 

 

 

 

 

..._ aw-

 

 

w

'1an

(v

.-.-.‘-.-.....m--

 

....\.‘.

 

This is to certify that the
dissertation entitled

Ozonation, Ultraﬂltration, and Bioﬁltration for the Control of
NOM and DBP in Drinking Water

presented by

Kuan-chung Chen

has been accepted towards fulﬁllment
of the requirements for the

 

PhD. degree in Environmental Engineering

-/ Major Profﬂsor’s Signature
@5441; J63, 02 00 3

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DaloDue.p65-p.15

OZONATION, ULTRAFILTRATION, AND BIOFILTRATION FOR THE
CONTROL OF NOM AND DBP IN DRINKING WATER

By

Kuan-chung Chen

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Civil and Environmental Engineering

2003

ABSTRACT

OZONATION, ULTRAFILTRATION, AND BIOFILTRATION FOR THE
CONTROL OF NOM AND DBP IN DRINKING WATER

By

Kuan-chung Chen

Ozone is an effective disinfectant that can inactivate microorganisms and
reduce the formation of disinfection by-products (DBPs). Ultraﬁltration and
nanoﬁltration membranes are also capable of removing DBP precursors and
pathogens. However, natural organic matter (NOM), which is ubiquitous in surface
water, can result in membrane fouling. A novelly designed ozonation/membrane
system followed by bioﬁltration, which is expected to enhance ozone mass transfer
and reduce membrane fouling, was used in this study. The effect of this system on the
rate of ozone mass transfer, the transformation of the NOM in terms of the variation
of UV-254, humic substances and non-humic substances, assimilable organic carbon
(ADC), and biodegradable organic carbon (BDOC) was assessed. The effect of
operational parameters on the control of DBPs and ozonation by-products was also
investigated. Tubular ceramic membranes with a molecular weight cut-off of 15 kD
were used in the system.

A mathematical model was developed and used to calculate the volumetric
mass transfer coefﬁcient of ozone for the bench-scale system. This model was
successfully predicted the dissolved ozone concentration in the water tank. The
effects of two different mixers, simple Y mixer (SYM) and high efﬁciency mixer

(HEM), on the rate of ozone mass transfer were investigated. With HEM, greater

values of ozone volumetric mass transfer coefficient were obtained at lower ozone
doses.

Lake Lansing water was used to study the transfomiation of NOM. Water
temperature, ozone gas flow rate, and ozone dose were controlled. The
ozonation/membrane system removed 15-30% of the dissolved organic carbon (DOC).
The removal efficiency increased to 69-88%, when followed by bioﬁltration.
Concentrations of AOC and BDOC significantly increased after treatment using the
ozonation/membrane system but were effectively removed by subsequent bioﬁltration.
The UV—254 absorbance decreased by up to 85% and humic substances were
converted to non-humic substances after treated by the ozonation/membrane system.

The formation of trihalomethanes (THMs) and haloacetic acids (HAAs) was
effectively decreased by the system and by its subsequent bioﬁltration. Although the
concentrations of aldehydes and ketoacids increased significantly after ozonation,
these compounds were removed effectively by the subsequent biofiltration. The
decrease of chlorine demand was limited when treatment consisted of only
ozonation/membrane. With bioﬁltration, the chlorine demand of water decreased by
60-90%.

Correlations between AOC and ozonation by-products were also investigated.
The sum of ozonation by-products (aldehydes + 2-butanone + ketoacids), the
concentration of aldehydes, and the concentration of methyl glyoxal were well
correlated with AOC and could be good surrogate parameters for the AOC

measurement.

 

Copyﬁghtby
Kuan-chung Chen

2003

To my parents and my wife for their love and support.

ACKNOWLEDGEMENTS

I would like to express my deepest gratitude to my advisor, Dr. Susan J.
Masten, for the opportunity to start and complete this research. She freely shared her
wealth of knowledge on ozonation with me and offered valuable suggestions that
helped me overcome obstacles that I encountered in this research. This research
would not have been completed without her support. I would also like to thank my
committee members Dr. Mackenzie L. Davis, Dr. Syed A. Hashsham, and Dr. Simon
H. R Davies for seeing me through this long journey and for offering good advice at
critical points along the way.

I would like to thank Dr. Xianda Zhao who gave me valuable suggestions and
comments on this work with his expertise. I also acknowledge Mr. Yan-lyang Pan for
his technical support in the laboratory. 1 would also like to thank Mr. Hung-hsu Yen
for the time he spent discussing data analysis with me. Thanks also to all of my
friends who have listened to me moan, rejoice, complain, laugh, and ponder my way
through these past years.

Finally, I would like to thank my family. I wish to thank my wife, Ya-yun,
who not only gave me moral support but also helped me organize tons of data during
this research. She has accompanied me through the ups and downs in these years and
no one could ask for a better friend in life than she has been to me. With her love and
support, life is much more pleasant and colorful. I also thank my parents, my sisters,
and my in-laws, for their love, encouragement and support. Thank you all for giving

me this opportunity to do what I love.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... xii

LIST OF FIGURES ......................................................................... xiv

CHAPTER 1

INTRODUCTION

1.1 SIGNIFICANCE .......................................................................... 1

1.2 OBJECTIVES ............................................................................. 4

1.3 HYPOTHESES ........................................................................... 5

1.4 REFERENCES ........................................................................... 6

CHAPTER 2

BACKGROUND

2.1 ORGANIC MATTER IN DRINKING WATER .................................... 8
2.1.1 Natural Organic Matter ............................................................ 9
2.1.2 Humic Substances .................................................................. 10
2.1.3 Molecular Weight Distribution ................................................... 12
2.1.4 Biodegradable Dissolved Organic Carbon ...................................... 13

2.2 DISINFECTION BY-PRODUCTS ..................................................... 16

2.3 PATHOGENS .............................................................................. 20

2.4 TREATMENT ............................................................................. 22
2.4.1 Ozonation ............................................................................ 22
2.4.2 Membrane ........................................................................... 26
2.4.3 Ozonation/Biotreatment ............................................................ 32

vii

2.4.4 Ozonation/Membrane .............................................................. 35

2.5 REFERENCES ............................................................................ 37

CHAPTER 3

MASS TRANSFER OF OZONE IN THE OZONATION/MEMBRANE
PROCESS

3.1 INTRODUCTION ........................................................................ 50
3.2 MATERIALS AND METHODS ...................................................... 51
3.2.1 Permeate Flux Enhancement Experiment ....................................... 51
3.2.2 Mass Transfer Experiment ........................................................ 54
3.2.3 Ozone Concentration Measurement .............................................. 57
3.3 MATHEMATICAL MODEL OF OZONE MASS TRANSFER ................... 58
3.4 RESULTS AND DISCUSSION ......................................................... 65
3.4.1 Permeate flux Enhancement Experiment ........................................ 65
3.4.2 Decomposition of Ozone in the Proposed System ............................. 67
3.4.3 Ozone Mass Transfer Coefficient ................................................ 70
3.4.4 Simple Y Inline Mixer vs. High Efficiency Inline Mixer ..................... 75
3.4.5 Model Validation ................................................................... 80
3.4.6 Comparison with Other Studies .................................................. 85
3.5 CONCLUSIONS .......................................................................... 89
3.6 REFERENCES ............................................................................ 90
CHAPTER 4

THE EFFECT OF OZONATION/MEMBRANE PROCESS ON THE
CHARACTERISTICS OF NATURAL ORGANIC MATTER

viii

4.1 INTRODUCTION .......................................................................... 94

4.2 MATERIALS AND METHODS ......................................................... 96
4.2.1 Experimental Setup and Protocol .................................................. 96
4.2.2 Water Source ....................................................................... 100
4.2.3 Gas-phase Ozone Analysis................... .................................... 101
4.2.4 UV-254 Absorption ............................................................... 101
4.2.5 TOC and DOC Analyses .......................................................... 102
4.2.6 Humic Substances and Non-humic Substances Analyses ..................... 102
4.2.7 Biodegradable Dissolved Organic Carbon Analysis ........................... 103
4.2.8 Assimilable Organic Carbon Analysis .......................................... 103

4.3 RESULTS AND DISCUSSION ......................................................... 104
4.3.1 Permeate Flux Recovery .......................................................... 104
4.3.2 DOC ................................................................................. 109
4.3.3 UV—254 ............................................................................. 115
4.3.4 Humic and Non—humic Substances .............................................. 118
4.3.5 BDOC ................................................................................ 123
4.3.6 AOC .................................................................................. 126

4.4 CONCLUSIONS ......................................................................... 130

4.5 REFERENCES ........................................................................... 132

CHAPTER 5

THE USE OF OZONATION/MEMBRANE PROCESS FOR THE CONTROL
OF DISINFECTION BY-PRODUCTS

5.1 INTRODUCTION ........................................................................ 136

5.2 MATERIALS AND METHODS ....................................................... 137

ix

1“"-

5.2.1 Experimental Setup and Protocol ................................................ 137

5.2.2 Water Source ........................................................................ 140
5.2.3 Chlorine Residual .................................................................. 140
5.2.4 SDS THMs .......................................................................... 140
5.2.5 SDS HAAs .......................................................................... 141
5.2.6 Aldehydes and Ketones ............................................................ 141
5.2.7 Ketoacids ............................................................................ 142
5.2.8 Other Analyses ..................................................................... 143
5.3 RESULTS AND DISCUSSION ........................................................ 143
5.3.1 Chlorine Demand .................................................................. 143
5.3.2 THMs and HAAs .................................................................. 147
5.3.3 Aldehyde, Ketone, and Ketoacids ................................................ 159
5.4 CONCLUSIONS ......................................................................... 172
5.5 REFERENCES ........................................................................... 175
CHAPTER 6

CORRELATION BETWEEN THE FORMATION OF ASSIMILABLE
ORGANIC CARBON AND OZONATION BY-PRODUCTS

6.1 INTRODUCTION ........................................................................ 179
6.2 MATERIALS AND METHODS ....................................................... 180
6.2.1 Experimental Setup and Protocol ................................................ 180
6.2.2 Water Source ........................................................................ 180
6.2.3 Analytical Methods ................................................................ 180
6.3 RESULTS AND DISCUSSION ........................................................ 181
6.3.1 Ozonation By-products ............................................................ 181

 

6.3.2 Correlations Between AOC and Ozonation By-products ..................... 189

6.4 CONCLUSIONS ......................................................................... 192
6.5 REFERENCES ........................................................................... 193
CHAPTER 7

CONCLUSIONS AND RECOMMENDATIONS

7.1 CONCLUSIONS ......................................................................... 196
7.1.1 Mass Transfer of Ozone in the System .......................................... 196
7.1.2 Characteristics of NOM ........................................................... 197
7.1.3 Control of DBPs .................................................................... 198
7.1.4 Correlations Between AOC and Ozonation By-products ..................... 199

7.2 RECOMMENDATIONS FOR FUTURE RESEARCH ............................. 200

xi

 

LIST OF TABLES

Table 3.1 Details of the ceramic membrane ............................................... 53
Table 3.2 Experimental conditions for the ozone mass transfer measurement ........ 57
Table 3.3 Effects of water ﬂow rate and gas flow rate on permeate ﬂux .............. 65

Table 3.4 Decomposition of dissolved ozone in DD] water acidified to pH 2 under
different water temperature in the proposed system ........................ 68

Table 3.5 Calculated values of ozone mass transfer coefﬁcient (KLa) and
experimentally determined ozone concentrations under different gas ﬂow
rates. (Operating conditions: water temperature 20 C, water ﬂow rate 3.5
LPM, ozone dose 2.5 g O3/m3) ................................................ 72

Table 3.6 Calculated values of the ozone volumetric mass transfer coefficient (KLa)
and experimentally determined ozone concentrations under different water
ﬂow rates. (Operating conditions: water temperature, 20 C; gas ﬂow rate,
100 mL/min; ozone dose, 2.5 g O3/m3) ........................................ 72

Table 3.7 Calculated values of ozone mass transfer coefficient (KLa) and
experimentally determined ozone concentrations under different ozone
doses. (Operating conditions: water temperature 20 C, water ﬂow rate 3.5
LPM, gas ﬂow rate 100 mL/min) ............................................. 73

Table 3.8 Calculated values of ozone mass transfer coefficient (KLa) and
experimentally determined ozone concentrations under different water
temperatures. (Operating conditions: water flow rate 3.5 LPM, gas ﬂow
rate 100 mL/min, ozone dose 2.5 g O3/m3) .................................. 75

Table 3.9 Parameters comparison between Simple Y Mixer (SYM) and High
Efficiency Mixer (HEM) ........................................................ 76

Table 3.10 Comparison of model predictions and experimental data at steady-state of
dissolved ozone concentration in the pH 2 DDI water under various

operating conditions .............................................................. 84

Table 3.11 KLa values and experimental conditions from this study and literature on

ozone mass transfer .............................................................. 88
Table 4.1 Details of the ceramic membrane ............................................... 98
Table 4.2 Experimental conditions .......................................................... 99
Table 4.3 Typical quality characteristics of Lake Lansing water ....................... 101
Table 5.1 Experimental conditions ......................................................... 138

xii

 

Table 5.2 Chlorine demands and DOC concentrations for samples collected at various
treatment stages under different experimental conditions .................. 145

Table 5 .3 Concentrations of THMs and HAAs for Lake Lansing water treated by
continuous stirred reactor. (Operating conditions: hydraulic retention time,
12.5 min; ozone dosage, 1 mg O3/mg C) ..................................... 150

Table 5.4 Removal efficiencies of aldehydes by ozonation/membrane/bio-
filtration ........................................................................... 166

Table 6.1 Concentrations of 2-Butanone for Lake Lansing water measured in different
treatment stages at various experimental conditions in this study.

(pg/L) .............................................................................. 188
Table 6.2 Correlations between AOC (total), AOC (P17), AOC (NOX) and ozonation
by-products ....................................................................... l 90

xiii

 

LIST OF FIGURES

Figure 2.1 Reactions of ozone with bromide in aqueous media, M represents organic

matter. (Haag and Hoigné 1983) ............................................... 19
Figure 3.1 Experimental setup for measurement of permeate flux enhancement of

ozone injection .................................................................... 52
Figure 3.2 Simple Y inline mixer (SYM) ................................................. 54
Figure 3.3 Experimental setup of the ozone mass transfer study ....................... 56
Figure 3.4 High efﬁciency inline mixer (HEM) ........................................... 57
Figure 3.5 Two-ﬁlm theory ................................................................... 59

Figure 3.6 The ozonation/membrane system for ozone mass transfer modeling. . 62

Figure 3.7 Effects of ozone flux on the permeate ﬂux. Operating conditions: gaseous
03 concentration, 2.5 g/m3; water temperature, 27 °C; water flow rate
varied from 2 LPM (L/min) to 5 LPM ......................................... 66

Figure 3.8 Ozone decomposition in DDI water for ﬁrst-order kinetic at 10 °C.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM ..................... 69

Figure 3.9 Ozone decomposition in DDI water for ﬁrst-order kinetic at 20 °C.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM..... . . 69

Figure 3.10 Ozone decomposition in DDI water for ﬁrst-order kinetic at 30 °C.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM ................... 70

Figure 3.11 Method to ﬁt the ozone volumetric mass transfer coefﬁcients by linear
regression at different ozone gas ﬂow rate ................................. 71

Figure 3.12 Comparison of dissolved ozone concentration using SYM and HEM
mixers. (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 2.5 g/m3). . . . . .....77

Figure 3.13 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 5.8 g/m3). .........77

Figure 3.14 Comparison of dissolved ozone concentration using SYM and HEM

mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas flow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 12.5 g/m3) ........ 78

xiv

Figure 3.15 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water flow rate, 3.5 LPM; gas flow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 23.0 g/m3) ........ 78

Figure 3.16 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas flow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 36.0 g/m3). . ......79

Figure 3.17 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 10 °C; ozone dose, 2.5 g/m3). . . .......79

Figure 3.18 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 30 °C; ozone dose, 2.5 g/m3). . . .......80

Figure 3.19 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
2.75 LPM; gas flow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 2.5 g/m3 ................................................................... 81

Figure 3.20 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water flow rate,
3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 2.5 g/m3 ................................................................... 82

Figure 3.21 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
3.5 LPM; gas ﬂow rate, 50 mL/min; water temperature, 20 °C; ozone
dose, 2.5 g/m3 ................................................................... 82

Figure 3.22 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water flow rate,
3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 5.8 g/m3 ................................................................... 83

Figure 3.23 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 36.0 g/m3 ................................................................. 83

Figure 4.1 Experimental setup for the ozonation/membrane process ................. 97

Figure 4.2 The variation of permeate ﬂux with respect to operating time. Operating
conditions: water ﬂow rate, 2.75 L/min; ozone gas ﬂow rate, 100 mL/min;
ozone dose, 12.5 g/m3; water temperature, 20 °C ............................ 106

Figure 4.3 The variation of permeate flux with respect to operating time. Operating

conditions: water flow rate, 2.75 L/min; ozone gas flow rate, 100 mL/min;
water temperature, 20 °C ......................................................... 106

XV

Figure 4.4 The atomic force microscope (AF M) scan of glass substrate with nothing
on it ................................................................................ 107

Figure 4.5 The atomic force microscope (AFM) scan of backwash of used ceramic
membrane without ozonation of NOM. ...................................... 107

Figure 4.6 The atomic force microscope (AF M) scan of backwash of used ceramic
membrane with ozonation of NOM ........................................... 108

Figure 4.7 C lose-up of area in white box in Figure 4.6 ................................... 108

Figure 4.8 The variation of DOC concentration of different treatment stages at various
water treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas flow rate, 100 mL/min; ozone dose, 2.5 g/m3....................111

Figure 4.9 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various water treatment temperatures.
Concentrations are compared to those in the raw water. For temperatures
10, 20, 30 °C, the raw water DOC concentrations were 9.5, 10.9, 9.0 mg/L,
respectively. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow
rate, 100 mL/min; ozone dose, 2.5 g/m3 ...................................... 1 1 1

Figure 4.10 The variation of DOC concentration of different treatment stages at
various ozone gas ﬂow rates. Operating conditions: water ﬂow rate,
2.75 L/min; ozone dose, 2.5 g/m3; water temperature, 20 °C ............ 112

Figure 4.1 1 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various ozone gas flow rates. Concentrations
are compared to those in the raw water. For gas flow rates 50, 75, 100
mL/min, the raw water DOC concentrations were 8.7, 10.4, 10.9 mg/L,
respectively. Operating conditions: water flow rate, 2.75 L/min; ozone
dose, 2.5 mg/L; water temperature, 20 °C ................................... 1 13

Figure 4.12 The variation of DOC concentration of different treatment stages at
various ozone doses. Operating conditions: water flow rate, 2.75 L/min;
gas flow rate, 100 mL/min; water temperature, 20 °C .................... 114

Figure 4.13 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various ozone doses. Concentrations are
compared to those in the raw water. For ozone doses 1.5, 2.5, 5.8 g/m3,
the raw water DOC concentrations were 8.7, 10.9, 8.9 mg/L, respectively.
Operating conditions: water flow rate, 2.75 L/min; gas ﬂow rate,
lOOmL/min; water temperature, 20 0C ...................................... 1 15

Figure 4.14 Efﬁciency of UV-254 removal at various water temperatures. Operating

conditions: water ﬂow rate, 2.75 L/min; gas flow rate, 100 mL/min;
ozone dose, 2.5 g/m3 .......................................................... 116

xvi

Figure 4.15 Efﬁciency of UV-254 removal at various ozone gas flow rates. Operating
conditions: water ﬂow rate, 2.75 L/min; ozone dose, 2.5 g/m3; water
temperature, 20 °C ............................................................. 117

Figure 4.16 Efﬁciency of UV-254 removal at various ozone doses. Operating
conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100 mL/min;
water temperature, 20 °C ...................................................... 117

Figure 4.17 The ratio of HS/DOC of different treatment stages at various water
treatment temperatures. Operating conditions: water flow rate, 2.75
L/min; gas flow rate, 100 mL/min; ozone dose, 2.5 g/m3 ................. 1 19

Figure 4.18 The ratio of NHS/DOC of different treatment stages at various water
treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas flow rate, 100 mL/min; ozone dose, 2.5 g/m3. ............... 120

Figure 4.19 The ratio of HS/DOC of different treatment stages at various ozone gas
ﬂow rates. Operating conditions: water flow rate, 2.75 L/min; ozone
dose, 2.5 g/m ; water temperature, 20 °C ................................... 120

Figure 4.20 The ratio of NHS/DOC of different treatment stages at various ozone gas
ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m ; water temperature, 20 °C ................................... 121

Figure 4.21 The ratio of HS/DOC of different treatment stages at various ozone doses.
Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100
mL/min; water temperature, 20 °C .......................................... 121

Figure 4.22 The ratio of NHS/DOC of different treatment stages at various ozone
doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C ..................................... 122

Figure 4.23 Correlation between the concentration of humic substances and UV-254
for Lake Lansing water. (Correlation is signiﬁcant at the 0.01
leveL) ........................................................................... 122

Figure 4.24 The percentages of BDOC to DOC of F RW and P2 water at various water
treatment temperatures. Operating conditions: water flow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.................125

Figure 4.25 The percentages of BDOC to DOC of FRW and P2 water at various
ozone gas flow rates. Operating conditions: water ﬂow rate, 2.75 L/min;
ozone dose, 2.5 g/m3; water temperature, 20 °C ........................... 125

Figure 4.26 The percentages of BDOC to DOC of FRW and P2 water at various

ozone doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas
ﬂow rate, 100 mL/min; water temperature, 20 °C ........................ 126

xvii

 

Figure 4.27 The concentration of AOC of different treatment stages at various water
treatment temperatures. Operating conditions: water flow rate, 2.75
L/min; gas flow rate, 100 mL/min; ozone dose, 2.5 g/m3 ................. 128

Figure 4.28 The concentration of AOC of different treatment stages at various ozone
gas ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m3; water temperature, 20 °C ................................... 129

Figure 4.29 The concentration of AOC of different treatment stages at various ozone
doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C ..................................... 129

Figure 5.1 The percentage variation of chlorine demand (CDi/CDO) of different
treatment stages at various water temperatures. (Operating conditions:
water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5
g/m3) ................................................................................ 145

Figure 5.2 The percentage variation of chlorine demand (CDi/CDO) of different
treatment stages at various gas ﬂow rates. (Operating conditions: water
flow rate, 2.75 L/min; 03 dose, 2.5 g/m3; water temperature, 20 °C). . 146

Figure 5.3 The percentage variation of chlorine demand (CDi/CDO) of different
treatment stages at various ozone doses. (Operating conditions: water ﬂow
rate, 2.75 L/min; gas ﬂow rate, 100 mL/min; water temperature, 20
0C) ................................................................................. 146

Figure 5.4 Correlation between chlorine demand and the concentration of DOC of
Lake Lansing water treated by the ozonation/membrane/bioﬁltration
system. (Correlation is statistically signiﬁcant at 95% CI.) ............... 147

Figure 5.5 Concentrations of THMs of different treatment stages at various water
treatment temperatures. (Operating conditions: water ﬂow rate, 2.75
L/min; gas flow rate, 100 mL/min; 03 dose, 2.5 g/m3) ...................... 149

Figure 5.6 Concentrations of HAAs of different treatment stages at various water
treatment temperatures. (Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; 0; dose, 2.5 g/m3) ...................... 150

Figure 5.7 Concentrations of THMs of different treatment stages at various gas ﬂow
rates. (Operating conditions: water ﬂow rate, 2.75 L/min; 03 dose, 2.5
g/m3; water temperature, 20 °C) ................................................ 152

Figure 5.8 Concentrations of HAAs of different treatment stages at various gas ﬂow
rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose, 2.5
g/m3; water temperature, 20 °C) ................................................ 152

Figure 5.9 Concentrations of THMs of different treatment stages at various ozone

doses. (Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 0C) ...................................... 154

xviii

 

Figure 5.10 Concentrations of HAAs of different treatment stages at various ozone
doses. (Operating conditions: water flow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 0C) .................................... 154

Figure 5.11 Correlation between the THMs and HAAs of Lake Lansing water treated
by the ozonation/membrane/bioﬁ1tration system. (Correlation is
statistically signiﬁcant at 95% CI.) ......................................... 155

Figure 5.12 Correlation between THMs and the concentration of DOC of Lake
Lansing water treated by the ozonation/membrane/bioﬁltration system.
(Correlation is statistically signiﬁcant at 95% CI.) ....................... 156

Figure 5.13 Correlation between THMs and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% Cl.) ...................................... 156

Figure 5.14 Correlation between THMs and the concentration of humic substances of
Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% CI.) ....................... 157

Figure 5.15 Correlation between HAAs and the concentration of DOC of Lake
Lansing water treated by the ozonation/membrane/bioﬁltration system.
(Correlation is statistically signiﬁcant at 95% CI.) ....................... 157

Figure 5.16 Correlation between HAAs and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.) ...................................... 158

Figure 5.17 Correlation between HAAs and the concentration of humic substances of
Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% CI.) ....................... 158

Figure 5.18 Concentrations of aldehydes of different treatment stages at various water
treatment temperatures. (Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/m3) .................... 162

Figure 5.19 Concentrations of ketoacids of different treatment stages at various water
treatment temperatures. (Operating conditions: water flow rate, 2.75
L/min; gas flow rate, 100 mL/min; 0; dose, 2.5 g/m3) .................... 162

Figure 5.20 Concentrations of aldehydes of different treatment stages at various gas
ﬂow rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose,
2.5 g/m3; water temperature, 20 °C) ......................................... 163

Figure 5.21 Concentrations of ketoacids of different treatment stages at various gas

ﬂow rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose,
2.5 g/m3; water temperature, 20 0C) ......................................... 164

xix

Figure 5.22 Concentrations of aldehydes of different treatment stages at various ozone
doses. (Operating conditions: water ﬂow rate, 2.75 L/min; gas flow rate,
100 mL/min; water temperature, 20 0C) .................................... 165

Figure 5.23 Concentrations of aldehydes of different treatment stages at various ozone
doses. (Operating conditions: water flow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C) .................................... 166

Figure 5.24 Correlations between the aldehydes and ketoacids of Lake Lansing water
treated by the ozonation/membrane/bioﬁltration system. (Line 1
represents the correlation during the earlier stage of ozonation containing
F RW and P1 data; Line 2 represents the correlation during the later stage
of ozonation containing P1, P2, and WT data. Correlation is statistically
signiﬁcant at 95% CI. for Line 1 only.) .................................... 168

Figure 5.25 Correlation between aldehydes and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.) ...................................... 168

Figure 5.26 Correlation between ketoacids and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.) ...................................... 169

Figure 5.27 Correlation between aldehydes and the concentration of humic substances
of Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% CI.) ....................... 169

Figure 5.28 Correlation between ketoacids and the concentration of humic substances
of Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% Cl.) ....................... 170

Figure 5.29 Correlation between aldehydes and the concentration of non-humic
substances of Lake Lansing water treated by the ozonation/membrane
system. (Correlation is statistically signiﬁcant at 95% C1.) ............. 170

Figure 5.30 Correlation between ketoacids and the concentration of non-humic
substances of Lake Lansing water treated by the ozonation/membrane
system. (Correlation is statistically signiﬁcant at 95% CI) ............. 171

Figure 6.1 Concentrations of formaldehyde of different treatment stages at various
water treatment temperatures. (Operating condition: water ﬂowrate, 2.75
L/min; gas flowrate, 100 mL/min; O3 dose, 2.5 g/m3) ....................... 184

Figure 6.2 Concentrations of formaldehyde of different treatment stages at various gas
flowrates. (Operating condition: water ﬂowrate, 2.75 L/min; O3 dose, 2.5
g/m3; water temperature, 20 0C) ................................................ 184

Figure 6.3 Concentrations of formaldehyde of different treatment stages at various

ozone doses. (Operating condition: water ﬂowrate, 2.75 L/min; gas
ﬂowrate, 100 mL/min; water temperature, 20 °C) ........................... 185

XX

Figure 6.4 Concentrations of glyoxal of different treatment stages at various water
treatment temperatures. (Operating condition: water flowrate, 2.75 L/min;
gas ﬂowrate, 100 mL/min; O3 dose, 2.5 g/m3) ............................... 185

Figure 6.5 Concentrations of glyoxal of different treatment stages at various gas ﬂow
rates. (Operating condition: water flowrate, 2.75 L/min; O3 dose, 2.5 g/m3;
water temperature, 20 °C) ...................................................... 186

Figure 6.6 Concentrations of glyoxal of different treatment stages at various ozone
doses. (Operating condition: water flowrate, 2.75 L/min; gas ﬂowrate, 100
mL/min; water temperature, 20 0C) ........................................... 186

Figure 6.7 Concentrations of methyl glyoxal of different treatment stages at various
water treatment temperatures. (Operating condition: water ﬂowrate, 2.75
L/min; gas ﬂowrate, 100 mL/min; O3 dose, 2.5 g/m3) ....................... 187

Figure 6.8 Concentrations of methyl glyoxal of different treatment stages at various
gas ﬂowrates. (Operating condition: water ﬂowrate, 2.75 L/min; 03 dose,
2.5 g/m3; water temperature, 20 °C) ............................................ 187

Figure 6.9 Concentrations of methyl glyoxal of different treatment stages at various
ozone doses. (Operating condition: water ﬂowrate, 2.75 L/min; gas
ﬂowrate, 100 mL/min; water temperature, 20 °C) ........................... 188

Figure 6.10 Correlation between AOC and the sum of aldehydes, 2—butanone, and
ketoacids of Lake Lansing water treated by the ozonation/membrane and
bioﬁltration system. (Correlation is signiﬁcant at the 0.01 level.). 191

Figure 6.11 Correlation between Pseudomonasﬂuorescens strain P17 and the sum of
aldehydes, 2-butanone, and ketoacids of Lake Lansing water treated by
the ozonation/membrane and bioﬁltration system. (Correlation is
signiﬁcant at the 0.01 level.) ................................................. 191

Figure 6.12 Correlation between Spirilium strain NOX and the sum of aldehydes, 2-
butanone, and ketoacids of Lake Lansing water treated by the
ozonation/membrane and bioﬁltration system. (Correlation is signiﬁcant
at the 0.01 level.) ............................................................... 192

xxi

DBPs

NOM

UV-254

AOC

BDOC

DOC

THMs

HAAs

D/DBPR

TOC

HS

NHS

UF

MWCO

SDS THMS

SDS HAAs

SYM

HEM

ABBREVIATIONS

Disinfection by-products

Natural organic matter

Ultraviolet absorbance measured at 254 nm
Assimilable organic carbon

Biodegradable organic carbon

Dissolved organic carbon

Trihalomethanes

Haloacetic acids

Disinfectants/Disinfection By-Products Rule
Total organic carbon

Humic substances

Non-humic substances

Ultraﬁltration

Molecular weight cutoff

Simulated distribution system THMs
Simulated distribution system HAAs
Simple Y inline mixer

High efﬁciency inline mixer

xxii

CHAPTER 1

INTRODUCTION

1.1 SIGNIFICANCE

The production of disinfection by-products (DBPs) during the disinfection
process in drinking water treatment has become a very important issue since some
DBPs, such as trihalomethanes (THMs), have been found to have carcinogenic
properties (Kool et a1. 1985; Morris et a1. 1992; Mughal 1992). Because of their
perceived health risk, the United State Environmental Protection Agency (USEPA)
promulgated the Stage 1 of the Disinfectants/Disinfection By-Products Rule
(DfDBPR) in December of 1998 and issued the ﬁnal rule on January 16, 2001. The
maximum contaminant level (MCL) for total THM (TTHM) was lowered from 100
pg/L to 80 ug/L as a running annual average (RAA-MCLS) for the sum of ﬁve
haloacetic acids (HAAS) and for the bromate were set at 60 pg/L and 10 ug/L,
respectively (Pontius 2000; Pontius 2001; Pontius 2002; USEPA 1998). The draft
version of the Stage 2 DBPR, which was proposed on October 17, 2001, further
requests that the locational RAA (LRAA), which means the annual average at each
monitoring site, of TTHM/HAAS should be below 120/100 ug/L in Phase 1 (three
years for all systems after rule promulgation) and below 80/60 ug/L in Phase 2 (six
years for large and medium systems, eight and a half years for small systems required
to do Cryptosporidz’um monitoring, and seven and a half years for all other small
systems after rule promulgation) (Pontius 2001; Pontius 2002; Pontius 2003).
Therefore, the development of novel methods to effectively remove DBP precursors
in order to meet the more stringent regulations concerning DBPs is a major need of

the water treatment industry.

In response to the new D/DBP rule, the number of ozonation installations
continues to grow each year in the United States. In 1997, there were about 200
drinking water treatment plants using ozone (Naude 1997). In 2001, the number
increases to about 350 (Ozone News 2001). Ozone was ﬁrst applied in the drinking
water treatment process as a disinfectant in 1893 at Oudshoom, Netherlands. It is an
effective alternative disinfectant and a more powerful oxidant than chlorine, which is
now widely used as a disinfectant in the United States. Ozone is capable of
decreasing the numbers of microorganisms (Lee and Deininger 2000) and the
concentration of DBP precursors (Cipparone et a1. 1997; Yavich 1998). It also reacts
with organic substances and increases their biodegradability (Takeuchi et a1. 1997).
Ozonation has also been found to reduce the concentrations of DBPs in ﬁnished
water. Further reductions in DBPs concentrations could be achieved if ozonation is
followed by biological treatment because of the conversion of total organic carbon
(TOC) to more easily biodegradable organic carbon (BDOC) during ozonation
(Shukairy and Summers 1992). However, this conversion is limited because of the
low solubility of ozone in water and the slow reactivity of TOC with ozone, requiring
long contact time.

Another effective approach to remove DBP precursors and pathogens is the
application of membrane process. It is considered as one of the best available
technologies (BAT) for meeting the Stage 2 D/DBP requirements (Arora et a1. 1997).
During the last decade, better membranes have been developed and the
characterization of membrane surfaces and causes of membrane fouling are better
understood (Carroll et al. 2000; Clark et a1. 1998; Lin et al. 2001). Some membrane
processes, such as nanoﬁltration (NF), ultraﬁltration (UF), and microﬁltration (MF),

use pressure as the driving force to maintain constant flux across the membrane and

produce reliable ﬁnished water. These membrane ﬁltrations provide physical barriers
to turbidity, pathogens, and natural organic matter (NOM), which are rejected by
membranes and become the residuals. This technology is suitable for water treatment
plants that need to be upgraded to comply with the stringent drinking water standards
or for those to serve small communities (Cleveland 1999; Taylor and Wiesner 1999).

The objective of this research project was to study the feasibility of combined
ozonation and membrane technology for controlling DBP precursors in drinking water
and improving the quality of the treated water. Bioﬁltration following the
ozonation/membrane system was used to investigate the biodegradability of water
treated by the proposed system and evaluate the removal efﬁciency of DBP precursors
and biodegradable organic matter (BOM) by biotreatment. Several potential
advantages of ozonation/membrane/bioﬁltration treatment are:

1. The mixing of gas phase ozone and the water stream within the membrane is
expected to increase the ozone solubility and concentration in water, enhance
the ozone mass transfer, minimize the required ozone dosage, enhance organic
matter broken down by ozonation, and maintain stable permeate ﬂux of
membrane.

2. The ozonation/membrane process is expected to decrease the formation of
chlorinated DBPs in the ﬁnished water, inactivate the pathogens in the
permeate of membrane and residual stream, reduce membrane fouling,
increase the efﬁciency of NOM destruction by ozone, and enhance the
formation of BOM which can be readily removed by the subsequent
biotreatment of the proposed system.

3. The longer contact time between ozone and water in the proposed system than

that in conventional ozonation is expected to convert the NOM to a greater

 

 

 

percentage of ozonation by-products, which are more hydrophilic, polar, and
readily biodegraded. These ozonation by-products are good substrates for
microorganisms and can be removed by the following bioﬁltration process.
Therefore, the concentration of dissolved organic carbon (DOC) can be
reduced, which lessens the chlorine demand and the DBP level in the treated
water.

If the proposed ozonation/membrane system is successfully developed, it will

help drinking water treatment plants upgrade or retroﬁt their conventional systems to

comply with the D/DBPR.

1.2 OBJECTIVES

1)

2)

3)

4)

5)

Five particular objectives that were investigated are,

To assess the effect of ozonation on the properties (e.g. humic substances,
non-humic substances, biodegradable dissolved organic carbon, assimilable
organic carbon) of NOM in the proposed ozonation/membrane process.

To investigative the effect of ozonation/membrane system and biotreatment on
the formation of regulated DBPs and ozonation by-products at various
operational conditions.

To investigate the effect of control parameters (e. g. flow rate, ozone dose,
temperature) on the rate of ozone mass transfer in the proposed system.

To develop a mathematical model for ozone mass transfer that can be used to
characterize the rate of ozone mass transfer in the proposed system as a
function of operating conditions.

To establish the relationships between the properties of NOM and DBPs, and

between assimilable organic carbon (AOC) and ozonation by-products.

1.3 HYPOTHESES

1)

2)

3)

4)

5)

6)

7)

8)

The following hypotheses were tested in this study:

The mass transfer of ozone from the gaseous phase into the liquid phase can
be enhanced by combining two treatment processes, ozonation and membrane
ﬁltration, into one single unit. The high ozone mass transfer rate achieved by
using the proposed system will minimize the applied ozone dosage required to
treat drinking water to an acceptable quality.

The coefﬁcient of ozone mass transfer will increase as the water ﬂowrate, gas
flowrate, ozone dose, or water temperature increases.

The injection of ozone and water mixture into the membrane module will
increase the permeate flux of membrane and maintain a stable permeate ﬂux
during operation. The membrane-fouling problem will be reduced as ozone
bubbles scrub the surface of membrane.

The long hydraulic retention time of the proposed system is expected to
transform humic substances (HS) into non-humic substances (NHS), reduce
levels of TOC and ultraviolet absorption measured at 254 nm (UV-254), and
increase the biodegradability of treated water.

The ozonation/membrane system will effectively decrease the formation of
DBPs compared to that obtained with conventional ozonation.

Biotreatment following the proposed system will result in the improved
removal of DBP precursors and the reduction of ozonation by-products, TOC,
and AOC.

Ozonation by-products can be decomposed by ozonation and also degraded by
microorganisms.

The ultraﬁltration membrane used in this study can not effectively retain AOC.

9) There exist correlations between HS and UV-254, TOC and DBPs, and

ozonation by-products and AOC.

1.4 REFERENCES

Arora, H., LeChevallier, M. W., and Dixon, K. L. (1997). "DBP occurrence survey."
Journal American Water Works Association, 89(6), 60-68.

Carroll, T., King, 8., Gray, S. R., Bolto, B. A., and Booker, N. A. (2000). "The
fouling of microﬁltration membranes by NOM after coagulation treatment."
Water Research, 34(11), 2861-2868.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

Clark, M. M., Allgeier, 8., Amy, G., Chellam, S., DiGiano, F., Elimelech, M.,
Freeman, 8., Jacangelo, J ., Jones, K., Laine, J. M., Lozier, J., Marinas, B.,
Riley, R., Taylor, J ., Thompson, M., Vickers, J., Wiesner, M., and Zander, A.
(1998). "Committee report: Membrane processes." Journal American Water
Works Association, 90(6), 91-105.

Cleveland, C. T. (1999). "Big advantages in membrane ﬁltration." Journal American
Water Works Association, 91(6), 10-10.

Kool, H. J., van Kreijl, C. F., and Hrubec, J. (1985). "Mutagenic and Carcinogenic
Properties of Drinking Water." Water Chlorination: Chemistry, Environmental
Impacts, and Health Effects, R. L. Jolley, W. A. Brungs, and R. B. Cumming,
eds., Lewis Pub1., Chelsea, Mich.

Lee, J ., and Deininger, R. A. (2000). "Survival of bacteria after ozonation." Ozone-
Science & Engineering, 22(1), 65-75.

Lin, C. F., Liu, S. H., and Hao, O. J. (2001). "Effect of functional groups of humic
substances on UF performance." Water Research, 35(10), 2395-2402.

Morris, R. D., Audet, A. M., Angelillo, I. F., Chalmers, T. C., and Mosteller, F.
(1992). "Chlorination, Chlorination by-Products, and Cancer - a
Metaanalysis." American Journal of Public Health, 82(7), 955-963.

Mughal, F. H. (1992). "Chlorination of Drinking Water and Cancer: A Review." Jour.
Envir. Pathol, T oxicol. & Oncol, 11(5,6), 287-292.

Naude, A. (1997). "Finding Choices in Disinfection." Chemical Market Reporter,
252(15), FR5-FR6.

Ozone News. (2001). "Regulations/Cryptosporidium." B. L. Loeb, ed., lntemational
Ozone Association, Cincinnati, Ohio, 6.

Pontius, F. W. (2000). "Regulations in 2000 and beyond." Journal American Water
Works Association, 92(3), 40-+.

Pontius, F. W. (2001). "Regulatory update for 2001 and beyond." Journal American
Water Works Association, 93(2), 66-80.

Pontius, F. W. (2002). "Regulatory compliance planning to ensure water supply
safety." Journal American Water Works Association, 94(3), 52-+.

Pontius, F. W. (2003). "USEPA's drinking water regulations." Journal American
Water Works Association, 95(3), 57-+.

Shukairy, H. M., and Summers, R. S. (1992). "The Impact of Preozonation and

Biodegradation on Disinfection by-Product Formation." Water Research,
26(9), 1217-1227.

Takeuchi, Y., Mochidzuki, K., Matsunobu, N., Kojima, R., Motohashi, H., and
Yoshimoto, S. (1997). "Removal of organic substances from water by ozone
treatment followed by biological activated carbon treatment." Water Science
and Technology, 35(7), 171-178.

Taylor, J. S., and Wiesner, M. (1999). "Membranes." Water Quality and Treatment,
American Water Works Association, Ch.1 1.

USEPA. (1998). "Disinfectants and Disinfection By-products, Final Rule." Fed. Reg.,
69394.

Yavich, A. A. (1998). "The Use of Ozonation and Biological Fluidized Bed
Treatment for the Control of NOM in Drinking Water," Ph.D. Dissertation,
Michigan State University, East Lansing, MI.

CHAPTER 2

BACKGROUND

In this chapter, background literature relevant to this research is provided. The
important characteristics of NOM and the formation and inﬂuence of biodegradability
of organic matter in treated water are described in Section 2.1. The formation of
major DBPs is discussed in Section 2.2. The health risks caused by pathogens are
discussed in Section 2.3. Finally, four treatment technologies, ozonation, membrane
ﬁltration, ozonation/biotreatment, and ozonation/membrane, are reviewed and

discussed in Section 2.4.

2.1 ORGANIC MATTER IN DRINKING WATER

Organic matter, which occurs in the environment either as the result of natural
processes or human activities, is ubiquitous in aquatic systems. Typical
concentrations of NOM in aquatic systems are reported to range from 0.5 to 100 mg/L
as organic carbon (Frimmel 1998).

The constituents of aquatic organic matter include humic materials from plants
and algae, microorganisms and their metabolites, and aliphatic and aromatic
hydrocarbons. Some of this material can cause adverse health effects if they are not
removed during water treatment. Humic materials are recognized as the precursors of
DBPs; pathogenic microorganisms can, potentially, spread diseases; and other organic
substances present in treated water can consume disinfectant, thereby requiring higher
disinfectant dosages to achieve regulatory criteria, and possibly increasing the

pathogenic bacteria regrowth potential in the distribution system.

2.1.1 Natural Organic Matter

NOM is made up of a heterogeneous mixture of organic compounds. Diverse
types of NOM, which originate from various environmental systems, such as forests,
marshes, and swamps, are presence in surface waters (Krasner et al. 1996a). NOM
that originates from surface waters contains low fractions of phenolic and aromatic
compounds and mainly comes from algal and bacterial biomass, which is mostly
aliphatic and totally devoid of lignins (Goel et a1. 1995). The low molecular weight,
relatively soluble fulvic acids make up the signiﬁcant fraction of this material. In
contrast, NOM originating from soils is derived from terrestrial vegetation, which has
high aromatic constituents (Goel et a1. 1995).

NOM in drinking water supplies poses signiﬁcant concerns to the water utility
and may signiﬁcantly inﬂuence many aspects of water treatment. Conventional water
treatment processes cannot effectively remove NOM. The presence of NOM can
decrease the effectiveness of oxidants or disinfectants, and may cause the formation
of DBPs due to its reactivity with disinfectants (e.g., see Glaze et al. 1993; Krasner et
al. 1996b; Owen et a1. 1995; Rencken 1994). It also causes bacterial regrowth in the
distribution system, fouling of ﬁltration membranes, and reduces granular activated
carbon (GAC) adsorber bed life (Bouwer and Crowe 1988; Krasner et al. 1996a).
Only a small portion of the constituents of NOM are biodegradable since, by
deﬁnition, NOM is the recalcitrant fraction of organic material remaining after natural
decomposition (Goel et a1. 1995).

Several analytical techniques, such as '3 C magnetic nuclear resonance (Preston
1987; Wilson 1987), transmission electron microscopy (Leppard 1992), and NOM
fractionation on XAD resins (Leenheer 1981; Martin-Mousset et a1. 1997), have been

applied to characterize aquatic NOM. NOM is also typically quantiﬁed in terms of

TOC. Although NOM is heterogeneous in nature, an easily measured parameter,
speciﬁc ultraviolet absorbance (SUVA) at 254nm, was found to be a powerful
indicator for its reactivity toward oxidants (Westerhoff et al. 1999). A paucity of
information, however, exists describing the detailed characterization of NOM, which
includes determining its chemical composition and reactivity in a broad range of

waters (Owen et al. 1995).

2.1.2 Humic Substances

Dissolved organic carbon (DOC), the organic matter in water that passes
through a 0.45 pm membrane ﬁlter, can be divided into humic and nonhumic
fractions. These two fractions, also called humic substances and nonhumic
substances, are commonly classiﬁed on the basis of their water solubility at acidic or
basic pH values (Malcolm 1985).

Humic substances, which represent a complex mixture of molecules of various
sizes and shapes, are more hydrophobic in character and comprise humin and humic
and fulvic acids. Due to their heterogeneous nature and composition, no single
structural formula can be given to them.

Humic substances are the primary precursors of chlorination byproducts,
such as THMs (Collins et al. 1986; lchihashi et al. 1999; Manahan 1993; Reckhow et
al. 1990). Organic and inorganic pollutants, many of which are insoluble in water,
can sorb to humic substances and therefore, be transported in the water with the
humic substances (Collins et al. 1986; Stevenson 1985; Thurman 1985). They are
resistant to both microbiological and chemical breakdown. Malcolm (1985) found

that humic substances comprise approximately 50% of the DOC in uncolored streams

10

in the United States. In his study, fulvic acid accounted for approximately 90% of the
humic substances and only 10% or less was humic acid.

The composition of humic substances in surface water is different from that in
groundwater. Humic substances from surface water usually contain an average of
52% carbon and 42% oxygen, whereas, humic substances from groundwater contain
greater than 60% carbon and less than 30% oxygen (Thurman and Malcolm 1981).
The color per unit of carbon of humic substances from groundwater is also
considerably less than that from surface water (Thurman 1985).

Aquatic humic substances contain aromatic structures with large numbers of
carboxyl groups, some phenolic groups, alcohol OH groups, methoxyl groups,
ketones, and aldehydes (Liao et a1. 1982; Reckhow et al. 1990). The average
molecular weight of aquatic humic acids usually ranges from 2000 to 3000 daltons,
and that of aquatic fUIVlC acids is from 800 to 1000 daltons (MacCarthy and Suffet
1989; Wershaw and Aiken 1985a). Thurman et al. (1982) pointed out that the average
molecular weight for humic substances from surface water is 100 to 2000 daltons. In
contrast, the molecular weight of humic acids from soil is as large as several hundred
thousand daltons.

The characteristics of the functional groups on humic and fulvic acids, which
inﬂuence the solubility of humic substances (Perdue 1985), provide insight into the
applicability of various water treatment processes for the removal of NOM. Humic
substances with the highest carboxylic acidity and thus highest charge density are
generally more difﬁcult to remove by conventional treatment, direct ﬁltration, or
softening processes (Collins et a1. 1986). For example, fulvic acids have higher
carboxylic and phenolic acidity, therefore, have a higher charge density than humic

acids, and are more difﬁcult to coagulate by charge neutralization. However, the

11

humic fraction of DOC is more easily removed by adsorption or coagulation.
Granular activated carbon (GAC) removes humic and higher apparent molecular
weight (AMW) NOM. It also provides some reduction in the levels of nonhumic and
lower AWM DOC. Coagulation reduces SUVA and shows the preferential removal
of the humic fraction of NOM and higher-molecular-weight NOM (Owen et a1. 1995).
The nonhumic substances comparing with the humic substances are more
hydrophilic in character and are comprised of hydrophilic acids, proteins, amino
acids, and carbohydrates (Owen et a1. 1995). The nonhumic fraction is of concern in
water treatment because it forms a signiﬁcant percentage of the regulated or
potentially regulated DBPs. In addition, the nonhumic fraction of DOC may
contribute to a higher fraction of the BDOC, which may cause of bacteria regrowth in
the distribution systems (Mogren et a1. 1990). Therefore, decreasing the nonhumic
fraction concentrations in drinking water to levels that are as low as technically and

economically feasible is also an important issue (Leenheer 1985).

2.1.3 Molecular Weight Distribution

Molecular weight is a valuable tool to characterize complex organic
compounds, such as humic substances (Wershaw and Aiken 1985b). Some
researchers have shown that molecular weight characterization of the DOC provides a
means for identifying potential treatment strategies (Amy et a1. 1988; Chadik and
Amy 1987).

Low molecular weight and nonhumic material are difﬁcult to remove by either
coagulation or adsorption (Owen et a1. 1995). Collins and colleagues ( 1986) found
that conventional treatment processes were rather ineffective in removing DOC

having a molecular weight of less than 500 daltons. In contrast, higher molecular

12

weight material proved to be more amenable to removal than lower molecular weight
material. Owen et al. (1995) and Amy et al. (1992) found that chemical coagulation
could effectively remove higher molecular weight material. Water sources with
material of medium molecular weight (e.g., 1000 to 5000 daltons) can also be
effectively adsorbed by activated carbon (Amy et al. 1992).

Molecular weight distributions may shift during the treatment processes.
Yavich (1998), Owen et al. (1995), and Amy et al. (1988) observed that ozonation did
not result in NOM destruction, but converted material with higher molecular weight
to lower molecular weight. This shift of molecular weight distribution not only
changes the properties of NOM (Owen et al. 1995) but also reduces the
trihalomethane formation potential (THMFP) (Amy et al. 1988; Amy et al. 1992).

Two main methods, gel permeation Chromatography (GPC) and ultraﬁltration
(UF), are used to estimate the molecular weight. These two techniques separate the
humic substances based on molecular size rather than molecular weight. Thus, the
results are usually expressed as apparent molecular weight (AMW) distribution. Kim
and colleagues (1989) compared the AMW values for several aquatic samples all
measured by the GPC and UF methods and found that the values assigned by UF were
higher than those assigned by GPC. Similar observations were reported by Thurman

et al. (1982).

2.1.4 Biodegradable Dissolved Organic Carbon

Biodegradable dissolved organic carbon (BDOC) is a fraction of DOC that can
be mineralized by heterotrophic microorganisms. Some researchers (Bablon et al.
1991a; Van der kooij et a1. 1989; Volk et al. 1993) found that the use of ozonation as

the disinfectant increases the regrowth potential of bacteria of a water by converting

13

part of the refractory carbon into BDOC. These unwanted bacteria have the potential
to spread disease, to cause unpleasant tastes, order, and color in the distribution
system, and to increase pipe corrosion problems (Korshin et al. 1996; Levy et al.
1986). Removing BDOC from the ﬁnished water has several advantages. It reduces
the extent of bacteria regrowth in the distribution system. It reduces the fraction of
the NOM having a high disinfectant reactivity, hence reducing the chlorine demand of
treated water and increasing the likelihood that a sufﬁciently high chlorine residual
can be maintained in the distribution system. Removing BDOC also results in the
elimination of some precursors of DBPs, thereby lowering the concentration of DBPs
in the treated water (Bablon et al. 1991a; Cipparone et al. 1997; Shukairy and
Summers 1992).

Servais and colleagues (1993) investigated the threshold value of BDOC for
which there is no BDOC consumption within the distribution system. They
concluded that if the BDOC value is less than 0.15 mg C/L, water can be considered
as biologically stable in the absence of free chlorine, which means the bacterial
growth in the distribution system is very limited. However, Rittmann (1995)
indicated that BDOC could cause problems during distribution even its concentration
is as low as 0.02 mg C/L.

Siddiqui and colleagues (1997) found that the conversion of DOC to BDOC
reaches a maximum at an applied O3/DOC ratio of 1:1 (mg/mg) for Silver Lake water
containing DOC levels ranging from 3 to 6 mg/L. This result is consistent with other
researchers’ results (Cipparone et al. 1997; Volk et a1. 1993). Moreover, the DOC
concentrations were not signiﬁcantly reduced when the applied ozone dose was
increased beyond 1.0 mg O3/mg DOC ratio (Paode et al. 1997; Siddiqui et al. 1997;

Volk et al. 1993).

14

The measurement of BDOC has been used mainly to indicate the quality of
raw and ﬁnished waters and evaluate the performance of biotreatment processes (e.g.,
biological ﬁltration) in water treatment plants. BDOC measurement assesses the
organic carbon that can be potentially biologically oxidized during the biotreatment
processes (Huck 1990). The concentrations of BDOC cannot be measured by simple
chemical methods, because a great variety of compounds comprise BDOC and many
of them are difﬁcult to quantify and cumbersome to isolate (Siddiqui et a1. 1997). As
such, several methods for measuring biodegradable organic matter in water have been
introduced (Huck 1990). There are two main and conceptually different approaches
to measuring it. One approach assesses easily assimilable organic carbon (AOC),
which represents the regrowth potential of heterotrophic bacteria in water and was
developed by van der Kooij, Visser, and Hijnen (1982). The method is based on
measuring the maximum growth level (colony forming units, CFU) of a selected pure
culture or cultures of bacteria, inoculated into autoclaved samples of drinking water
(Van der Kooij 1995; Van der kooij et a1. 1982). For the AOC determination, two
selected microorganisms, Pesudomonasﬂuorescens strain P17 and Spirillum strain
NOX, are both able to reproduce at low substrate concentrations and utilize different
groups of compounds (Van der Kooij 1995). The Pesudomonasﬂuorescens strain
P17 mainly utilizes amino acids, carbohydrates, and aromatic acids. The Spirillum
strain NOX mainly utilizes carboxylic acids.

The other approach was developed by Servais, Billen, and Hascoét (1987).
This method is based either on the reduction of DOC during incubation or on an
estimate of the ﬂux of organic matter utilized by bacteria. The inoculum bacteria
used in this method is obtained from the natural environment from which the water

sample originated. Incubation occurs in the dark at room temperature for a period of

15

10 to 30 days. Joret and colleagues (1988; 1991) using an aerated bioﬁlm on sand
particles to measure BDOC, reduced the incubation time to 3 to 5 days. The principal
BDOC method that is currently widely used is the recirculating acclimated-sand
column technique (Mogren et a1. 1990). A glass column partially ﬁlled with
biologically active sand serves as the main body of the reactor. The sample is
recirculated through the column, and DOC is measured daily until the concentration
of DOC no longer changes with time.

Some researchers (Siddiqui et al. 1997) found that a positive correlation was
obtained between BDOC and AOC; both parameters had a tendency to plateau at an
applied O3/DOC dosage of 2 mg O3/mg DOC. However, Chamock and KjOnnG
(2000) found that the correlations between AOC and BDOC for both raw and
drinking water were not signiﬁcant. The authors suggest that the results may indicate

that the AOC and BDOC analyses target different fractions of BOM.

2.2 DISINFECTION BY-PRODUCTS

In 1974, Rook ﬁrst reported that THMs were formed when chlorine reacted
with naturally occurring humic substances in water treatment plants and water
distribution systems. Subsequently, several researchers also reported the production
of THMs and other halogenated organics such as HAAs, haloacetronitriles,
haloalcohols, and haloketones during the chlorination process (e.g., see Christman et
a1. 1983; lchihashi et a1. 1999; Richardson et al. 1999; Singer 1994; Stevenson 1994).
Generally, the DBPs can be divided into two groups: organic DBPs and inorganic
DBPs. Organic DBPs include aldehydes, ketones, ketoaldehydes, carboxylic acids,

aldo acids, keto acids, hydroxyl acids, alcohols, esters, and alkanes (Anderson et a1.

16

1985; Haag and Hoigné 1983). Reported inorganic DBPs include bromate,

hypobromite, and hydrogen peroxide (Haag and Hoigné 1983).

A. Chlorinated DBPs

THMs and HAAs are considered to be the major chlorination byproducts in
water treatment (Oxenford 1996). Four chemicals, chloroform,
dichlorobromomethane, dibromochloromethane, and bromofonn, are well known as
the predominant THMs formed during chlorination (McGuire and Meadow 1988). A
number of factors that affect the formation of THMs are the nature of aquatic humic
substances, chlorine dose, reaction time, bromide ion concentration, preozonation,
temperature, and pH (Trussell and Umphres 1978).

Five major HAAs includ monochloroacetic acid (MCAA), dichloroacetic acid
(DCAA), trichloroacetic acid (TCAA), monobromoacetic acid (MBAA), and
dibromoacetic acid (DBAA). Their formation can be appreciable if drinking water is
chlorinated under slightly acidic pH values, distributed under neutral distribution
system pH values, and containing low bromide concentrations (Singer et a1. 1995).

The other DBPs that have been observed during the chlorination process of
drinking waters are organic halides, haloacetonitriles, haloketones, chloropicrin,

cyanogens chloride, chlorite, and chlorate (Andrews and Ferguson 1996).

B. Ozonation DBPs

Some researchers found that both TH Ms and HAAs are also formed when
chlorine or chloramines were applied following the ozonation of waters containing
NOM (Reckhow and Singer 1990; Richardson et al. 1999; Uden and Miller 1983).

Richardson and colleagues (1999) found that the majority of the identiﬁed ozone

DBPs contained oxygen in their structures. No halogenated DBPs were observed
unless high natural levels of bromide (Br') existed in the raw water and chlorine or
chloramines were used as a secondary disinfectant.

Other reported ozonation byproducts are aldehydes (formaldehyde,
acetaldehyde, glyoxal, and methylglyoxal), ketones, glyoxylic, and pyruvic acids

(Paode et al. 1997; Weinberg and Glaze 1996).

C. Bromate

Bromate (BrO3') is a DBP that results when bromide containing waters are
ozonated. Reactions of ozone with bromide in aqueous media are shown in Figure 3.1
(Haag and Hoigné 1983). Hypobromous acid (BrOH) is formed when ozone,
chlorine, or chloramine oxidizes bromide ion during the water treatment. Then the
BrOH in turn reacts with DOC to form brominated DBPs (Haag and Hoigné 1983;
Rook 1974; Rook et al. 1978; Song et al. 1997). NOM, ozone dose, bromide
concentration, pH, and temperature are factors that inﬂuence bromate formation.

The concentration of bromide in water is important because it not only reacts
with ozone and forms bromate, but also affects the formation of other DBPs. In some
instances, bromide is passed conservatively through the treatment processes (e.g.,
coagulation, GAC, ion exchange) resulting in higher Br'-to-NOM ratios than are
present in bulk waters prior to treatment (Owen et al. 1995). The Br'-to-NOM ratios

affect the DBP speciation (Krasner et al. 1996b).

l8

k1=160M ‘S‘

,/

/ q // mucosa .‘

Br BrO‘

\ ;\“H CHBr
. \ M 7' 3
\\ ..z’

\ HBrO :
NH ‘L NH3Br
3
k2=330 M 's '

Figure 2.1 Reactions of ozone with bromide in aqueous media, M represents organic
matter. (Haag and Hoigné 1983)

D. Treatment

Chemical coagulation, a process that is already operational at many treatment
plants, is capable of removing both turbidity and THM precursors when the process is
optimized. Reckhow and Singer (1990) reported that alum coagulation effectively
removes signiﬁcant fractions of the DBP precursors. However, some researchers
pointed out that either bromide or brominated haloforms cannot be removed
effectively through the coagulation process (Chadik and Amy 1983; Owen et al.
1995)

Ozonation followed by biotreatment is able to remove ozonation DBPs
efﬁciently because of the high biodegradable characteristics of those DBPs. However,
Siddiqui and co-workers (1997) reported that a drawback of the ozonation of high

bromide containing waters is once the bromate formed ozonation-bioﬁltration

l9

treatment scheme cannot effectively remove it during bioﬁltration. The combination
of ozonation/biotreatment process will be discussed in Section 3.4.3.

The other promising technique for removing DBPs precursors is the
membrane process. Membranes are good barriers of turbidity causing material,
pathogens, and DBP precursors. DBP precursors removal of up to 99% can be
achieved depending on the characteristics of the applied membrane. A more detailed

investigation will be discussed in Section 2.4.2.

2.3 PATHOGEN S

Pathogens commonly exist in many types of water sources. They can enter
surface water sources through surface runoff from numerous sources including,
livestock grazing areas and feedlots, from sewage efﬂuent, or untreated or poorly
treated wastewater. Pathogens are also found in ground water sources, which are
naturally ﬁltered by the geological material. Septic tank efﬂuent, runoff from
agricultural areas, and wastewater are often the sources of contamination.

The Microbiological Contaminants Research Committee of the AWWA
Research Division determined that four bacterial, three viral, and three protozoan
microorganisms, and one set of bacterial toxins that are of greatest concern to the
water industry (LeChevallier et al. 1999a; LeChevallier et al. 1999b). These
microorganisms and toxins can cause disease, infecting the human lungs, stomach,
upper gastrointestinal tract, liver, kidney, and central nervous system. Typical
symptoms of infection are cough, fatigue, dehydration, weight loss, fever, diarrhea,
abdominal cramps, nausea, permanent nerve damage, headache, and muscle pain.
Whether an individual becomes illness depends on the number of organisms ingested,

the health status of the individual, and the resistance of the person to the organism or

20

toxin. Generally, children, the elderly, and other immuno-compromised or immuno-
deﬁcient individuals are at particular risk.

Drinking water treatment plants provide barriers between people and
pathogens to protect public health. Pathogens are removed or inactivated in different
treatment units. Nieminski and Ongerth (1995) compared the effectiveness of
removing Giardia and Cryptosporidium by conventional treatment and direct
ﬁltration. They found that both properly operated conventional treatment and direct
ﬁltration can achieve 3-log removal of Giardia cysts, but only about 2.5-log removal
of Cryptosporidium oocysts. If the performance of treatment changed and raw water
turbidity ﬂuctuated, a high variability in cyst concentration was observed in efﬂuent.

In the past decade, several waterborne outbreaks have drawn the attention of
water suppliers, public health professionals, and researchers. Solo-Gabriele and
Neumeister (1996) reviewed numerous outbreaks of Cryptosporidiosis. They found
those outbreaks had occurred in drinking water treatment plants that were meeting
existing regulatory requirements. Thus, they suggest that ﬁltration system of
conventional treatment process should be operated at optimum level exceeding
existing regulatory requirements, because chlorination process alone is incapable of
providing adequate protection of public health from Cryptosporidium outbreaks.
Also, Fox and Lytel (1996) reported that Cryptosporidium oocysts were able to pass
through one of two treatment plants in Milwaukee, Wisconsin, with other particles
during a high turbidity period in March and April 1993. This waterborne outbreak
caused illness in more than 400,000 people.

Coliforrn bacteria have been used as the indicator to access the
microbiological quality of drinking water for many decades. Craun and colleagues

found that coliform bacteria are adequate indicators for the potential existence of

21

bacteria and viruses in drinking water, but they are inadequate as an indicator for
waterborne protozoa. That is because oocysts are more resistant to water disinfectants
than coliform bacteria.

Several researchers (Frost et a1. 1996; Kramer et al. 1996; Moore et a1.
1994) discussed waterborne disease surveillance systems and point out that outbreaks
are probably underreported because of current reporting systems. Most current
surveillance programs are based on disease reporting by health care providers and
clinical laboratories. Because of the same symptoms of enteric infections, the
unawareness of infection, and insufﬁciently speciﬁc symptoms to identify pathogens,

outbreaks are probably inadequately investigated or unrecognized.

2.4 TREATMENT
2.4.1 Ozonation

Ozone has been used widely as an alternative disinfectant to chlorine,
chloramines, and chlorine dioxide. It is a powerful oxidant and preferentially oxidizes
electron-rich moieties, which contain carbon-carbon double bonds, and aromatic
alcohols (Bablon et al. 1991b). Ozone also acts as an electrophil reacting with
molecular sites with a strong electronic density, and as a nucleophil reacting with
molecular sites with electronic deﬁciencies (Bablon et al. 1991b). Due to the
reactivity of the most commonly occurring contaminants and the relatively short
contact time in ozonation contactors, complete oxidation of these chemicals rarely
happen. Therefore, partially oxidized organic compounds, such as aldehydes, ketones,
organic acids, and alcohols, are produced during the ozonation of drinking waters.

Ozone is partially soluble in water. Its solubility decreases with an increase in

water temperature and increases with an increase in partial pressure of ozone (or,

22

correspondingly, in the ozone concentration in the gas stream) according to Henry’s
Law. The higher the concentration or partial pressure of ozone in the gas stream, the
higher the ozone solubility in the water and the higher mass transfer efﬁciencies of
ozone. One other factor that affects ozone mass transfer is the existence of reactive
material such as NOM in the water. If the concentrations of reactive species are high,
then the mass transfer efﬁciencies are increased and the demand for ozone will be
greater.

Ozonation reactions in aqueous solution involve either direct reactions with
molecular ozone or indirect ones with the hydroxyl radical (Bablon et al. 1991b;
Hoigné and Bader 1975; Peleg 1976). During the ozonation of water, a portion of the
ozone reacts directly with DOC and the remaining amount of ozone decomposes to
form the hydroxyl (HO') radical. The decomposition reaction is catalyzed by
hydroxide ions and other dissolved compounds such as NOM. OH radicals can react
with DOC and accelerate the decomposition of ozone (Bablon et al. 1991b; Hoigné
and Bader 1975; Staehelin and Hoigné 1985).

Factors that favor the molecular ozone reaction pathway are pH values below
6, the absence of redox metal ions, such as F e(II) and Fe(III), and the presence of
compounds having unsaturated carbon-carbon bonds. Studies show that reactions
with molecular ozone are comparatively slow and highly selective compared with
reactions involving the hydroxyl radical, which are faster and relatively nonselective
(Haag and Yao 1992; Hoigné and Bader 1983). The net result of the molecular ozone
reactions is usually the formation of aldehydes, ketones, and carboxylic acids (Glaze
et a1. 1988; Maggiolo 1978; Miltner et al. 1992; Schechter and Singer 1995).

In contrast, ozone rapidly decomposes into hydroxyl radicals at pH values

above 8. When targeting compounds such as NOM, which react slowly with

23

molecular ozone, the hydroxyl radical is the preferred oxidant. The reaction of
organic compounds with hydroxyl radicals will produce organic free radicals that
ultimately result in the formation of aldehydes, ketones, alcohols, and carboxylic
acids. Some of the aldehydes including formaldehyde, acetaldehyde, glyoxal, and
methylglyoxal are of particular concern due to their mutagenicity and carcinogenicity
(Bull and McCabe 1984).

Molecular ozone and OH radical reactions, both of which occur during
ozonation, can result in the cleavage of larger molecules. This results in lower
molecular weight material and the formation of more polar compounds, which are
more hydrophilic (e.g., see Amy et a1. 1988; Amy et al. 1992; Flogstad and Odegaard
1985; Kaastrup and Halmo 1989; Koechling et al. 1996; Owen et al. 1995). A
decrease in the concentration of UV absorbing compounds of NOM was also
observed during ozonation (Amy et a1. 1988; Kaastrup and Halmo 1989; Koechling et
a1. 1996; Shukairy et al. 1994; Yavich 1998). Some researchers found that DOC
concentration was not affected at ozone-to —DOC ratios ranging between 0 and 2.5
mg O3/mg C (Owen et al. 1995; Shukairy et al. 1994). However, Cipparone et al.
(1997) observed increasing TOC removal with increasing ozone dosage
concentration. Takeuchi et al (1997) observed that only partial breakage of C-C
chains occurred during ozonation but that DOC did not signiﬁcantly decrease. Some
researchers observed that the biodegradability of ozonated water increased with
increasing the ozone dose (Siddiqui et a1. 1997; Somiya et a1. 1986). Takahashi et al
(1995) found the ratio of BOD5/COD increased from 0.02~0.03 to 0.25 as the ratio of
O3/TOC increased from 0 to 3.5.

Replacing chlorination with ozonation as the primary disinfectant can

signiﬁcantly reduce the formation of THMs and HAAs. This was observed when

24

using the chlorine or chloramine as a secondary disinfectant (Richardson et al. 1999).
In the presence of NOM, ozonation results in the formation of partial oxidation
compounds, which are less reactive with chlorine in forming THMs (Amy et al.
1988). The required ozone dosage and the formation of aldehydes and ketoacids
increased with increasing the NOM concentration (Najm and Krasner 1995). As the
ozone dosage increased, THMs and HAAs formation decreased (Cipparone et al.
1997). The ozonation of water containing humic substances results in the formation
of hydroxyl, carbonyl, and carboxyl groups, and aliphatic and alicyclic ketones
(Anderson et al. 1985; Glaze et a1. 1989). Gracia and colleagues (1996) obatined
similar results as they identiﬁed 110 different organic compounds resulting from
ozonation of humic substances.

One drawback of ozonation is the formation of bromate (BrO3') ion in waters
containing bromide (Br') ion (Siddiqui and Amy 1993). Either a molecular ozone or a
hydroxyl radical pathway can form bromate ion, which is mainly dependent on ozone
dosage, alkalinity, and pH of the water (Najm and Krasner 1995; Siddiqui et a1. 1995).
Ozone can oxidize the bromide ion to form hypobromous acid (HOBr), which in turn
reacts with NOM to form brominated DBPs (Richardson et al. 1999).

Factors that inﬂuence the bromate formation are initial bromide concentration,
the concentration of NOM, the presence of ammonia, the pH of water, the addition of
hydrogen peroxide, ozone dosage, temperature, and alkalinity (e.g., see Croue et al.
1996; Minear and Amy 1996; Siddiqui and Amy 1993; Song et al. 1997; Von Gunten
and Hoigné 1996). Increasing the initial bromide concentration and ozone dosage
increases the bromate concentration in treated water (Croue et al. 1996; Minear and
Amy 1996; Najm and Krasner 1995). Higher dosages of ozone cause signiﬁcant

production of bromate ion and shifts in the speciation from brominated THMs toward

25

chlorinated THMs (Miltner et a1. 1992; Speitel et al. 1993). The formation of bromate
ion also increased as the concentration of bicarbonate ion increased because of the
increase of ozone stability (Croue et a1. 1996). Other factors that favor bromate ion
formation are low OH radical scavengers, high temperature, high alkalinity and high
pH value (Siddiqui and Amy 1993; Song et a1. 1997; Symons et a1. 1994). Increasing
the pH value results in increasing the OBr'/HOBr ratio and promoting bromate
formation. Song et al. (1997) reported that pH value is a more important factor than
ozone dosage that affects bromate formation.

There have been some conﬂicting reports relating to the effect of DOC
concentration on bromate ion formation. Najm and Krasner (1995) reported that the
formation of bromate ion increased with the DOC concentration. However, Siddiqui
and colleagues (1995) found that DOC has been shown to decrease bromate ion

formation. Similar results were obtained by Song et al (1997).

2.4.2 Membrane

The membrane process has been considered to be an effective means of
removing turbidity, organics, microorganisms, and DBP precursors and to comply
with the more stringent regulatory controls. It offers several advantages, compared to
conventional treatment, such as i) providing high quality water, ii) minimizing
disinfectant demand, iii) being more compact, iv) easier control of operation and

maintenance, and v) less production of sludge (Nakatsuka et al. 1996).

A. Microﬁltration
Microﬁltration (MF) is a ﬁltration process that use porous membranes to

separate particles in water with pore diameters between 0.1 and 10 pm. Ultraﬁltration

26

(UF) uses ﬁner porous membranes, having average pore diameters between 10 and
1000 A, to separate microsolutes and water. While MF and UF membranes can
achieve effective removal of particles and microorganisms, they are not as effective in
removing color and dissolved constituents.

Wiesner et a1. (1991) used ceramic MF membranes with and without
coagulation as a pretreatment step for waters with moderate to high turbidities and
concentrations of TOC and THMF P. They reported that the removal of turbidity,
UV254, TOC and THMF P were 99%, 60-87%, 48-59% and 40-65%, respectively.
Signiﬁcant removal of TOC could be achieved by MP alone. Scanlan et al. (1997)
studied the effectiveness of MP and UP of a low turbidity, moderate color and TOC
water source. They found that UF was more effective at particle removal than MF,
with or without chemical addition; however, the operation times of MP were much
longer than those of UP. The concentrations of THMs and HAAs in treated water did

not consistently meet the Stage 2 D/DBP requirements.

B. Ultraﬁltration

Two studies (Fu et al. 1994; Taylor et al. 1987) showed that NOM can be
effectively rejected during ﬁltration by low and medium pressure membranes, such as
UP and NF. As NOM was removed by UP and NF, the chlorine demand of the
permeate was decreased (Escobar and Randall 1999; Jacangelo et al. 1995).

Chadik et a1. (1991) used several UF membranes with different nominal pore
size to separate the NOM into several AMW fractions. The “reactivity”, deﬁned as
the ratio of the THMF P to the nonpurgeable organic carbon (NPOC), was used to
estimate the THMF P of the NOM in each AMW fraction. They found that there was

no trend as to one fraction being more reactive than another and the THMF P was

27

found to be principally associated with the concentrations of NPOC. A substantial
portion of THMF P (up to 45%) was associated with the AMW fraction of less than
5000 daltons.

Jacangelo et al. (1995) evaluated the efﬁcacy of combined powdered activated
carbon (PAC) and UF for removing DBP precursors. They found that UF alone could
not remove DBP precursors effectively because of its high molecular weight cutoff
(MWCO). Less than 15% of the TOC and less than 10% of the THM and HAA
precursors were removed in their study. However, DBP precursor removal increased
with increasing PAC dosages in the combined PAC-UF treatment. Since bromide
was not removed by the combined treatment, increasing of the bromide-to-TOC ratio
was observed as TOC was removed by the PAC. They also reported that membrane
fouling was retarded at a 30 mg/L PAC dose, but further study of the effect of PAC on

membrane fouling is needed.

C. Nanoﬁltration

Nanoﬁltration (NF) is a low-pressure-driven membrane process. Compared
with reverse osmosis (RO), NF yields very good rejection rates of large molecules and
is much more energy efﬁcient than RO. Several studies have demonstrated large
removals of DBP precursors from ground and surface water using NF. Lainé et al.
(1993) reported that NF membranes were effective for DBP precursor removal but
ineffective for removing bromide, thereby resulting in increases in the bromide to
TOC ratio and the proportion of brominated to chlorinated THMs.

Siddiqui and colleagues (2000) evaluated NF, UF, and MF to determine the
rejection of DBP precursors from low turbidity surface waters in Colorado. They

found that the average rejections of DOC, THMFP, HAAF P, and CHF P (chloral

28

hydrate formation potential) of NF were all over 85%. MF was only moderately
effective in particle removals, with virtually no DBP precursors removal provided.

UF alone removed less than 30% of DOC. NF membranes with MWCO less than 300
daltons removed an average of 90% DOC. Only 40% reduction of DOC was
observed on permeates of NF with 500 daltons of MWCO. Bromide rejections were
among 40 and 80% in the study. Similar observations were reported by Allgeier and
Summers (1995).

Lozier and Carlson (1991) evaluated ultra-low pressure membrane processes,
deﬁned as typical transmembrane pressure requirements of 150 psi or less, to treat
three surface waters in the eastern US. with TOC and THMF P concentrations ranging
from 5 to over 50 mg/L and from 400 jig/L to greater than 2,000 pg/L, respectively.
They found that the NF (MWCO < 300 daltons) and the UP (MWCO = 3,000 daltons)
were capable of reducing THMFP to less than 60 pg/L and TOC levels by more than
90 percent.

NF needs pretreatment to reduce its cleaning frequency and lengthen its
operation time. Therefore, some utilities are considering NF as an “add-on” option to
their existing facilities for DBP precursor removal before ﬁnal disinfection. Chellam
et al. (1997a) treated water from the Occoquan Reservoir in Virginia with MF, UF,
and conventional treatment as pretreatment followed by NF. MF removed 11% and
24% of the THM and HAA precursors, respectively. NF achieved greater than 90%
rejection of TOC that resulted in greater than 95% rejection of THMs and HAAs
precursors. These high removals of TOC and DBP precursors were observed to be
independent of membrane operating conditions, suggesting that these compounds
were predominantly rejected by mechanical sieving. There was no statistical

difference between TOC concentrations in the MF and UP ﬁltrates.

29

Mulford et al. (1999) used similar systems that incorporated MF or UF before
NF to treat water containing 11mg/L TOC, 336 jug/L SDS THMFP, 227 ug/L SDS
HAAF P, and 24mg/L chlorine demand. High removal efﬁciencies were achieved and
the average concentrations of TOC, SDS THMFP, SDS HAAFP, and chlorine
demand of ﬁnished water were 0.4 mg/L, 35 pg/L, 28 rig/L, and 6.5 mg/L,
respectively. Chellam et al. (1997a) observed similar results and reported that surface
water treated by a dual membrane system (e.g., MF followed by NF) can meet current
and anticipated THMs and HAAs regulations. DiGiano and colleagues (1993)
suggested that pretreatment of the raw water before NF would be necessary in order to
remove THM precursors effectively if the concentrations of THMFP in the raw water
is higher than 100 jig/L.

Taylor et al. (1987) tested NF and R0 membranes ranging in MWCOs from
100 to 40,000 daltons with high-color groundwaters from Florida. They found that 3
RO membrane with MWCO of 100 daltons removed 98% of THMF P precursors. Its
removal efﬁciency was just a little bit higher than that of a NF membrane with
MWCO of 400 daltons, which removed 96% of THMFP precursors. But the RO
membrane required 60% greater pressure and produced a 50% lower ﬂux than NF.
Amy et al. (1990) observed only a 65 to 70% reduction in THMF P by using NF
membrane (MWCO < 500 daltons) for treating Colorado River water. The difference
in the reductions of DBP in these studies may be caused by the different

characteristics of NOM.

D. Membrane Fouling
One of the major problems in membrane processes is the decrease in permeate

ﬂux to far below its theoretical capacity, resulting from membrane fouling. Decreases

30

in ﬂux may occur as a result of concentration polarization, formation of gel layer on
the membrane surface, or adsorption or plugging of organic and inorganic materials in
the membrane pores. The fouling rates are inﬂuenced by the nature of the solutes,
their concentrations, membrane type and pore size distribution, water quality,
hydrodynamics and surface characteristics of the membrane (Field 1996). Membrane
fouling lowers the economic efﬁciency of membrane treatment by reducing the
quantity of treated water, produced shortening membrane life, and increasing the
frequency of membrane cleaning.

Organic matter is often found to be a primary source of ﬂux decline due to
fouling of membrane systems (Nilson and DiGiano 1996; Ravindran et al. 1993).
Fouling may be caused by the interaction of NOM with membrane surface or
incorporation into its porous support. DiGiano et al. (1994) found the fouling of UP
was attributed to TOC having a molecular weight greater than 30,000 daltons.

Bonner et al. (1991) found that a hydrophobic membrane will foul to a much
greater extent relative to a hydrophilic one. Larger dissolved molecules caused more
fouling on hydrophilic membrane than the smaller humic material, while on the
hydrophobic membrane the fouling depended upon the solution conditions.

Pretreatment of raw water can reduce membrane fouling. Moulin et al. (1991)
found that without pretreatment, the membrane fouling occurred within 15 min on MF
but occurred on UF after 180 min of operation time. Some researchers suggest that
combining MF or UF with NF could be a feasible way to reduce NF fouling (Chellam
et al. 1997b). Carroll et al. (2000) demonstrated that coagulation pretreatment
improves NOM removal and reduces fouling. Siddiqui et a1. (2000) and Amy et al.
(1993) showed that MF followed by NF could readily meet both turbidity and DBP

regulations. Other researchers ((Laine et al. 1990) observed that combined PAC and

31

membrane decreased irreversible membrane fouling, regardless of membrane
hydrophobicity. However, Lin et a1. (1999) found that the combined PAC-UF system

signiﬁcantly decreased the ﬂux and increased the membrane fouling.

2.4.3 Ozonation/Biotreatment

Biological treatment is frequently employed downstream of ozonation since
many of the ozonation by-products are biodegradable. Biodegradation of ozonated
water was shown to be an effective treatment method for increasing water biostability,
removing some DBP precursors, and increasing the chlorine residual (Cipparone et a1.
1997). The goals of combined ozonation and biodegradation process are not only the
direct destruction of DBP precursors through ozonation, but also the conversion of
precursors to more biodegradable products and the transformation of higher molecular
weight compounds to lower molecular weight compounds for subsequent removal
during biological treatment (Leisinger and et al. 1981; Speitel et al. 1993).

Aldehydes, which are not easily removed during ﬂocculation/coagulation by
complexation with aluminum or iron salts, and other ozonation by-products, such as
organic acids and ketoacids, are biodegradable and can serve as a carbon source for
bacteria for growth and cell maintenance, which potentially cause regrowth problems
in distribution system. The regrowth of bacteria in distribution systems can result in
the formation of bioﬁlms, unpleasant taste, odors, and color, and also result in
biologically mediated corrosion and interference with microbiological monitoring
(Lee et al. 1980; Miltner et a1. 1992; Schechter and Singer 1995). In order to remove
BDOC and prevent the occurrence of these problems, biotreatment methods such as
biological activated carbon (BAC) and bioﬁltration following ozone treatment are

employed in the water treatment process. Kasarabada (1997), Mellema (1998) and

32

Yavich (1998) used a combination of ozonation and biological ﬂuidized bed treatment
(F BT) system to investigate the removal efﬁciency of NOM. They found that the
biodegradability of NOM increased after ozonation. They also reported that using the
combined ozonation/FBT system for controlling the THM precursors in drinking
water is technically and economically feasible.

Several researchers investigated the correlation between ozone dosage and
biodegradability. An increase in biodegradability with increasing ozone dose was
observed (Goel et al. 1995; Kasarabada 1997; Yavich 1998). Somyia and coworkers
(1986) used the ratio of BOD/TOC versus the ratio of O3/TOC to represent the change
in biodegradability during ozonation. They ozonated humic acids samples and found
that the amount of BOD degraded by microbes increased with an increase in
ozonation dosage and gradually leveled off after an ozone dose of 2 mg O3/mg TOC.
Other researchers (Cipparone et al. 1997; Miltner et a1. 1992; Shukairy et al. 1995;
Siddiqui et a1. 1997) observed the same trend but the extent of removal of DOC
improved with increasing ozone dosage up to an O3-to-DOC ratio of almost 1 mg
O3/mg TOC, after which the biodegradability was not signiﬁcantly improved by
ozonation. This may be caused by the different composition of NOM in different
water sources.

As a result of the combined ozonation/biotreatment process, DBP formation
potential is signiﬁcantly reduced because the DOC concentration is reduced by
biodegradation, and the reactivity of organic matter to subsequent chlorination can
also be decreased (e.g., see Amy et a1. 1988; Cipparone et al. 1997; Kasarabada 1997;
Miltner et a1. 1992; Shukairy et al. 1995; Yavich 1998). Yavich (1998) reported that
the pilot scale ozonation/FBT system was able to remove up to 70% of THMF P at an

EBCT of 180 minutes and ozone dose at 2 mg O3/mg C. Siddiqui and colleagues

33

(1997) found the formation potentials of THMs, HAAs, and chloral hydrate were
reduced by as much as 70 to 80% using ozonation-bioﬁltration treatment at an ozone
dosage of 1.5 03 mg/mg DOC. They also observed a 40 to 50% decrease in DOC,
and a 90 to 100% reduction in aldehydes after 5 days of bioﬁltration of ozonated
water. Similar aldehydes removal efﬁciencies have been observed by several other
researchers (Sketchell et al. 1995; Van Hoof and et al. 1986). Richardson and
colleagues (1999) studied the effect of bioﬁltration on ozonation by-products. They
found that 45 to 90% of carboxylic acids, approximately 50% of the AOC, 35 to 95%
of aldehydes, ketones, and aldo and keto acids were removed by bioﬁltration.
Combined ozonation and biodegradation process removed up to 50% of THM
precursors and up to 70% of HAA precursors (Speitel et al. 1993).

The bromate ion, formed during ozonation, is a toxic inorganic compound.
Miltner and colleagues (1992) found that biodegradation was less effective for
controlling the formation of brominated compounds. However, Kirisits and Snoeyink
(1999) showed that bromate removal efﬁciencies of up to 95% can be achieved if the
inﬂuent dissolved oxygen (DO) concentration, nitrate concentration, and empty bed
contact time (EBCT) are suitably controlled. In their study, bromate removal in the
BAC column increased as the inﬂuent DO concentration was reduced from 8.0 to 4.5
mg/L or as the inﬂuent nitrate concentration was reduced. Higher bromate reduction
was also observed at longer EBCTs. This may be caused by decreasing the hydraulic
loading on the BAC column, then inducing an increase in the numbers of

microorganisms.

34

2.4.4 Ozonation/Membrane

Few studies have been performed with the combination of ozonation and
membrane processes since organic membranes, which have relatively large surface
area per volume and are widely used in water and wastewater treatment, are prone to
destruction by ozone. Several researchers (Castro and Zander 1995; Shanbhag et al.
1998; Shen et al. 1990) reported that membranes were damaged during ozone
contacting and high replacement costs would be likely required. Thus, the use of
ozonation processes in conjunction with membrane devices is an area that has yet to
see much development. However, the ozonation/membrane process was found to be
able to maintain high and stable permeate ﬂux, reduce membrane fouling, and
increasing ozone mass transfer efﬁciency.

In order to prevent the damage of membrane by ozone, a retention or aeration
tank is often used between ozonation and the membrane module, allowing for
stripping of residual ozone. Hyung and colleagues (2000) investigated the effect of
ozonation on membrane ﬂux and water quality in an ozonation/UP hybrid system.
Water samples were collected from two unidentiﬁed locations in the Han River.
Ozonated water was retained for one hour before being used as feed water to UF.
They found that with the upstream water ozonation enhanced membrane ﬂux
regardless of ozone dose; whereas, for the downstream water, the membrane ﬂux was
increased or decreased depending on the ozone dose. About 22% of TOC, 64% of
UV254, and 36-53% of TOC precursors were removed by this hybrid system and the
ratio of THMFP/TOC was also decreased from 31.3 ug/mg of raw water to 19.7
ug/mg of permeate of the hybrid system at 9.4 mg/L ozone dose. O’Connell and
Danos (1997) used a similar system to treat well water with elevated levels of iron

and manganese. High removal rates, 97% of the iron and 91% of the manganese,

35

were achieved and the quality of permeate remained consistent despite the ﬂuctuation
of feed quality.

Ozonation has proven to prevent foulants from adhering to the membrane
surface, thereby decreasing membrane fouling. Hashino et al. (2000) reported that
using a combined ozonation with an ozone resistant MF which is made of
polyvinylideneﬂuoride (PVDF), high dissolved ozone concentrations on the
membrane surface were necessary to obtain high permeate ﬂuxes and to prevent
membrane fouling. Kim and colleagues (1999) investigated the effect of ozone
bubbling on ﬂux recovery of ceramic membrane. The results showed that intermittent
ozonation is effective for maintaining high permeation ﬂux and to prevent membrane
fouling caused by particle accumulation on the membrane surface.

Increasing the ozone mass transfer efﬁciency from gas to liquid phase is one
of the beneﬁts of the combined ozonation/membrane system. Usually, bubble
columns, packed towers, staged towers, spray towers, and venturi scrubbers are
conventionally used to achieve ozone mass transfer but their efﬁciencies are not very
good. Shen and Semens (1990) investigated the ozone mass transfer coefﬁcient (KL)
and speciﬁc interfacial area (a) of an ozonation/MF system. Comparing the
volumetric mass transfer coefﬁcient (KLa) values of bubble columns collected from
the literature, they found that the hybrid system had the highest KLa value (by up to a
factor of 110). This means that the hybrid system shows promise for efﬁciently
transferring ozone.

In recent years, ceramic membranes have gained more attention because they
can resist chemical corrosion and oxidation, extreme pH values, and have a high
range of operating temperatures. Several researchers found that the combined

ozonation and ceramic membrane process can achieve a high permeate ﬂux (Kim et al.

36

1999; Moulin et a1. 1991). No membrane damage was observed. Ozone transfer
through various ceramic membranes was also investigated by Janknecht and
colleagues (2000). They observed that the ceramic membrane with hydrophobic
coating has much higher ozone mass transfer rates than that without hydrophobic
coating. The mass transfer rates, surprisingly, were not adversely affected by the
decay of ozone within the membrane. The energy demand of this system is in the
same order of magnitude as the bubble column but the required contactor volume is

smaller.

2.5 REFERENCES

Allgeier, S. C., and Summers, R. S. (1995). "Evaluating Nf for Dbp Control with the
Rbsmt." Journal American Water Works Association, 87(3), 87-99.

Amy, G., Siddiqui, M., Ryan, J ., Odem, W., Hendricks, D., and Champlin, T. (1993).
"Membrane Separation of DBP Precursors from Low-Turbidity Surface
Waters." Proc. 1993 Membrane Technology Conf., Baltimore, Md., 651-663.

Amy, G. L., Alleman, B. C., and Cluff, C. B. (1990). "Removal of Dissolved Organic-
Matter by Nanoﬁltration." Journal of Environmental Engineering-Asce,
116(1), 200-205.

Amy, G. L., Kuo, C. J., and Sierka, R. A. (1988). "Ozonation of Humic Substances -
Effects on Molecular-Weight Distributions of Organic-Carbon and
Trihalomethane Formation Potential." Ozone-Science & Engineering, 10(1),
39-54.

Amy, G. L., Sierka, R. A., Bedessem, J., Price, D., and Tan, L. (1992). "Molecular-
Size Distributions of Dissolved Organic-Matter." Journal American Water
Works Association, 84(6), 67-75.

Anderson, L. J., Johnson, J. D., and Christman, R. F. (1985). "The Reaction of Ozone
with Isolated Aquatic F ulvic-Acid." Organic Geochemistry, 8(1), 65-69.

Andrews, R. C., and Ferguson, M. J. (1996). "Minizing Disinfection By-Product
Formation while Ensuring Giardia Control." Disinfection By-products in
Water Treatment, R. A. Minear and G. L. Amy, eds., Lewis Publisher, Boca
Raton, FL., 17-55.

37

Bablon, G., Bellamy, W. D., Bourbigot, M.-M., Daniel, F. B., Dore, M., Erb, F.,
Gordon, G., Langlais, B., Laplanche, A., Legube, B., Martin, G., Masschelein,
W. J., Pacey, G., Reckhow, D. A., and Ventresque, C. (1991a). "Practical
Application of Ozone: Principles and Case Studies." Ozone in Water
Treatment: Application and Engineering, B. Langlais, D. A. Reckhow, and D.
R. Brink, eds., Lewis Publisher, Chelsea, M1, 133-316.

Bablon, G., Bellamy, W. D., Bourbigot, M.-M., and et. al. (1991b). "Fundamental
Aspects." Ozone in Water Treatment: Application and Engineering, B.
Langlais, D. A. Reckhow, and D. R. Brink, eds., Lewis Publishers, Chelsea,
MI, 11-132.

Bonner, F. A., II, and O'Melia, C. R. (1991). "Some Aspects of the Fouling of
Ultraﬁltration Membranes by Natural Organic Matter in Water Treatment."
Proc. 1991 Membrane Technologies in the Water Industry, Orlando, Florida,
239-251.

Bouwer, E. J., and Crowe, P. B. (1988). "Biological Processes in Drinking-Water
Treatment." Journal American Water Works Association, 80(9), 82-93.

Bull, R. J., and McCabe, L. J. (1984). "Risk Assessment Issues in Evaluating the
Health Effects of Alternate Means of Drinking Water Disinfection." Water
Chlorination: Chemistry, Environmental Impact, and Health Effects, R. L.
Jolley and e. al., eds., Lewis Publishers, Chelsea, Mich.

Carroll, T., King, 8., Gray, S. R., Bolto, B. A., and Booker, N. A. (2000). "The
fouling of microﬁltration membranes by NOM after coagulation treatment."
Water Research, 34(1 1), 2861-2868.

Castro, K., and Zander, A. K. (1995). "Membrane Air-Stripping - Effects of
Pretreatment." Journal American Water Works Association, 87(3), 50-61.

Chadik, P. A., and Amy, G. L. (1983). "Removing Trihalomethane Precursors from
Various Natural-Waters by Metal Coagulants." Journal American Water
Works Association, 75(10), 532-536.

Chadik, P. A., and Amy, G. L. (1987). "Coagulation and Adsorption of Aquatic
Organic Matter and Humic Substances: An Analysis of Surrogate Parameters

for Predicting Effects on Trihalomethane Formation Potential." Envir. Tech.
Letters, 261.

Chadik, P. A., Pujals, V. J ., and Fisher, J. H. (1991). "Removal of Natural Organic
Matter and Reduction of THM Formation Potential in Florida Waters with

Ultraﬁltration." Proc. 1991 Membrane Technologies in the Water Industry,
Orlando, Florida, 289-303.

Chamock, C., and Kjonno, O. (2000). "Assimilable organic carbon and biodegradable

dissolved organic carbon in Norwegian raw and drinking waters." Water
Research, 34(10), 2629-2642.

38

Chellam, S., Jacangelo, J. G., Bonacquisti, T. P., and Long, B. W. (1997a). "Effect
Operating Conditions and Pretreatment for Nanoﬁltration of Surface Water."
Proc. 1997 Membrane Technology Conf., New Orleans, LA, 215-231.

Chellam, S., Jacangelo, J. G., Bonacquisti, T. P., and Schauer, B. A. ( 1997b). "Effect
of pretreatment on surface water nanoﬁltration." Journal American Water
Works Association, 89(10), 77-89.

Christman, R. F., Norwood, D. L., Millington, D. S., Johnson, J. D., and Stevens, A.
A. (1983). "Identity and Yields of Major Halogenated Products of Aquatic

Fulvic-Acid Chlorination." Environmental Science & Technology, 17(10),
625-628.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

Collins, M. R., Amy, G. L., and Steelink, C. (1986). "Molecular-Weight Distribution,
Carboxylic Acidity, and Humic Substances Content of Aquatic Organic-
Matter - Implications for Removal During Water-Treatment." Environmental
Science & Technology, 20(10), 1028-1032.

Croue, J. P., Koudjonou, B. K., and Legube, B. (1996). "Parameters affecting the
formation of bromate ion during ozonation." Ozone-Science & Engineering,
18(1), 1-18.

DiGiano, F. A., Braghetta, A., Uten, B., and Nilson, J. (1993). "Nanoﬁltration Fouling
by Natural Organic Matter and Role of Particles in Flux Enhancement." Proc.
1993 Membrane Technology Conf., Baltimore, Md, 273-291.

DiGiano, F. A., and et a1. (1994). "Fouling of Nanoﬁltration Membranes by Natural
Organic Matter." Proc. 1994 ASCE Natl. Conf. On Envir. Engrg., Boulder,
Colo.

Escobar, I. C., and Randall, A. A. (1999). "Inﬂuence of NF on distribution system
biostability." Journal American Water Works Association, 91(6), 76-89.

Field, R. W. (1996). "Mass Transport and the Design of Membrane Systems."
Industrial Membrane Separation Technology, K. Scott and R. Hughes, eds.,
Blackie Academic & Professional, Glasgow, UK.

Flogstad, H., and Odegaard, H. (1985). "Treatment of Humic Waters by Ozone."
Ozone-Science & Engineering, 7(2), 121-136.

Fox, K. R., and Lytle, D. A. (1996). "Milwaukee's crypto outbreak: Investigation and
recommendations." Journal American Water Works Association, 88(9), 87-94.

Frimmel, F. H. (1998). "Characterization of natural organic matter as major
constituents in aquatic systems." Journal of Contaminant Hydrology, 35(1-3),
201-216.

39

Frost, F. J ., Craun, G. F ., and Calderon, R. L. (1996). "Waterborne disease
surveillance." Journal American Water Works Association, 88(9), 66-75.

Fu, P., Ruiz, H., Thompson, K., and Spangenberg, C. (1994). "Selecting Membranes
for Removing Nom and Dbp Precursors." Journal American Water Works
Association, 86(12), 55-72.

Glaze, W. H., Koga, M., Cancilla, D., Wang, K., McGuire, M. J., Sun, L. A., Davis,
M. K., Tate, C. H., and Aieta, E. M. (1989). "Evaluation of Ozonation by-
Products from 2 California Surface Waters." Journal American Water Works
Association, 81(8), 66-73.

Glaze, W. H., Koga, M., Ruch, E. C., and Cancilla, D. (1988). "Application of Closed
Loop Stripping and XAD Resin Adsorption for the Determination of Ozone
By-products from Natural Water." Biohazards of Drinking Water Treatment,
201-210.

Glaze, W. H., Weinberg, H. S., and Cavanagh, J. E. (1993). "Evaluating the
Formation of Brominated Dbps During Ozonation." Journal American Water
Works Association, 85(1), 96-103.

Goel, S., Hozalski, R. M., and Bouwer, E. J. (1995). "Biodegradation of Nom - Effect
of Nom Source and Ozone Dose." Journal American Water Works Association,
87(1), 90-105.

Gracia, R., Aragues, J. L., and Ovelleiro, J. L. (1996). "Study of the catalytic
ozonation of humic substances in water and their ozonation byproducts."
Ozone-Science & Engineering, 18(3), 195-208.

Haag, W. R., and Hoigné, J. (1983). "Ozonation of Bromide-Containing Waters -
Kinetics of Formation of Hypobromous Acid and Bromate." Environmental
Science & Technology, 17(5), 261-267.

Haag, W. R., and Yao, C. C. D. (1992). "Rate Constants for Reaction of Hydroxyl
Radicals with Several Drinking-Water Contaminants." Environmental Science
& Technology, 26(5), 1005-1013.

Hashino, M., Mori, Y., Fujii, Y., Motoyama, N., Kadokawa, N., Hoshikawa, H.,
Nishijima, W., and Okada, M. (2000). "Pilot plant evaluation of an ozone-
microﬁltration system for drinking water treatment." Water Science and
Technology, 41(10-11), 17-23.

Hoigné, J ., and Bader, H. (1975). "Ozonation of Water: Role of Hydroxyl Radicals as
Oxidizing Agents." Science, 190, 782-784.

Hoigné, J ., and Bader, H. (1983). "Rate Constants of Reactions of Ozone with
Organic and Inorganic-Compounds in Water .1. Non-Dissociating Organic-
Compounds." Water Research, 17(2), 173-183.

Huck, P. M. (1990). "Measurement of Biodegradable Organic-Matter and Bacterial-

Growth Potential in Drinking-Water." Journal American Water Works
Association, 82(7), 78-86.

40

Hyung, H., Lee, 8., Yoon, J ., and Lee, C. H. (2000). "Effect of preozonation on ﬂux
and water quality in ozonation- ultraﬁltration hybrid system for water
treatment." Ozone-Science & Engineering, 22(6), 637-652.

lchihashi, K., Teranishi, K., and lchimura, A. (1999). "Brominated trihalomethane
formation in halogenation of humic acid in the coexistence of hypochlorite
and hypobromite ions." Water Research, 33(2), 477-483.

Jacangelo, J. G., Laine, J. M., Cummings, E. W., and Adham, S. S. (1995). "Uf with
Pretreatment for Removing Dbp Precursors." Journal American Water Works
Association, 87(3), 100-1 12.

Janknecht, P., Wilderer, P. A., Picard, C., Larbot, A., and Sarrazin, J. (2000).
"Investigations on ozone contacting by ceramic membranes." Ozone-Science
& Engineering, 22(4), 379-392.

Joret, J. C., Levi, Y., Dupin, T., and Gilbert, M. (1988). "Rapid Method for
Estimating Bioeliminable Organic Carbon in Water." Proc. 1988 AWWA Ann.
Conf., Orlando, Florida, 1715-1725.

Joret, J. C., Levi, Y., and Volk, C. (1991). "Biodegradable Dissolved Organic-Carbon
(Bdoc) Content of Drinking-Water and Potential Regrowth of Bacteria."
Water Science and Technology, 24(2), 95-101.

Kaastrup, E., and Halmo, T. M. (1989). "Removal of Aquatic Humus by Ozonation
and Activated Carbon Adsorption." Aquatic Humic Substances: Inﬂuence on
Fate and Treatment of Pollutants, I. H. Suffet and P. MacCarthy, eds.,
American Chemical Society, Washington, DC., 697-726.

Kasarabada, A. N. (1997). "Use of THMFP, TOC and Chlorine Demand as Control
Parameters for Monitoring the Performance of an Ozonation/FBT System,"
MS. Thesis, Michigan State University.

Kim, J. O., Somiya, 1., and Fujii, S. (1999). "Fouling Control of Ceramic Membrane
in Organic Acid F ermenter by Intermittent Ozonation." Proceedings of the
14th Ozone World Congress, Dearbom, MI, 131-143.

Kim, J. S., Chian, E. S. K., Saunders, F. M., Michael Perdue, E., and Giabbai, M. F.
(1989). "Characteristics of Humic Substances and Their Removal Behavior in
Water Treatment." Aquatic Humic Substances: Inﬂuence on Fate and
Treatment of Pollutants, l. H. Suffet and P. MacCarthy, eds., American
Chemical Society, Washington, DC., 473-497.

Kirisits, M. J., and Snoeyink, V. L. (1999). "Reduction of bromate in a BAC ﬁlter."
Journal American Water Works Association, 91(8), 74-84.

Koechling, M. T., Shukairy, H. M., and Summers, R. S. (1996). "Effect of Ozonation
and Biotreatment on Molecular Size and Hydrophilic Fractions of Natural
Organic Matter." Water Disinfection and Natural Organic Matter:
Characterization and Control, R. A. Minear and G. L. Amy, eds., American
Chemical Society, Washington, DC., 196-210.

41

Korshin, G. V., Perry, S. A. L., and Ferguson, J. F. (1996). "Inﬂuence of NOM on
copper corrosion." Journal American Water Works Association, 88(7), 36-47.

Kramer, M. H., Herwaldt, B. L., Craun, G. F., Calderon, R. L., and Juranek, D. D.
(1996). "Waterbome disease: 1993 and 1994." Journal American Water Works
Association, 88(3), 66-80.

Krasner, S. W., Croue, J. P., Bufﬂe, J., and Perdue, E. M. (1996a). "Three approaches
for characterizing NOM." Journal American Water Works Association, 88(6),
66-79.

Krasner, S. W., Sclimenti, M. J., Chinn, R., Chowdhury, Z. K., and Owen, D. M.
(1996b). "The Impact of TOC and Bromide on Chlorination By-Product
Formation." Disinfection By-products in Water Treatment, R. A. Minear and
G. L. Amy, eds., Lewis Publisher, Boca Raton, F L., 59-90.

Laine, J. M., Clark, M. M., and Mallevialle, J. (1990). "Ultraﬁltration of Lake Water -
Effect of Pretreatment on the Partitioning of Organics, Thmfp, and Flux."
Journal American Water Works Association, 82(12), 82-87.

Laine, J. M., Jacangelo, J. G., Cummings, E. W., Cams, K. E., and Mallevialle, J.
(1993). "Inﬂuence of Bromide on Low-Pressure Membrane Filtration for
Controlling Dbps in Surface Waters." Journal American Water Works
Association, 85(6), 87-99.

LeChevallier, M. W., Abbaszadegan, M., Camper, A. K., Hurst, C. J ., Izaguirre, G.,
Marshall, M. M., Naumovitz, D., Payment, P., Rice, E. W., Rose, J., Schaub,
S., Sliﬂ(o, T. R., Smith, D. B., Smith, H. V., Sterling, C. R., and Stewart, M.
(1999a). "Committee report: Emerging pathogens - bacteria." Journal
American Water Works Association, 91(9), 101-109.

LeChevallier, M. W., Abbaszadegan, M., Camper, A. K., Hurst, C. J., Izaguirre, G.,
Marshall, M. M., Naumovitz, D., Payment, P., Rice, E. W., Rose, J., Schaub,
S., Sliﬂco, T. R., Smith, D. B., Smith, H. V., Sterling, C. R., and Stewart, M.
(1999b). "Committee report: Emerging pathogens - viruses, protozoa, and
algal toxins." Journal American Water Works Association, 91(9), 110-121.

Lee, S. H., Oconnor, J. T., and Banerji, S. K. (1980). "Biologically Mediated
Corrosion and Its Effects on Water- Quality in Distribution-Systems." Journal
American Water Works Association, 72(1 1), 636-645.

Leenheer, J. A. (1981). "Comprehensive Approach to Preparative Isolation and
Fractionation of Dissolved Organic-Carbon from Natural-Waters and
Wastewaters." Environmental Science & Technology, 15(5), 578-587.

Leenheer, J. A. (1985). "Fractionation Techniques for Aquatic Humic substances."
Humic Substances in Soil, Sediment, and Water, G. Aiken and e. al., eds.,
Wiley-Interscience, New York, 409—429.

Leisinger, T., and et al. (1981). Microbial Degradation of Xenobiotics and
Recalcitrant Compounds, Academic Press, London, England.

42

Leppard, G. G. (1992). "Evaluation of Electron Microscope Techniques for the
Description of Aquatic Colloids." Environmental Particles, J. Bufﬂe and H. P.
van Leeuwen, eds., Lewis Publisher, Boca Raton, F L., 231-290.

Levy, R. V., Hart, F. L., and Cheetham, R. D. (1986). "Occurrence and Public-Health
Signiﬁcance of Invertebrates in Drinking-Water Systems." Journal American
Water Works Association, 78(9), 105-110.

Liao, W., Christman, R. F., Johnson, J. D., Millington, D. S., and Hass, I. R. (1982).
"Structural Characterization of Aquatic Humic Material." Environmental
Science & Technology, 16(7), 403-410.

Lin, C. F., Huang, Y. J., and Hao, I. J. (1999). "Ultraﬁltration processes for removing
humic substances: Effect of molecular weight fractions and PAC treatment."
Water Research, 33(5), 1252-1264.

Lozier, J. C., and Carlson, M. (1991). "Organics Removal from Eastern US. Surface
Waters Using Ultra-Low Pressure Membranes." Proc. 1991 Membrane
Technologies in the Water Industry, Orlando, Florida, 521-543.

MacCarthy, P., and Suffet, I. H. (1989). "Aquatic Humic Substances and Their
Inﬂuence on the Fate and Treatment of Pollutants." Aquatic Humic Substances:
Inﬂuence on Fate and Treatment of Pollutants, I. H. Suffet and P. MacCarthy,
eds., American Chemical Society, Washington, DC., xvii-xxx.

Maggiolo, A. (1978). "Ozone's Radical and Ionic Mechanisms of Reaction with
Organic Compounds in Water." Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, R. G. Rice and J. A. Cotruvo, eds, Ozone Press Intl.,
Cleveland, Ohio.

Malcolm, R. L. (1985). "Geochemistry of Stream fulvic and Humic Substances."
Humic Substances in Soil, Sediment, and Water, G. R. Aiken, D. M.
McKnight, R. L. Wershaw, and P. MacCarthy, eds., John Wiley & Sons., New
York, 181-209.

Manahan, S. E. (1993). "Environmental Chemistry of Water." Fundamentals of
Environmental Chemistry, Lewis Publisher, Chelsea, M1., 388-390.

Martin-Mousset, B., Croue, J. P., Lefebvre, E., and Legube, B. (1997). "Distribution
and characterization of dissolved organic matter of surface waters." Water
Research, 31(3), 541—553.

McGuire, M. J., and Meadow, R. G. (1988). "Awwarf Trihalomethane Survey."
Journal American Water Works Association, 80(1), 61-68.

Mellema, J. A. (1998). "The Use of Apparent Molecular Weight Distribution to
Evaluate the Transformation of Natural Organic Matter During Ozonation and

Biological Treatment," MS. Thesis, Michigan State University, East Lansing,
MI.

43

Miltner, R. J., Shukairy, H. M., and Summers, R. S. (1992). "Disinfection by-Product
Formation and Control by Ozonation and Biotreatment." Journal American
Water Works Association, 84(11), 53-62.

Minear, R. A., and Amy, G. L. (1996). "Water Disinfection and Natural Organic
Matter: History and Overview." Water Disinfection and Natural Organic
Matter: Characterization and Control, R. A. Minear and G. L. Amy, eds.,
American Chemical Society, Washington, DC., 1-9.

Mogren, E. M., Scarpino, P., and Summers, R. S. (1990). "Measurement of
Biodegradable Dissolved Organic Carbon in Drinking Water." Proc. 1990
AWWA Ann. Conf., Cincinnati, Ohio, 573.

Moore, A. C., Herwaldt, B. L., Craun, G. F., Calderon, R. L., Highsmith, A. K., and
Juranek, D. D. (1994). "Waterbome Disease in the United-States, 1991 and
1992." Journal American Water Works Association, 86(2), 87-99.

Moulin, C., Bourbigot, M. M., Tazi-Pain, A., and F aivre, M. (1991). "Potanilisation of
Surface Waters by Crossﬂow Ultra- and Microﬁltration on Mineral
Membranes: Interest of Ozone." Proc. 1991 Membrane Technologies in the
Water Industry, Orlando, Florida, 729-737.

Mulford, L. A., Taylor, J. S., Nickerson, D. M., and Chen, S. S. (1999). "NF
performance at full and pilot scale." Journal American Water Works
Association, 9 1 (6), 64-75.

Najm, I. N., and Krasner, S. W. (1995). "Effects of Bromide and Nom on by-Product
Formation." Journal American Water Works Association, 87(1), 106-115.

Nakatsuka, S., Nakate, 1., and Miyano, T. (1996). "Drinking water treatment by using
ultraﬁltration hollow ﬁber membranes." Desalination, 106(1-3), 55-61.

Nieminski, E. C., and Ongerth, J. E. (1995). "Removing Giardia and Cryptosporidium
by Conventional Treatment and Direct-Filtration." Journal American Water
Works Association, 87(9), 96-106.

Nilson, J. A., and DiGiano, F. A. (1996). "Inﬂuence of NOM composition on
nanoﬁltration." Journal American Water Works Association, 88(5), 53-66.

O'Connell, J ., and Danos, S. (1997). "An Innovative Combination of Ozonation and
Ultraﬁltration." Proc. 1997 Membrane Technologies Conf., New Orleans, LA.,
1127-1145.

Owen, D. M., Amy, G. L., Chowdhury, Z. K., Paode, R., McCoy, G., and Viscosil, K.
(1995). "Nom - Characterization and Treatability." Journal American Water
Works Association, 87(1), 46-63.

Oxenford, J. L. (1996). "Disinfection By-Products: Current Practices and Future

Directions." Disinfection By-products in Water Treatment, R. A. Minear and
G. L. Amy, eds., Lewis Publisher, Boca Raton, FL., 3-16.

44

Paode, R. D., Amy, G. L., Krasner, S. W., Summers, R. S., and Rice, E. W. (1997).
"Predicting the formation of aldehydes and BOM." Journal American Water
Works Association, 89(6), 79-93.

Peleg, M. (1976). "The Chemistry of Ozone in the Treatment of Water." Wat. Res,
361-365.

Pemberton, S. G., and Jones, B. (1999). "Ichnology of the pleistocene ironshore
formation, Grand Cayman Island, British West-Indies." Journal of
Paleontology, 62(4), 495.

Perdue, E. (1985). "Acidic Functional Groups of Humic Substances." Humic
Substances in Soil, Sediment, and Water, G. Aiken and e. al., eds., Wiley-
Interscience, New York.

Preston, C. M. (1987). NMR of Humic Substances and Coal. Techniques, Problems
and Solutions, R. L. Wershaw and M. A. Mikita, eds., Lewis, Chelsea, 3-32.

Ravindran, V., Badriyha, B. N., and Pirbazari, M. (1993). "Crossﬂow Membrane
Filtration for the Removal of Natural Organic Matter." Proc. 1993 Membrane
Technology Conf., Baltimore, Md., 587-599.

Reckhow, D. A., and Singer, P. C. (1990). "Chlorination by-Products in Drinking
Waters - from Formation Potentials to Finished Water Concentrations."
Journal American Water Works Association, 82(4), 173-180.

Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990). "Chlorination of Humic
Materials - by-Product Formation and Chemical Interpretations."
Environmental Science & Technology, 24(11), 1655-1664.

Rencken, G. E. (1994). "Ozonation at Wiggins-Water-Puriﬁcation-Works, Durban,
South- Africa." Ozone-Science & Engineering, 16(3), 247-260.

Richardson, S. D., Thruston, A. D., Caughran, T. V., Chen, P. H., Collette, T. W.,
Floyd, T. L., Schenck, K. M., Lykins, B. W., Sun, G. R., and Majetich, G.
(1999). "Identiﬁcation of new ozone disinfection byproducts in drinking
water." Environmental Science & Technology, 33(19), 3368-3377.

Rittmann, B. E. (1995). "Fundamentals and Application of Bioﬁlm Processes in
Drinking Water Treatment." The Handbook of Environmental Chemistry, 61-
87.

Rook, .I. J. (1974). "Formation of Haloforms During Chlorination of Natural Waters."
J. Water Treat. Exam., 234.

Rook, J. J., Gras, A. A., Vanderheijden, B. G., and Wee, J. D. (1978). "Bromide
Oxidation and Organic Substitution in Water Treatment." Journal of
Environmental Science and Health Part a- Environmental Science and
Engineering & Toxic and Hazardous Substance Control, 13(2), 91-116.

45

Scanlan, P., Pohlman, B., Freeman, S., Spillman, B., and Mark, J. (1997). "Membrane
Filtration for the Removal of Color and TOC from Surface Water." Proc. 1997
Membrane Technology Conf., New Orleans, LA., 127-141.

Schechter, D. S., and Singer, P. C. (1995). "Formation of Aldehydes During
Ozonation." Ozone-Science & Engineering, 17(1), 53-69.

Servais, P., Billen, G., and Hascoet, M. C. (1987). "Determination of the
Biodegradable Fraction of Dissolved Organic-Matter in Waters." Water
Research, 21(4), 445-450.

Servais, P., Laurent, P., and Randon, G. (1993). "Impact of Biodegradable Dissolved
Organic Carbon (BDOC) on Bacterial Dynamics in Distribution Systems."
Proc. 1993 AWWA WQTC, Miami, Florida, 963-979.

Shanbhag, P. V., Guha, A. K., and Sirkar, K. K. (1998). "Membrane-based ozonation
of organic compounds." Industrial & Engineering Chemistry Research, 37(1 1),
4388-4398.

Shen, Z. S., Semmens, M. J., and Collins, A. G. (1990). "A Novel-Approach to Ozone
Water Mass-Transfer Using Hollow Fiber Reactors." Environmental
Technology, 1 1(7), 597-608.

Shukairy, H. M., Miltner, R. J., and Summers, R. S. (1994). "Bromides Effect on Dbp
Formation, Speciation, and Control .1. Ozonation." Journal American Water
Works Association, 86(6), 72-87.

Shukairy, H. M., Miltner, R. J., and Summers, R. S. (1995). "Bromides Effect on Dbp
Formation, Speciation and Control .2. Biotreatment." Journal American Water
Works Association, 87(10), 71-82.

Shukairy, H. M., and Summers, R. S. (1992). "The Impact of Preozonation and
Biodegradation on Disinfection by-Product Formation." Water Research,
26(9), 1217-1227.

Siddiqui, M., Amy, 0., Ryan, J., and Odem, W. (2000). "Membranes for the control
of natural organic matter from surface waters." Water Research, 34(13), 3355-
3370.

Siddiqui, M. S., and Amy, G. L. (1993). "Factors Affecting Dbp Formation During
Ozone Bromide Reactions." Journal American Water Works Association,

85(1), 63-72.

Siddiqui, M. 8., Amy, G. L., and Murphy, B. D. (1997). "Ozone enhanced removal of
natural organic matter from drinking water sources." Water Research, 31(12),
3098-3106.

Siddiqui, M. S., Amy, G. L., and Rice, R. G. (1995). "Bromate Ion Formation - a
Critical-Review." Journal American Water Works Association, 87(10), 58-70.

Singer, P. C. (1994). "Control of Disinfection by-Products in Drinking-Water."
Journal of Environmental Engineering-Asce, 120(4), 727-744.

46

Singer, P. C., Obolensky, A., and Greiner, A. (1995). "Dbps in Chlorinated North-
Carolina Drinking Waters." Journal American Water Works Association,
87(10), 83-92.

Sketchell, J., Peterson, H. G., and Christoﬁ, N. (1995). "Disinfection by-Product
Formation after Biologically Assisted Gac Treatment of Water-Supplies with
Different Bromide and Doc Content." Water Research, 29(12), 2635-2642.

Solo-Gabriele, H., and Neumeister, S. (1996). "US outbreaks of cryptosporidiosis."
Journal American Water Works Association, 88(9), 76-86.

Somiya, 1., Yamada, H., Nozawa, E., and Mohri, M. (1986). "Biodegradability and
Gac Adsorbability of Micropollutants by Preozonation." Ozone-Science &
Engineering, 8(1), 11-26.

Song, R., Westerhoff, P., Minear, R., and Amy, G. (1997). "Bromate minimization
during ozonation." Journal American Water Works Association, 89(6), 69-78.

Speitel, G. E., Jr., Symons, J. M., Diehl, A. C., Sorensen, H. W., and Cipparone, L. A.
(1993). "Effect of Ozone Dosage and Subsequent Biodegradation on Removal
of Dbp Precursors." Journal American Water Works Association, 85(5), 86-95.

Staehelin, J., and Hoigné, J. (1985). "Decomposition of Ozone in Water in the
Presence of Organic Solutes Acting as Promoters and Inhibitors of Radical
Chain Reactions." Environmental Science & Technology, 19(12), 1206-1213.

Stevenson, F. J. (1985). "Geochemistry of Soil Humic Substances." Humic
Substances in Soil, Sediment, and Water, G. R. Aiken, D. M. McKnight, R. L.
Wershaw, and P. MacCarthy, eds., John Wiley & Sons., New York, 13-52.

Stevenson, F. J. (1994). "Humus Chemistry: Genesis, Composition, Reactions." John
Wiley & sons. Inc., 48-53.

Symons, J. M., Speitel, G. E., Diehl, A. C., and Sorensen, H. W. (1994). "Precursor
Control in Waters Containing Bromide." Journal American Water Works
Association, 86(6), 48-60.

Takahashi, N., Nakai, T., Satoh, Y., and Katoh, Y. (1995). "Ozonolysis of Humic-
Acid and Its Effect on Decoloration and Biodegradability." Ozone—Science &
Engineering, 17(5), 51 1-525.

Takeuchi, Y., Mochidzuki, K., Matsunobu, N., Kojima, R., Motohashi, H., and
Yoshimoto, S. (1997). "Removal of organic substances from water by ozone

treatment followed by biological activated carbon treatment." Water Science
and Technology, 35(7), 171-178.

Taylor, J. S., Thompson, D. M., and Carswell, .l. K. (1987). "Applying Membrane
Processes to Groundwater Sources for Trihalomethane Precursor Control."
Journal American Water Works Association, 79(8), 72-82.

Texier, J. P., Huxtable, J., Rhodes, E., Miallier, D., and Ousmoi, M. (1998). "New
data on the chronological situation of the aterian of Morocco and their

47

implications." Comptes Rendus De L Academie Des Sciences Serie 1i

Mecanique Physique Chimie Sciences De L Univers Sciences De La T erre,
88(8), 124-140.

Thurman, E. M. (1985). "Humic Substances in Groundwater." Humic Substances in
Soil, Sediment, and Water, G. R. Aiken, D. M. McKnight, R. L. Wershaw, and
P. MacCarthy, eds., John Wiley & Sons., New York, 87-103.

Thurman, E. M., and Malcolm, R. L. (1981). "Preparative Isolation of Aquatic Humic
Substances." Environmental Science & Technology, 15(4), 463-466.

Thurman, E. M., Wershaw, R. L., Malcolm, R. L., and Pinckney, D. (1982). "
Molecular Size of Aquatic Humic Substances." J. Org. Geochem., 27-35.

Trussell, R. R., and Umphres, M. D. (1978). "Formation of Trihalomethanes." Journal
American Water Works Association, 70(1 1), 604-612.

Uden, P. C., and Miller, J. W. (1983). "Chlorinated Acids and Chloral in Drinking
Warer." Jour. AWWA., 524-527. .

Van der Kooij, D. (1995). "Signiﬁcance and Assessment of the Biological Stability of
Drinking Water." The Handbook of Environmental Chemistry, 89-102.

Van der kooij, D., Hijnen, W. A. M., and Kruithof, J. C. (1989). "The Effects of
Ozonation, Biological Filtration and Distribution on the Concentration of
Easily Assimilable Organic-Carbon (Aoc) in Drinking-Water." Ozone-Science
& Engineering, 11(3), 297-31 1.

Van der kooij, D., Visser, A., and Hijnen, W. A. M. (1982). "Determining the
Concentration of Easily Assimilable Organic- Carbon in Drinking-Water."
Journal American Water Works Association, 74(10), 540-545.

Van Hoof, F., and et a1. (1986). "Formation of Oxidation By-products in Surface
Water Pre-ozonation and Their Behavior in Water Treatment." Wat. Supply,
93.

Volk, C., Renner, C., Roche, P., Paillard, H., and Joret, J. C. (1993). "Effects of
Ozone on the Production of Biodegradable Dissolved Organic-Carbon (Bdoc)
During Water-Treatment." Ozone-Science & Engineering, 15(5), 389-404.

Von Gunten, U., and Hoigné, J. (1996). "Ozonation of Bromide-Containing Waters:
Bromate Formation through Ozone and Hydroxyl Radicals." Disinfection By-
products in Water Treatment: The Chemistry of Their Formation and Control,
R. A. Minear and G. L. Amy, eds., CRC Press, Inc., Boca Raton, PL, 187-206.

Weinberg, H. S., and Glaze, W. H. (1996). "An Overview of Ozonation Disinfection
By-Products." Disinfection By-products in Water Treatment, R. A. Minear and
G. L. Amy, eds., Lewis Publisher, Boca Raton, FL., 165-186.

Wershaw, R. L., and Aiken, G. R. (1985a). "Molecular Size and Weight
Measurements of Humic Substances." Humic Substances in Soil, Sediment,

48

and Water, G. R. Aiken, D. M. McKnight, R. L. Wershaw, and P. MacCarthy,
eds., John Wiley & Sons., New York, 477-492.

Wershaw, R. L., and Aiken, G. R. (1985b). "Molecular Size and Weight
Measurements of Humic Substances." Humic Substances in Soil, Sediment,
and Water, G. R. Aiken, D. M. McKnight, R. L. Wershaw, and P. MacCarthy,
eds., John Wiley & Sons., New York, 181-209.

Westerhoff, P., Aiken, G., Amy, G., and Debroux, J. (1999). "Relationships between
the structure of natural organic matter and its reactivity towards molecular
ozone and hydroxyl radicals." Water Research, 33(10), 2265-2276.

Wiesner, M. R., Schmelling, D. C., and Pickering, K. (1991). "Permeation Behavior
and F iltrate Quality of Tubular Ceramic Membranes Used for Surface Water

Treatment." Proc. 1991 Membrane Technologies in the Water Industry,
Orlando, Florida, 371 -383.

Wilson, M. A. (1987). NMR Techniques and Applications in Geochemistry and Soil
Chemistry, M. A. Wilson, ed., Pergamon Press, Oxford, 182-216.

Yavich, A. A. (1998). "The Use of Ozonation and Biological Fluidized Bed
Treatment for the Control of NOM in Drinking Water," Ph.D. Dissertation,
Michigan State University, East Lansing, MI.

49

CHAPTER 3
MASS TRANSFER OF OZONE IN THE OZONATION/MEMBRANE

PROCESS

3.1 INTRODUCTION

Since the discovery of potentially carcinogenic disinfection by-products in
chlorinated water, the use of ozone as an alternative disinfectant has gradually
increased in water treatment facilities. Although ozone effectively reduces the
formation of THMs and HAAs (Amy et al. 1988; Cipparone et al. 1997; Richardson et
a1. 1999), its low aqueous solubility and the high production cost of ozone gas have
limited ozone treatment applications. Thus, it is important to improve the rate of
ozone mass transfer from gas phase to aqueous phase. Consequently, the
development of an ozonation reactor with a high ozone mass transfer rate is needed.

Devices, such as stirred reactors, venturi scrubbers, bubble columns, and static
mixers, which achieve efﬁcient gas-liquid mass transfer in the ozonation of water and
wastewater, are well documented in literature (e.g., see Barberis and Howarth 1991;
Hsu et a1. 2002; Ledakowicz et al. 2001; Richards et a1. 1976; Roustan et al. 1987;
Roustan et al. 1996; Sheffer and Esterson 1982; Wright et al. 1998). However, only
three publications discuss the use of ozonation processes in conjunction with
membrane ﬁlters (Janknecht et a1. 2000; Shanbhag et al. 1998; Shen et a1. 1990).

Membrane ﬁlters are effective for the removal of turbidity, organic matter,
microorganisms, and some DBP precursors. Several researchers (Cui and Wright
1994; Klijn et al. 2000; Laborie et a1. 1997; Laborie et a1. 1998; Verberk et a1. 2000)
reported that injecting air directly into the feed water stream of membrane can

enhance the permeate ﬂux and control membrane fouling. Others observed similar

50

results when ozonation preceded membrane ﬁltration (Hashino et a1. 2000; Kim et al.
1999).

We hypothesize that the combination of ozonation with membrane ﬁltration
will result in a reduction in the DBPs formation potential and the membrane fouling
potential, maintain a stable permeate ﬂux, and retain and inactive microorganisms as
compared to either ozone or membrane ﬁltration alone. As part of this work we
investigated the effect of membrane ﬁltration on the mass transfer of ozone.

In this study, a new ozone/membrane system was set up and the effects of
various operational parameters on the ozone volumetric mass transfer coefﬁcient were
determined. The water ﬂow rate, ozone gas ﬂow rate, inlet ozone concentration, and
temperature were varied. An ozone mass transfer model that predicts the ozone
concentration proﬁle during the period of mass transfer was developed. The ability of
ozone gas injection to enhance the permeate ﬂux of the membrane was studied. The
ozone gas and water ﬂow rates were varied during this portion of the study.
Furthermore, due to the importance of volumetric mass transfer coefﬁcient in the
design and speciﬁcation of ozonation equipment, the coefﬁcient obtained from this
study and from different ozonation contactors were compared. The efﬁciency of two
ozone inline mixers, a simple Y inline mixer and a high efﬁciency inline mixer, were

also investigated.

3.2 MATERIALS AND METHODS
3.2.1 Permeate Flux Enhancement Experiment

A specially designed ozonation/membrane system was used for investigating
the enhancement of permeate ﬂux by ozone injection. The schematic diagram of this

system is shown in Figure 3.1. The membrane module and water tank served as

51

ozone contactor. Deionized distilled water (DDI) was prepared and used in this study.

Teﬂon tubing and stainless steel joints and valves were used throughout the system.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

M . Waste Gas
Oisture
A A dsorben t , ,3 Spectrophotometer l
" . Ozone w p__3
Generator ,
Oxygen Cylinder / t._____
K1
1
__ @— E
A
M e. ‘9
/\ [\l/<l l Membrane
ER 7 Permeate / ,\
:[ t.__5
CO
CO
® Presure gauge
CO ® Flow meter
0 Z: Pump
X Valve
____- _l
Water Tank

Figure 3.1 Experimental setup for measurement of permeate ﬂux enhancement of
ozone injection.

52

Tubular ceramic membranes (Clover, CéRAM Inside, Tami North America, St.
Laurent, Québec, Canada) with a molecular weight cut-off of 15 kD were used in the
system. The characteristics of the laboratory-scale ultraﬁltration membrane are

summarized in Table 3.1.

Table 3.1 Details of the ceramic membrane.

 

 

External Active Filtering Number Operating Cut-off
diameter length area of pH range
(mm) (cm) (cmz) channels range (kD)
10 25.0 95.2 3 0-14 15

 

 

 

 

 

 

 

Ozone was generated by feeding pure oxygen gas from a gas cylinder to the
ozone generator (Model OZZPCS, Ozotech Inc., Yreka, Calif). The oxygen gas
stream was dried using a molecular sieve trap prior to ozone generation. The ozone
gas concentration was controlled using the electric current of the ozone generator.
The wasted ozone gas was vented after passing the gas through a 2% potassium iodide
(KI) solution to destroy any residual ozone gas.

Twenty-three liters of acidiﬁed DDI water were prepared in the water-jacked
water tank, which was used to adjust the temperature of the DDI water. The
temperature of the water was monitored by a thermometer. The water tank was
equipped with a power mixer with four impellers to completely mix the water. The
acidiﬁed DDI water was introduced with a centrifugal pump (Model 4RH 12, Dayton
Electric Mfg. Co., Niles, Illinois) through the housing of the membrane. Ozone gas

was added into the water stream through a simple Y inline mixer (SYM) (see Figure

53

 

3.2) just before entering the membrane module. The DDI water circulated
tangentially on the surface of the membrane. The permeate ﬂowed out from the
membrane housing through two permeate outlets and was collected to measure the
ﬂux. The rejected DDI water was continuously recirculated back to the water tank.
The water ﬂow rate and gas ﬂow rate were varied by control valves from 2 to 5 L/min
and 0 to 500 mL/min, respectively. The water temperature was controlled at 27 °C

and the inlet ozone gas concentration was ﬁxed at 2.5 g/m3.

Simple ”Y" In line Mlxer

30

 

l~——°mm ‘ l

Figure 3.2 Simple Y inline mixer (SYM).

3.2.2 Mass Transfer Experiment
The experimental setup for the ozone mass transfer experiment was similar to
that shown in Figure 3.1 but modiﬁed slightly. Since the retention time in the twenty-

three-liter water tank was too long to avoid ozone decomposition and to reach steady

54

state conditions, it was replaced by a 3.5-liter water-jacked glass reservoir made of
Pyrex glass in this study. A schematic diagram of the experimental setup is shown in
Figure 3.3. In order to prevent any interference from impurities and ozone self-
decomposition, DDI water acidiﬁed to pH 2 using sulfuric acid was prepared and used
in this study. Two ozone-water inline mixers, the simple Y inline mixer (SYM) and
the high efﬁciency inline mixer (HEM) shown in Figures 3.2. and 3.4., respectively
(Ozone Service, Burton, BC, Canada), were used in the mass transfer study to
compare their mass transfer efﬁciency.

The experimental procedure for mass transfer experiments was similar to that
described in Section 3.2.1 with slight modiﬁcation. A magnetic stirrer at the bottom
of the water reservoir was used to completely mix the water. A peristaltic pump
(Masterﬂex Model 7520-35, Cole-Parmer Co., Chicago, Illinois) was used to
continuously pump the acidiﬁed DDI water that was prepared in a 1.5-liter water tank
into the 3.5-liter water reservoir to compensate the water loss to permeate and due to
sampling. Dissolved ozone concentrations in the permeate and in the water reservoir
were monitored spectrophotometrically at 254 nm at appropriate time intervals.
Ozone gas phase concentrations in the headspace of water reservoir were measured in
the same manner. When the dissolved ozone concentration in water reservoir reached
steady state conditions, the system operation was stopped.

Several operational factors, such as ozone concentration in the inlet gas phase,
water ﬂow rate, gas ﬂow rate, and water temperature, which inﬂuence the ozone mass
transfer and solubility were investigated. Table 3.2. summarizes the experimental

conditions that were varied for the ozone mass transfer measurements.

55

Moisture
Adsorbent

 

—_4

——E~—~

Waste Gas

 

 

Spectrophotometer T

Ozone
Generator

 

 

 

 

 

 

Oxygen Cylinder

vi
A

 

 

/“ i l ‘— '—"

j LIME<ETTYTT"M“W’1: Membrane J——————j
CF) --__ ._

3 HM PermeatelJ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' J\
L
Water J
Tank C? (a Presure gauge
Ll. \
6:) Flow meter
O (At
L H Pump
O 9"

 

 

 

A Valve

Figure 3.3 Experimental setup of the ozone mass transfer study.

56

High Efficiency In line Mixer

 

 

7 mm DIM T
Ozone Port 35
.‘1 mm
Diffuser J—
l l
< 75 mm =
l I

Figure 3.4 High efﬁciency inline mixer (HEM).

Table 3.2 Experimental conditions for the ozone mass transfer measurement.

 

 

 

 

 

 

Parameters Values
Inlet ozone dose (g/m’) 1.5, 2.5, 5.8, 12.5, 23.0, 36.0
Water ﬂow rate (L/min) 2, 2.75, 3.5
Ozone gas ﬂow rate (mL/min) 25, 50, 75, 100
Water temperature (°C) 10, 20, 30

 

 

3.2.3 Ozone Concentration Measurement

The concentration of ozone in the gas phase was measured at the wavelength
254 nm with an UV/V IS spectrophotometer (Milton Roy Genesis-5, Milton Roy, Inc.,
Rochester, NY) by passing ozone gas through a 2-mm quartz ﬂow-through cell. An
extinction coefﬁcient of 3000 M'lcm'1 at 254 nm (Hoigné 1988) was used for
calculating ozone concentrations.

Dissolved ozone concentrations were also determined with an UV/VIS
spectrophotometer (Milton Roy Genesis-5, Milton Roy, Inc., Rochester, NY).

Samples were collected in a 1 cm quartz cell wrapped with paraﬁlm to prevent ozone

57

 

releasing to air. Then the quartz cell was placed immediately into the cell holder in
the spectrophotometer to measure the absorbance of ozone at 258 nm. An extinction
coefﬁcient of 2950 M'lcm'l at 258 nm (Bablon et al. 1991) was used for calculating

aqueous OZOI'IC concentrations.

3.3 MATHEMATICAL MODEL OF OZONE MASS TRANSFER

In the bench-scale ozonation/membrane system, the transfer of ozone from the
gas phase into the water can be described by the two-ﬁlm theory as shown in Figure
3.5. If there is no gas accumulation in the gas-liquid interface, the mass ﬂux from gas

phase must equal that from liquid phase and can be described as follows (Canale and

Weber 1972):

F:kgi(p_pi):_kLi(C_Ci) Eq.3-l

where F is the ﬂux [M/LZT]; kg,- is a convective ozone mass transfer coefﬁcient in the
gas phase [L/T]; p and p,- represents the ozone partial pressure in the gas phase and at
the interface [M/L3], respectively; ku is a convective ozone mass transfer coefﬁcient
in the liquid phase [UT], and C and C,- are concentrations of dissolved ozone in the
bulk liquid phase and at the interface [M/L3], respectively. M, L, and T are the

dimensions of mass, length, and time, respectively.

58

Gas Side Liquid Side
Film Film

\ Bulk Liquid

 

Bulk Gas

 

 

 

 

lnte rface

Figure 3.5 Two-ﬁlm theory.

Since it is impossible to measure the partial pressure of ozone and the
concentration of the dissolved ozone at the gas-liquid interface, the mass transfer
coefﬁcients are deﬁned based on the ozone partial pressure in the bulk gas phase and
dissolved ozone concentration in the bulk solution in equilibrium. Then, the ﬂux can

be expressed as:

F=kg(p—pe)=—kL(C—CS) Eq.3-2

where kg and kL are overall mass transfer coefﬁcients based on the bulk gas and liquid
phase concentrations, respectively; p, is the partial pressure of ozone in equilibrium

with C, and C5 is the concentration of dissolved ozone in equilibrium with p.

59

According to the Henry’s law, the relationship between the molar

concentration of ozone in water (C) and the ozone partial pressure (p), expressed as

atm, in the gas phase can be written as (Weber Jr. and DiGiano 1996):

p = H x C Eq.3-3

where H is the Henry’s constant in units of atm-L/mole.

From Eq. 3-3, the following relationships can be deduced in solutions at
equilibrium:

pi
CS = % Eq.3-5
[9
c = —H-‘?— Eq.3-6

Substituting Eq.3-4, Eq.3-5, and Eq.3-6 into Eq.3-1 and Eq. 3-2, and express kL in

terms of H, kg,- and k“, we can get the Eq.3-7:

 

Eq.3-7
L gi Li

For sparingly soluble gases such as ozone, whose Henry’s constant at 20 °C is 67.7
atm-L/mole (Weber Jr. and DiGiano 1996), H is sufﬁciently large that kL can be

approximated by kg. The mass transfer, therefore, is controlled by diffusion through

60

the liquid ﬁlm. Thus, the resistance of the gas ﬁlm is very small and can be neglected,
the liquid ﬁlm resistance is considered to dominate ozone transport in the membrane

contactor, and the Eq.3-7 can be rewritten as:

1+ 2 2.1— Eq.3-8
1. Li

Thus, the overall mass transfer coefﬁcient can be described by the liquid ﬁlm mass
transfer coefﬁcient.
The conﬁned volume of the ozonation/membrane system used to evaluate the
ozone mass transfer mathematical model is depicted in Figure. 3.6.
The rate of mass transfer of ozone in the aqueous phase in the

ozonation/membrane system can be described by the following equation.

V£=Q

d, in in—Q QSC+kLA(CS—C)—kVC

C _
011! out

Eq.3-9

where C is the dissolved ozone concentration in the system [M/L3]; C 5 is the
equilibrium dissolved ozone concentration in the system [M/L3]; C ,-,, and C0“, are the
dissolved ozone concentration in water entering and leaving the system [M/L3],
respectively; Q,-,,, QM, and Q, are the volumetric ﬂow rate of water entering, leaving
the system and sampling form the system [L3/T], respectively; A is the interfacial area
through which mass transfer occurs [L2]; V is the volume containing interfacial area

[L3]; k is the ozone decomposition rate in water [UT].

61

Q.

l.

1,c

Qout’ Cout Q ,2, C77

g1

 

 

 

 

 

 

 

Figure 3.6 The ozonation/membrane system for ozone mass transfer modeling.

The dissolved ozone concentration in water entering the system was zero, thus

C,-,, = 0. Therefore, the Eq.3-9 can be simpliﬁed and rearranged as:

Let,

dC A l

a_t.=kLF(CS—C)—;(Q0mC0m +QSC)—kC Eq.3-10
dC l
03 _ E+?(Qoutcout +QSC)+kC Eq.3-ll

62

Then Eq.3-10 can be rewritten as:

z =k —u7

03 l. V —C) Eq.3-12

S

Therefore, the ozone volumetric mass transfer coefﬁcient can be obtained by plotting
203 against (Cs-C).

Let a = A/ V, then Eq.3-10 can be rewritten and rearranged as:

dC
_: —k
dz ( L

1

S —7QoutCout) Eq.3-l3

Q5 kc: k C
a——V——- ) +( La

where a is the speciﬁc interfacial area [LZ/L3].

Let,

A k QS 13314

"La‘77‘ q'
and
1

B —kLaCS —;Q0utC0m Eq.3-15
Substituting Eq.3-14 and Eq.3-15 into Eq.3-13 yields:

id€=A*C+B Eq.3-16

63

Let,

D = C + — Eq.3-17
Substituting Eq.3-l7 into Eq.3-16, rearranging and integrating yields,
jidD=Ajdt Eq.3-l8
D

where the boundary conditions are: D0 = C0+B/A at t = 0 and D, = C,+B/A at any

later time, t. Integrating Eq.3-18 with the boundary conditions given, yield:
B At B
Ct = C +— *e -— E .3-19
( ) ( 0 A) A q

Substituting Eq.3-14 and Eq.3-15 into Eq.3-19 yields an equation for the

dissolved ozone concentration at any time in the ozonation/membrane system,

 

 

1 Q5 1
kLaCS -jT/‘Qoutcout (—kLa—T—k)t kLaCS _I/_QoutCout
C(l) = (C0 + )e _
Qs Qs
-k a-—— —k a————
L V L V
Eq.3-20

The ozone volumetric mass transfer coefﬁcient (kLa) was calculated and
reported rather than the overall mass transfer coefﬁcient (kL) because of the difﬁculty

in measuring the speciﬁc interfacial area (a) in the ozonation/membrane system.

64

3.4 RESULTS AND DISCUSSION

3.4.1 Permeate Flux Enhancement Experiment

The experimental results obtained from this study in which ozone was injected

into the tubular ceramic membrane are summarized in Table 3.3. A plot of permeate

ﬂux vs. ozone gas ﬂow rate for water ﬂow rates varied from 2 to 5 L/min is shown in

Figure 3.7.

Table 3.3 Effects of water ﬂow rate and gas ﬂow rate on permeate ﬂux.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water ﬂow Ozone gas ﬂow rate Permeate ﬂux Increased permeate
rate (LPM) (mL/min) (mL/hr) ﬂux (%)
0 49 0
100 55 12.2
2 200 60 22.4
300 64 30.6
400 76 55.1
500 80 63.3
0 82 0
100 90 9.8
3 200 100 22.0
300 106 29.3
400 115 40.2
500 123 50.0
0 140 0
100 147 5.0
4 200 156 11.4
300 163 16.4
400 172 22.9
500 178 27.1
0 184 0
100 199 8.2
5 200 210 14.1
300 218 18.5
400 228 23.9
500 235 27.7

 

 

 

 

 

 

300 ' +4 LPM
25o - —o—2LPM

200 - M

Egg/d
100 ;

 

 

 

 

Permeate ﬂux (mL/hr)

 

 

 

 

O 100 200 300 400 500
Ozone gas ﬂowrate (mL/min)

Figure 3.7 Effects of ozone gas ﬂow rate on the permeate ﬂux. Operating conditions:
gaseous 0;; concentration, 2.5 g/m3; water temperature, 27 °C; water ﬂow
rate varied from 2 LPM (L/min) to 5 LPM.

As shown above, increase in permeate ﬂux can be achieved by either
increasing water ﬂow rate or increasing ozone gas ﬂow rate. This result is consistent
with the observation of other researchers (Cui and Wright 1994; Ducom et al. 2002;
Laborie et a1. 1997; Laborie et al. 1998; Mercier et al. 1995). The increase in water
ﬂow rate and ozone gas ﬂow rate are thought to increase the transmembrane pressure,
which results in the increase of permeate ﬂux. Cheng and Wu (2003) reported that
the permeate ﬂux increases with the increase in water ﬂow rate, gas ﬂow rate, and the
transmembrane pressure. Cui and Wright (1994) concluded that the injected air
promotes turbulence near the membrane surface, increasing the superﬁcial cross-ﬂow
velocity of the ﬂuid, suppressing the polarisation layer and enhancing the
ultraﬁltration process. In this study, the rates of permeate ﬂux increased by 5.0 and
63.3 %, depending on the system operating conditions. In addition, it is interesting to

notice that at low water ﬂow rates, the effect of ozone addition on permeate ﬂux was

66

more signiﬁcant than that at higher water ﬂow rates. Other researchers observed

similar trend (Cui and Wright 1994; Laborie et al. 1998; Mercier et a1. 1995).

3.4.2 Decomposition of Ozone in the Proposed System

Although the DDI water acidiﬁed to pH 2 with sulfuric acid was used in this
study to minimize the effect of dissolved ozone decomposition on the mass transfer
study, experiments of ozone decomposition in the novelly designed system were
conducted to examine its effect on the rate constant of the decomposition of dissolved
ozone. Decomposition rate constants of dissolved ozone at three different water
temperatures are summarized in Table 3.4. Experimental results shown in Figures 3.8
to 3.10 indicate that the decomposition of dissolved ozone follows ﬁrst-order kinetics

and the decomposition rate can be expressed as:

__ = kd [03] Eq.3-21

As seen in Table 3.4, the rate of ozone decomposition increased as the water
temperature increased. The decomposition rate constant increased by ca. 180% and
ca. 178% as the water temperature increased from 10 °C to 20 °C and from 20 °C to
30 °C, respectively. It appears that water temperature is an important factor in this
proposed system and the decomposition rate constant can not be neglected even at
low pH level when using the ozonation/membrane system. However, several
researchers observed that the decomposition rate of dissolved ozone in acidic
conditions is quite small under water temperature at 20 °C (Gurol and Singer 1982;

Sotelo et a1. 1987; Sugimitsu et al. 1989). Therefore, the increase of decomposition

67

rate constant obtained in this study can not be explained only by the factor of
increasing water temperature. Sotelo et al. (1987) observed that the decomposition
rate of dissolved ozone increased when the agitation speed in the ozone contactor (a
bubble column) increased. They explained that this was probably due to the
occurrence of ozone desorption. In the proposed system used in this study, water was
circulated by a centrifugal pump and the transportation of water through the chamber
of pump head could offer a very violent agitation. Therefore, the desorption of
dissolved ozone could contribute to the decomposition of dissolved ozone in this

bench-scale system.

Table 3.4 Decomposition of dissolved ozone in DDI water acidiﬁed to pH 2 under
different water temperatures in the proposed system.

 

 

 

 

 

Water temperature Decomposition rate constant R2
1°C) (kd. min")
10 0.0132 0.9979
20 0.0240 0.9938
30 0.0427 0.9946

 

 

 

* water ﬂow rate: 3.5 L/min.

68

 

 

 

 

Q 0.
0')
E o.
8'
8 0.4
8 02 y=0.8817e'0'°132x
R2=O.9979
0 - J
o 15 30 45 60
Time (min)

Figure 3.8 Ozone decomposition in DDI water for ﬁrst-order kinetic at 10 °C.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM.

 

 

 

’3
B)
E,
2'
O
O
5 . = -0.024x
D 02 y 02.6225e
0.1 R =0.9938
0 l
0 10 20 30 40

Time (min)

Figure 3.9 Ozone decomposition in DDI water for ﬁrst-order kinetic at 20 OC.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM.

69

    

 

 

t y=0.4183e'°'0427" :
Q 0'4 ‘ R2=0.9946
C)
E- 0.3
8
8 0.2
6
D 0.1
O 1
o 5 1o 15 2o 25

Time (min)

Figure 3.10 Ozone decomposition in DDI water for ﬁrst-order kinetic at 30 oC.
Operating conditions: pH 2; water ﬂow rate, 3.5 LPM.

3.4.3 Ozone Mass Transfer Coefﬁcient

The effects of variation in the ozone gas ﬂow rate, water ﬂow rate, inlet ozone
concentration, and water temperature on the mass transfer of ozone in the
ozonation/membrane system were determined. The SYM mixer shown in Figure 3.2
was used in this study. Dissolved ozone concentrations in the water tank and in the
permeate and the permeate ﬂux were determined experimentally. These data and
Eq.3-12 were used to calculate the of volumetric mass transfer coefﬁcient of ozone
(Figure 3.11). It was assumed that the dissolved ozone was completely mixed in the
water tank.

Table 3.5 lists the ozone volumetric mass transfer coefﬁcients and dissolved
ozone concentrations in the water tank and permeate for different inlet gas ﬂow rates.
From the experimental results, the volumetric mass transfer coefﬁcient of ozone
increased as the inlet gas ﬂow rate increased. This is because the increase of gas ﬂow

rate further increases the turbulence on the gas-liquid interface, which increases the

70

interfacial area and results in an increase of ozone absorption in water. Many
researchers reported similar observations (Hsu et al. 2002; Roustan et a1. 1987;
Stankovic 1988; Wu and Masten 2001; Zhou and Smith 2000). Higher ozone
concentrations in the water tank were also observed for higher inlet gas ﬂow rates.
Since the amount of ozone added into the system is larger at higher gas ﬂow rates
than at lower gas ﬂow rates, the ozone concentrations in water tank follow similar

trends as KLa.

    

 

 

0.05 -~—~—-~,——--— -
:075mUmm . I
0-04 l - 50 mL/min
3 ' . i
a. 0.03 .-tZEmE/mm _ .
g l
0 i
c}, 0.02
o
0.01
O l
0 0.1 0.2 0.3 0.4 0,5
203 (mg/L x min")

Figure 3.1 1 Method to ﬁt the ozone volumetric mass transfer coefﬁcients by linear
regression at different ozone gas ﬂow rates.

71

Table 3.5 Calculated values of ozone mass transfer coefﬁcient (KLa) and
experimentally determined ozone concentrations under different gas ﬂow
rates. (Operating conditions: water temperature 20 C, water ﬂow rate 3.5

LPM, ozone dose 2.5 g Og/m3)

 

 

 

 

 

 

Inlet gas ﬂow rate KLa R;2 D03 in water tank D03 in permeate
(mL/min) (mini (mg/L)* (mg/L)*
25 0.049 0.98 0.462 0.21
50 0.075 0.95 0.560 0.24
75 0.109 0.94 0.602 0.31
100 0.138 0.84 0.661 0.38

 

 

 

 

 

* Experimental data.

Table 3.6 Calculated values of the ozone volumetric mass transfer coefﬁcient (KLa)
and experimentally determined ozone concentrations under different water
ﬂow rates. (Operating conditions: water temperature, 20 C; gas ﬂow rate,
100 mL/min; ozone dose, 2.5 g O3/m3)

 

 

 

 

 

Water ﬂow rate KLa R2 D03 in water tank D03 in permeate
(L/min) (min'l) (mg/L)* (mg/L)*
2.0 0.090 0.97 0.61 0.45
2.75 0.117 0.95 0.68 0.48
3.5 0.138 0.84 0.66 0.38

 

 

 

 

 

* Experimental data.

The calculated values of ozone volumetric mass transfer coefﬁcient at

different water ﬂow rates and other parameters are given in Table 3.6. As shown in

Table 3.6, the volumetric mass transfer coefﬁcient increases as the water ﬂow rate

increases. In this ozonation/membrane system, the water was circulated between

membrane unit and the water tank before it was ﬁltered out of the system, which

indicates that the water-gas contact time increases as the water ﬂow rate increases and

results in the increase of the ozone mass transfer rate. Several researchers reported

 

 

that the injection of air into the liquid stream promotes turbulence in the membrane
ﬁltration unit (Mercier et al. 1997). Therefore, the increase of turbulence by
increasing water ﬂow rate also contributes to the increase of ozone volumetric mass
transfer coefﬁcient.

Table 3.7 presents the effect of ozone dose on the ozone volumetric mass
transfer coefﬁcient and dissolved ozone concentration. It shows a trend that the value
of KLa ﬁrst increases with increasing ozone doses and then decreases after it reaches a
maximum value at ozone dose of 23 g/m3. It is apparent that at high ozone doses, the
diffusion of gaseous ozone becomes rate limiting, resulting in poor pass transfer
efﬁciencies. Therefore, the value of KLa decreases for ozone doses exceeding 23
g/m3. This also indicates that this bench-scale ozonation/membrane system has higher

ozone mass transfer efﬁciency at low ozone doses than that at high ozone doses.

Table 3.7 Calculated values of ozone mass transfer coefﬁcient (KLa) and
experimentally determined ozone concentrations under different ozone
doses. (Operating conditions: water temperature 20 C, water ﬂow rate 3.5
LPM, gas ﬂow rate 100 mL/min)

 

7

 

 

 

 

 

 

 

03 dose conc. KLa R“ D03 in water tank D03 in permeate
(2.2/ma min") (mg/L>* (mgmr
1.5 0.133 0.82 0.379 0.211
2.5 0.138 0.84 0.661 0.379
5.8 0.152 0.97 1.365 0.801
12.5 0.161 0.95 2.988 1.587
23.0 0.173 0.87 5.907 3.152
36.0 0.128 0.99 9.501 4.536

 

 

 

 

 

* Experimental data.

73

 

The ozone mass transfer coefﬁcients and dissolved ozone concentrations in
water at different water temperatures are listed in Table 3.8. It can be seen that the
higher the water temperature, the higher value of KLa. Temperature affects the ozone
mass transfer coefﬁcient by: 1) changing the ozone solubility, 2) changing the rate of
ozone molecular diffusion through the gas-liquid interface, and 3) changing the rate of
ozone decomposition. An increase in temperature results in a decrease in the ozone
solubility and an increase in the ozone molecular diffusion rate thus enhancing the
ozone concentration gradient between gas phase and aqueous phase. Also the
decomposition rate increases may increase ozone mass transfer. This promotes the
ozone mass transfer rate and results in an increase in the value of KLa. However, the
level of ozone mass transfer coefﬁcient only slightly increased for the range of
operating conditions studied. This is probably due to the low ozone dose (2.5 g/m3)
applied in this experiment or the conﬁguration of this bench-scale system. The
experimental results also indicate that the water temperature does not strongly affect
the Km value of ozone under the operating conditions used in this study.

Temperature also has strong inﬂuence on Henry’s constant. An increase in
temperature results in an increase in the Henry’s constant (Faust and Aly 1999; Roth
and Sullivan 1981). According to Equation 3-3, the ozone solubility decreases when
the Henry’s constant increases and the ozone partial pressure remains constant.
Consequently, the dissolved ozone concentration in the water tank and in the

permeate decreases.

74

Table 3.8 Calculated values of ozone mass transfer coefﬁcient (KLa) and
experimentally determined ozone concentrations under different water
temperatures. (Operating conditions: water ﬂow rate 3.5 LPM, gas ﬂow
rate 100 mL/min, ozone dose 2.5 g 03/m3)

 

 

 

 

 

Water temperature KLa R2 D03 in water tank D03 in permeate
(°C) (minl) ling/L)" (mg/LT“
10 0.135 0.92 0.881 0.571
20 0.138 0.84 0.661 0.379
30 0.141 0.92 0.404 0.232

 

 

 

 

 

* Experimental data.

3.4.4 Simple Y Inline Mixer vs. High Efﬁciency Inline Mixer

The SYM is a Y shape Pyrex glass tube without diffuser. The HEM is similar
to the SYM but with a build—in fused glass diffuser, which produces bubbles
approximately 40 to 50 pm in diameter. Table 3.9 lists the ozone volumetric mass
transfer coefﬁcients and the dissolved ozone concentrations in the water tank under
different operating conditions using the SYM and HEM. The effects of different
mixers on the dissolved ozone concentration in the water tank under different
operating conditions are shown in Figures 3.12 to 3.18.

As can be seen in the experimental results (Table 3.9), the values of KLa
obtained using HEM are generally higher than those obtained using SYM except
those of ozone doses between 5.8 and 36.0 g/m3. The HEM allows for the injection of
smaller-sized bubbles, which contain a higher interfacial area, into the feed water

stream than does the SYM. However, KLa values obtained using HEM were observed

75

 

only higher than those of SYM at lower ozone doses. It was also found that the

dissolved ozone concentrations in the water tank when using the HEM were less than

those with the SYM. In running the experiments, it was observed that with the HEM

some bubbles generated from the diffuser accumulated on the top of it and did not

mix with the feed water stream and move into the membrane module. This design

problem may result in the low ozone mass transfer at higher ozone doses and reduced

dissolved ozone concentrations in the water tank when using the HEM.

Table 3.9 Parameters comparison between Simple Y Mixer (SYM) and High

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Efﬁciency Mixer (HEM).
Simple Y Mixer High Efﬁciency Mixer
Water ﬂow KLa D03 in water KLa D03 in water
rate (LPM) (min’l) tank Qig/ L) (min'l) tank (mg/L)
3.5 0.138 0.661 0.162 0.560
Simple Y mixer High efﬁciency mixer
Ozone dose KLa D03 in water KLa D03 in water
(g/m3) (min'l) tank (mg/L) (min'l) tank (mg/LL
2.5 0.138 0.661 0.162 0.560
5.8 0.152 1.365 0.161 1.217
12.5 0.161 2.988 0.159 2.843
23.0 0.173 5.907 0.136 5.223
36.0 0.128 9.501 0.134 8.898
Simple Y mixer High efﬁciency mixer
Temperature KLa D03 in water KLa D03 in water
( C) (min'l) tank (mg/L) (min‘l) tank (mg/L)
10 0.135 0.881 0.148 0.780
20 0.138 0.661 0.162 0.560
30 0.141 0.404 0.179 0.342

 

 

 

 

 

76

 

 

0.8

  
 

 

g Simple Y Mixer

3’ 0.6 - C G O .. o
g 0.4— . .. .
8 ngh EffICIency Mixer
m 0.2 -

C

S 0.0 -

O

 

 

 

010 20 30 4O 50 60 7O
Time(min)

Figure 3.12 Comparison of dissolved ozone concentration using SYM and HEM
mixers. (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 2.5 g/m3)

 

  
  

 

2
S, Simple Y Mixer
E U n 0 a o
“T 1 3
8
8 High Efﬁciency Mixer
2 0.
o
N
O

 

 

 

010 20 30 40 5O 60 70

Time (min)

Figure 3.13 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 5.8 g/m3)

77

 

    
 
 

3

E5 3 , Simple Y Mixer

:5 2 ;

C . . . .

8 High EffiCiency Mixer
0) 1 J

8

N a

O 0

 

 

 

010 20 30 4O 50 60 70
Time (min)
Figure 3.14 Comparison of dissolved ozone concentration using SYM and HEM

mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 12.5 g/m3)

 

  
 

: 7

3, 5 Simple Y Mixer

E 5 «

E; 4 ‘ High Efficiency Mixer
0 3 ’

a) 2 ~

C

S 1 ‘

o 0 ~

 

 

 

010 20 30 40 50 60 70
Time(min)

Figure 3.15 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 23.0 g/m3)

78

A

DMD-036100

 

Simple Y Mixer

   
 

High Efficiency Mixer

Ozone Conc. (mg/L)

 

 

 

0 10 20 3O 40 50 60 70
Time (m in)
Figure 3.16 Comparison of dissolved ozone concentration using SYM and HEM

mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 20 °C; ozone dose, 36.0 g/m3)

 

    
 

A 2

d

0')

E . .

U. 1 . Simple Y Mixer
C

o

0 H' h Eff' ' M'
2 01 lg iCiency ixer
o

N

O

 

 

 

010 20 30 40 50 60 70
Time (min)

Figure 3.17 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 10 °C; ozone dose, 2.5 g/m3)

79

 

O
01

Simple Y Mixer

  
 
 

0.4 3

High Efficiency Mixer

Ozone Conc. (mg/L)
0 O O O
'o “.1 N 00

 

 

 

010 20 30 4O 50 60 70

Time (min)

Figure 3.18 Comparison of dissolved ozone concentration using SYM and HEM
mixers (Operating conditions: water ﬂow rate, 3.5 LPM; gas ﬂow rate,
100 mL/min; water temperature, 30 °C; ozone dose, 2.5 g/m3)

3.4.5 Model Validation

The mathematical model developed in Section 3.3 was validated by comparing
with the experimental data obtained in this study. Figures 3.19 to 3.23 present some
of the comparison of simulations obtained using the mathematical model (Eq.3-20)
and experimentally determined data for the dissolved ozone concentration in the water
tank. Similar results are obtained for the other experimental conditions. The results
indicate that the concentrations of dissolved ozone predicted by the model were
slightly greater than the levels of experimental observation in the ﬁrst 30 minutes of
ozonation but were slightly lower after that. This may be due to the turbulent mixing

of gas ozone and water in the early operating stage, which depresses the adsorption of

80

ozone, and a gradually accumulation of dissolved ozone in the aqueous phase in the

latter stage. Table 3.10 presents the steady-state concentrations of dissolved ozone

obtained from experiments and model prediction under different operating conditions.

The relative percentage differences between these two data are also provided. It

shows good agreement between the experimental data and model predictions.

0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1

Dissolved 03 Conc. (mg/L)

+ Experimental Data
—--— Model prediction

1

 

0 20 4O 60

Time (min)

Figure 3.19 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
2.75 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 2.5 g/m3.

81

0.8
0.7
0.6
0.5
0.4

0.2
0.1

Dissolved 03 Cone. (mg/L)

0.3 .

 

+ Experimental Data
-»-— Model prediction

1 i

20 4O 60
Time (min)

Figure 3.20 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone

dose, 2.5 g/m3.

9.0.0999
Amuse-(no:

Dissolved 03 Conc. (mg/L)

0

l

 

 

+ Experimental Data
—---— Model prediction

20 40 60
Time (min)

Figure 3.21 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
3.5 LPM; gas ﬂow rate, 50 mL/min; water temperature, 20 °C; ozone

dose, 2.5 g/m3.

82

Dissolved 03 Conc. (mg/L)

1.5
1.2
0.9
0.6

0.3

 

 

+ Experimental Data
+ Model prediction 3

20 4O 60
Time (min)

Figure 3.22 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,
3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 5.8 g/m3.

Dissolved 03 Cone. (mg/L)

A—L

 

 

2 ..
0
8 r/‘fl
6 //
4 f,
2 + Experimental Data
0 - -- Model prediction
0 20 40 60
Time (min)

Figure 3.23 Comparison of experimental data and modeling data of dissolved ozone
concentration in the water tank. Operating conditions: water ﬂow rate,

3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 °C; ozone
dose, 36.0 g/m3.

83

Table 3.10 Comparison of model predictions and experimental data at steady-state of

dissolved ozone concentration in the pH 2 DDI water under various
operating conditions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water ﬂow rate Model D03 cone. Exper. D03 cone. Difference
(L/min)l (mg/L) (mg/L) (%)
2 0.52 0.61 14.0
2.75 0.62 0.68 8.2
3.5 0.63 0.66 4.0
Gas ﬂow rate Model D03 conc. Exper. D03 conc. Difference
(mL/min)2 (mg/L) (mg/L) (%)
25 0.38 0.46 18.5
50 0.49 0.56 12.2
75 0.55 0.60 8.7
100 0.63 0.66 4.0
Ozone dose Model D03 conc. Exper. D03 conc. Difference
(any (mg/LL (mg/L) (%)
1.5 0.37 0.38 3.0
2.5 0.63 0.66 4.0
5.8 1.32 1.36 3.1
12.5 2.91 2.99 2.5
23.0 5.75 5.90 2.7
36.0 9.10 9.5 4.3
Water temperature Model D03 conc. Exper. D03 conc. Difference
(°C)4 (mg/L) (mg/L) (%)
10 0.80 0.88 9.0
20 0.63 0.66 4.0
30 0.35 0.40 12.6

 

 

 

 

1 Gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3; water temperature, 20 °C.

2 Water ﬂow rate, 3.5 LPM; ozone dose, 2.5 g/m3; water temperature, 20 °C.

3 Water ﬂow rate, 3.5 LPM; gas ﬂow rate, 100 mL/min; water temperature, 20 0C.
4 Water ﬂow rate, 3.5 LPM; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

84

 

3.4.6 Comparison with Other Studies

Many factors, such as the conﬁguration of the reactor, the interfacial area, the
characteristics of water, the gas ﬂow rate, the water ﬂow rate, and temperature, can
inﬂuence the volumetric mass transfer coefﬁcient of ozone. Since the value of KLa is
also related to the operating conditions, water source, and the design of the contactor,
it is not possible to compare the mass transfer coefﬁcient between research projects
without considering these parameters. Therefore, the experimental results and
operating conditions from this study have been compared to those in the literature
(Table 3.11).

The value of KLa calculated from this study is the smallest among all values
presented in Table 3.11. But as mentioned above, other parameters need to be taken
into account. Many researchers reported that an increase of gas ﬂow rate results in
an increase of KLa value (Hsu et al. 2002; Roustan et al. 1987; Stankovic 1988; Zhou
and Smith 2000). This is due to the greater turbulence achieved at higher ﬂow rates,
and a concomitant thinning of the gas-liquid ﬁlm. As can be seen in Table 3.11, the
gas ﬂow rate used in this study is much smaller than the others. In addition, our
system is the only one that does not have physical diffuser, which makes produce ﬁne
bubbles impossible. These two factors may result in lower values of KLa. However,
the advantages of using low gas ﬂow rate and no diffuser in the system are lowering
operating cost and maintenance cost.

Wu and Masten (2001) investigated the KLa obtained in a semi-batch stirred
reactor by using a smaller 1.5 liter water tank, higher gas ﬂow rate, higher ozone dose,
and ﬁne bubble diffuser. The KLa values from their study are only slightly greater
than those obtained in this study. Bin and colleagues (2001) used a bubble column

(0.15 m ID, 5.5 m high) as an ozone contactor and applied high gas ﬂow rate and ﬁner

85

bubbles. It is noticeable that they used pH 7 to 8 water and did not consider the effect
of ozone auto-decomposition on the Km. As such, the KLa values they obtained do
not reﬂect the real values for their experimental system.

Zhou and Smith (2000) treated pulp mill wastewater using a bubble column.
Since the chemical oxygen demand (COD) of the wastewater is 609 mg/L, it is
reasonable to assume that KLa values they obtained were predominately due to the
reaction of ozone with the COD. Roustan et al. (1996) applied the highest gas ﬂow
rate and water ﬂow rate among all studies listed in Table 3.1 1 to treat tap water at pH
7.5 to 8.1. Although high KLa values were obtained from their study, their KLa values
were not much higher than those obtained by Bin et al. (2001) who used lower water
ﬂow rate and applied ﬁner gas bubbles. Sheffer and Esterson (1982) found that the
mass transfer coefﬁcient of ozone in tap water (pH 7.5 to 8.1) increased as the water
ﬂow rate increased. Thus, high gas ﬂow rate, high water ﬂow rate, and high pH of
the water applied in other studies resulted in KLa values greater than what were
obtained in this study.

Shen and Semmens (1990) used the membrane as an ozone contactor rather
than as a physical barrier to investigate the ozone mass transfer. This was done
because of the membrane’s high interfacial area. Although, they reported the highest
KLa value of ozone in Table 3.11, almost half (44%) of the ozone dosed in their
experiments was estimated to be consumed by the experimental system. This is
because of the membrane used in their experiment. The membrane was made of
polypropylene, which was attacked by ozone and seriously damaged during the
experiment.

Bollyky (1981) indicated that the general requirement for an ozone contactor

include: 1) low cost, 2) minimal energy consumption, 3) minimal maintenance, 4)

86

automated operation, 5) efﬁcient mass transfer, 6) simple design, 7) preferably no
moving parts, 8) nothing to clog, 9) ﬂexibility to respond to the variations in operating
parameters and water quality. The ozonation/membrane system studied can achieve
reasonable ozone mass transfer, requires less gas ﬂow rate and water ﬂow rate (than
comparable systems), and does not require diffuser, thereby reducing clogging.

These may result in low energy consumption, low maintenance, and have economical

advantage over other ozone contactors.

87

Table 3.1 1 KLa values and experimental conditions from this study and literature on

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ozone mass transfer
Reference This study Wu and Bin et a1 Zhou and Roustan Shen et a1
Masten (2001) Smith et a1 (1990)
(2001) (2000) (1996)
KLa 0.05 ~ 0.12 ~ 0.18 ~ 0.5 ~ 1.5 0.31~ 3.4 ~ 4.4
(min'l) 0.17 0.39 1.2 1.15
Temp 10, 20, 30 14,22, 30 20~23 21 i2 12~19 25
(°C)
Gas
ﬂowrate 0.025 ~ 0.1 ~ 0.5 5 ~ 20 0.5 ~ 2.0 2 ~ 22 0.22
(L/min) 0.1
Water 0.063 ~
ﬂowrate 2.0 ~ 3.5 NA 1.7 ~ 7.5 NA 5 ~ 22 0.269
(L/min)
Type of Ozone/ Semibatch Bubble Bubble Bubble Hollow
Reactor membrane (stirred) column column column ﬁber
Diffuser Symple Y Fritted Porous Porous Porous Membrane
Type mixer w/o glass diffuser diffuser diffuser
diffuser or venturi
injector
Bubble NA < 1mm 100 ~ NA 1.75 ~ NA
Size 150nm 5.25 mm
pH 2.0 2.1 7 ~ 8 NA 7.5 ~ 8.1 2
Water DDI DDI D1 or tap Lagoon Tap DI
Source (3.2 L) (1.5 L) water efﬂuent water (1.5 L)
(Volume) (97 L) (4 L) (NA)
03 Dose 1.5 ~ 36.0 46 8 ~ 60 0 ~ 134 10 ~ 60 7.86 ~ 9.97
(ta/"13)

 

 

88

 

3.5 CONCLUSIONS

1. Ozone gas was injected into the feed water stream of the tubular ceramic
membrane to investigate its effect on the permeate ﬂux. It was observed that an
increase in permeate ﬂux can be achieved by either increasing the gas ﬂow rate or
increasing the water ﬂow rate. The increase in the permeate ﬂux ranged from 5.0
to 63.3 %, depending on the system operating conditions. In addition, the effect
of ozone addition on permeate ﬂux at low water ﬂow rate is more signiﬁcant than
that at high water ﬂow rate.

2. The decomposition of dissolved ozone follows the ﬁrst-order kinetics. The rate of
ozone decomposition increased as the water temperature increased. The
experimental results showed that the decomposition rate constant can not be
neglected even at low pH level when using the ozonation/membrane system.

3. A mathematical model was developed and used to calculate the volumetric mass
transfer coefﬁcient of ozone in the ozonation/membrane system. This model was
also veriﬁed to be able to predict the dissolved ozone concentration in the water
tank.

4. The experimental results show that as the water ﬂow rate increases, the volumetric
mass transfer coefﬁcient increases. The increase of ozone gas ﬂow rate also
increases the volumetric mass transfer coefﬁcient of ozone.

5. Increases in the ozone dose resulted in an increase in the value of KLa. However,
the KLa value ﬁrst increased with increasing ozone doses and then decreased after
it reached a maximum value.

6. The increase of KLa value is limited as the water temperature increased from 10
0C to 30 °C. This is probably due to the low ozone dose (2.5 g/m3) applied in this

experiment or the conﬁguration of this bench-scale system.

89

7. Two different mixers, SYM and HEM, were used to investigate their effect on the
performance of the system. The HEM produces ﬁner bubbles than does the SYM,
which resulted in greater values of KLa using the HEM. However, with the HEM,
the dissolved ozone concentration was less than that obtained with the SYM.
With the HEM, it was found that some bubbles accumulated on the top of the
diffuser and did not mix with the feed water stream and pass into the membrane
module. This design problem may result in the lower dissolved ozone
concentrations observed.

8. Considering some important factors, such as gas ﬂow rate, water ﬂow rate, and
the characteristics of the source water, which inﬂuence the KLa value, the
ozonation/membrane system developed in this study can achieve reasonable ozone
mass transfer by requiring lesser ozone dose, gas ﬂow rate, and water ﬂow rate
than comparable systems. This system has the added advantage that it does not

include diffuser thus reducing clogging problems.

3.6 REFERENCES

Amy, G. L., Kuo, C. J ., and Sierka, R. A. (1988). "Ozonation of Humic Substances —
Effects on Molecular-Weight Distributions of Organic-Carbon and

Trihalomethane Formation Potential." Ozone-Science & Engineering, 10(1),
39-54.

Bablon, G., Bellamy, W. D., Bourbigot, M.-M., and et. al. (1991). "Fundamental
Aspects." Ozone in Water Treatment: Application and Engineering, D. R.
Brink, ed., Lewis Publishers, Chelsea, MI, 11-132.

Barberis, K., and Howarth, C. R. (1991). "Reactivity Studies in the Ozonolysis of

Pollutants Using a Perforated Spinning Disk Reactor to Enhance Mass-
Transfer." Ozone-Science & Engineering, 13(5), 501-519.

90

Bin, A. K., Duczmal, B., and Machniewski, P. (2001). "Hydrodynamics and ozone
mass transfer in a tall bubble column." Chemical Engineering Science, 56(21-
22), 6233-6240.

Bollyky, L. J. (1981). "The Mass-Transfer of Ozone into Water - Energy-
Requirements State of the Art." Ozone-Science & Engineering, 3(3), 181-210.

Canale, R. P., and Weber, W. J. J. (1972). "Aeration and Gas Transfer."
Physicochemical Processes - For Water Quality Control, W. J. Weber, ed.,
John Wiley & Sons., New York, 503-531.

Cheng, T. W., and Wu, J. G. (2003). "Quantitative ﬂux analysis of gas - Liquid two-
phase ultraﬁltration." Separation Science and Technology, 38(4), 817-835.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

Cui, Z. F., and Wright, K. I. T. (1994). "Gas-Liquid 2-Phase Cross-F low
Ultraﬁltration of Bsa and Dextran Solutions." Journal of Membrane Science,
90(1-2), 183-189.

Ducom, G., Matamoros, H., and Cabassud, C. (2002). "Air sparging for ﬂux
enhancement in nanoﬁltration membranes: application to O/W stabilised and
non—stabilised emulsions." Journal of Membrane Science, 204(1-2), 221-236.

Faust, S. D., and Aly, O. M. (1999). Chemistry of Water Treatment 2ND Edition,
Lewis Publishers, Ann Arbor, MI.

Gurol, M. D., and Singer, P. C. (1982). "Kinetics of Ozone Decomposition - a
Dynamic Approach." Environmental Science & Technology, 16(7), 377-383.

Hashino, M., Mori, Y., Fujii, Y., Motoyama, N., Kadokawa, N., Hoshikawa, H.,
Nishijima, W., and Okada, M. (2000). "Pilot plant evaluation of an ozone-

microﬁltration system for drinking water treatment." Water Science and
Technology, 41(10-1 1), 17-23.

Hoigne, J. (1988). "The Chemistry of Ozone in Water." Process Technologies for
Water Treatment, Plenum Publishing Corp., New York.

Hsu, Y. C., Chen, T. Y., Chen, J. H., and Lay, C. W. (2002). "Ozone transfer into
water in a gas-inducing reactor." Industrial & Engineering Chemistry
Research, 41(1), 120-127.

Janknecht, P., Wilderer, P. A., Picard, C., Larbot, A., and Sarrazin, J. (2000).
"Investigations on ozone contacting by ceramic membranes." Ozone-Science
& Engineering, 22(4), 379-392.

Kim, J. 0., Somiya, 1., and Fujii, S. (1999). "Fouling Control of Ceramic Membrane

in Organic Acid Fermenter by Intermittent Ozonation." Proceedings of the
14th Ozone World Congress, Dearbom, MI, 131-143.

91

Klijn, R. B., van der Meer, W. G. J ., Vriezen, H., and van Ekkendonk, F. H. J. (2000).
"Surface Water Treatment with Zenon Microﬁltration Membranes:
Minimisation of Energy and Chemical Use." Proceedings of the Conference

on Membranes in Drinking and Industrial Water Production, L'Aquila, Italy,
557-563.

Laborie, S., Cabassud, C., Durand-Bourlier, L., and Laine, J. M. (1997). "Flux
Enhancement by a Continuous Tangential Gas Flow in Ultraﬁltration Hollow
Fibres for Drinking Water Production: Effects of Slug Flow on Cake
Structure." Filtration & Separation, 34(8), 887-891.

Laborie, S., Cabassud, C., Durand-Bourlier, L., and Laine, J. M. (1998). "Fouling
Control by Air Sparging Inside Hollow Fibre Membranes-Effects on Energy
Consumption." Desalination, 1 18(1-3), 189-196.

Ledakowicz, S., Maciejewska, R., Perkowski, J., and Bin, A. (2001). "Ozonation of
Reactive Blue 81 in the bubble column." Water Science and Technology, 44(5),
47-52.

Mercier, M., Fonade, C., and Lafforgue-Delorme, C. (1995). "Inﬂuence of the Flow
Regime on the Efﬁciency of a Gas-Liquid Two-Phase Medium Filtration."
Biotechnology Techniques, 9(12), 853-858.

Mercier, M., Fonade, C., and Lafforgue-Delorme, C. (1997). "How slug ﬂow can
enhance the ultraﬁltration ﬂux in mineral tubular membranes." Journal of
Membrane Science, 128(1), 103-113.

Richards, D. A., F leischman, M., and Ebersold, L. P. (1976). "Mass Transfer
Coefﬁcients for the Ozone-Water System." AIChE Symposium Series, 73(166),
213-224.

Richardson, S. D., Thruston, A. D., Caughran, T. V., Chen, P. H., Collette, T. W.,
Floyd, T. L., Schenck, K. M., Lykins, B. W., Sun, G. R., and Majetich, G.
(1999). "Identiﬁcation of new ozone disinfection byproducts in drinking
water." Environmental Science & Technologi, 33(19), 3368-3377.

Roth, J. A., and Sullivan, D. E. (1981). "Solubility of Ozone in Water." Industrial &
Engineering Chemistry Fundamentals, 20(2), 137-140.

Roustan, M., Duguet, J. P., Brette, B., Brodard, E., and Mallevialle, J. (1987). "Mass
Balance Analysis of Ozone in Conventional Bubble Contactors." Ozone
Science & Engineering, 9(3), 289-297.

Roustan, M., Wang, R. Y., and Wolbert, D. (1996). "Modeling hydrodynamics and
mass transfer parameters in a continuous ozone bubble column." Ozone-
Science & Engineering, 18(2), 99-1 15.

Shanbhag, P. V., Guha, A. K., and Sirkar, K. K. (1998). "Membrane-based ozonation
of organic compounds." Industrial & Engineering Chemistry Research, 37(1 1),
4388-4398.

92

Sheffer, S., and Esterson, G. L. (1982). "Mass-Transfer and Reaction-Kinetics in the
Ozone Tap Water- System." Water Research, 16(4), 383-389.

Shen, Z. S., Semmens, M. J ., and Collins, A. G. (1990). "A Novel-Approach to Ozone
Water Mass-Transfer Using Hollow Fiber Reactors." Environmental
Technology, 1 1(7), 597-608.

Sotelo, J. L., Beltran, F. J., Benitez, F. J., and Beltranheredia, J. (1987). "Ozone
Decomposition in Water - Kinetic-Study." Industrial & Engineering
Chemistry Research, 26(1), 39-43.

Stankovic, I. (1988). "Comparison of Ozone and Oxygen Mass-Transfer in a
Laboratory and Pilot-Plant Operation." Ozone-Science & Engineering, 10(3),
321-338.

Sugimitsu, H., Okazaki, S., and Moriwaki, T. "Decomposition of Ozone in Aqueous
Solutions." Ninth Ozone World Congress, New York, 97-105.

Verberk, I. Q. J. C., Worm, G. I. M., Futselaar, H., and van Dijk, J. C. (2000).
"Combined Air-Water Flush in Dead-End Ultraﬁltration." Proceedings of the
Conference on Membranes in Drinking and Industrial Water Production,
L'Aquila, Italy, 655-663.

Weber Jr., W. J ., and DiGiano, F. A. (1996). "Microtransport Processes." Process
Dynamics in Environmental Systems, John Wiley & Sons., New York, 147-
230.

Wright, P. C., Meeyoo, V., and Soh, W. K. (1998). "A study of ozone mass transfer in
a cocurrent downﬂow jet pump contactor." Ozone-Science & Engineering,
20(1), 17-33.

Wu, J. J ., and Masten, S. J. (2001). "Mass transfer of ozone in semibatch stirred
reactor." Journal of Environmental Engineering-Asce, 127(12), 1089-1099.

Zhou, H. D., and Smith, D. W. (2000). "Ozone mass transfer in water and wastewater
treatment: Experimental observations using a 2D laser particle dynamics
analyzer." Water Research, 34(3), 909-921.

93

CHAPTER 4
THE EFFECT OF OZONATION/MEMBRANE PROCESS ON THE

CHARACTERISTICS OF NATURAL ORGANIC MATTER

4.1 INTRODUCTION

Natural organic matter (NOM), which is made up of a heterogeneous mixture
of organic compounds and occurs either as results of natural processes or human
activities in the environment, is ubiquitous in aquatic systems. The composition of
NOM can be divided into two fractions, humic substances, which consist of humin,
fulvic acid and humic acid, and non-humic substances. Fractionation is accomplished
based on the difference of the aqueous solubilities of the different components at
different pH values (MacCarthy and Suffet 1989; Malcolm 1985). The molecular
weight distribution of NOM ranges from a few hundred daltons for fulvic acids to tens
of thousands daltons for the humic acids (Manahan 1993).

In water treatment, the presence of NOM is problematic because the reaction
of NOM with disinfectants results in the formation of DBPs. Humic substances are
recognized as the primary precursors of chlorination byproducts (Collins et al. 1986;
lchihashi et a1. 1999; Manahan 1993; Reckhow et a1. 1990). However, non-humic
substances also result in the formation of a signiﬁcant percentage of the regulated or
potentially regulated DBPs. In addition, the non-humic fraction of NOM contributes
to a higher fraction of BDOC, which may cause bacteria regrowth in water
distribution systems (Mogren et al. 1990; Yavich 1998).

BDOC is a fraction of DOC that can be mineralized by heterotrophic
microorganisms. The ozonation of water has been found to increase the regrowth

potential of bacteria by converting part of the non-biodegradable carbon into BDOC

94

(Kuo et al. 1977; Narkis and Schneiderrotel 1980; Van der kooij et al. 1989; Yavich
1998). Moreover, the growth of bacteria in the distribution system may pose tastes,
odor, color, and pipe corrosion problems (Levy et a1. 1986).

The measurement of BDOC has been used predominantly to indicate the
biostability of water and to evaluate the performance of biotreatment processes (e.g.,
biological ﬁltration). One approach to measuring the biodegradability of water is to
assess the AOC, which represents the regrowth potential of heterotrophic bacteria in
water and was developed by van der Kooij, Visser, and Hijnen (1982). The other
approach is based either on the reduction of DOC during incubation or on an estimate
of the ﬂux of organic matter utilized by bacteria and was developed by Servais, Billen,
and Hascoét (1987). The correlation between BDOC and AOC depends
predominately on the characteristics of the tested water (Chamock and Kjonno 2000;
Siddiqui et al. 1997).

The application of ozonation prior to membrane ﬁltration effectively reduces
membrane fouling and enhances permeate ﬂux (Hashino et al. 2000; Hyung et al.
2000; Kim et al. 1999). Ozonation of water containing NOM results in more lower
molecular weight material, a decrease in the concentration of UV absorbing
compounds, an increase in the biodegradability, a reduction in the formation of DBPs,
and an increase in the formation of more polar compounds. In addition, membrane
ﬁltration provides a good physical barrier to turbidity, pathogens, and NOM, although
the effectiveness of the ﬁltration process depends very much on the size of the
membrane ﬁlters employed. Based upon earlier research, we hypothesize that the
ozonation/membrane process will effectively remove DBP precursors, reduce DBPs
formation, and inactivate pathogens and produce biologically stable water if followed

by biotreatment.

95

The objective of this study was to investigate the effect of the combined
ozonation/membrane process on the characteristics of NOM and to evaluate the
biodegradability of the treated water. Several experiments were conducted and their
results are discussed. We considered: 1) the effect of ozonation on the permeate ﬂux;
2) the effect of ozonation on DOC, UV-254, humic substances, non-humic substances,
AOC, and BDOC; 3) the effect of biotreatment on the raw water and treated water
quality following the ozonation/membrane process; 4) and the effect of operational
parameters, such as water ﬂow rate, gas ﬂow rate, ozone dose, and water temperature,
which affect the ozone mass transfer, on the efﬁcacy of the ozonation/membrane

system.

4.2 MATERIALS AND METHODS
4.2.1 Experimental Setup and Protocol

A specially designed ozonation/membrane system was used to investigate the
permeate ﬂux recovery as a result of ozonation and variation in the NOM resulting
from the treatment process. NOM was assessed in terms of DOC, UV-254, humic
substances, non-humic substances, AOC, and BDOC. A schematic diagram of this
system is shown in Figure 4.1. Tubular ceramic membranes (Clover-leaf design
(containing three channels), CéRAM Inside, Tami North America, St. Laurent,
Quebec, Canada) with a molecular weight cut-off of 15 kD were used in the system.
The characteristics of the laboratory-scale ultraﬁltration membrane are summarized in
Table 4.1. Teﬂon tubing and stainless steel joints and valves were used throughout
the system. Other components included: 3.5-liter and 1.5-liter water-jacked glass

reservoirs made of Pyrex glass, and a simple Y inline mixer (Ozone Service, Burton,

BC, Canada).

96

Waste Gas

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Moisture Spectrophotometer
Adsorbent
. I: Ozone -1 I
Generator L—T—l
KLJ 1
Oxygen Cylinder Kl

 

 

 

 

 

 

 

E
__ @—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V
2
ii
/\ M Membrane
~® i
E? PermeateJIL
Water.
Tank a (P) Presuregauge
® Flow meter
0
Q Pump

 

 

 

X Valve

Figure 4.1 Experimental setup for the ozonation/membrane process.

97

Table 4.1 Details of the ceramic membrane.

 

 

External Active Filtering Number Operating Cut-off
diameter length area of pH range
(mm) (cm) (cmz) channels range (kDL
10 25.0 95.2 3 0-14 15

 

 

 

 

 

 

 

Ozone was generated by feeding pure oxygen gas from a gas cylinder to the
ozone generator (Model OZZPCS, Ozotech Inc., Yreka, Calif). The oxygen gas
stream was dried using a molecular sieve trap prior to ozone generation. The ozone
gas concentration was controlled using the electric current of the ozone generator.

The wasted ozone gas was vented after passing the gas through a 2% potassium iodide
(KI) solution to destroy any residual ozone gas.

For the experiments investigating permeate ﬂux recovery, a centrifugal pump
(Model 4RH12, Dayton Electric Mfg. Co., Niles, Illinois) was used to inject water
from the 3.5-liter water reservoir into the membrane module. A magnetic stirrer at the
bottom of the reservoir was used to completely mix the water. The water ﬂow rate
was ﬁxed at 2.75 L/min and the water temperature was maintained at 20 °C. The
permeate ﬂux was monitored by measurement of the ﬂux in a volumetric cylinder
over speciﬁed time intervals. A peristaltic pump (Masterﬂex Model 7520—35, Cole-
Parmer Co., Chicago, Illinois) was used to continuously pump the water that was
prepared in a 1.5-liter water tank into the 3.5-liter water reservoir to maintain the
constant volume of water. The system was kept running under these operating
conditions until the permeate ﬂux decreased to 60% of the initial ﬂux without

ozonation. Then ozone gas was injected through the simple Y inline mixer into the

98

 

feed water stream of the membrane module. The ozone gas ﬂow rate was 100
mL/min and the ozone dose was 12.5 g/m3. The system was stopped as soon as the
permeate ﬂux reached a constant value. Same experimental procedure was used to
investigate the effect of oxygenation on the permeate ﬂux. Pure oxygen was used
instead of ozone and was injected into the feed water stream after the permeate ﬂux
decreased to 75% of the initial ﬂux with out oxygenation.

The same system used in permeate ﬂux recovery experiment was used to
investigate the effect of the ozonation/ﬁltration on the characteristics of NOM.
Several operational factors, such as ozone concentration in the inlet gas phase, water
ﬂow rate, gas ﬂow rate, and water temperature, which inﬂuence the ozone mass
transfer and solubility, were investigated. Table 4.2. summarizes the experimental

conditions used in this study.

Table 4.2. Experimental conditions.

 

 

 

 

 

 

Parameters Values
Inlet ozone dose (g/m3) 1.5, 2.5, 5.8
Water ﬂow rate (L/min) 2.75
Ozone gas ﬂow rate (mL/min) 50, 75, 100
Water temperature (°C) 10, 20, 30

 

 

The natural water and ozone gas were continuously fed into the membrane
module. Prior to ozonation, raw water samples were pro-ﬁltered through a 0.45 pm

ﬁlter to remove coarse suspended solids. Since the collection of permeate from the

99

 

membrane required several hours, a collection bottle with a cap was used and placed
in an ice bath to chill and preserve the sample. The ﬁrst 500 mL of permeate was
collected and marked as “Permeate 1”. The next 1.5 L of permeate was collected and
marked as “Permeate 2”. Permeate 1 and 2 represent the characteristics of the
permeate collected at different time intervals and provide information of treated water
quality on the performance of the ozonation/membrane system. The concentrate was
recirculated back to the raw water reservoir. The ﬁltered raw water and the Permeate
2 water were further treated by bioﬁltration to determine their biodegradability.
Water samples were collected from the ﬁltered raw water, the permeate, the water
tank, and the ﬁltrate of bioﬁltration columns. DOC, UV-254, humic substances, non-
humic substances, AOC and BDOC were monitored. The dissolved ozone residual in
collected samples was quenched by sparging the solution with a high purity helium
gas (99.999 %) before analysis. Earlier studies show that quenching can be achieved

in this manner in several seconds (Hemer 1999).

4.2.2 Water Source

Experiments were carried out on Lake Lansing (Haslett, MI) water. Lake
Lansing water samples were collected at the boat ramp at the Lake Lansing Park-
South, Haslett, Michigan. Water quality characteristics of Lake Lansing are
summarized in Table 4.3. Lake Lansing waters were collected in ﬁve-gallon tanks,
properly labeled, and stored at 4°C. The maximum storage period was seven days.

All waters were ﬁltered through a 0.45-um ﬁlter before testing.

100

4.2.3 Gas-phase Ozone Analysis

The concentration of ozone in gas phase was measured at UV 254 nm with a
Milton Roy Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) using a
2-mm quartz ﬂow-through cell. An extinction coefﬁcient of 3000 M'lcm'1 at 254 nm

(Hoigné 1988) was used for calculating ozone concentrations.

Table 4.3 Typical quality characteristics of Lake Lansing water.

 

 

 

 

 

 

 

 

 

 

 

Parameters Lake Lansing
TOC (mg/L) 8.6 ~ 11.0
pH 7.7 ~ 8.0
Alkalinity (mg/L as CaCO3) 140
UV-254 (abs.) 0.160 ~ 0.180
SDS THM (pg/L) 240
SDS HAA (pg/L) 75
BDOC (mg/L) 1.0 ~ 4.0

 

4.2.4 UV-254 Absorption

UV absorption was measured at a wavelength of 254 nm with a Milton Roy
Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) at room temperature.
A quartz cell providing a light path of 1 cm was used. Samples were preﬁltered

through a 0.45-um ﬁlter before measurement.

101

4.2.5 TOC and DOC Analyses

TOC was determined by the heated-persulfate oxidation method (Standard
Methods 1998) using an 01 Analytical Model 101 analyzer with auto sampler. DOC
was determined the same way but samples were preﬁltered through a 0.45 pm glass-
ﬁber ﬁlter before measurement.

Each sample was performed with three replicates. A blank (DDI water) and a
standard calibration series of 2.5, 5.0, 7.5, and 10.0 mg C/L were prepared and run

before analyzing water samples.

4.2.6 Humic Substances and Non-humic Substances Analyses

The concentrations of HS and NHS in water samples were measured by
adsorption on XAD-8 resin using Method 5510C (Standard Methods 1998) with slight
modiﬁcation. The XAD-8 resin was prepared and cleaned following the procedure
described in the Method 5510 C and then was packed in a 1.0 x 15-cm glass column.

A 100 mL water sample was ﬁrst acidiﬁed to pH 2 with concentrated
phosphoric acid and then pumped onto the top of the resin-packed column at a rate of
3 mL/min. The efﬂuent from the column was collected and then analyzed for TOC,
which represented the non-humic fraction of the dissolved organic matter in the water
sample. Then the resin-packed column was back eluted with 100 mL of 0.1 N sodium
hydroxide at a ﬂow rate of 3 mL/min. The eluent was collected and acidiﬁed with
concentrated phosphoric acid to a pH less than 4, purged with high-purity helium for
3 minutes to remove inorganic carbon, and analyzed for TOC. The organic content of

the eluent represented the concentration of HS.

102

4.2.7 Biodegradable Dissolved Organic Carbon Analysis

BDOC was determined by recirculating raw and treated water through a
bioﬁltration column that contains Lake Lansing water inoculum. The procedure
described by Cipparone et a1. (1997) was followed. The bioﬁltration system consisted
of a glass chromatography column with a diameter of 2.5 cm and a total volume of
100 cm3, a peristaltic pump (Masterﬂex Model 7553-50, Cole-Partner Co., Chicago,
Illinois), and a l-liter glass water tank. Half of the column volume was ﬁlled with
non-activated carbon. Untreated Lake Lansing water was used to seed the column
and the organisms were acclimated by feeding ozonated Lake Lansing water into the
column continuously over a period of four months before use of the column in this
study. One liter of water sample in the feed water tank was recirculated at a rate of 10
mL/min at room temperature. The TOC concentration was monitored daily for seven
days. The BDOC concentration was calculated as the difference between the initial

TOC concentration in the sample and that in the ﬁnal biodegradation sample.

4.2.8 Assimilable Organic Carbon Analysis

The concentration of AOC was measured using the Method 9217 of Standard
Methods (1998). Water samples were collected in organic-carbon-free glass bottles,
mixed well and poured into several test tubes. These test tubes were capped
immediately and pasteurized in a 70 °C water bath for 30 minutes to inactivate
bacteria originally present in the sample. After pasteurized samples cooled to room
temperature, they were inoculated with 500 colony-forming units (CFU)/mL each of
Pseudomonasﬂuorescens strain P17 and Spirillum strain NOX. The samples were
incubated at room temperature (20 °C) for seven days, then a Heterotrophic Plate

Count (HPC) was performed by distributing samples on predried R2A agar plates

103

(Standard Methods 1998). The colonies were counted on the plates after three days of

incubation at room temperature and the concentration of AOC was calculated.

4.3 RESULTS AND DISCUSSION
4.3.1 Permeate Flux Recovery

The permeate ﬂuxes with and without ozonation are shown in Figure 4.2. The
permeate ﬂuxes with and without oxygenation are presented in Figure 4.3. The
preﬁltered Lake Lansing water was used in the experiment. The ozonation/membrane
system was operated at 2.75 L/min water ﬂow rate and the water temperature was
maintained at 20 °C. For the ozonation experiment, the initial permeate ﬂux was 1.9
mL/min and gradually decreased to 1.15 mL/min (60% of the initial ﬂux) in 12 hours.
As soon as the permeate ﬂux reached 60% of the initial ﬂux, ozone gas was prepared
from pure oxygen and injected into the water stream at 100 mL/min gas ﬂow rate and
12.5 g/m3 dose. It was observed that the permeate ﬂux rebounded immediately and
reached 1.85 mL/min (97% of the initial ﬂux) in just one hour.

For the oxygenation experiment, the initial ﬂux was 2.1 mL/min and
decreased to 1.54 mL/min (75% of the initial ﬂux) in 13.5 hours. After oxygen gas
was injected to the feed water stream at 100 mL/min ﬂow rate, the permeate ﬂux
increased immediately. The permeate ﬂux reached 1.71 mL/min (84% of the initial
ﬂux) within just 10 minutes after the addition of oxygen gas and maintained at this
ﬂux level. Kim et al. (1999) observed similar results by using a ceramic membrane to
investigate the effect of ozone and oxygen bubbling on permeate ﬂux recovery. They
found that ozone bubbling can result in recoveries up to 94% of the initial ﬂux but
oxygen bubbling resulted in recoveries up to 60%. The more efﬁcient permeate ﬂux

recovery obtained with ozone compared to that with oxygen is thought to be due to

104

the reaction of ozone with the organic matter that deposits on the membrane surface
during normal operation. This reaction not only removes the deposition layer but also
breaks down large molecular weight molecules. The injection of both oxygen and
ozone is thought to increase the turbulence of the water stream thereby removing or
thinning the deposition layer on the membrane surface and, with ozone, also breaking
down large molecules and particles.

The hypothesis that organic substances are destroyed during ozonation is
supported by atomic force microscope (AF M) scans obtained during this study (see
Figures 4.4 to 4.7). The height images on the left mean topography and the phase
images on the right show the phase difference between the resonance frequency of the
cantilever and the altered oscillation frequency induced by the interaction of the tip
with the sample. The phase image for the residuals of non-ozonation backwash (see
Figure 4.5) is all one color. This implies that the residuals are basically the same
substance, and that it doesn’t have sharp edges. The phase images for the residuals of
ozonation backwash (see Figures 4.6 and 4.7) show quite a bit of variation that

resulted from the ozonation of NOM.

105

120 e

11'

 

 

w/o ozonation w/ozonation i

H i l

 

E
g 1004
E
:3 80
.s
3 60
“5
8) 40
i
e 20
(D
Q
0
0

2 4 6 8 10 12 14 16
Time(hr)

Figure 4.2 The variation of permeate ﬂux with respect to operating time. Operating
conditions: water ﬂow rate, 2.75 L/min; ozone gas ﬂow rate, 100
mL/min; ozone dose, 12.5 g/m3; water temperature, 20 0C.

 

 

4 N bl
w/o oxygenation w/ oxygenatiion
l

 

A 120
e‘:
5 1001
E
a 80
.5
3 50
"5
g, 40
g 20
Q)
0- 0

1'

0 2 4 6 8 10 12 14 16

Time (hr)

Figure 4.3 The variation of permeate ﬂux with respect to operating time. Operating
conditions: water ﬂow rate, 2.75 L/min; oxygen gas ﬂow rate, 100
mL/min; water temperature, 20 °C.

106

Height Image Phase Image

 

0 5.00pm 0 5.00pm
Data type Height Data type Phase
Z range 94.2 nm Z range 158 de

Figure 4.4 The atomic force microscope (AFM) scan of glass substrate with nothing
on it. (Courtesy: Dr. Virginia Ayres, Michigan State University)

Height Image Phase Image

 

0 5.00pm 0 5.00pm
Data type Height Data type Phase
Z range 94.2 nm Z range 158 de

Figure 4.5 The atomic force microscope (AF M) scan of backwash of used ceramic
membrane without ozonation of NOM. (Courtesy: Dr. Virginia Ayres,
Michigan State University)

107

Height Image Phase Image

    

0 5.00pm 0 5.00pm
Data type Height Data type Phase
Z range 94.2 nm Z range 158 de

Figure 4.6 The atomic force microscope (AFM) scan of backwash of used ceramic

membrane with ozonation of NOM. (Courtesy: Dr. Virginia Ayres,
Michigan State University)

Height Image Phase Image

 

Data type Height Data type Phase
Z range 94.2 nm Z range 158 de

Figure 4.7 Close-up of area in white box in Figure 4.6. (Courtesy: Dr. Virginia Ayres,
Michigan State University)

108

4.3.2 DOC

Before considering the removal of DOC in the ozonation/membrane system, it
is important to recognize that the characteristics of NOM in the raw water can vary
signiﬁcantly every time when water samples are collected from Lake Lansing. This
makes the direct comparison of treatment efﬁciency between different operational
conditions challenging. Therefore, careful interpretation of the data obtained from
this study is necessary.

The effect of water temperature on the DOC concentration is shown in Figure
4.8. Water samples were collected from ﬁltered raw water (F RW), the ﬁrst 500 mL
of the permeate (P1), the next 1.5 liters of the permeate (P2), the 3.5-liter water tank
(WT), and the ﬁltrate (B-P2) of bioﬁltration column used to treat P2 water. The
permeate ﬂux of 10, 20, and 30 °C operating temperature was 1.9, 2.3, and 2.8
mL/min, respectively. It is apparent from Figure 4.4 that the DOC concentration
decreased in P1, P2, and WT compared to that in the F RW water at 20 °C and 30 °C
(p < 0.01). At 10 °C, the DOC concentration decreased by ca. 17% in P1 and P2 but
only slightly decreased (2.5%) in WT. This suggests that, on one hand, ozonation
partially removed DOC at all three temperatures, but that more organic matter was
retained in the water tank at 10 °C because of low permeate yield. On the other hand,
it has been established that the rate of ozone decomposition and the production of
hydroxyl radical increase with increasing temperature (Sotelo et al. 1987).
Consequently, the increased effectiveness of the process in degrading DOC in P2 and
WT at higher temperatures may be due to the reaction of DOC with the hydroxyl
radical, which is more powerful, faster, and relatively less selective oxidant than is

molecular ozone.

109

 

Figure 4.9 presents the removal efﬁciency of DOC of P1, P2, and B-P2 at
various water temperatures. For P2 water samples, about 15-33% of DOC was
removed from the ozonation/membrane system. Since ozonation mainly converts
humic to non-humic substances and higher- to lower-molecular weight, only a
fraction of the DOC is oxidized. Galapate et al. (2001) used a semi-batch ozonation
reactor to treat Minaga Reservoir water at ozone dose of 3 mg 03/mg DOC and
observed a 16% removal of DOC. Owen et a1. (1995) used similar set up and
observed DOC reduction up to 24% at ozone dose of 1 mg O3/mg DOC.

The removal efﬁciency of DOC, as expected, achieved 70 to 80% when the
bioﬁltration followed the ozonation/membrane system to treat P2 water (see Figure
4.9). This is because ozonation of water containing NOM can enhance the
biodegradability of water due to the increase of BDOC (Volk et a1. 1993; Volk et al.
1997; Yavich 1998). Siddiqui et al. (1997) observed a 40 to 50% removal of DOC at
a ratio of O3/DOC = 2 mg/mg by ozonation and bioﬁltration, regardless of the source
of water or initial DOC. Similar results were reported by other researchers
(Cipparone et al. 1997; Goel et a1. 1995). Yavich (1998) compared the efﬁciency of
two systems, ﬂuidized bed treatment (FBT) /ozonation/bioﬁltration and
ozonation/FBT, in the removal of DOC from Huron River water. He observed an up
to 46% removal of DOC at an ozone dose of 3 mg/mg C. Comparing our results with
reports mentioned above, the ozonation/membrane system followed by bioﬁltration
used in this study seems to have higher DOC removal. The may be due to the extent
of reaction resulting from the surface catalysis of ozone decomposition or the

increased ozone contact time in our system.

110

12.0
10.0

DOC (mg/L)
0)
O

 

 

10 20 30
Temperature (°C)

 

Figure 4.8 The variation of DOC concentration of different treatment stages at various
water treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

100 ‘ :1 B-P2

DOC removal (%)

 

 

1 0 20 30
Temperature (°C)

Figure 4.9 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various water treatment temperatures.
Concentrations are compared to those in the raw water. For temperatures
10, 20, 30 °C, the raw water DOC concentrations were 9.5, 10.9, 9.0 mg/L,
respectively. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow
rate, 100 mL/min; ozone dose, 2.5 g/m3.

lll

Figure 4.10 shows the concentration of DOC in the samples as a function of
ozone gas ﬂow rate. At all the ﬂow rates used, the DOC concentration decreased in
the permeate of the membrane. Figure 4.11 shows the percentage of DOC removal in
the P2 and bio-treated B-P2 water. The efﬁciency of removal was essentially
independent of gas ﬂow rate. Without bioﬁltration, the system removed about 30% of
DOC. With bioﬁltration, the removal of DOC was further increased to about 70%.
These results indicate that the gas ﬂow rate of ozone had little effect on the system
performance in terms of DOC removal. This is thought to be due to the
characteristics of the raw water quality or the long hydraulic retention time of this

ozonation/membrane system.

12.0 ~ I FRW
10.0
8.0
6.0
4.0
2.0
0.0

DOC (mg/L)

 

 

 

50 75
Gas ﬂow rate (mL/min)

Figure 4.10 The variation of DOC concentration of different treatment stages at
various ozone gas ﬂow rates. Operating conditions: water ﬂow rate,
2.75 L/min; ozone dose, 2.5 g/m3; water temperature, 20 °C.

112

100 — ~ -—IP2

 

 

 

 

 

 

 

   

i:iB-P2

c; 80 l
e
g 50
5
- 40
O
o
D 20 r l—

0 __i‘

 

50 75 1 00
Gas ﬂow rate (mL/min)
Figure 4.1 1 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various ozone gas ﬂow rates. Concentrations
are compared to those in the raw water. For gas ﬂow rates 50, 75, 100
mL/min, the raw water DOC concentrations were 8.7, 10.4, 10.9 mg/L,

respectively. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m3; water temperature, 20 °C.

The effect of ozone dose on the DOC removal is shown in Figure 4.12 and the
removal efﬁciency is shown in Figure 4.13. The DOC concentration decreased in the
permeate at all three ozone doses. As mentioned earlier, ozonation dose not remove
much of the DOC but increases its biodegradability. Based on Figure 4.12, the DOC
concentration in the permeate was not signiﬁcantly reduced (p > 0.05), even at an
ozone dose of 5.8 g/m3. However, its DOC concentration after bioﬁltration was
obviously less than those obtained at lower ozone doses (1.5 and 2.5 g/m3). It
indicates that even though the high ozone dose was not more effctive at removing

DOC from the system than that observed at low ozone dose, more biodegradable

113

organic material was produced at the 5.8 g/m3 ozone dosage. Ozonation followed by
bioﬁltration of P2 water removed 88% of DOC at 5.8 mg O3/L but only removed 70%
to 77% of DOC at 1.5 and 2.5 mg 03/L (see Figure 4.13). This is consistent with the

ﬁndings of Lee (2001) in terms of AOC.

 

12.0 .- IFRW
10.0 1
2 8.0
O)
5- 5.0
8
D 4.0
2.0
0.0
1.5 2.5 5.8
03 dose (g/ma)

Figure 4.12 The variation of DOC concentration of different treatment stages at
various ozone doses. Operating conditions: water ﬂow rate, 2.75 L/min;
gas ﬂow rate, 100 mL/min; water temperature, 20 °C.

114

100 r , I P2

 

 

 

A UB-PZ
§ 80 ’

g 50

E

g 40

D

 

 

 

 

 

20
0 1 _I__.
1.5 2.5 5.8

03 dose (g/m3)

 

Figure 4.13 The percentage of DOC removal for permeate 2 (P2) and biotreated
permeate 2 (B-P2) water at various ozone doses. Concentrations are
compared to those in the raw water. For ozone doses 1.5, 2.5, 5.8 g/m3,
the raw water DOC concentrations were 8.7, 10.9, 8.9 mg/L, respectively.
Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
lOOmL/min; water temperature, 20 °C.

4.3.3 UV-254

The removal efﬁciency of UV-254 at various operational conditions is shown
in Figures 4.14, 4.15 and 4.16. As expected, the UV-absorbing compounds were
effectively removed, but not eliminated, at all tested conditions. The removal
efﬁciencies for P1, P2, and WT ranged from 55% to 71%, 74 to 84%, and 75 to 85%,
respectively. The maximum removal of UV-254 was 85% in our bench-scale system.
No further removal was observed after the 85% removal efﬁciency was achieved.

The percentage removal of UV-254 of P1, which was the ﬁrst 500 mL of

collected permeate of membrane, was less than that of P2, which was the 1.5 L

115

collection of permeate after P1. It appears that with increased ozone contact times,
the removal efﬁciency of UV-254 increases. However, only minimal changes in the
UV-254 removal efﬁciencies were observed between P2 and WT. This indicates that
the contact time for P2 were sufﬁciently long to remove or break down most of the
UV-254 absorbing compounds. Although the contact time of WT was longer than P2,
only slight changes of UV-254 were observed, suggesting the remaining UV-254

absorbing substances are recalcitrant towards ozone.

 

  

IP1

A100 i:iP2
,\° awr
: 80
g
E 60 ..
9 40 4“}
a «i
5- 20 U

0

 

1 0 20 30
Temperature (°C)

Figure 4.14 Efﬁciency of UV-254 removal at various water temperatures. Operating
conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100 mL/min;
ozone dose, 2.5 g/m3.

116

IP1

 

100 13132
:3 80 UWT
g
g 50
e
4 40
R
>- 20
D

0

 

 

 

Gas ﬂow rate (mL/min)

Figure 4.15 Efﬁciency of UV-254 removal at various ozone gas ﬂow rates. Operating
conditions: water ﬂow rate, 2.75 L/min; ozone dose, 2.5 g/m3; water
temperature, 20 °C.

 

 

 

 

   

l P1
A 100 D P2
i“ 80 DWT
E
E 60
9
V 40
L!)
N
>' 20
D
0
2.5
03 dose (g/m3)

Figure 4.16 Efﬁciency of UV-254 removal at various ozone doses. Operating
conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100 mL/min;
water temperature, 20 °C.

117

4.3.4 Humic and non-humic substances

The ratios of HS to DOC and of NHS to DOC for F RW, P1, P2, and WT water
samples at various water temperatures are shown in Figures 4.17 and 4.18. The
experimental results show that the NOM of Lake Lansing water originally consisted
of about 60% of HS. At all three tested water temperatures, the concentration of HS

decreased and the concentration of NHS increased as the contact time increased. As

 

the water temperature at 20 °C and 30°C, the ratio of NHS to DOC ﬁrst increased

rapidly in P1 and then slightly increased in P2 and WT. In contrast, the ratio

 

gradually increased in P1, P2, and WT at 10°C. This is thought to be because that the
concentration of hydroxyl radicals increased with increasing water temperature due to

rapid decomposition of ozone with increasing temperature. The 0H radicals would

 

react faster with HS than would molecular ozone, converting HS to NHS. However,
at lower water temperature, fewer hydroxyl radicals were produced and the direct
reaction between molecular ozone and HS was slower than that between hydroxyl
radicals and HS.

Figures 4.19 and 4.20 present the ratios of HS to DOC and NHS to DOC at
various ozone gas ﬂow rates, respectively. It appears that the ratio of HS/DOC
decreased as contact time increased for ﬂowrates of 50 and 75 mL/min. At a ﬂowrate
of 100 mL/min, the ratio of HS/DOC in P1, P2, and WT were essentially constant at
approximately 30%. Similarly, for gas ﬂow rates of 50 and 75 mL/min, the ratios of
NHS/DOC increased with contact time. With the lOOmL/min samples, the results
correspond well to those for the HS/DOC ratios: the greatest increase occurred in P1.
A slight increase in NHS increased from P1 to P2 and the ratio was essentially equal

(no statistical difference at p = 0.05 level) in P2 and WT samples.

118

The effect of ozone dose on the ratio of HS and NHS to DOC is shown in
Figures 4.21 and 4.22, respectively. The ozone dose at 1.5 g/m3 converted less HS to
NHS comparing with that obtained at 2.5 and 5.8 g/m3. However, the ratios of
HS/DOC and NHS/DOC obtained for an ozone dose of 2.5 g/m3 were very close to
those of 5.8 g/m3 at P1, P2, and WT. This may suggest that increasing the ozone dose
over 2.5 g/m3 does not further convert HS to NHS in the tested conditions. .

Figure 4.23 shows the relationship between the concentration of HS at F RW,
P1, P2, and WT and their UV-254 data for Lake Lansing water. As expected, the
concentration of HS and the adsorption of UV-254 at different treatment stages and
under various operational conditions were well correlated. This is because the aquatic
HS contains aromatic functional groups and conjugated double bonds that absorb light
at 254 nm in the ultraviolet wavelength region. The value of UV-254 increased as the

concentration of HS increased (correlation is signiﬁcant at the 0.01 level at 95% C .1.).

 

100 I FRW

HS/DOC (%)

 

 

 

Temperature (°C)

Figure 4.17 The ratio of HS/DOC of different treatment stages at various water
treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

119

 

 

100

NHS/DOC (%)

 

 

 

 

 

20 30
Temperature (°C)

Figure 4.18 The ratio of NHS/DOC of different treatment stages at various water
treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

 

100 I FRW
1: P1
80 :1 P2

HS/DOC (%)

 

 

50 75 100
Gas ﬂow rate (mL/min)

Figure 4.19 The ratio of HS/DOC of different treatment stages at various ozone gas
ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m ; water temperature, 20 °C.

120

I FRW
100 CI P1
80 Cl P2

[1 WT
60
40
20
0
50 75 100

Gas ﬂow rate (mL/min)

NHS/DOC (%)

 

 

 

Figure 4.20 The ratio of NHS/DOC of different treatment stages at various ozone gas
ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m ; water temperature, 20 °C.

 

 

100
A 80
o\°
o 60
O
Q 40
‘13

20

O

1.5 2.5 5.8
03 dose(g/m3)

Figure 4.21 The ratio of HS/DOC of different treatment stages at various ozone doses.
Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100
mL/min; water temperature, 20 °C.

121

100 -lFRW

 

 

 

 

 

 

 

 

a; 80
o\
o 60
8
a 40
I
Z 20

0

1.5 2.5 5.8
03 dose (g/ms)

Figure 4.22 The ratio of NHS/DOC of different treatment stages at various ozone
doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C.

y=25.47x{i.’333 ”
R2=0.90]4_ 3 3 7.

    

 

Humic substances (mg/L)

0.0 ‘ 4
0.000 0.050 0.100 0.150 0.200
uv-254 (Abs.)

Figure 4.23 Correlation between the concentration of humic substances and UV-254
for Lake Lansing water. (Correlation is signiﬁcant at the 0.01 level at
95% CI.)

122

4.3.5 BDOC

For measuring the BDOC concentration, water samples were recirculated
through a bioﬁltration column as described in Section 4.2.7. The BDOC
concentration was measured by subtracting the day seven DOC from day zero DOC
of the water sample. In this study, BDOC was only measured for FRW water, as the
inﬂuent, and P2 water, as the efﬂuent, of the ozonation/membrane system.

The percentages of BDOC to DOC of F RW and P2 water at various

 

experimental conditions are shown in Figures 4.24, 4.25, and 4.26. As can be seen in

l‘?‘ ‘

these ﬁgures, the fraction of BDOC to DOC in the F RW ranged from 10 to 38%.
After treated by the ozonation/membrane system, the percentage increased to 55 to
86%. A similar trend was observed by Escobar and Randall (2001) when they treated
a low DOC raw water (1 mg/L) with ozone. They observed a 49% increment of the
efﬂuent BDOC when ozonation was used to treat the raw water.

The BDOC of permeate 2 water treated at 10 °C was approximately 77% of
the DOC (see Figure 4.24). At 20 and 30 °C, the BDOC was only about 57% of the
DOC. This may be due to the higher concentration of hydroxyl radicals at higher
temperature than at lower temperature as was discussed in Section 4.3.2. The
observation that the percentage of BDOC of the permeate 2 (B-P2) was lesser at the
higher two temperatures is because one would expect that the OH radical would be
less efﬁcient than ozone at degrading the DOC into low molecular weight compounds
that would provide substrate material for the microorganisms. On the other hand, the
high percentage of BDOC in the FRW water at 10 0C may contribute to the higher
percentage of BDOC in its P2 water.

The effect of ozone gas ﬂow rate on the ratio of BDOC to DOC is shown in

Figure 4.25. The ratio of BDOC:DOC of FRW water ranged from 14 to 35% and

123

 

 

increased to 55 to 66% after treatment at various gas ﬂow rates. The ozone gas ﬂow
rate did not have a signiﬁcant inﬂuence on the BDOCzDOC ratio in the P2 water.
This conﬁrms the previous observations discussed in Sections 4.3.2.and 4.3.4. It also
suggests that the conversion of non-BDOC to BDOC is as efﬁcient at a gas ﬂow rate
of 50 mL/min as it is at gas ﬂow rates of 75 and 100 mL/min.

Figure 4.26 shows the effect of ozone dose on the BDOCzDOC ratio. At an
ozone dose of 5.8 g/m3 most of the non-BDOC was converted to BDOC, with a
BDOC:DOC of 86%. At ozone doses of 1.5 and 2.5 g/m3, the BDOCzDOC ratios
were only 67% and 57%, respectively. However, the ozone dose of 1.5 g/m3 may be a
better operational condition than the other two doses because of its lower cost of
ozone production and acceptable efﬁciency in terms of conversion of non-BDOC.

The experimental results show that using the ozonation/membrane system can
dramatically increase the percentage of BDOC in the permeate water, which can be
easily removed by subsequent biotreatment. Urfer et al. (1997) suggest that ozonation
and bioﬁltration should be considered as a coupled process rather than two
independent process steps, since the ozonation results in the increase of BDOC and
leads to regrowth problems in the water distribution system. Similar results were

observed by Yavich and Masten (2003).

124

I FRW

100 '
CIPZ

80 ‘
60

 

40:

BDOC/DOC (%)

20 :-

 

 

 

     

 

1 0 20 30
Temperature (°C)

Figure 4.24 The percentages of BDOC to DOC of FRW and P2 water at various water
treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

 

 

 

 

 

 

 

 

 

100 _ MIFRW
UPZ
g 80 ’
U
o 50
S 40
O
C)
m 20 -
o l._-__ J
50 75 100

Gas flow rate (mL/min)

Figure 4.25 The percentages of BDOC to DOC of FRW and P2 water at various
ozone gas ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min;
ozone dose, 2.5 g/m3; water temperature, 20 °C.

125

100 _ . ..IFRW
DPZ

 

80

 

BDOC/DOC (%)

 

 

 

 

 

1 .5 2.5
03 dose (g/m3)

Figure 4.26 The percentages of BDOC to DOC of FRW and P2 water at various
ozone doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas
ﬂow rate, 100 mL/min; water temperature, 20 °C.

4.3.6 AOC

Figures 4.27, 4.28, and 4.29 show the concentration of AOC of different
treatment stages under various operational conditions. The abbreviation of B-FRW
represents the ﬁltrate of bioﬁltration column from the treatment of ﬁltered raw water.
Other notation is the same as described in Section 4.3.2. In this study, the AOC
concentrations of FRW ranged from 1 10 to 470 jig/L and averaged 273 pg/L. After
treatment by the ozonation/membrane system, the AOC concentration signiﬁcantly
increased in P1, P2 and WT waters. The average concentrations of AOC for P1, P2,
and WT were 1038, 1365, and 1377 ug/L, respectively. The increase in the AOC
concentration is thought to result from the conversion of large molecules to small
molecules during ozonation. The smaller molecules are easily biodegraded, during
ozonation (Escobar and Randall 2001; Volk et a1. 1993; Volk et al. 1997). However,

there were the formation of AOC could not be correlated to operational conditions

126

 

used in our experimental results. Since it is well known that NOM is made up of a
heterogeneous mixture of organic compounds and has complex composition, the
amount and species of biodegradable compounds, which can be utilized by
Pseudomonasﬂuorescens strain P17 and Spirillum strain NOX in terms of the
measurement of AOC, may vary signiﬁcantly after a long contact time with ozone
under different operational conditions. Consequently, this induces that the AOC
concentration and operational conditions are not correlated.

Although the ultraﬁltration might be expected to remove some AOC, in our
experiments we did not observe any signiﬁcant removal of AOC by ultraﬁltration, as
measured by the difference in the AOC concentrations in P2 and WT waters. Several
researchers found that membrane ﬁltration is ineffective in reducing AOC
concentrations. Escobar and Randall (1999; 2001) reported that nanoﬁltration (NF)
effectively removed BDOC but no signiﬁcant removal of AOC was observed.
Chamock and Kjonno (2000) indicated that this is due to the fact that the AOC and
BDOC analyses target different fractions of the biodegradable organic matter (BOM).
Therefore, it is apparent that the ultraﬁltration ceramic membrane used in our system
did not effectively remove the AOC from the treated water.

Experiments of FRW and P2 water treated by bioﬁltration columns were
conducted to evaluate the effect of bioﬁltration in reducing AOC concentrations. As
expected, the AOC concentrations of B-FRW and B-P2 water decreased under most
experimental conditions. The average concentration of AOC for B-FRW and B-P2
water was 106 and 77 ug/L, respectively. It suggests that bioﬁltration of F RW and P2
water can effectively remove much of the AOC concentration. The AOC
concentration of B-P2 water was always less than that of B-FRW water, regardless the

AOC concentration in P2 water. For FRW and P2 water samples, about 52 to 75%

127

 

 

 

and 91 to 97% of AOC were removed by bioﬁltration, respectively. Only one case of

FRW water showed slight increase of AOC (from 139 t0185 ug/L) after bioﬁltration.

 

 

 

 

 

 

 

 

 

 

 

 

I FRW
2500 I P1
._ El
A 2000 E1 B-FRW

<1 B-P2
C) 1500 re
8 1000 -
<1: f
500 g

1 0 20 30
Temperature (°C)

Figure 4.27 The concentration of AOC of different treatment stages at various water
treatment temperatures. Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; ozone dose, 2.5 g/m3.

 

 

128

 

 

 

 

 

 

 

 

 

 

 

 

 

I FRW
2500 . p1
1:1
2°00 -- a B-FRW "‘
3 a B-P2
B) 1500 r
a
8 1000 » T
<(
500
0
50 75 100

Gas ﬂow rate (mL/min)

Figure 4.28 The concentration of AOC of different treatment stages at various ozone
gas ﬂow rates. Operating conditions: water ﬂow rate, 2.75 L/min; ozone
dose, 2.5 g/m3; water temperature, 20 °C.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IFww
2500 I P1
:1 P2
2000 - _. :1 WT
. . a B-FRW
3 B-P2
a, 1500 .
s r“ .—
8 1000
<
500
0 - ~1..
1.5 2.5 5.8

03 dose (g/m3)

Figure 4.29 The concentration of AOC of different treatment stages at various ozone
doses. Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C.

129

4.4 CONCLUSIONS

The characteristics of NOM changed in terms of DOC, UV-254, humic and

non-humic substances, BDOC, and AOC when using the ozonation/membrane system

to treat Lake Lansing water. The overall conclusions of the study are the following:

1.

The addition of ozone in the feed water of membrane resulted in a recovery of the
permeate ﬂux from 60% to 97% of the initial ﬂux. The increase in permeate ﬂux
may have been the result of an increase in the turbulence in the water stream as a
result of the injection of ozone. The increased turbulence would be expected to
remove or thin the deposition layer on the membrane surface. Additionally, the
ozone would react and break down large molecules and particles. This latter
phenomenon was important since ozone was more effective than air.

The water temperature inﬂuenced the DOC concentrations of permeate of the
membrane. As the water temperature increased, the DOC concentration decreased.
This may be due to the high permeate ﬂux and/or the high concentration of
hydroxyl radicals at elevated water temperatures.

The ozonation/membrane system removed 15 to 33% DOC. If followed by
bioﬁltration, the percentage of DOC removal achieved 69 to 88% because the
ozonation of NOM enhanced the biodegradability of water, measured as an
increase in the BDOC.

The experimental results show that the gas ﬂow rate of ozone had little effect on
the system performance in terms of the removal of DOC, the conversion of humic
and non-humic substances, and the production of BDOC.

There were no correlations between the formation of AOC and operational
conditions found in our experimental results. This could be due to the amount and

species of biodegradable compounds, which can be utilized by Pseudomonas

130

 

 

 

ﬂuorescens strain P17 and Spirillum strain NOX in terms of the measurement of
AOC, vary signiﬁcantly after a long contact time with ozone under different
operational conditions.

. The concentration of UV-absorbing compounds decreased at all experimental
conditions. The removal efﬁciency for the ozonation/membrane system ranged

from 55 to 85% at various operating conditions. Increased ozone contact time

 

resulted in an increase in the removal efﬁciency. The maximum removal of UV-

254 was 85%.

 

The concentration of HS decreased and the concentration of NHS increased as the
contact time increased. At water temperatures of 20°C and 30°C, the ratio of
NHS/DOC ﬁrst increased rapidly in P1 and then slightly increased in P2 and WT.
In contrast, the ratio gradually increased in P1, P2, and WT at 10°C. This is
thought to be because the higher concentration of hydroxyl radicals that were

produced at higher water temperature reacted faster with HS than did molecular

 

ozone to convert HS to NHS. However, at lower water temperature, fewer
hydroxyl radicals were produced and the direct reaction between molecular ozone
and HS was slower than that between hydroxyl radicals and HS.

. The concentration of HS and the adsorption of UV-254 were well correlated at
different treatment stages and various operational conditions.

. Approximately 10 to 38% of the DOC in FRW was BDOC. After treatment using
the ozonation/membrane system, the percentage increased, with ranges from 55 to
86%. The experimental results show that using ozonation/membrane system can
dramatically increase the percentage of BDOC in the permeate water, which will

improve performance of the subsequent bioﬁlter.

131

10. The ozonation/membrane system increased the concentration of AOC
signiﬁcantly at all experimental conditions. The AOC concentration also
decreased signiﬁcantly after subsequent biotreatment, suggesting that a simple
biotreatment of raw and ozonated water can effectively remove AOC
concentration. The AOC concentration of the biotreated ozonated water was less
than that of biotreated raw water, regardless the AOC concentration in ozonated
water.

1 1. According to our experiment results, at operational conditions of 50 mL/min gas
ﬂow rate and an ozone dose of 1.5 g/m3 effective removal of DOC, UV-254, and
conversion of HS to NHS was possible. These conditions should reduce the cost

of system operation over more extensive ozonation.

4.5 REFERENCES

Chamock, C., and Kjonno, O. (2000). "Assimilable organic carbon and biodegradable
dissolved organic carbon in Norwegian raw and drinking waters." Water
Research, 34(10), 2629-2642.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

Collins, M. R., Amy, G. L., and Steelink, C. (1986). "Molecular-Weight Distribution,
Carboxylic Acidity, and Humic Substances Content of Aquatic Organic-

Matter - Implications for Removal During Water-Treatment." Environmental
Science & Technology, 20(10), 1028-1032.

Escobar, I. C., and Randall, A. A. (1999). "Inﬂuence of NF on distribution system
biostability." Journal American Water Works Association, 91(6), 76-89.

Escobar, I. C., and Randall, A. A. (2001). "Assimilable organic carbon (AOC) and
biodegradable dissolved organic carbon (BDOC): Complementary
measurements." Water Research, 35(18), 4444-4454.

132

Galapate, R. P., Baes, A. U., and Okada, M. (2001). "Transformation of dissolved
organic matter during ozonation: Effects on trihalomethane formation
potential." Water Research, 35(9), 2201-2206.

Goel, S., Hozalski, R. M., and Bouwer, E. J. (1995). "Biodegradation of Nom - Effect
of Nom Source and Ozone Dose." Journal American Water Works Association,
87(1), 90-105.

Hashino, M., Mori, Y., Fujii, Y., Motoyama, N., Kadokawa, N., Hoshikawa, H.,
Nishijima, W., and Okada, M. (2000). "Pilot plant evaluation of an ozone-

microﬁltration system for drinking water treatment." Water Science and
Technology, 41(10-11), 17-23.

Hemer, H. A. (1999). "The Reaction of Ozone with Pyrene: The By-products and
Their Toxicological Implications," Michigan State University, East Lansing.

Hoigne, J. (1988). "The Chemistry of Ozone in Water." Process Technologies for
Water Treatment, Plenum Publishing Corp., New York.

Hyung, H., Lee, 8., Yoon, J., and Lee, C. H. (2000). "Effect of preozonation on ﬂux
and water quality in ozonation- ultraﬁltration hybrid system for water
treatment." Ozone-Science & Engineering, 22(6), 637-652.

lchihashi, K., Teranishi, K., and lchimura, A. (1999). "Brominated trihalomethane
formation in halogenation of humic acid in the coexistence of hypochlorite
and hypobromite ions." Water Research, 33(2), 477-483.

Kim, J. 0., Somiya, 1., and Fujii, S. "Fouling Control of Ceramic Membrane in
Organic Acid F ermenter by lnterrnittent Ozonation." Proceedings of the 14th
Ozone World Congress, Dearbom, MI, 131-143.

Kuo, P. P. K., Chian, E. S. K., and Chang, B. J. (1977). "Identiﬁcation of End
Products Resulting from Ozonation and Chlorination of Organic-Compounds

Commonly Found in Water." Environmental Science & Technology, 11(13),
1177-1 181.

Lee, K.-H. (2001). "The effects of ozonation pathways on the formation of ketoacids
and assimilable organic carbon (AOC) in drinking water," Michigan State
University, East Lansing, MI 48823.

Levy, R. V., Hart, F. L., and Cheetham, R. D. (1986). "Occurrence and Public-Health
Signiﬁcance of Invertebrates in Drinking-Water Systems." Journal American
Water Works Association, 78(9), 105-110.

MacCarthy, P., and Suffet, I. H. (1989). "Aquatic Humic Substances and Their

Inﬂuence on the Fate and Treatment of Pollutants." Aquatic Humic Substances:
Inﬂuence on Fate and Treatment of Pollutants, I. H. Suffet and P. MacCarthy,
eds., American Chemical Society, Washington, DC., xvii-xxx.

Malcolm, R. L. (1985). "Geochemistry of Stream fulvic and Humic Substances."
Humic Substances in Soil, Sediment, and Water, G. R. Aiken, D. M.

133

McKnight, R. L. Wershaw, and P. MacCarthy, eds., John Wiley & Sons., New
York, 181-209.

Manahan, S. E. (1993). "Environmental Chemistry of Water." Fundamentals of
Environmental Chemistry, Lewis Publisher, Chelsea, MI., 388-390.

Mogren, E. M., Scarpino, P., and Summers, R. S. (1990). "Measurement of
Biodegradable Dissolved Organic Carbon in Drinking Water." Proc. 1990

AWWA Ann. Conf., Cincinnati, Ohio, 573.

Narkis, N., and Schneiderrotel, M. (1980). "Evaluation of Ozone Induced
Biodegradability of Wastewater- Treatment Plant Efﬂuent." Water Research,

14(8), 929-939.

Owen, D. M., Amy, G. L., Chowdhury, Z. K., Paode, R., McCoy, G., and Viscosil, K.
(1995). "Nom - Characterization and Treatability." Journal American Water

Works Association, 87(1), 46-63.

Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990). "Chlorination of Humic
Materials - by-Product Formation and Chemical Interpretations."
Environmental Science & Technology, 24(11), 1655-1664.

Servais, P., Billen, G., and Hascoet, M. C. (1987). "Determination of the
Biodegradable Fraction of Dissolved Organic-Matter in Waters." Water
Research, 21(4), 445-450.

Siddiqui, M. 8., Amy, G. L., and Murphy, B. D. (1997). "Ozone enhanced removal of
natural organic matter from drinking water sources." Water Research, 31(12),

3098-3106.

Sotelo, J. L., Beltran, F. J., Benitez, F. J., and Beltranheredia, J. (1987). "Ozone
Decomposition in Water - Kinetic-Study." Industrial & Engineering

Chemistry Research, 26(1), 39-43.

Standard Methods. (1998). Standard Methods for the Examination of Water and
Wastewater, A. E. Greenberg, L. S. Clesceri, and A. D. Easton, eds., APHA,

AWWA, WEF, Washington, DC.

Urfer, D., Huck, P. M., Booth, S. D. J., and Coffey, B. M. (1997). "Biological
ﬁltration for BOM and particle removal: a critical review." Journal American
Water Works Association, 89(12), 83-98.

Van der kooij, D., Hijnen, W. A. M., and Kruithof, J. C. (1989). "The Effects of

Ozonation, Biological Filtration and Distribution on the Concentration of
Easily Assimilable Organic-Carbon (Aoc) in Drinking-Water." Ozone-Science

& Engineering, 1 1(3), 297-31 1.

Van der kooij, D., Visser, A., and Hijnen, W. A. M. (1982). "Determining the
Concentration of Easily Assimilable Organic- Carbon in Drinking-Water."

Journal American Water Works Association, 74(10), 540-545.

134

 

 

Volk, C., Renner, C., Roche, P., Paillard, H., and Joret, J. C. (1993). "Effects of
Ozone on the Production of Biodegradable Dissolved Organic-Carbon (Bdoc)
During Water-Treatment." Ozone-Science & Engineering, 15(5), 389-404.

Volk, C., Roche, P., Joret, J. C., and Paillard, H. (1997). "Comparison of the effect of
ozone, ozone-hydrogen peroxide system and catalytic ozone on the

biodegradable organic matter of a fulvic acid solution." Water Research, 31(3),
650-656.

Yavich, A. A. (1998). "The Use of Ozonation and Biological Fluidized Bed
Treatment for the Control of NOM in Drinking Water," Ph.D. Dissertation,
Michigan State University, East Lansing, MI.

Yavich, A. A., and Masten, S. J. (2003). "Use of Ozonation and F BT to Control THM
Precursors." Journal American Water Works Association, 95(4), 159-171.

135

nﬁmagl

CHAPTER 5
THE USE OF OZONATION/MEMBRANE PROCESS FOR THE

CONTROL OF DISINFECTION BY-PRODUCTS

5.1 INTRODUCTION

The production of disinfection by-products (DBPs) during the chlorination of
drinking water is an important issue, arising from the discovery that some DBPs, such
as trihalomethanes (THMs), are carcinogenic (Kool et al. 1985; Morris et al. 1992;
Mughal 1992). Several oxidants, e.g. ozone, chloramines, and chlorine dioxide, may
be used instead of chlorine to reduce the formation of chlorinated DBPs.

Ozone is a powerful oxidant and preferentially oxidizes electron-rich moieties
and aromatic alcohols. It also acts as an electrophil reacting with molecular sites
having a strong electronic density and as a nucleophil reacting with molecular sites
with electronic deﬁciencies (Bablon et a1. 1991). It was found that increasing the
ozone dosage in water treatment processes decreases the formation of THMs and
haloacetic acids (HAAs) (Cipparone et al. 1997; Lee 2001). Amy et al. (1988) also
reported that in the presence of natural organic matter (NOM), ozonation results in the
formation of partially oxidized compounds, which are less reactive with chlorine in
forming THMs.

Ozone decomposes quickly when dissolved organic carbon (DOC) and
hydroxyl radicals are present in water (Bablon et a1. 1991; Hoigné and Bader 1975;
Staehelin and Hoigné 1985). A secondary disinfectant is usually required when using
ozone as the primary disinfectant because of the rapid disappearance of ozone residual

(Oxenford 1996). The use of chlorine or chloramines as a secondary disinfectant

136

 

 

 

following ozonation reduces the formation of THMs and HAAs compared to that
observed when using only chlorine or chloramines (Richardson et al. 1999).

While ozonation can help to minimize the formation of THMs and HAAs, the
reactions that occur during ozonation produce other by-products. Reported ozonation
by-products include aldehydes (formaldehydes, acetaldehydes, glyoxal, and

methylglyoxal), ketone, glyoxylic acid, and pyruvic acid (Paode et a1. 1997; Weinberg

 

and Glaze 1996). Although some of them are of particular concern due to their

mutagenicity and carcinogenicity (Bull and McCabe 1984), these by-products are

 

easily biodegradable and may serve as substrates for microbial regrowth in the
distribution system. High removal efﬁciencies of aldehydes and ketoacids were
observed when using ozonation-biological ﬁltration processes to treat ozonated water
(Grifﬁni et al. 1999).

The purpose of this study was to evaluate the effect of ozonation/membrane
process on the formation of DBPs by treating Lake Lansing water under various

operational conditions. Chlorine residual and the concentration of THMs, HAAs,

 

aldehydes, ketone, and ketoacids were monitored at different treatment stages.
Bioﬁltration followed the ozonation/membrane system, allowing for a study of the

effect of biotreatment on the formation of DBPs.

5.2 MATERIALS AND METHODS
5.2.1 Experimental Setup and Protocol

A specially designed ozonation/membrane system was used to investigate the
formation of several DBPs (i.e., THMs, HAAs, aldehydes, ketones, and ketoacids)
resulting from the treatment process. The system was the same one that was

described in Chapter 4 and a schematic diagram of this system is shown in Figure 4.1.

137

Tubular ceramic membranes (Clover-leaf design (containing three channels), CéRAM
Inside, Tami North America, St. Laurent, Quebec, Canada) with a molecular weight
cut-off of 15 kD were used in the system. The characteristics of the laboratory-scale
ultraﬁltration membrane are summarized in Table 4.1. Teﬂon tubing and stainless
steel joints and valves were used throughout the system. Other components included:
3.5-liter and 1.5-liter water—jacked glass reservoirs made of Pyrex glass, and a simple
Y inline mixer (Ozone Service, Burton, BC, Canada).

Ozone was generated by feeding pure oxygen gas from a gas cylinder to the
ozone generator (Model OZZPCS, Ozotech Inc., Yreka, Calif). The oxygen gas
stream was dried using a molecular sieve trap prior to ozone generation. The ozone
gas concentration was controlled using the electric current of the ozone generator.
The wasted ozone gas was vented after passing the gas through a 2% potassium iodide
(KI) solution to destroy any residual ozone gas.

Several operational factors, such as ozone concentration in the inlet gas phase,
gas ﬂow rate, and water temperature, which inﬂuence the ozone mass transfer and
solubility, were investigated. Table 5.1. summarizes the experimental conditions used

in this study.

Table 5.1 Experimental conditions.

 

 

 

 

 

 

Parameters Values
Inlet ozone dose (g/m3) 1.5, 2.5, 5.8
Water ﬂow rate (L/min) 2.75
Ozone gas ﬂow rate (mL/min) 50, 75, 100
Water temperature (°C) 10, 20, 30

 

 

138

 

 

 

 

Lake Lansing water and ozone gas were continuously fed into the membrane
module of the ozonation/membrane system. Prior to ozonation, the raw water was
pre-ﬁltered through a 0.45 pm ﬁlter to remove coarse suspended and colloidal solids.
Since the collection of permeate from the membrane required several hours, a
collection bottle with a cap was used and placed in an ice bath to chill and preserve
the sample. The ﬁrst 500 mL of permeate was collected and marked as “Permeate 1”.
The next 1.5 L of permeate was collected and marked as “Permeate 2”. Permeate 1
and 2 represent the characteristics of the permeate collected at different time intervals
and provide information on the performance of the ozonation/membrane system. The
concentrate was recirculated back to the raw water reservoir. Herein, the water
samples from ﬁltered raw water, permeate 1, and permeate 2 are noted as FRW, P1,
and P2, respectively. After the system was stopped, the water remained in the water
reservoir, denoted as WT. The F RW and the P2 water were further treated by
bioﬁltration at room temperature for seven days to investigate the efﬁciency of
biotreatment on the removal of DBPs. Water samples collected from the ﬁltrate of
bioﬁltration columns are noted as B-FRW and B-P2. The concentrations of DOC,
humic substances (HS), non-humic substances (NHS), chlorine residual, simulated
distribution system (SDS) THMs, SDS HAAs, aldehydes (formaldehyde,
acetaldehyde, propionaldehyde, glyoxal, and methyl glyoxal), ketones (acetone and 2-
butanone), and ketoacids (glyoxylic acid, pyruvic acid, and ketomalonic acid) were
monitored. The dissolved ozone residual in collected samples was quenched by
sparging the solution with a high purity helium gas (99.999 %) before analysis.
Earlier studies show that quenching can be achieved in this manner in several seconds

(Hemer 1999).

139

 

 

 

5.2.2 Water Source

Experiments were carried out on Lake Lansing (Haslett, MI) water. Water
quality characteristics of Lake Lansing are summarized in the previous chapter (see
Table 4.3). Lake Lansing waters were collected in ﬁve-gallon tanks, properly labeled,
and stored at 4°C. The maximum storage period was seven days. All waters were

ﬁltered through a 0.45-um ﬁlter and warmed to room temperature before testing.

5.2.3 Chlorine Residual

The chlorine residual was determined by the iodometric method as described
in Standard Methods 45008 (Standard Methods 1998). A 50 mL water sample was
acidiﬁed with the addition of 2.5 mL of acetic acid. After approximately 1 g of K1
and a few drops of starch indicator were added, the solution turned blue. Then the
solution was titrated against 0.01 N sodium thiosulfate (Na2S203). Titration was
continued until the blue color became colorless. The titration volume was recorded

and the chlorine residual was determined.

5.2.4 SDS THMs

Four THM compounds, chloroform (CHC13), bromodichloromethane
(CHBrClz), dibromochloromethane (CHBrzCl), and bromoforrn (CHBr3), were
monitored in this study. These compounds were measured in chlorinated FRW, P1,
P2, WT, B-FRW, and B-P2 waters. The analysis of SDS THMs used Method 5710
(Standard Methods 1998). Water samples were dosed a chlorine concentration that
allowed residual chlorine concentration in the range of 0.5 to 2 mg/L after 48 hours
incubation at room temperature according to the procedures in Method 2350

(Standard Methods 1998). Then samples were prepared according to the THMF P

140

 

 

 

method in Method 5710 (Standard Methods 1998). THMs were extracted by hexane
and analyzed by a gas chromatography (Hewlett-Packard 5890 Series II) that was
equipped with 63Ni electron capture detector (ECD). A 30 m, 0.53 mm ID. DB-624
column (J&W Scientiﬁc, Folsom, CA) was used. The oven temperature was
programmed from 50 °C increasing to 150 °C with a rate of 10 °C/min. The ﬂow rate
of the carrier gas (N2) was 12.0 mL/min. The injector temperature and detector

temperature were 275 and 350 °C, respectively.

5.2.5 SDS HAAs

HAAs were determined using USEPA Method 552.2. The compounds
determined were monochloroacetic acid (MCAA), monobromoacetic acid (MBAA),
dichloroacetic acid (DCAA), bromochloroacetic acid (BCAA), trichloroacetic acid
(TCAA), and dibromoacetic acid (DBAA). These compounds were measured in
FRW, P1, P2, WT, B-FRW, and B-P2 waters chlorinated according to the procedures
in Method 2350 (Standard Methods 1998). A Perkin Elmer Autosystem gas
chromatograph equipped with an ECD, an autosampler, and a DB-I column (J&W
Scientiﬁc, Folsom, CA) was used for the analysis. The oven temperature was
programmed to hold 15 minutes at 32 °C, then increase to 75 °C with a rate of 5
oC/min and hold 5 minutes, then increase to 100 °C with a rate of 5 °C/min. The
carrier ﬂow was 1.0 mL/min. The injector temperature and detector temperature were

200 and 260 °C, respectively.

5.2.6 Aldehydes and Ketones

Measured aldehydes and ketones included formaldehyde, acetaldehyde,

propionaldehyde, glyoxal, methyl glyoxal, acetone, and 2-butanone. These

141

 

 

 

compounds were measured in F RW, P1, P2, WT, B-FRW, and B-P2 waters. The
method used followed the instructions in USEPA Method 556 (Munch et al. 1998). A
Perkin Elmer Autosystem gas chromatograph equipped with an ECD, an autosampler,
and a DB-5 column (J&W Scientiﬁc, Folsom, CA) was used in the analysis. The
oven temperature was programmed to hold 1 minute at 50 °C, then increase to 220 °C
at a rate of 4 °C/min, then increase to 250 °C at a rate of 20 °C/min and hold for 5
minutes. The carrier ﬂow was 1.0 mL/min. The injector temperature and detector

temperature were 180 and 300 °C, respectively.

5.2.7 Ketoacids

Glyoxylic acid, pyruvic acid, and ketomalonic acid were monitored in this
study. In a 30 mL vial containing 20 mL water sample, 200 mg potassium hydrogen
phthalate (KHP) and 1 mL of a 15 mg/mL freshly prepared aqueous 0-(2,3,5,6-
Pentaﬂuorobenzyl)-hydroxylamine hydrochloride (PFBHA) reagent were added.
After incubation at 45 C for 1.75 hours, the sample was cooled to room temperature.
Following the addition of 0.05 mL of concentrated sulfuric acid, 4 mL of methyl-tert-
butyl ether (MTBE) containing 3 mg/L of dibromopropane was added and the sample
was shaken by hand for 3 minutes. After allowing the phases to separate for
approximately 5 minutes, approximately 3 mL of the upper MTBE layer was
transferred to a 15 mL glass vial and 2 mL 10% sulfuric acid in methanol was added.
The vial was capped with PTFE faced septa and a screw cap, then placed in a heating
block at 50 C and maintained for 1.5 hours. After cooling to room temperature, 4
mL of 10% sodium sulfate (NaZSOa) solution was added and the tube was shaken by
hand for 3 minutes. After allowing the phases to separate for approximately 5

minutes, the upper MTBE layer was then submitted for analysis. A Perkin Elmer

142

EW'

 

 

Autosystem gas chromatograph equipped with an ECD, an autosampler, and a DB-5
column (J&W Scientiﬁc, Folsom, CA) was used in the analysis. The settings of GC

were the same as used in aldehydes analyses (see Section 5.2.5).

5.2.8 Other Analyses
The concentration of gas-phase ozone, DOC, UV-254, humic substances, and
non-humic substances were measured by using the same method as described in

Section 4.2.

5.3 RESULTS AND DISCUSSION
5.3.1 Chlorine Demand

Chlorine demand values and DOC concentrations for water samples that were
collected at various treatment stages under different experimental conditions are
summarized in Table 5.2. The value of chlorine demand was calculated by
subtracting the concentration of chlorine residual from the initial chlorine dosage.
Figures 5.1 to 5.3 show the variation of chlorine demand of different treatment stages
at various experimental conditions. According to the experimental results shown in
Table 5.2, it is interesting to ﬁnd that the levels of chlorine demand and DOC
concentrations for water samples of FRW, P1, P2 and WT ﬂuctuate between 7.0 ~ 8.5
mg/L and 6.0 ~ 10.9 mg/L, respectively. The maximum reduction of the chlorine
demand was less than 10% (see the P1 water sample at water temperature 30 °C) and
the demand increased over its initial value in some water samples (see the P2 and WT
samples at water temperature 30 °C, the WT samples at 75 mL/min gas ﬂowrate, and
the WT sample at 5.8 g/m3 O3 dose). The removal of chlorine demand by the

ozonation/membrane system was limited in this study. An explanation for this

143

 

 

 

observation is that two competing reactions were occurred during ozonation. One
reaction is the attack of nucleophilic sites on DOC by molecular ozone, thereby
decreasing the reactivity of DOC with chlorine. On the contrary, ozone can generate
new chlorine-consuming sites by the hydroxylation of aromatic moieties in humic
material. Jadashecart et al. (1991) observed a signiﬁcant increase of chlorine demand
when fulvic acid was ozonated at pH 8 and in the absence of bicarbonate, i.e., at
conditions where the radical-type pathway (hydroxyl radical reactions) predominates.
However, they also found that the ozonation of fulvic acid in the presence of
bicarbonate led to a small decrease in the chlorine demand, as a result of the direct
reaction with molecular ozone. Thus it is reasonable to presume that the competition
of these two reactions resulted in the limited increase and decrease of chlorine
demand in our study. The variation of DOC concentration has been discussed in the
previous chapter (see Section 4.3.2, Chapter 4).

Figures 5.1 to 5.3 also show that the biotreatment can effectively remove the
chlorine demand of water. Bioﬁltration alone without pre-ozonation (i.e., B-FRW
waters) removed 30 to 55% of chlorine demand. The removal efﬁciency of chlorine
demand with bioﬁltration following the ozonation/membrane system ranged from
60% up to 90% for B-P2 waters. This indicates that a large fraction of the chlorine-
consuming material was biodegradable and combining ozonation and biotreatment
processes can reduce the chlorine demand of water.

Figure 5.4 presents the correlation between the concentration of DOC and
chlorine demand of Lake Lansing water tested in this study. According to the
experimental results, chlorine demand and DOC were positively and linearly
correlated under the conditions we tested. The correlation is statistically signiﬁcant (p

= 0.01 at 95% CI).

144

 

 

 

Table 5.2 Chlorine demands and DOC concentrations for samples collected at various
treatment stages under different experimental conditions.

 

Temp. FRW P1 P2 WT B-FRW B-P2
(°C)l DOC CD DOC CD DOC CD DOC CD DOC CD DOC CD

 

 

10 9.6 8.5 7.7 8.2 8.1 8.5 9.3 8.3 5.8 4.5 1.9 1.4

 

 

20 10.9 7.5 8.7 7.3 7.6 7.4 6.6 7.3 7.6 5.2 3.3 3.1

 

30 9.0 7.7 7.3 7.0 6.0 7.9 6.3 7.9 8.0 4.4 2.6 2.0

 

Gas FRW P1 P2 WT B-FRW B-P2
ﬂowrate2 DOC CD DOC CD DOC CD DOC CD DOC CD DOC CD

 

 

50 8.7 7.4 7.0 7.2 6.1 7.0 6.7 7.2 7.5 3.4 2.8 0.7

 

75 10.4 7.5 8.3 7.5 7.6 7.2 8.3 7.6 6.8 5.2 2.6 2.4

 

100 10.9 7.5 8.7 7.3 7.6 7.4 6.6 7.3 7.6 5.2 3.3 3.1

 

O3 FRW P1 P2 WT B-FRW B-P2
dose3 DOC CD DOC CD DOC CD DOC CD DOC CD DOC CD

 

 

1.5 8.7 8.0 6.4 7.9 6.1 7.4 7.3 7.5 6.9 4.0 2.0 1.5

 

2.5 10.9 7.5 8.7 7.5 7.6 7.2 8.3 7.6 6.8 5.2 2.6 2.4

 

 

5.8 8.9 8.0 6.9 7.9 7.7 7.9 7.9 8.2 6.4 4.5 1.1 1.7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100mL/min; O3 dose, 2.5 g/m .

Water ﬂow rate, 2.75 L/min; O3 dose, 2.5 g/m3; water temperature, 20 °C.

Water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100mL/min; water temperature, 20 °C.

DOC, CD (chlorine demand), and O3 dose: g/ms; Gas ﬂow rate: mL/min.

F RW: ﬁltered raw water; Pl: permeate 1; P2: permeate 2; WT: samples from the
water tank; B-FRW: bioﬁltrated F RW water; B-P2: bioﬁltrated P2 water.

9:591"?

 

I FRW

120
100
80
60
40
20

% of Chlorine demand

   

 

 

10 20 30

Temperature (°C)

Figure 5.1 The percentage variation of chlorine demand (CDi/CDO) of different
treatment stages at various water temperatures. (Operating conditions:
water ﬂow rate, 2.75 L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/ml)

145

I FRW

120
100
80
60
40
20

% of Chlorine demand

 

 

 

50 75 1 00
Gas ﬂow rate (mL/min)
Figure 5.2 The percentage variation of chlorine demand (CDi/CDo) of different

treatment stages at various gas ﬂow rates. (Operating conditions: water
ﬂow rate, 2.75 L/min; 03 dose, 2.5 g/m3; water temperature, 20 °C)

120
100
80
60
40
20

 

% of Chlorine demand

 

 

03 dose (g/ma)

Figure 5.3 The percentage variation of chlorine demand (CDi/CDO) of different
treatment stages at various ozone doses. (Operating conditions: water ﬂow
rate, 2.75 L/min; gas ﬂow rate, 100 mL/min; water temperature, 20 °C)

146

12.0 ’ --——-— i

    

 

 

:3; y = 0.8183x + 0.5034 .
g 9.0 R2=0.6611 i
'8 O
E
m 5.0
'0
g 3.0
_q
5 ' .
0.0 5' J
0.0 3.0 5.0 9.0 12.0

ooc (mg/L)

Figure 5.4 Correlation between chlorine demand and the concentration of DOC of
Lake Lansing water treated by the ozonation/membrane/bioﬁltration
system. (Correlation is statistically signiﬁcant at 95% 01.)

5.3.2 THMs and HAAs

The effects of water temperature on the concentration of THMs and HAAs of
different treatment stages are shown in Figures 5.5 and 5.6, respectively. As seen in
Figure 5.5, the decreases of the formation of THMs in the P1, P2, and WT water at 10
°C water treatment temperature were less than those at 20 °C and 30 °C water
treatment temperature. However, the concentration of HAAs in the P1, P2, and WT
water decreased at 20 °C and 30 °C but increased at 10 °C (see Figure 5.6). In a
previous work, the effects of temperature on the formation of THMs and HAAs of
Lake Lansing water were investigated by using a continuous stirred reactor. Its
results are summarized in Table 5.3. It was observed that the THMs and HAAs
concentrations increased slightly as the temperature increased from 5 °C to 35 °C.

However, the experimental results in this study show that the THMs concentrations

147

 

under all three operating conditions and the HAAs concentrations of treated water at
20 °C and 30 °C decreased signiﬁcantly as the water temperature increased (p < 0.01).
This was probably the result of a longer hydraulic retention time used in this study,
and higher ozone mass transfer rate into the water and higher reaction rate constant at
higher water temperature. Similar observations were reported by Zhang et al. (2001).

The measured concentrations of HAAs in P1, P2, and WT water were greater
than that of F RW water when Lake Lansing water was treated at 10 °C (Figure 5.6).
It may be that HAAs precursors were generated by the partial oxidation of humic
substances more quickly than they were destroyed by ozonation. On the other hand,
the increase in HAA concentrations could result from reactions of ozone, exposing
more active sites in the NOM matrix when ozonation broke down large molecules.
Some researchers observed the same trend when measuring THM concentration
during ozonation. Yamada et al. (1986) observed that the THM formation potential
(THMF P) increased in the initial stage of low ozone dose and then gradually
decreased with respect to the increase in the amount of ozone consumed when humic
acids were ozonated. Rice (1980) explained that ozone can destroy some types of
THM precursors and produce THM precursors from some types of naturally occurring
humic and fulvic materials. Several researchers reported similar observations
(Duguet et al. 1985; Glaze et al. 1982; Wallace et a1. 1988).

Biotreatment alone or following ozonation before chlorination effectively
reduced the formation of THMs and HAAs. From Figures 5.5 and 5.6, it can be seen
that the concentrations of THMs and HAAs decreased signiﬁcantly in biotreated
F RW waters (B-FRW) (except the HAAs concentration of B-FRW water at 10 °C)
and biotreated P2 waters (B-P2) (p < 0.01). The removal efﬁciency of THMS

precursors by ozonation/membrane/bioﬁltration (B-P2 water) increased as the water

148

temperature increased (p < 0.05), however, that of HAAs increased as the water
temperature increased from 10 °C to 20 °C but decreased as the water temperature
further increased from 20 °C to 30 °C (p < 0.01). The experimental results indicates
that biological ﬁltration following ozonation can effectively decrease the formation of
TH Ms and HAAs, but the removal efﬁciency of HAAs precursors by bioﬁltration
(either alone or following ozonation/membrane system) was independent to the
treated water temperature and may depend on the characteristics of raw water, the

characteristics of P2 water, or the condition of microorganisms in the bioﬁlter.

I FRW

300 c

 

THMs (jig/L)

   

 

 

 

20 30
Temperature (°C)

Figure 5.5 Concentrations of THMs of different treatment stages at various water

treatment temperatures. (Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/m3)

149

   

80
Q .
360j
‘0
$40,
20’
0

 

10 20 30
Temperature (°C)

Figure 5.6 Concentrations of HAAs of different treatment stages at various water
treatment temperatures. (Operating conditions: water ﬂow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/m3)

Table 5.3 Concentrations of THMs and HAAs for Lake Lansing water treated by

continuous stirred reactor. (Operating conditions: hydraulic retention time,
12.5 min; ozone dosage, 1 mg O3/mg C)

HAAs*
5 l 1

1 1
35 1 108.5
* including: monochloroacetic acid (MCAA), monobromoacetic acid (MBAA),

dichloroacetic acid (DCAA), trichloroacetic acid (TCAA), and dibromoacetic acid
(DBAA).

 

150

The effects of ozone gas ﬂowrate on the concentration of THMs and HAAs of
different treatment stages are shown in Figures 5.7 and 5.8, respectively. As seen
from Figure 5.7, the THMs concentrations in water treated by the
ozonation/membrane system (P2 water) decreased from 177.52 i 6.67 rig/L to 1 10.24
i 1.96 ug/L and the increase in removal efﬁciency as the gas ﬂow rate increased from
50 mL/min to 100 mL/min was statistically signiﬁcant (p < 0.01). The decrease of
THMs levels can be explained by the increase of ozone dosage at high gas ﬂowrate.
Figure 5.8 shows that HAAs concentrations of P2 water decreased with increasing the
gas ﬂowrate. However, the removal efﬁciency with increasing the gas ﬂow rate was
not statistically signiﬁcantly decreased (p > 0.05). The THM precursors may have
higher reaction rates with ozone than those of HAAs, which results in the different
effect of ozone on the removal efﬁciencies.

The data in Figures 5.7 and 5.8 also illustrate the performance of the
biotreatment on the removal of THMs and HAAs precursors. The experimental
results showed that biological ﬁltration alone or following ozonation/membrane
decreased the formation of THMs and HAAs. The percent removal of THMs by
bioﬁltration alone and following ozonation/membrane system ranged from 22% to
33% and 78% to 89%, respectively. For HAAs, the removal efﬁciency ranged 16% to
63% for bioﬁltration alone and 79% to 92% for following the ozonation/membrane

system.

151

 

 

 

100
Gas ﬂow rate (mL/min)

Figure 5.7 Concentrations of THMs of different treatment stages at various gas ﬂow
rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose, 2.5
g/m3; water temperature, 20 °C)

HAAs (pg/L)

 

 

 

75 100
Gas ﬂow rate (mL/min)

Figure 5.8 Concentrations of HAAs of different treatment stages at various gas ﬂow
rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose, 2.5
g/m3; water temperature, 20 °C)

152

and HAAs, respectively. As expected, the removal efﬁciency of THMs precursors
signiﬁcantly increased as the ozone dose increased (p < 0.01). The
ozonation/membrane system removed 28%, 51%, and 60% of THMs precursors at
ozone dose of 1.5, 2.5, and 5.8 g/m3, respectively. If, at higher ozone doses, more
ozone reacted with the sites on the NOM molecule that are very reactive to chlorine, a

Figures 5.9 and 5.10 show the effects of ozone dose on the formation of THMs
reduction in their ability to generate THMs during the following chlorination would
131‘

 

occur. The indirect action, being less selective, will degrade certain molecular sites
very reactive to chlorine, but will also react on other sites not very reactive to chlorine.
However, the removal of HAAs precursors did not follow the same trend. Although
ozonation decreased the formation of HAAs, the removal efﬁciency of HAAs
precursors was independent of ozone doses and ranged between 29% and 48%. The
bioﬁltration following ozonation/membrane signiﬁcantly decreased the formation of
THMs. The removal efﬁciency of THMs ranged from 65% to 89% and increased as
the ozone dose increased from 1.5 g/m3 to 5.8 g/m3 (p < 0.01). However, the removal
efﬁciency of HAAs precursors of B-P2 water was independent of ozone doses (p =

0.29) and ranged between 86% and 92%.

 

According to our experimental results, chloroform was the predominant THM
species and DCAA and TCAA were the predominant HAA species in all water
samples, including ﬁltered raw water, ozonation/membrane efﬂuent and biotreated
water. Increases in the percentage of bromodichloromethane, dibromochloromethane,
BCAA, and DBAA were observed in biotreated water. This was probably caused by
the increase of bromide-to-DOC ratio in the water sample after bioﬁltration. This
observation is consistent with the ﬁnding of other researchers (Coleman et al. 1992;

Miltner et al. 1992; Wobma et al. 2000).

153

 

250 ,
200
150 ,
100
50 1

THMs (pg/L)

    

 

   

2.5
03 dose (g/ma)

Figure 5.9 Concentrations of THMs of different treatment stages at various ozone
doses. (Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C)

100 ,-

HAAs (jig/L)

 

 

 

2.5 5.8
03 dose (g/m3)

Figure 5.10 Concentrations of HAAs of different treatment stages at various ozone
doses. (Operating conditions: water ﬂow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C)

154

The correlation between THMs and HAAs based on the data obtained from
this study is shown in Figure 5.11. The relationship between THMs and HAAs was
statistical signiﬁcance (at p = 0.01 level at 95% 01.). Several researchers have
indicated that surrogate parameters such as TOC, UV-254, humic substances can be
used as indicators of DBP precursors (Chiang et al. 2002; Edzwald et al. 1985;

Nieminski et a1. 1993). Therefore, the relationships between DBPs and surrogates

 

such as DOC, UV-254, and humic substances of Lake Lansing water tested in this

3:1
l :—

study were also investigated. Figures 5.12 to 5.14 show the relationships between

I‘D)!!—
I

THMs and surrogates and Figures 5.15 to 5.17 present that of HAAs. As seen from
the experimental results, both THMs and HAAs were positively correlated with
surrogate parameters under the conditions we tested. They are all linearly related and

the correlations are statistical signiﬁcance (at p = 0.01 level at 95% CI).

 

100 - n-a-

y = 0.2757x + 9.8604
80
50

4O

HAAs (pg/L)

20

 

 

 

0 100 200 300
THMs (pg/L)

Figure 5.1 1 Correlation between the THMs and HAAs of Lake Lansing water treated
by the ozonation/membrane/bioﬁltration system. (Correlation is
statistically signiﬁcant at 95% CI)

155

 

300 1

    

25 y=21.989x-7.3042 .
A R2=0.6368 e .
d 200 i
3
a, 150
2
E 100

50 ~

0
0

O 3 6 9 12

DOC (mg/L)

Figure 5.12 Correlation between THMs and the concentration of DOC of Lake
Lansing water treated by the ozonation/membrane/bioﬁltration system.
(Correlation is statistically signiﬁcant at 95% CI.)

300 — ~
250

—|~l\)
010
cc

 

THMs (pg/L)

100 w

y = 802.23x + 103.04
R2 = 0.6908

U1
O

 

O

0 0.05 0.1 0.15 0.2
uv-254 (Abs)

Figure 5.13 Correlation between THMs and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.)

156

 

300 —
250
200
150
100 .
50

 

 

THMs (pg/L)

y = 27.948x + 72559
R2 = 0.5994

 

 

Humic substances (mg/L) ~

'1‘:

Figure 5.14 Correlation between THMs and the concentration of humic substances of
Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% CI)

    

 

 

 

100 — * ‘
y = 8.0481X - 5.6296 0 e o

80 R2 = 0.7074
g l
m :
3 60
(I)
g 40

20

O
O l
0 3 6 9 12
DOC (mg/L)

Figure 5.15 Correlation between HAAs and the concentration of DOC of Lake
Lansing water treated by the ozonation/membrane/bioﬁltration system.
(Correlation is statistically signiﬁcant at 95% CI)

157

100

20*

80
Q
0')
3 60
(I)
g 40
0

 

 

y = 202.35x + 42.59
R2 = 0.4037

0.05 0.1 0.15 0.2
uv-254 (Abs)

Figure 5.16 Correlation between HAAs and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.)

100
80
60

4O

HAAs (ﬁg/L)

20

 

 

 

00
.1
. y = 8.3432x + 30.696
R2 = 0.4906
2 4 6 8

Humic substances (mg/L)

Figure 5.17 Correlation between HAAS and the concentration of humic substances of
Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% Cl.)

158

5.3.3 Aldehydes, Ketones, and Ketoacids

Five aldehydes (formaldehyde, acetaldehyde, propionaldehyde, glyoxal, and
methyl glyoxal), two ketones (acetone and 2-butanone) and three ketoacids (glyoxylic
acid, pyruvic acid, and ketomalonic acid) were monitored in this study. Among these
compounds, acetaldehyde, propionaldehyde, and acetone were not detected in all
water samples and the concentrations of 2-butanone were too variable to allow ﬁrm
conclusions. These more volatile compounds may have been lost when ozonated
sample were purged with high purity nitrogen to quench residual ozone and stop the
reactions. Alternatively, their loss could have been the result of volatilization during
the lengthy times required to collect samples of the P1 and P2 waters. Acetaldehyde,
propionaldehyde, and acetone were also not detected in all non-ozonated water
samples such as FRW and B-FRW. Therefore, only the experimental results of total
aldehydes including formaldehyde, glyoxal, and methyl glyoxal and total ketoacids
are discussed in this section.

Formaldehyde and pyruvic acid were the two principal products of ozonation
in our study. This observation was consistent with the ﬁndings of Coleman et al.
(1992) in studies where they ozonated model humic acid solutions and of Schechter
and Singer (1995) and Paode et al. (1997) in studies where they ozonated raw surface
waters. Figures 5.18 and 5.19 show the effects of water temperature on the
concentrations of aldehydes and ketoacids, respectively. Aldehydes levels and
ketoacids concentrations increased signiﬁcantly as a result of ozonation at all three
experimental conditions (p < 0.01). Interestingly, the variations in the aldehyde and
ketoacid levels with time were different. The formation of aldehydes increased to a
maximum then was followed by a gradual decrease with the progress of ozonation

time in all three water temperatures. Ko et al. (1998) and Yamada and Somiya (1980)

159

 

 

reported the same observation. It is assumed that the production of aldehydes during
the early stages of ozonation is greater than the rate of their decomposition by
oxidation. The aldehydes are slowly destroyed by molecular ozone or the hydroxyl
radical. As a result of the complicated production/decomposition reactions, the
concentration of aldehydes reached a maximum then declines. However, the
formation of ketoacids increased with the progress of ozonation time, i.e., the
ketoacids concentrations in the FRW < Pl < P2 < WT. It is likely that the methyl
glyoxal was oxidized to pyruvic acid and the decomposition rate of pyruvic acid was
lower than that of methyl glyoxal (Weinberg and Glaze 1996; Yamada and Somiya
1980). The concentrations of ketoacids were much higher than those of the aldehydes
in ozonated water samples. This is consistent with other reports (Carlson and Amy
1997; Grifﬁni et al. 1999; Xie and Reckhow 1992).

The concentration of total aldehydes in P2 water increased from 154.8 i 15.3
ug/L at 10 °C water temperature to 21 1.3 i 5.0 ug/L at 20 °C and then decreased to
123.0 :1: 4.9 ug/L at 30 °C. The correlation between water temperature and aldehydes
level was not signiﬁcant (p > 0.05). Hoigné (1982) indicated that the direct reaction
of molecule ozone is the primarily pathway for the formation of aldehydes. Schechter
and Singer (1995) also reported this observation. It is reasonable to assume that more
molecular ozone is involved in the direct reaction pathway at low water temperature.
Also, at low water temperatures, the permeate ﬂux of the membrane system was
smaller than that at high temperatures which resulted in a longer retention time at low
temperatures than high temperatures. At low water temperatures, the aldehyde
concentration reached a maximum earlier, followed by the destruction of aldehydes,
resulting in the lower aldehyde levels in water. In contrast, at higher water

temperatures, it is hypothesized that the production of aldehydes through direct

160

"r!

1:1-La

 

 

reaction was less signiﬁcant (than at lower temperatures) and the destruction of
aldehydes by the hydroxyl radical was more signiﬁcant than at lower temperatures.
This could explain the experimental results in Figure 5.18.

The correlation between water temperature and ketoacids level was also not
signiﬁcant (p > 0.05) (see Figure 5.19). The effects of water temperature on the
formation of ketoacids may be explained as the same reason of aldehydes. Yamada
and Somiya (1980) observed that pyruvic acid, which was the principle species of
ketoacids in our experiments, gradually increased to a maximum and then very slowly
decreased. They also reported that pyruvic acid increased with reaction time and did
not react appreciably with ozone. According to our data, the rate of production of
ketoacids appeared to be faster than its destruction, which resulted in the
accumulation of ketoacids in WT water. The variation of initial F RW water quality
may also contribute to the difference in ketoacids formation at different water

temperature.

161

 

_ IFRW

 

250 —- - --—— — IP1
DPZ

g 200 awr
c» BB-FRW
f 150 »9 __ IB-P2
a) .3. l
E 100 :
OJ
'2
< 50 i _

 

 

 

 

 

 

 

 

 

1 0 20 30
Temperature (°C)

Figure 5.18 Concentrations of aldehydes of different treatment stages at various water
treatment temperatures. (Operating conditions: water flow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/m3)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1400 IFRW _
=— ; IP1
A 1200 ~ I am
<1 - I EIWT _
E’ 1000 - {EB-FRW -
7; 800 ~ ‘ IB-P2
.12
8 600 -
*6 400 - s
x 200 I  
10 20 30

Temperature (°C)

Figure 5.19 Concentrations of ketoacids of different treatment stages at various water
treatment temperatures. (Operating conditions: water flow rate, 2.75
L/min; gas ﬂow rate, 100 mL/min; O3 dose, 2.5 g/m3)

162

The effects of ozone gas flow rate on the formation of aldehydes and ketoacids
are shown in Figures 5.20 and 5.21, respectively. When a gas ﬂow rate of 100
mL/min was used, the formation of aldehydes reached a maximum at the Pl stage.
Under these conditions a greater amount of ozone was delivered than that at 75 and 50
mL/min. As less gas flow rate was applied, the aldehydes level reached its maximum
later. According to Figure 5.20, the maximum aldehyde concentration increased
signiﬁcantly as the ozone gas flow rate increased (p < 0.01). The formation of

ketoacids also signiﬁcantly increased as the gas ﬂow rate increased (p < 0.01) (see

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5.21).

250 - - * - - - ~ -"“:E§W
A 0P2
a 200 awr
g '5 £- EBB-FRW
m 150 ' _, IB-PZ
0’ !
E 100
3
< 50

50 75 100

Gas ﬂow rate (mL/min)

Figure 5.20 Concentrations of aldehydes of different treatment stages at various gas
flow rates. (Operating conditions: water ﬂow rate, 2.75 L/min; O3 dose,
2.5 g/m3; water temperature, 20 °C)

163

 

1400 I FRW

 

 

 

 

 

 

 

 

 

 

 

 

 

I P1

1200 ﬂ :1 pz
3 7 E ii ClWT
3 1°00 _ i EBB-FRW
a, 800 xi IB-P2
:9 '1,
g 600
.9
:02 400

200

O l ”571 l
50 75 100

Gas ﬂow rate (mL/min)

Figure 5.21 Concentrations of ketoacids of different treatment stages at various gas
flow rates. (Operating conditions: water ﬂow rate, 2.75 L/min; 0;, dose,
2.5 g/m3; water temperature, 20 °C)

Figure 5.22 presents the effects of ozone dose on the formation of aldehydes at
different treatment stages. The maximum concentration of aldehydes increased as the
ozone dose increased from 2.5 to 5.8 g/m3 (p < 0.01), but the difference of the level
between ozone doses of 1.5 g/m3 and 2.5 g/m3 was not signiﬁcant (p > 0.05). This
probably was due to the characteristics of the F RW water. However, the time needed
for aldehydes to reach maximum was shorter at higher ozone doses than at lower ones.

The effects of ozone dose on the ketoacids concentrations are shown in Figure
5.23. The maximum concentration of ketoacids signiﬁcantly increased as the ozone
dose increased (p < 0.01).

The concentrations of biotreated P2 water (B-PZ) of aldehydes and ketoacids
are also illustrated in Figures 5.18 to 5.23. The removal efﬁciency by bioﬁltration of
aldehydes is shown in Table 5.4. It appears that very high removal efﬁciency, ranged

between 85.4% and 94.8%, was achieved in the bioﬁlter, which followed the

164

 

ozonation/membrane system. The biotreatment process also removed ketoacids very
effectively. Only two samples were found to have residual ketoacids concentrations.
The concentration of ketoacids in P2 waters ranged between 632 ug/L and 1476 ug/L
and the ketoacids residuals in B-P2 waters were 38.1 ug/L and 42.6 ug/L. The

percent removal of ketoacids by bioﬁltration was between 94.1% and 100%. Several
researchers have reported that both aldehydes and ketoacids are very biodegradable,
with removal efﬁciencies between 80 and 100% (Grifﬁni et al. 1999; Price et al. 1993;

Wobma et al. 2000).

 

 

 

 

 

 

 

 

 

 

300 4 . . .._ I FRW
l P1
3 250 , 1:] P2
‘5, E] WT
s 200 a B-FRW
8 150 _ B-P2
l
E 100 T
2
< 50
0 i ' . “312714
1.5 2.5 5.8
03 dose (g/m3)

Figure 5.22 Concentrations of aldehydes of different treatment stages at various ozone
doses. (Operating conditions: water flow rate, 2.75 L/min; gas ﬂow rate,
100 mL/min; water temperature, 20 °C)

165

 

180‘ I FRW
El P1
160‘ 1:: P2 if}
140' [3 WI"
120. El B-FRW *
100,.E B'PZ
800
600
400
200 ~
O . ' 1 -—1 l
1.5 2.5 3 5.8
03 dose (g/m )

 

 

Ketoacids (pg/L)

 

 

 

 

 

 

 

 

Figure 5.23 Concentrations of ketoacids of different treatment stages at various ozone

doses. (Operating conditions: water ﬂow rate, 2.75 L/min; gas flow rate,
100 mL/min; water temperature, 20 °C)

Table 5.4 Removal efﬁciencies of aldehydes by ozonation/membrane/bioﬁltration.

 

 

 

 

 

Temperature Efﬁciency Gas flowrate Efﬁciency Ozone dose Efﬁciency
(°C) (%) (mL/min) (%) (g/m3) (%)
10 88.6 50 89.9 1.5 85.4
20 90.9 75 91.1 2.5 90.9
30 91.0 100 90.9 5.8 94.8

 

 

 

 

 

 

 

The relationships between aldehydes, ketoacids, and several parameters were
also investigated in this study. Figure 5.24 shows the correlation between aldehydes
and ketoacids. The relationship between aldehydes and ketoacids was statistical

signiﬁcance (at p = 0.01 level at 95% CI.) in the early stage of ozonation (see Line 1

166

in Figure 5.24). However, as we discussed earlier, the concentrations of aldehydes
decreased as the ozone contact time increased but the formation of ketoacids
increased with increasing the ozone contact time. The decrease of aldehydes and the
increase of ketoacids also depended on the operational conditions as described earlier.
Thus, the correlation between aldehydes and ketoacids was not statistically signiﬁcant
(p > 0.05) after long contact time (see Line 2 in Figure 5.24). Grifﬁni et al. (1999)
reported that a strong relationship between aldehydes and ketoacids was observed.
They also found that the formation of aldehydes and ketoacids was proportional to the
DOC concentration. However, in our study, neither aldehydes nor ketoacids were
correlated with DOC at a signiﬁcant level (p > 0.05). This may be due to variations in
the production of aldehydes and ketoacids as the raw water quality changed with
season and environmental conditions.

The relationships between ozonation byproducts (aldehydes and ketoacids)
and surrogate parameters (UV-254, humic substances, and non-humic substances) are
plotted in Figures 5.25 to 5.30. Correlations are statistically signiﬁcant at 95% CI.
UV-254 and humic substances are negatively correlated with aldehydes and ketoacids
and the non-humic substances are correlated positively with them. This is because the
aromatic moieties of DOC were degraded by ozone, which resulted in the reduction of
the UV-254 absorbance, and the humic substances decreased as they were converted
to non-humic substances during ozonation. In addition, the formation of aldehydes
and ketoacids increased due to the decomposition of large molecules of DOC by

ozonaﬁon.

167

 

 

 

300

100

Aldehydes (ug/L)

50‘

250 -

200 1

150 1

 

-1

 

 
   

0 R2: 0.0072
Line 2

O
\ Line 1

R2=0.9539

 

 

200 400 600 80010001200140016001800
Ketoacids (ug/L)

Figure 5.24 Correlations between the aldehydes and ketoacids of Lake Lansing water
treated by the ozonation/membrane/bioﬁltration system. (Line 1
represents the correlation during the earlier stage of ozonation containing
F RW and P1 data; Line 2 represents the correlation during the later stage
of ozonation containing P1, P2, and WT data. Correlation is statistically
signiﬁcant at 95% CI. for Line 1 only.)

3
2
2
1
1

Aldehydes (pg/L)

00
50
00
50
00
50

. y= -1053.1x + 207.8

 

 

. ,. . R2 = 0.6591
L . .

O

O

. ‘ O '

O
. .
0 0.05 0.1 0.15 0.2
uv-254 (Abs.)

Figure 5.25 Correlation between aldehydes and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.)

168

 

 

 

 

R2=0.8321

   

 

 

   

 

Q
U)
3 1200 '
U)
.9 1000
8 8001
0
+5, 6001
X 4004
200
0 2 . . . . . T .4 .
0.00 0.05 0.10 0.15 0.20

uv-254 (Abs.)

Figure 5.26 Correlation between ketoacids and the absorbance of UV-254 of Lake
Lansing water treated by the ozonation/membrane system. (Correlation
is statistically signiﬁcant at 95% CI.)

300 4
o y = -36.837x + 248.16

250 2
, .o R =0.5765
200 ° ,

150
100
50

Aldehydes (pg/L)

O.

'8

O 2 4 6 8
Humic substances (mg/L)

 

 

Figure 5.27 Correlation between aldehydes and the concentration of humic substances
of Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% CI.)

169

 

1800

A 1600 1 '
_l 0
B, 1400 1
a '0 2-
v 1200 1 . R - 0.6825
E 10004
0
8 800 ~
*5; 600 ~
¥ 400«

200 1

O Y . . . c. . .

 

 

 

 

0 1 2 3 4 5 6 7
Humic substances (m g/L)

 

Figure 5.28 Correlation between ketoacids and the concentration of humic substances
of Lake Lansing water treated by the ozonation/membrane system.
(Correlation is statistically signiﬁcant at 95% Cl.)

300 ~
y = 38.502x - 33.458 ,
250 -

R2 = 0.3307
200

150
100

Aldehydes (pg/L)

01
O

 

 

 

0 2 4 6 8
Non-Humic substances (mg/L)

Figure 5.29 Correlation between aldehydes and the concentration of non-humic
substances of Lake Lansing water treated by the ozonation/membrane
system. (Correlation is statistically signiﬁcant at 95% C .I.)

170

1800 _ y=375.07x-953.62

 

 

 

R2=O.6153 .
23 1350
CD
3
U)
:9 900
0
(U
.9
g 450
O l
0 2 4 6 8

Non-Humic substances (mg/L)

 

Figure 5.30 Correlation between ketoacids and the concentration of non-humic

substances of Lake Lansing water treated by the ozonation/membrane
system. (Correlation is statistically signiﬁcant at 95% Cl.)

171

5.4 CONCLUSIONS

1.

The ozonation/membrane system resulted in limited decreases in chlorine demand.
In some cases, the treated water had a greater chlorine demand than that of the
initial ﬁltered raw water. Two competing reactions may have contributed to this
phenomena. First, a decrease in the number of chlorine-consuming sites may have
occurred as a result of reactions of NOM with molecular ozone, by direct reaction
pathway. Secondly and simultaneously to the ﬁrst reaction, new chlorine-
consuming sites may have been generated through direct or radical pathway
involving hydroxyl radicals and other radical species produced by the
decomposition of molecular ozone. The competition of these two reactions
resulted in the variations in the chlorine demand observed in this study.
Bioﬁltration alone decreased 30 to 55% of chlorine demand from Lake Lansing
water. The percent removal of chlorine demand with bioﬁltration following the
ozonation/membrane system ranged from 60% up to 90%. This indicates that a
great fraction of chlorine-consuming material was biodegradable and combining
ozonation and biotreatment processes can signiﬁcantly reduce the level of chlorine
demand of water.

The THMs concentrations decreased signiﬁcantly as the water temperature
increased. The HAAs concentrations in the treated water decreased signiﬁcantly
as the water temperature increased from 20 °C to 30 °C. This was probably the
result of a longer hydraulic retention time used in this study, and higher ozone
mass transfer rate into the water and higher reaction rate constants at higher water
temperature. When treating the water at 10 °C, the observed concentrations of
HAAs in treated water were greater than that in the ﬁltered raw water. HAAs

precursors generated by partial oxidation of humic substances may have reacted

172

were more quickly than they were destroyed by ozonation. On the other hand, this
observation could have been the result of the exposure of additional active sites in
the NOM matrix when ozonation decomposed large molecules.

Biological ﬁltration following ozonation can effectively decrease the formation of
THMs and HAAs. The removal efﬁciency of THMs precursors increased as the
water temperature increased. But the removal efﬁciency of HAAs precursors by
bioﬁltration (either alone or following ozonation/membrane system) was
independent to the treated water temperature and may depend on the
characteristics of raw water, the characteristics of water treated by
ozonation/membrane system, or the condition of microorganisms in the bioﬁlter.
The removal efﬁciency of THMs precursors by ozonation/membrane system or
subsequent bioﬁltration increased as the gas flow rate or the ozone dose increased.
HAAs concentrations under the same experimental conditions decreased with
increasing the gas flow rate and ozone dose. However, decreases in their removal
efﬁciency were not statistically signiﬁcant. As such, precursors of THMs may
have greater reaction rates with ozone and secondary radicals than those of HAAs,
resulting in differences in the removal efﬁciencies.

Aldehydes and ketoacids concentrations increased signiﬁcantly during ozonation
under all experimental conditions. The concentrations of ketoacids were much
greater than those of aldehydes in ozonated water samples. The formation of
aldehydes increased to a maximum, followed by a gradual decrease with
ozonation time. However, the formation of ketoacids consistently increased with
ozonation time.

The correlations between ketoacids and ozone gas flowrate, ketoacids and ozone

dose, and aldehydes and ozone gas flowrate were signiﬁcant, which indicates that

173

as these control parameters increased the formation of aldehydes and ketoacids
increased. However, it was found that the formation of aldehydes was not
signiﬁcantly correlated with either water temperature or ozone dose. The
formation of ketoacids and water temperature were also not signiﬁcantly
correlated. This was probably due to the variations in the characteristics of the
DOC in the tested water (since water was collected over the fall and winter
seasons) or because of the competitive reactions involving the generation and
destruction of aldehydes and ketoacids in the ozonation/membrane system.

. Very high removal efﬁciencies of aldehydes and ketoacids were achieved by the
bioﬁltration following the ozonation/membrane system. It was observed that the
concentrations of aldehyde were reduced by 85.4% to 94.8% using the
ozonation/membrane/bioﬁltration system. In the entire study, only two of nine
samples were found to have ketoacids residual. The percent removal of ketoacids
by bioﬁltration was between 94.1% and 100%.

. The relationship between THMs and HAAs was statistically signiﬁcant (at p =
0.01 level at 95% CI). The concentrations of aldehydes and ketoacids were also
statistically related. THMs and HAAs were positively correlated with surrogate
parameters (DOC, [IV-254, and humic substances) under the experimental
conditions tested. The correlctions between ozonation byproducts (aldehydes and
ketoacids) and surrogate parameters (UV-254, humic substances, and non-humic
substances) were also found to be statistically related. However, neither
aldehydes and DOC nor ketoacids and DOC were signiﬁcantly correlated in this
study. This may be due to variations in the production of aldehydes and ketoacids

in the different raw waters.

174

10. According to these experimental results, chloroform was the predominant species
of THMs. DCAA and TCAA were the predominant species of HAAs, and
formaldehydes and pyruvic acid were two principal products of ozonation in all
water samples in this study. Increases in the percentage of bromodichloromethane,

dibromochloromethane, BCAA, and DBAA were observed in biotreated water.

5.5 REFERENCES

Amy, G. L., Kuo, C. J., and Sierka, R. A. (1988). "Ozonation of Humic Substances -
Effects on Molecular-Weight Distributions of Organic-Carbon and

Trihalomethane Formation Potential." Ozone-Science & Engineering, 10(1),
39-54.

Bablon, G., Bellamy, W. D., Bourbigot, M.-M., and et. al. (1991). "Fundamental
Aspects." Ozone in Water Treatment: Application and Engineering, B.
Langlais, D. A. Reckhow, and D. R. Brink, eds., Lewis Publishers, Chelsea,
MI, ll-132.

Bull, R. J., and McCabe, L. J. (1984). "Risk Assessment Issues in Evaluating the
Health Effects of Alternate Means of Drinking Water Disinfection." Water
Chlorination: Chemistry, Environmental Impact, and Health Effects, R. L.
Jolley and e. al., eds., Lewis Publishers, Chelsea, Mich.

Carlson, K., and Amy, G. (1997). "The formation of ﬁlter-removable biodegradable
organic matter during ozonation." Ozone-Science & Engineering, 19(2), 179-
199.

Chiang, P. C., Chang, E. E., and Liang, C. H. (2002). "NOM characteristics and
treatabilities of ozonation processes." Chemosphere, 46(6), 929-936.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

Coleman, W. E., Munch, J. W., Ringhand, H. P., Kaylor, W. H., and Mitchell, D. E.
(1992). "Ozonation Post-Chlorination of Humic-Acid - a Model for Predicting
Drinking-Water Disinfection by-Products." Ozone-Science & Engineering,
14(1), 51-69.

175

Duguet, J. P., Brodard, E., Dussert, B., and Mallevialle, J. (1985). "Improvement in
the Effectiveness of Ozonation of Drinking-Water through the Use of
Hydrogen-Peroxide." Ozone—Science & Engineering, 7(3), 241-258.

Edzwald, J. K., Becker, W. C., and Wattier, K. L. (1985). "Surrogate Parameters for
Monitoring Organic-Matter and Thm Precursors." Journal American Water
Works Association, 77(4), 122-132.

Glaze, W. H., Peyton, G. R., Lin, 8., Huang, R. Y., and Burleson, J. L. (1982).
"Destruction of Pollutants in Water with Ozone in Combination with

Ultraviolet-Radiation .2. Natural Trihalomethane Precursors." Environmental
Science & Technology, 16(8), 454-458.

Grifﬁni, O., Bao, M. L., Barbieri, K., Burrini, D., Santianni, D., and Pantani, F.
(1999). "Formation and removal of biodegradable ozonation by-products

during ozonation-bioﬁltration treatment: Pilot-scale evaluation." Ozone-
Science & Engineering, 21(1), 79-98.

Hemer, H. A. (1999). "The Reaction of Ozone with Pyrene: The By-products and
Their Toxicological Implications," Michigan State University, East Lansing.

Hoigne, J. (1982). "Mechanisms, Rates and Selectivities of Oxidation of Organic
Compounds Initiated by Ozonation of Water." Handbook of Ozone
Technology and Applications, R. G. Rice and A. Netzer, eds., Ann Arbor
Science, Ann Arbor, MI, 307-379.

Hoigne, J., and Bader, H. (1975). "Ozonation of Water: Role of Hydroxyl Radicals as
Oxidizing Agents." Science, 190, 782-784.

Jadashecart, A., Ventresque, C., Legube, B., and Dore, M. (1991). "Effect of
Ozonation on the Chlorine Demand of a Treated Surface-Water and Some
Macromolecular Compounds." Ozone-Science & Engineering, 13(2), 147-160.

K0, Y. W., Chiang, P. C., and Chang, E. E. (1998). "Ozonation of p-hydroxybenzoic
acid solution." Ozone-Science & Engineering, 20(5), 343-360.

[(001, H. J., van Kreijl, C. F ., and Hrubec, J. (1985). "Mutagenic and Carcinogenic
Properties of Drinking Water." Water Chlorination: Chemistry, Environmental
Impacts, and Health Effects, R. L. Jolley, W. A. Brungs, and R. B. Cumming,
eds., Lewis Publ., Chelsea, Mich.

Lee, K.-H. (2001). "The effects of ozonation pathways on the formation of ketoacids
and assimilable organic carbon (AOC) in drinking water," Michigan State
University, East Lansing, MI 48823.

176

 

 

 

Miltner, R. J., Shukairy, H. M., and Summers, R. S. (1992). "Disinfection by-Product
Formation and Control by Ozonation and Biotreatment." Journal American
Water Works Association, 84(11), 53-62.

Morris, R. D., Audet, A. M., Angelillo, 1. F ., Chalmers, T. C., and Mosteller, F.
(1992). "Chlorination, Chlorination by-Products, and Cancer - a
Metaanalysis." American Journal of Public Health, 82(7), 955-963.

Mughal, F. H. (1992). "Chlorination of Drinking Water and Cancer: A Review." Jour.
Envir. Pathol, T oxicol. & 0ncol., 11(5,6), 287-292.

Munch, J. W., Munch, D. J., Winslow, S. D., Wendelken, S. C., and Pepich, B. V.
(1998). "Determination of Carbonyl Compounds in Drinking Water by
Pentafluorobenzylhydroxylamine Derivatization and Capillary Gas
Chromatography with Electron Capture Detection." USEPA, Cincinnati, Ohio.

Nieminski, C. E., Chaudhuri, S., and Lamoreaux, T. (1993). "The Occurrence of Dbps
in Utah Drinking Waters." Journal American Water Works Association, 85(9),
98-105.

Oxenford, J. L. (1996). "Disinfection By-Products: Current Practices and Future
Directions." Disinfection By-products in Water Treatment, R. A. Minear and
G. L. Amy, eds., Lewis Publisher, Boca Raton, FL., 3-16.

Paode, R. D., Amy, G. L., Krasner, S. W., Summers, R. S., and Rice, E. W. (1997).
"Predicting the formation of aldehydes and BOM." Journal American Water
Works Association, 89(6), 79-93.

Price, M. L., Bailey, R. W., Enos, A. K., Hook, M., and Herrnanowicz, S. W. (1993).
"Evaluation of Ozone Biological Treatment for Disinfection by- Products

Control and Biologically Stable Water." Ozone-Science & Engineering, 15(2),
95-130.

Rice, R. G. (1980). "The Use of Ozone to Control Trihalomethanes in Drinking-Water
Treatment." Ozone-Science & Engineering, 2(1), 75-99.

Richardson, S. D., Thruston, A. D., Caughran, T. V., Chen, P. H., Collette, T. W.,
Floyd, T. L., Schenck, K. M., Lykins, B. W., Sun, G. R., and Majetich, G.
(1999). "Identiﬁcation of new ozone disinfection byproducts in drinking
water." Environmental Science & Technology, 33(19), 3368—3377.

Schechter, D. S., and Singer, P. C. (1995). "Formation of Aldehydes During
Ozonation." Ozone-Science & Engineering, 17(1), 53-69.

Staehelin, J ., and Hoigne, J. (1985). "Decomposition of Ozone in Water in the
Presence of Organic Solutes Acting as Promoters and Inhibitors of Radical
Chain Reactions." Environmental Science & Technology, 19(12), 1206-1213.

177

Standard Methods. (1998). Standard Methods for the Examination of Water and
Wastewater, A. E. Greenberg, L. S. Clesceri, and A. D. Easton, eds., APHA,
AWWA, WEF, Washington, DC.

Wallace, J. L., Vahadi, B., Femandes, J. B., and Boyden, B. H. (1988). "The
Combination of Ozone Hydrogen-Peroxide and Ozone Uv- Radiation for
Reduction of Trihalomethane Formation Potential in Surface-Water." Ozone-
Science & Engineering, 10(1), 103-112.

Weinberg, H. S., and Glaze, W. H. (1996). "An Overview of Ozonation Disinfection
By-Products." Disinfection By-products in Water Treatment, R. A. Minear and
G. L. Amy, eds., Lewis Publisher, Boca Raton, F L., 165-186.

Wobma, P., Pemitsky, D., Bellamy, B., Kjartanson, K., and Sears, K. (2000).
"Biological ﬁltration for ozone and chlorine DBP removal." Ozone—Science &
Engineering, 22(4), 393-413.

5“".ng

Xie, Y. F., and Reckhow, D. A. (1992). "Formation of Ketoacids in Ozonated
Drinking-Water." Ozone-Science & Engineering, 14(3), 269-275.

Yamada, H., and Somiya, I. (1980). "Identiﬁcation of Products Resulting from
Ozonation of Organic- Compounds in Water. " Ozone-Science & Engineering,
2(3), 251-260.

Yamada, H., Somiya, I., and Inanami, F. (1986). "The Effect of Preozonation on the
Control of Trihalomethane Formation." Ozone-Science & Engineering, 8(2),
129-150.

Zhang, X. 2., Ni, Y., and van Heiningen, A. (2001). "Effect of temperature on the
kinetics of pulp ozonation." Journal of Pulp and Paper Science, 27(8), 279-
283.

178

CHAPTER 6
CORRELATIONS BETWEEN THE FORMATION OF ASSIMILABLE

ORGANIC CARBON AND OZONATION BY-PRODUCTS

6.1 INTRODUCTION

The use of ozone as a disinfectant results in a decrease in the concentration of
halogenated disinfectant by-products (DBPs) (Amy et al. 1988; Cipparone et al. 1997).
However, the ozonation of water containing natural organic matter (NOM) can
increase the concentration of biodegradable organic matter (BOM) by decomposing
larger molecules, resulting in the formation of lower molecular weight material and
more polar and hydrophilic compounds (e.g., see Amy et al. 1988; Amy et al. 1992;
F logstad and Odegaard 1985; Kaastrup and Halmo 1989; Koechling et al. 1996;
Owen et al. 1995). The presence of BOM in treated water can result in the formation
of bioﬁlms in water distribution systems, unpleasant taste, odors, and color, in
biologically mediated corrosion and in interferences with microbiological monitoring
(Lee et al. 1980; Miltner et al. 1992; Schechter and Singer 1995).

Assimilable organic carbon (AOC) is one approach to estimating the
concentration of BOM. It represents the degradable fraction of BOM that can be
consumed by microorganisms. A signiﬁcant correlation between the AOC
concentration and the level of regrowth of heterotrophic bacteria in water distribution
systems has been reported (Vanderkooij 1992). Several researchers found that the
correlation between AOC and ozonation by-products such as aldehydes was
signiﬁcant (Paode et al. 1997; Schechter and Singer 1995). Therefore, the objective
of this study was to determine whether signiﬁcant correlations exist between the

concentration of AOC and ozonation by-products.

179

6.2 MATERIALS AND METHODS
6.2.1 Experimental Setup and Protocol

A specially designed ozonation/membrane system and four bioﬁltration
columns were used to investigate the formation of AOC and ozonation DBPs (i.e.,
aldehydes, ketones, and ketoacids) resulting from the treatment process. The system
was the same one that was described in Chapter 4. A schematic diagram of this
system is shown in Figure 4.1. Operational and experimental conditions were
described in Section 5.2.1. AOC, aldehydes (formaldehyde, acetaldehyde,
propionaldehyde, glyoxal, and methyl glyoxal), ketones (acetone and 2-butanone),
and ketoacids (glyoxylic acid, pyruvic acid, and ketomalonic acid) were monitored in

all water samples in this study.

6.2.2 Water Source

Experiments were carried out on Lake Lansing (Haslett, MI) water. Water
quality characteristics of Lake Lansing are summarized in the previous chapter (see
Table 4.3). Lake Lansing waters were collected in ﬁve-gallon tanks, properly labeled,
and stored at 4°C. The maximum storage period was seven days. All waters were

ﬁltered through a 0.45-um ﬁlter and warmed to room temperature before testing.

6.2.3 Analytical Methods

The measurement of AOC followed the Method 9217 of Standard Methods
(1998) as described in Section 4.2.8. Aldehydes and ketones were measured by using
the USEPA Method 556 (Munch et al. 1998) as described in Section 5.2.5. The

concentration of ketoacids was measured by using the same method as described in

Section 5.2.6.

180

 

6.3 RESULTS AND DISCUSSION
6.3.1 Ozonation By-products

The experimental results describing total aldehydes and ketoacids were
presented in the previous chapter. As mentioned in Chapter 5, acetaldehyde,
propionaldehyde, and acetone were not detected in all water samples. Therefore, only
the results of three species of aldehydes (formaldehyde, glyoxal, and methyl glyoxal)
and one species of ketone (2-Butanone) were discussed in this section. All species
were found to be signiﬁcantly correlated with AOC in our study (at p = 0.01 level at
95% CI).

The concentrations of formaldehyde, which is the principle aldehydes species,
were determined at different treatment stages under various experimental conditions.
These results are shown in Figures 6.1 to 6.3. Although the maximum concentration
of formaldehyde was different at different experimental conditions, likely due to the
varying composition of NOM in the raw water and the different operating conditions,
the level of formaldehyde usually reached a maximum level in P1 or P2, and then
decreased in WT. This indicates that the formation of formaldehyde in the early
stages of ozonation was faster than its decomposition by oxidation, but it was
gradually destroyed by molecular ozone or hydroxyl radical after all its precursors
were converted to formaldehyde. Similar results were reported by several researchers
(K0 et al. 1998; Yamada and Somiya 1980). Correlations between the maximum
level of formaldehyde and ozone gas ﬂow rate and between the maximum level of
formaldehyde and ozone dose were signiﬁcant (p < 0.01). As the gas flow rate or
ozone dose increased, the formation of formaldehyde increased. However, the

correlation between the formation of formaldehyde and water temperature was not

181

signiﬁcant (p > 0.05). This observation is the same as that observed for total
aldehydes.

Figures 6.4 to 6.6 show the concentrations of glyoxal at various treatment
stages under different experimental conditions. Figures 6.7 to 6.9 show of the
variations in the concentrations of methyl glyoxal. The formation of glyoxal and
methyl glyoxal follow similar trends to that observed for formaldehyde. The
concentrations of glyoxal and methyl glyoxal increased early during treatment and
then decreased after reaching the maximum level. At 10 °C water temperature,
glyoxal concentration of different treatment stage increased as the ozonation time
increased. It appears that the difference in the formation and decomposition
potentials of glyoxal were higher at lower water temperature. It could be due to the
direct reaction of molecular ozone is the primarily pathway that favors the formation
of aldehydes (Hoigné 1982).

The maximum concentration of glyoxal and methyl glyoxal increased as the
gas ﬂowrate increased but decreased as the water temperature increased (p < 0.01).
Their concentrations also increased as the ozone dose increased but the correlations
between concentration and ozone dose were not statistical signiﬁcance (p > 0.05).

Figures 6.1 to 6.9 also illustrate the performance of the biotreatment on the
removal of ozonation by-products. Bioﬁltration effectively and signiﬁcantly removed
the concentrations of all aldehydes species in P2 water, which was treated by
ozonation/membrane system (p < 0.01). The percentage of formaldehyde, glyoxal,
and methyl glyoxal removed ranged between 72.9% and 88.3%, 93.0% and 100%,
and 93.2% and 98.1%, respectively.

Table 6.1 summarizes the concentration of 2-butanone measured in waters

from the different treatment stages at various experimental conditions. It shows that

182

 

the concentrations of 2-butanone increased signiﬁcantly during ozonation (p < 0.01)
except for one raw water sample treated at ozone gas ﬂowrate of 75 mL/min.
However, no ﬁrm conclusions can be made for the relationships between 2-butanone
concentrations and experimental conditions because of the variability of the results.
The vapor pressure of 2-butanone is 90.6 mmHg at 25 °C, and the Henry’s law
constant is 5.77 x 10'5 atm m3/mol at 25 °C (Agency for Toxic Substances and
Disease Registry 1992). This suggests that volatilization may result in losses of 2-
butanone to the air during the purging ozonated water (to quench ozone). Moreover,
the variations in the composition of NOM in FRW water (the concentration of 2-
butanone in FRW water varied between 2.6 ug/L and 297.7 ug/L) and the complex
competing reactions during ozonation in the water may also result in the variability of
2-butanone concentrations.

The results obtained after biotreatment of 2-butanone (B-FRW and B-P2) were
also highly variable. The highest removal efﬁciency was over 99% but the level of 2-
butanone was also found to increase from 2.6 ug/L to 646.0 ug/L in one sample after
bioﬁltration. The 2-butanone concentration decreased in 9 out of 14 biotreated
samples (p < 0.01) but increased in the other 5 ones (p < 0.01). Several researchers
have studied the biodegradation of 2-butanone and reported that bioﬁltration can
remove 2-butanone (Agathos et al. 1997; Amanullah et al. 2000; Deshusses 1997;
Deshusses et al. 1995; Edwards et al. 1994). Several factors may result in the
variation of 2-butanone concentration: 1) the decomposition rate of 2-butanone by the
microorganisms in the bioﬁltration column, 2) the losses of 2-butanone from the
bioﬁltration system to the air, and 3) the sample handling during analysis. Therefore,
further studies on the biodegradation of 2-butanone by using ozonation/membrane and

subsequent bioﬁltration should be conducted.

183

 

160
140
120
100
80
60
4O
20

Formaldehyde (pg/L)

 

 

Temperature (°C)
Figure 6.1 Concentrations of formaldehyde of different treatment stages at various

water treatment temperatures. (Operating condition: water ﬂowrate, 2.75
L/min; gas ﬂowrate, 100 mL/min; 03 dose, 2.5 g/m3)

160 I FRW

 

140
120
100
80
60
40
20

Formaldehyde (pg/L)

 

 

 

 

50 75 1 00
Gas ﬂow rate (mL/min)

Figure 6.2 Concentrations of formaldehyde of different treatment stages at various gas

flowrates. (Operating condition: water flowrate, 2.75 L/min; O3 dose, 2.5
g/m3; water temperature, 20 °C)

184

160
140 r
120
100 ~
80
60 .
40
20

Formaldehyde (pg/L)

 

   

 

 

 

 

 

 

2.5 3 5.8
03 dose (g/m )

Figure 6.3 Concentrations of formaldehyde of different treatment stages at various
ozone doses. (Operating condition: water ﬂowrate, 2.75 L/min; gas
ﬂowrate, 100 mL/min; water temperature, 20 °C)

 

 

 

 

 

 

 

 

 

 

 

 

 

70 I FRW
I P1
60 E 5%
,2 - [:1
S, 50 El B-FRW
s 40 B-P2
T9 30
_g e-
o 20
10
0
10 20 30

Temperature (°C)

Figure 6.4 Concentrations of glyoxal of different treatment stages at various water
treatment temperatures. (Operating condition: water ﬂowrate, 2.75 L/min;
gas flowrate, 100 mL/min; O3 dose, 2.5 g/m3)

185

 

 

 

 

 

 

 

 

 

 

 

 

50 .
IERW
I
2 4° .4. “‘ ”5’6
‘ [:1
3 30 - *- :EiB-FRW
g s -. lB—P2
g 20 ‘
o
10 f
50 75 100

Gas ﬂow rate (mL/min)

Figure 6.5 Concentrations of glyoxal of different treatment stages at various gas ﬂow
rates. (Operating condition: water ﬂowrate, 2.75 L/min; 0; dose, 2.5 g/m3;
water temperature, 20 °C)

 

 

 

 

 

 

 

 

 

 

 

 

 

60 1
l FRW
50 I P1
A ~1- P2
3 4° » 1:: WT
£3 30 _ 1:3 B-FRW
92 B-P2
(D 20
10
O I
1.5 2.5 5.8
0;; dose (g/m3)

Figure 6.6 Concentrations of glyoxal of different treatment stages at various ozone
doses. (Operating condition: water ﬂowrate, 2.75 L/min; gas flowrate, 100
mL/min; water temperature, 20 °C)

186

 

 

60 1
A IFRW
g, 50 IP1
3 4o [9P2
T3 5 , ElWT
9, 30 - ElB-FRW
a g IB-P2
g 20
a:
12 10

0

 

 

 

 

 

 

 

 

 

 

 

1 0 20 30
Temperature (°C)

Figure 6.7 Concentrations of methyl glyoxal of different treatment stages at various

water treatment temperatures. (Operating condition: water ﬂowrate, 2.75
L/min; gas flowrate, 100 mL/min; O3 dose, 2.5 g/m3)

 

 

 

 

 

 

 

 

 

 

 

60 l
A 50 I FRW
a I P1
3 40 4 P2
g 30 L EBB-FRW
“g: j? B-P2
5 20 P .
(D
2 10

O I

50 75 100

Gas flow rate (mL/min)

Figure 6.8 Concentrations of methyl glyoxal of different treatment stages at various

gas ﬂow rates. (Operating condition: water ﬂowrate, 2.75 L/min; O3 dose,
2.5 g/m3; water temperature, 20 °C)

187

 

 

 

 

 

 

 

 

 

 

 

 

 

70 ‘
A I FRW
3 60 . : IP1
3 50 1:1 P2
__ EIWT
a 40 _E ‘ EIB-FRW
a 30 . aB-Pz
E 20
CD
2 10
0 ,qugr ,_- 1 '- .2].- 1
1.5 2.5 5.8
03 dose (g/m3)

Figure 6.9 Concentrations of methyl glyoxal of different treatment stages at various
ozone doses. (Operating condition: water ﬂowrate, 2.75 L/min; gas
ﬂowrate, 100 mL/min; water temperature, 20 °C)

Table 6.1 Concentrations of 2-Butanone for Lake Lansing water measured in different
treatment stages at various experimental conditions in this study. (pg/L)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TenperatureT FRW P1 P2 WT 13-an B-P2
10 195.9 1048.2 112.8 179.0 115.8 448.9
20 179.1 599.6 906.7 1105.4 484.8 1.3
30 114.0 261.0 580.9 873.2 60.7 33.9
Gas ﬂowratez FRW P1 P2 WT B-FRW B-PZ
50 297.7 832.9 982.1 1427.7 288.7 256.7
75 297.5 158.7 224.9 229.6 475.5 3.3
100 179.1 599.6 906.7 1105.4 484.8 1.3
Ozone dose3 FRW P1 P2 WT B-FRW B-P2
1.5 208.8 382.7 198.2 158.6 484.8 112.6
2.5 179.1 599.6 906.7 1105.4 484.8 1.3
5.8 2.6 245.4 405.8 300.5 646.0 594.6
1. unit: °C

2. unit: mL/min
3. unit: g/m3

 

6.3.2 Correlations Between AOC and Ozonation By-products

The Pearson correlation coefﬁcient of similarity (r value) is used to investigate
the correlations between AOC(total), AOC(P17), and AOC(NOX) and ozonation by-
products. The AOC(total) means that the density of viable cells of Pseudomonas
ﬂuorescens strain P17 and Spirillum strain NOX was converted to AOC
concentrations. The AOC(P17) and AOC(NOX) mean that only the density of viable
cells of Pseudomonasﬂuorescens strain P17 and only that of Spirillum strain NOX
was converted to AOC concentrations, respectively. The Pearson coefﬁcient was
generated by using the statistical software, SPSS for Windows (Release 10.0, SPSS
Inc.). The absolute value of the Pearson coefﬁcient indicates the strength of the linear
relationship between AOC and ozonation by-products, with larger absolute values
indicating stronger relationships. The experimental results for formaldehyde, glyoxal,
methyl glyoxal, and 2-butanone were discussed in the previous section. The results
for AOC(total), aldehydes and ketoacids were presented in Chapters 4 and 5.

Table 6.2 summarizes the Pearson correlation coefﬁcients of AOC(total),
AOC(P17), and AOC(NOX) that were determined by using ozonation by-products as
independent variables. Signiﬁcant and positive correlations were found between
independent variables and AOC(total), AOC(P17), AOC(NOX) at the p = 0.01 level
(n=3 8, at 95% CI). According to the Pearson coefﬁcients in Table 6.2, total
concentration of ozonation by-products (aldehydes + 2-butanone + ketoacids) and the
concentration of AOC behaved similarly in the treatment system in this study and the
total concentration of ozonation by-products could be used as a surrogate of
AOC(total). It also shows the same for AOC(P17) and AOC(NOX). Figures 6.10 to

6.12 show correlations between AOC(total), AOC(P17), AOC(NOX) and the sum of

189

1 f‘m

aldehydes, 2-butanone, and ketoacids treated by the ozonation/membrane and
bioﬁltration system in our study.

Schechter and Singer (1995) reported that the concentration of aldehydes
could be used as a surrogate for the more complex AOC analysis. The concentration
of aldehydes is the sum of formaldehyde, glyoxal, and methyl glyoxal in our study.

Our results conﬁrmed their observation. The concentration of methyl glyoxal was
another good candidate as a surrogate for AOC(total), AOC(Pl7), and AOC(NOX).

However, it should be noted that the correlation between AOC and ozonation E
by-products would be affected by the nature of NOM. Small differences in the ‘
composition of NOM might result in a signiﬁcant difference of the formation of
ozonation by-products. Therefore, waters obtained from different sources might react
differently in the treatment system. As such, further investigation is necessary to
assess the more general correlations between the formation of AOC and ozonation by-

products.

Table 6.2 Correlations between AOC(total), AOC(Pl7), AOC(NOX) and ozonation
by-products.

 

 

 

 

 

 

 

 

 

 

 

 

Independent variables AOC (total) AOC(P l 7) AOC(NOX)
(r value) (r value) (r value)

Aldehydes 0.863 0.850 0.850
Formaldehyde 0.722 0.684 0.716
Glyoxal 0.821 0.873 0.794
Methyl glyoxal 0.885 0.902 0.865
2-Butanone 0.780 0.641 0.796
Ketoacids 0.752 0.783 0.731
Aldehydes + 2-Butanone

+Ketoacids 0.899 0.856 0.892
Aldehydes + 2-Butanone 0.860 0.736 0.871
Aldehydes + Ketoacids 0.804 0.841 0.781
2-Butanone + Ketoacids 0.876 0.829 0.870

 

 

 

 

* n = 38, the correlation is signiﬁcant at the 0.01 level at 95% CI.

** r is the correlation between the observed and predicted values of AOC(total),

AOC(Pl7), and AOC(NOX).

 

2500

l

y = 0.824x + 122.29
R2 =0.8081 .

 
 

2000

  

1500

1000

AOC(total) (pg/L)

500 *

 

1

0 1000 2000 3000
Aldehydes + 2-Butanone + Ketoacids (pg/L)

 

Figure 6.10 Correlation between AOC(total) and the sum of aldehydes, 2-butanone,
and ketoacids of Lake Lansing water treated by the ozonation/membrane
and bioﬁltration system. (Correlation is signiﬁcant at the 0.01 level.)

  

 

 

500 ~
a
0" 0
ﬁ 300 ,
L‘. l
‘5, 200 l
g y = 0.1434x + 48.099 .
100 R2 = 0.7321
0 v I
0 1000 2000 3000

Aldehydes + 2-Butanone + Ketoacids (pg/L)

Figure 6.11 Correlation between AOC(P17) and the sum of aldehydes, 2-butanone,
and ketoacids of Lake Lansing water treated by the ozonation/membrane
and bioﬁltration system. (Correlation is signiﬁcant at the 0.01 level.)

191

   

 

 

2000 —

A 8

_J

,5 1500

.3

6 1000

‘3':

8 l

< 500 y = 0.6803x + 74.142;

R2 = 0.7954 .
0 1 1
0 1000 2000 3000

Aldehydes + 2-Butanone + Ketoacids (pg/L)

 

Figure 6.12 Correlation between AOC(NOX) and the sum of aldehydes, 2-butanone,
and ketoacids of Lake Lansing water treated by the ozonation/membrane
and bioﬁltration system. (Correlation is signiﬁcant at the 0.01 level.)

6.4 CONCLUSIONS

1. The concentrations of formaldehyde, glyoxal, and methyl glyoxal decreased after
they reached their maximum level, which indicated that their formation in the
early stages of ozonation was faster than their decomposition by oxidation.
However, these ozonation by-products were gradually destroyed by ozonation
after the precursors were oxidized.

2. As the gas ﬂow rate or ozone dose increased, the maximum concentration of
formaldehyde increased. However, the correlation between the maximum level of
formaldehyde and water temperature was not signiﬁcant. On the other hand, the

maximum concentration of glyoxal and methyl glyoxal increased as the gas flow

192

rate increased but decreased as the water temperature increased. The maximum
concentrations of glyoxal and methyl glyoxal also increased as the ozone dose
increased but the increases were not statistical signiﬁcance.

3. Bioﬁltration effectively removed the concentrations of all aldehydes in water
treated by ozonation/membrane system.

4. The biotreatment of 2-butanone shows variable results. Further studies on the
biodegradation of 2-butanone by using ozonation/membrane and subsequent
bioﬁltration should be conducted.

5. Signiﬁcant and positive correlations were found between ozonation by-products
and AOC(total), AOC(Pl7), and AOC(NOX).

6. The total concentration of ozonation by-products (aldehydes + 2-butanone +
ketoacids), aldehydes, and methyl glyoxal are three good surrogates for
AOC(total), AOC(Pl7), and AOC(NOX) for Lake Lansing water treated by
ozonation/membrane system and bioﬁltration under experimental conditions

tested in this study.

6.5 REFERENCES

Agency for Toxic Substance and Disease Registry (1992). Toxicological Proﬁle for 2-
Butanone, Agency for Toxic Substances and Disease Registry, US. Public
Health Service.

Agathos, S. N., Hellin, E., Ali-Khodja, H., Deseveaux, S., Vandermesse, F., and
Naveau, H. (1997). "Gas-phase methyl ethyl ketone biodegradation in a

tubular bioﬁlm reactor: microbiological and bioprocess aspects."
Biodegradation, 8(4), 251-264.

193

Amanullah, M., Viswanathan, S., and Farooq, S. (2000). "Equilibrium, kinetics, and
column dynamics of methyl ethyl ketone biodegradation." Industrial &
Engineering Chemistry Research, 39(9), 3387-3396.

Amy, G. L., Kuo, C. J., and Sierka, R. A. (1988). "Ozonation of Humic Substances -
Effects on Molecular-Weight Distributions of Organic-Carbon and

Trihalomethane Formation Potential." Ozone-Science & Engineering, 10(1),
39-54.

Amy, G. L., Sierka, R. A., Bedessem, J ., Price, D., and Tan, L. (1992). "Molecular-
Size Distributions of Dissolved Organic-Matter." Journal American Water
Works Association, 84(6), 67-75.

Cipparone, L. A., Diehl, A. C., and Speitel, G. E. (1997). "Ozonation and BDOC
removal: Effect on water quality." Journal American Water Works Association,
89(2), 84-97.

 

Deshusses, M. A. (1997). "Transient behavior of bioﬁlters: Start-up, carbon balances,
and interactions between pollutants." Journal of Environmental Engineering-
Asce, 123(6), 563-568.

Deshusses, M. A., Hamer, G., and Dunn, l. J. (1995). "Behavior of Bioﬁlters for
Waste Air Biotreatment .2. Experimental Evaluation of a Dynamic-Model."
Environmental Science & Technology, 29(4), 1059-1068.

Edwards, D. B, Adams, W. J ., and Heitkamp, M. A. (1994). "Laboratory-Scale
Evaluation of Aerobic F luidized-Bed Reactors for the Biotreatment of a
Synthetic, High-Strength Chemical- Industry Waste Stream." Water
Environment Research, 66(1), 70-83.

Flogstad, H., and Odegaard, H. (1985). "Treatment of Humic Waters by Ozone."
Ozone-Science & Engineering, 7(2), 121-136.

Hoigné, J. (1982). "Mechanisms, Rates and Selectivities of Oxidation of Organic
Compounds Initiated by Ozonation of Water." Handbook of Ozone
Technology and Applications, R. G. Rice and A. Netzer, eds., Ann Arbor
Science, Ann Arbor, MI, 307-379.

Kaastrup, E., and Halmo, T. M. (1989). "Removal of Aquatic Humus by Ozonation
and Activated Carbon Adsorption." Aquatic Humic Substances: Inﬂuence on
Fate and Treatment of Pollutants, I. H. Suffet and P. MacCarthy, eds.,
American Chemical Society, Washington, DC., 697-726.

Ko, Y. W., Chiang, P. C., and Chang, E. E. (1998). "Ozonation of p-hydroxybenzoic
acid solution." Ozone-Science & Engineering, 20(5), 343-360.

194

Koechling, M. T., Shukairy, H. M., and Summers, R. S. (1996). "Effect of Ozonation
and Biotreatment on Molecular Size and Hydrophilic Fractions of Natural
Organic Matter." Water Disinfection and Natural Organic Matter:
Characterization and Control, R. A. Minear and G. L. Amy, eds., American
Chemical Society, Washington, DC., 196-210.

Lee, S. H., Oconnor, J. T., and Banerji, S. K. (1980). "Biologically Mediated
Corrosion and Its Effects on Water- Quality in Distribution-Systems." Journal
American Water Works Association, 72(11), 636-645.

Miltner, R. J., Shukairy, H. M., and Summers, R. S. (1992). "Disinfection by-Product
Formation and Control by Ozonation and Biotreatment." Journal American
Water Works Association, 84(11), 53-62.

Munch, J. W., Munch, D. J., Winslow, S. D., Wendelken, S. C., and Pepich, B. V.
(1998). "Determination of Carbonyl Compounds in Drinking Water by
Pentaﬂuorobenzylhydroxylamine Derivatization and Capillary Gas
Chromatography with Electron Capture Detection." USEPA, Cincinnati, Ohio.

 

Owen, D. M., Amy, G. L., Chowdhury, Z. K., Paode, R., McCoy, G., and Viscosil, K.
(1995). "Nom - Characterization and Treatability." Journal American Water
Works Association, 87(1), 46-63.

Paode, R. D., Amy, G. L., Krasner, S. W., Summers, R. S., and Rice, E. W. (1997).
"Predicting the formation of aldehydes and BOM." Journal American Water
Works Association, 89(6), 79-93.

Schechter, D. S., and Singer, P. C. (1995). "Formation of Aldehydes During
Ozonation." Ozone-Science & Engineering, 17(1), 53-69.

Standard Methods. (1998). Standard Methods for the Examination of Water and
Wastewater, A. E. Greenberg, L. S. Clesceri, and A. D. Easton, eds., APHA,
AWWA, WEF, Washington, DC.

Vanderkooij, D. (1992). "Assimilable Organic-Carbon as an Indicator of Bacterial
Regrowth." Journal American Water Works Association, 84(2), 57-65.

Yamada, H., and Somiya, I. (1980). "Identiﬁcation of Products Resulting from
Ozonation of Organic- Compounds in Water." Ozone-Science & Engineering,
2(3), 251-260.

195

CHAPTER 7

CONCLUSIONS AND RECOMMENDATIONS

7.1 CONCLUSIONS

A novel ozonation/membrane system was used in this study to investigate its
ability to control NOM and DBPs in drinking water under various operating
conditions. This research was conducted by using a bench-scale system to determine
the efﬁciency of ozone mass transfer in the system, to investigate the effect of the
system followed by bioﬁltration on the characteristics of NOM and the formation of
DBPs, and to establish the correlation between NOM, DBPs, and AOC. Considering
important factors such as gas flow rate and water flow rate which influence the Km
value, the ozonation/membrane system can achieve reasonable ozone mass transfer by
requiring lesser ozone dose, gas ﬂow rate, and water flow rate than comparable
systems. This system has the added advantage that it does not include diffuser thus

reducing clogging problems.

7.1.1 Mass Transfer of Ozone in the System

The volumetric mass transfer coefﬁcient of ozone was determined under
different operating conditions. Water ﬂow rate, gas flow rate, inlet gaseous ozone
concentration, and water temperature were the operational parameters controlled in
this study. The KLa value of ozone increased with increasing the water flow rate and
gas ﬂow rate over the range of operating conditions studied. However, as the ozone

dose increased, the KLa value ﬁrst increased to a maximum value and then decreased.

196

 

The water temperature was found to have little effect on the KLa value over the
temperature range from 10 °C to 30 °C.

A mathematical model was developed to describe the concentration proﬁles of
dissolved ozone in the ozonation/membrane system. The simulations ﬁt reasonably
well with the experimental data under different Operating conditions. Although the
DDI water used in the mass transfer study was acidiﬁed to pH 2 to avoid ozone
decomposition in the water, the ozone decomposition rate constant, k, was considered
in the model because the ozone decomposition experiment showed that the k value is

not negligible when using the ozonation/membrane system.

7.1.2 Characteristics of NOM

The characteristics of NOM in terms of DOC, UV-254, humic and non-humic
substances, BDOC, and AOC were monitored in the treatment process of the
ozonation/membrane system. The DOC concentration in the permeate water of the
membrane decreased as the temperature increased and 15 to 33% of DOC was
removed. Ozonation effectively decreased the absorbance of UV-254 by 55 to 85%,
converted humic substances to non-humic substances, and signiﬁcantly increased the
biodegradability of water in terms of BDOC and AOC. A good correlation was
established between the humic substances and UV-254.

After bioﬁltration, the removal efﬁciency of DOC increased to between 69
and 88%. Between 91 to 97% of the AOC was also removed by ozonation/membrane
ﬁltration/bioﬁltration. This suggests that biotreatment following the
ozonation/membrane system can maximize the removal of organic carbon, which may
react with chlorine and result in the formation of DBPs or result in the bacteria

regrowth problem in the distribution system if remains in the treated water.

197

 

Adding ozone gas at the membrane inlet resulted in a recovery of the permeate
ﬂux from 60% to 97% of the initial ﬂux. The experimental results also show that
effective removal of DOC, UV-254 and conversion of humic substances to non-humic
substances can be achieved at low gas flow rate and low ozone dose. Therefore, these
conditions should reduce the cost of system operation over more extensive ozonation
and decrease the frequency of membrane cleaning or replacement, which should also,

indirectly, reduce the cost of maintenance.

7.1.3 Control of DBPs

The decrease in chlorine demand was not signiﬁcant for the range of operating
conditions studied by using the ozonation/membrane system. Some experimental
results even show an increase in the chlorine demand. This implies that ozonation
create chlorine-reacting sites on DOC. Biotreatment was shown to decrease the
chlorine demand effectively. The removal efﬁciency ranged from 60% to 90% in the
ozonation/membrane system followed by bioﬁltration. This indicates that the
chlorine reactive material is biodegradable.

The formation of THMs decreased with increasing temperature, gas ﬂow rate,
and ozone dose regardless of whether or not bioﬁltration followed the bench-scale
system. Similarly, the level of HAAs decreased with increasing the gas flow rate or
the ozone dose, regardless of whether or not bioﬁltration followed
ozonation/membrane ﬁltration. On the contrary HAAs was independent of the water
temperature. The formation of ozonation by-products, i.e., aldehydes and ketoacids,
signiﬁcantly increased after ozonation under all operating conditions. Aldehydes and

ketoacids are easily biodegradable. The use of bioﬁltration following

198

ozonation/membrane system resulted in an 85 to 95% reduction in aldehydes
concentrations and a 94 to 100% reduction in ketoacids levels.

Chlorine demand and DOC, THMs and HAAs, aldehydes and ketoacids,
THMs and surrogate parameters (DOC, UV-254, HS), HAAs and surrogate
parameters (DOC, UV-254, HS), aldehydes and NHS, ketoacids and NHS were
positively correlated and their correlations were statistically signiﬁcant. Aldehydes
and surrogate parameters (UV-254, HS), ketoacids and surrogate parameters (UV-254,

HS) were negatively correlated and their correlations were statistically signiﬁcant.

7.1.4 Correlations Between AOC and Ozonation By-products

The level of formaldehyde, glyoxal, and methyl glyoxal increased in the early
stage of ozonation and then decreased after they reached their maximum level. It was
observed that the maximum concentration of formaldehyde increased with increasing
gas flow rate and ozone dose. The maximum concentration of glyoxal and methyl
glyoxal increased as the gas flow rate increased but decreased as the water
temperature increased.

There exists statistically signiﬁcant and positive correlations between
ozonation by-products and AOC. The experimental results also show that the total
concentration of ozonation by-products (aldehydes + 2-butanone + ketoacids),
aldehydes, and methyl glyoxal are three good surrogates for AOC, P17, and NOX for
Lake Lansing water treated by ozonation/membrane system and bioﬁltration under

experimental conditions studied in this research.

199

7.2 RECOMMENDATIONS FOR FUTURE RESEARCH

Based upon the experimental results of this study, the ozonation/membrane
system followed by biotreatment should be able to effectively decrease the
concentrations of DOC, AOC, BDOC, and UV-254-absorbing compounds, reduce the
formation of THMs and HAAs, remove most of the aldehydes and ketoacids, decrease
the chlorine demand of water, and reduce the bacteria regrowth potential in the
distribution system. Recommendations for future research include:

1. Tubular ceramic membranes with a molecular weight cut-off of 15 kD were used
in this study. Ceramic membranes with different molecular weight cut-offs and
different types of ceramic membranes (when they are available) should be
investigated to determine their effect on the treated water quality.

2. Only a simple membrane ﬂux recovery test was conducted in this study. In order
to better understand the mechanism of the fouling process, the prevention of
membrane fouling by ozonation should be investigated. In addition, the
transmembrane pressure, which is an operational parameter that inﬂuences
permeate flux and can be an indicator of membrane fouling, should be monitored.

3. Carboxylic acids, such as formic acid and acetic acid, need to be monitored in
order to better understand their formation and transformation of ozonation by-
products.

4. The Lake Lansing water contains low bromide concentration. As bromide not
only reacts with ozone and forms bromate, but also affects the formation of other
DBPs, it is important to investigate the applicability of this system on waters
containing high bromide concentrations.

5. The economic analysis needs to be conducted when cost related information of

ceramic membrane is available.

200