: 309-<br />
335.<br />
Lloyd D., L. Boddy, <strong>and</strong> K. J. P. Davies. 1987. Persistence <strong>of</strong> Bacterial Denitrification Capacity<br />
under Aerobic Conditions: The Rule Rather than the Exception. FEMS Microbiology Ecology, 45:<br />
185-190.<br />
Logan, Bruce E. et al. 1999. * Remediation <strong>of</strong> Contam<strong>in</strong>ated Waters Us<strong>in</strong>g Innovative<br />
Bioreactors" Presented at the <strong>Perchlorate</strong> Conference, Ontario, California, March 18-19.<br />
Ludwig, W., G. Mittenhuber, <strong>and</strong> C. G. Friedrich. 1993. Transfer <strong>of</strong> Thiosphaera pantotropha to<br />
Paracoccus denitrificans. Int. J. Syst. Bacteriol., 43: 363-367.<br />
Meiberg, J. B. M., P. M. Bru<strong>in</strong>enberg, <strong>and</strong> W. Harder. 1980. Effect <strong>of</strong> Dissolved Oxygen Tension<br />
on the Metabolism <strong>of</strong> Methylated Am<strong>in</strong>es <strong>in</strong> Hyphomicrobium X <strong>in</strong> the Absence <strong>and</strong> Presence <strong>of</strong><br />
Nitrate: Evidence for x Aerobic' Denitrification. Journal <strong>of</strong> General Microbiology, 120: 453 - 463.<br />
132
Miller, J. A. 1996. AThe Reduction <strong>of</strong> <strong>Bromate</strong> <strong>in</strong> Dr<strong>in</strong>k<strong>in</strong>g Water by Activated Carbon. @ PhD<br />
Thesis. University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Champaign.<br />
Mills, A., A. Belghazi, <strong>and</strong> D. Rodman. 1996a. <strong>Bromate</strong> <strong>Removal</strong> from Dr<strong>in</strong>k<strong>in</strong>g Water by<br />
Semiconductor Photocatalysis. Water Research, 30 (9): 1973-1978.<br />
Mills, A., A. Belghazi, D. Rodman, <strong>and</strong> P. Hitch<strong>in</strong>s. 1996b. The <strong>Removal</strong> <strong>of</strong> <strong>Bromate</strong> from Potable<br />
Water Us<strong>in</strong>g Granular Activated Carbon. Journal <strong>of</strong> the Institution <strong>of</strong> Water <strong>and</strong> Environmental<br />
Management, 10 (3): 215-217.<br />
Najm, L, R. Trussell, L. Boulos, B. Gallagher, R. Bowcock, C. Williams, <strong>and</strong> D. Clifford. 1999. In<br />
Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works Association. Chicago, Ill<strong>in</strong>ois: AWWA.<br />
Pfaff, J. D., <strong>and</strong> C. A. Brockh<strong>of</strong>f. 1990. Determ<strong>in</strong><strong>in</strong>g Inorganic Dis<strong>in</strong>fection By-Products by Ion<br />
Chromatography. Journal AWWA, 82:4:192.<br />
Renner, Rebecca. 1999. EPA Draft Almost Doubles Safe Dose <strong>of</strong> <strong>Perchlorate</strong> <strong>in</strong> Water.<br />
Environmental Science <strong>and</strong> Technology. 33 (5): 110A-111 A.<br />
Rikken, G.B., A.G.M. Kroon, <strong>and</strong> C.G. van G<strong>in</strong>kel. 1996. Appl. Microb. Biotech., 45: 420.<br />
Rittmann, B.E., R. Nurenberg, <strong>and</strong> I. Najm. 2000. "Autohydrogenotrophic <strong>Perchlorate</strong><br />
Reduction." Paper presented at the AWWA/AWWARF Inorganic Contam<strong>in</strong>ants Workshop <strong>in</strong><br />
Albuquerque, NM, February 27-29.<br />
Robertson, L.A. <strong>and</strong> J. G. Kuenen. 1990. Comb<strong>in</strong>ed Heterotrophic Nitrification <strong>and</strong> Ae<br />
<strong>in</strong> Thiosphaera pantotropha <strong>and</strong> Other Bacteria. Antonie van Leeuwenhoek, 57:139-152.<br />
Robertson L. A. <strong>and</strong> J. G. Kuenen. 1984. Aerobic Denitrification: A Controversy Revived.<br />
Archives <strong>of</strong> Microbiology, 139: 351-354.<br />
133
Rogers, K. 1997. Water Contam<strong>in</strong>ated. Las Vegas Review Journal. August 19: 1-3.<br />
Sclimenti, M.J., S.W. Krasner, W.H. Glaze, <strong>and</strong> H.S. We<strong>in</strong>berg, 1990. <strong>Ozone</strong> Dis<strong>in</strong>fection By-<br />
Products: Optimization <strong>of</strong> the PFBHA Derivatization Method for the Analysis <strong>of</strong> Aldehydes. In<br />
Proceed<strong>in</strong>gs <strong>of</strong> the American Water Works Association, San Diego, California: AWWA.<br />
Siddiqui, M., W. Zhai, G. Amy, <strong>and</strong> C. Mysore. 1996. <strong>Bromate</strong> Ion <strong>Removal</strong> by Activated Carbon.<br />
Water Research, 30 (7): 1651-1660.<br />
Siddiqui, M.J., G. Amy, K. Ozek<strong>in</strong>, W. Zhai, <strong>and</strong> P. Westerh<strong>of</strong>f. 1994. Alternative Strategies for<br />
Remov<strong>in</strong>g <strong>Bromate</strong>. Journal AWWA, 86 (10): 81-96.<br />
Simpk<strong>in</strong> T. J. <strong>and</strong> W. C. Boyle. 1988. The Lack <strong>of</strong> Repression by Oxygen <strong>of</strong> the Denitrify<strong>in</strong>g<br />
Enzymes <strong>in</strong> Activated Sludge. Water Research, 22 (2): 201-206.<br />
Skoog, D. A. <strong>and</strong> J. J. Leary. 1992. Pr<strong>in</strong>ciples <strong>of</strong> Instrumental Analysis. 4*ed. Fort Worth, TX:<br />
Harcourt Brace College Publishers.<br />
Song R., C. Donohoe, R. M<strong>in</strong>ear, P. Westerh<strong>of</strong>f, K. Ozek<strong>in</strong>, <strong>and</strong> G. Amy. 1996. Empirical<br />
Model<strong>in</strong>g <strong>of</strong> <strong>Bromate</strong> Formation dur<strong>in</strong>g Ozonation <strong>of</strong> Bromide-Conta<strong>in</strong><strong>in</strong>g Waters. Water Research,<br />
30(5): 1161-1168.<br />
SoucieW. 1999. November 9. Personal communication. Central Lake County Jo<strong>in</strong>t Action Water<br />
Agency, Paul M. Neal Water Treatment Facility, 200 Rockl<strong>and</strong> Road, Lake Bluff, Ill<strong>in</strong>ois 60044.<br />
St<strong>and</strong>ard Methods for the Exam<strong>in</strong>ation <strong>of</strong> Water <strong>and</strong> Wastewater, 19th Edition, 1995.<br />
Tiedje,J. M. 1988. Ecology <strong>of</strong> enitrification <strong>and</strong> Dissimilatory Nitrate Reduction to Ammonium.<br />
In Biology <strong>of</strong> Anaerobic Microorganisms. Edited by A. J. B. Zehnder. New York: John Wiley <strong>and</strong><br />
Sons.<br />
134
USEPA. 1998. Health Risk Assessment/Characterization <strong>of</strong> the Dr<strong>in</strong>k<strong>in</strong>g Water Dis<strong>in</strong>fection<br />
Byproduct <strong>Bromate</strong>. Office <strong>of</strong> Science <strong>and</strong> Technology, Office <strong>of</strong> Water. March 13,1998. Quoted<br />
<strong>in</strong> Fed. Reg. 63:61:15673, FR Document 98-8215.<br />
V<strong>and</strong>enabeele, J., D. de Beer, R. Germonpre, R. Van de S<strong>and</strong>e, <strong>and</strong> W. Verstraete. 1995. Influence<br />
<strong>of</strong> Nitrate on Manganese Remov<strong>in</strong>g Microbial Consortia from S<strong>and</strong> Filters. Water Research, 29 (2):<br />
579-587.<br />
von Gunten, U. <strong>and</strong> J. Hoigne. 1996. Ozonation <strong>of</strong> Bromide-Conta<strong>in</strong><strong>in</strong>g Waters: <strong>Bromate</strong><br />
Formation through <strong>Ozone</strong> <strong>and</strong> Hydroxyl Radicals. In Dis<strong>in</strong>fection By-Products <strong>in</strong> Water Treatment<br />
- The Chemistry <strong>of</strong> Their Formation <strong>and</strong> Control. Edited by R. A. M<strong>in</strong>ear <strong>and</strong> G. L. Amy. Boca<br />
Raton, Florida: Lewis Publishers.<br />
von Gunten, U. <strong>and</strong> J. Hoigne. 1994. <strong>Bromate</strong> Formation Dur<strong>in</strong>g Ozonation <strong>of</strong> Bromide-Conta<strong>in</strong><strong>in</strong>g<br />
Waters: Interaction <strong>of</strong> <strong>Ozone</strong> <strong>and</strong> Hydroxyl Radical Reactions. Environmental Science <strong>and</strong><br />
Technology, 28: 1234-1242.<br />
Waterweek. 1997. <strong>Perchlorate</strong> Detected <strong>in</strong> Lake Mead Water. AWWA, Denver (Sept. 1, 1997).<br />
Waterweek. 1998. EPA F<strong>in</strong>alizes M/DBP Rules. AWWA, Denver (December 4,1998).<br />
Westerh<strong>of</strong>f, P., G. Amy, R. M<strong>in</strong>ear, M. Siddiqui, R. Song, <strong>and</strong> K. Ozek<strong>in</strong>. 1993. <strong>Bromate</strong> Formation<br />
<strong>in</strong> Laboratory Studies: Influence <strong>of</strong> Water Quality <strong>and</strong> Treatment Variables. In Proceed<strong>in</strong>gs <strong>of</strong> the<br />
American Water Works Association. Denver, Colorado: AWWA.<br />
9<br />
Ye R. W., B. A. Averill, <strong>and</strong> J. M. Tiedje. 1994. Denitrification: Production <strong>and</strong> Consumption <strong>of</strong><br />
Nitric Oxide. Appl. Envir. Microbiol. 60 (4): 1053-1058.<br />
135
ABBREVIATIONS<br />
AWWA<br />
AWWARF<br />
American Water Works Association<br />
American Water Works Research Foundation<br />
BAG<br />
biologically active carbon<br />
CDHS California Department <strong>of</strong> Health Services<br />
°C degrees Celsius<br />
C<br />
concentration at time, t<br />
C/C0 fraction rema<strong>in</strong><strong>in</strong>g<br />
cm centimeter<br />
C0<br />
<strong>in</strong>itial concentration<br />
CUW Champaign-Urbana water<br />
DDW<br />
DFE<br />
DO<br />
DOC<br />
distilled deionized water<br />
Danville filter effluent<br />
dissolved oxygen<br />
dissolved organic carbon<br />
EBCT<br />
empty bed contact time<br />
<strong>GAC</strong><br />
granular activated carbon<br />
HPC<br />
heterotrophic plate count<br />
L<br />
LOD<br />
liter<br />
limit <strong>of</strong> detection<br />
M<br />
molar<br />
137
MCL maximum contam<strong>in</strong>ant level<br />
mg milligram<br />
mg/L milligrams per liter<br />
mL/m<strong>in</strong> milliliters per m<strong>in</strong>ute<br />
MWDSC Metropolitan Water District <strong>of</strong> Southern California<br />
pg/L microgram per liter<br />
NOM natural organic matter<br />
PAC Project Advisory Committee<br />
t time<br />
TNTC too numerous to count<br />
U.S. United States<br />
USEPA United States Environmental Protection Agency<br />
138
AWWA<br />
Research<br />
Foundation<br />
Advanc<strong>in</strong>g the Sderm <strong>of</strong> Water*<br />
6666 W. Qu<strong>in</strong>cy Avenue, Denver, CO 80235<br />
(303)347-6100<br />
1 P-3.25C-90836-3/01 -CM ISBN 1-58321-104-7