an: augumutw.
. .......... 3% .
fig“;

fluafis

, 3.3.3 4

~
.3,

fix

Rm .

53a:

("1‘ 'I

., 7 firm...
fié w

t
t

.F
r‘
a}, .

3
’3
‘~ X

.. . v A.

“WE.

5". .l
a

Law;

nmfiirdm .5 fine

{I

n 1" oul ‘.vavfilu
_fl§m¥ 3.? V . .6...
. Lea. , . ‘ . .
Lyn... . , ,. ‘_

5 n4 .7
51.1“?! 417
1;. :{Jfinail

1.....9 :

 

 

 

 

.7 4....(1. : f
... .I: .

THF‘TlQ

UTA

This is to certify that the

dissertation entitled

The Effects of Ozonation Pathways on the
Formation of Ketoacids and Assimilable Organic
Carbon (AOC) in Drinking Water

presented by

Kyung-Hyuk Lee

has been accepted towards fulfillment
of the requirements for

PhD degree in Environmental Engineering

 

 

/ Jaw GWEN

Mflr professor

Date April 30, 2001

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

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 cJCIRC/DatoDquGG-MS

 

THE EFFECTS OF OZONATION PATHWAYS ON THE
FORMATION OF KETOACIDS AND ASSIMILABLE ORGANIC
CARBON (AOC) IN DRINKING WATER
By

Kyung-Hyuk Lee

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

2001

ABSTRACT

THE EFFECTS OF OZONATION PATHWAYS ON THE FORMATION OF
KETOACIDS AND ASSIMILABLE ORGANIC CARBON (AOC) IN DRINKING
WATER
By

Kyung-Hyuk Lee

In the United States most drinking water plants use chlorine for disinfection (98%).
However, concern relating to the formation of disinfection byproducts has resulted in 332
(OzoneNews, 2000) facilities having replaced chlorination with ozonation. Although
ozonation is effective at removing pathogens and organic pollutants, ozonation can result
in an increase in the amount of biodegradable organic matter in the treated water, which
can enhance the regrowth of bacteria in the distribution system. Moreover, the reaction of
ozonation by-products with chlorine during post-disinfection leads to the formation of
potentially toxic disinfection by-products (DBPS). The reaction of ozone with NOM
(Natural Organic Matter) can occur by two different pathways: reaction of molecular
ozone and the reaction of OH radicals, which are produced from the decomposition of
molecular ozone. The molecular ozone reaction with NOM would preferentially form
ADC and decrease precursors of DBPS rather than that of OH radical reaction.As such,
the pathways for the reaction of ozone with NOM were controlled by the factors that

affect the pathways of ozone decomposition to control the formation of ADC and DBPs.

The ratios of [OH-]/[O3] under the various ozonation conditions were determined to
investigate the pathways forming ADC and DBPs. The effect of operational parameters
such as temperature, pH, HRT (Hydraulic Retention Time) and initial ozone
concentration on the formation of ADC (Assimilable Organic Carbon), ketoacids, which
is a portion of ADC as well as ozonation by-product, during ozonation and DBPS during
post-chlorination were investigated. In addition, inorganic additives such as hydrogen
peroxide, bicarbonate ion and phosphate ion, which affect the pathways of ozone
decomposition, were used to control the formation of ADC, ketoacids and DBPs during
ozonation. The conditions where the molecular ozone reaction predominated resulted in

an increase in the formation of ADC and resulted in the formation of DBPs precursors.

ACKNOWLEDGMENTS

I would like to express my deep appreciation to my advisor, Dr. Susan Masten for
her guidance and supports during the years I have been in Michigan State University. I
have learned a lot more other than academic research from her. Also, I would like to
thanks my committee members, Dr. Alex Yavich, Dr. Mackenzie Davies, Dr. Syed
Hashsham, and Dr. Jon Bartholic for their guidance and fruitful discussion.

A special thanks also goes to my friends, Kuan-Chung Chen and other colleagues,
who took part in this research. I also would like to thanks to my friends, who shared A10
Lab. and Dr. Jeong-Hoon Park. We have had a nice time whenever I frustrated or I had
good events. They also gave me a special memory in MSU golf courses.

I need one more sentence for my Korean friends to express my appreciation. I
owe a lot of debt from them since they encouraged me to pursuit a degree in US and
exchanged academic information from Colorado, Kentucky, Texas and Korea while I
studied in MSU. I could never have completed my Ph.D. without them. I recognize that
my degree is not only for me but also all others who were always with me.

More than research work, I have had a lot of experiences that I could not have had
if I were not at MSU. During my past years in Michigan, my experiences in CB 280 and
Korean Student Organization, Catholic Church as well as a graduate student would be
precious fortunes in my life. I would not forget my memory in Michigan.

Finally, I would like to thank my family, my parents and brother, for their

heartfelt love. They were physically far from me. However, they always were near from

me in my mind.
I especially thank to God who is leading me to the way where God needs me and I
want to be.
Thanks be to God.
2001. 5.

Kyung—Hyuk Lee

TABLE OF CONTENTS

 

1. OBJECTIVE ................................................................................................................. l
2. SIGNIFICANCE ........................................................................................................... 2
3. BACKGROUND - - .................................................. 6
3.1 . Chlorination ............................................................................................................. 6
A. Usual practices in US. ........................................................................................... 6

B. NOM ....................................................................................................................... 7

3 .2. Ozonation ................................................................................................................. 9
3.3. Ozone by-products ................................................................................................. 10
A. BDOC formation .................................................................................................. 11

B. BDOC problems ................................................................................................... 12

C. AOC as a parameter to measure regrth ........................................................... 12

3.4. BDOC control ........................................................................................................ 16
A. BDOC/AOC formation by molecular ozone. ....................................................... 16

B. BDOC/AOC formation by ozone/hydrogen peroxide .......................................... 16

C. BDOC/AOC formation by hydrogen peroxide/UV ............................................. 18

D. Gamma ray/electron beam irradiation .................................................................. 19

3.5. Ozone pathway ....................................................................................................... 22
A. Ozone decomposition reaction ............................................................................. 22

B. Factors affecting ozone decomposition ................................................................ 25

3.6. Biodegradation of additives for ozonation ............................................................. 33

vi

4. RESEARCH APPROACH ......................................................................................... 35

 

4.1. Experimental methods ........................................................................................... 40
4.2. Water sources and materials .................................................................................. 50
4.3. Analytical method .................................................................................................. 50
5. EFFECT OF OZONE DOSAGE ............................................................................... 59
5.1 RCT values for ozone dosages. ................................................................................ 59
5.2. Formation of ketoacids with ozone dosage ............................................................ 62
5.3. Formation of AOC for ozone dosages. .................................................................. 65
5.4. Formation of THM and HAA for ozone dosages. ................................................. 68
6. EFFECT OF HYDRAULIC RETENTION TIME (HRT) ...................................... 71
6.1. Effect of HRT on RCT values. ................................................................................ 71
6.2. Formation of ketoacids for HRTs .......................................................................... 76
6.3 Formation of AOC for HRTS. ................................................................................. 78
6.4 Formation of THMs and HAAs for HRTs. ............................................................. 80
7. EFFECT OF TEMPERATURE . ........ . - ...... 83
7.1. RC1 values for temperatures ................................................................................... 83
7.2. Formation of ketoacids for temperatures. .............................................................. 86
7.3. Formation of AOC for temperatures. ..................................................................... 89
7.4. Formation of THM and HAA for temperatures. .................................................... 91
8. EFFECT OF pH .......................................................................................................... 93
8.1. RCT values for pH ................................................................................................... 93
8.2. Formation of ketoacids for pH ............................................................................... 96
8.3. Formation of AOC for pH ...................................................................................... 99

Vii

8.4. Formation of THMS and HAAS for pH ................................................................ 101

 

 

 

9. EFFECT OF BICARBONATE - -- _ - - - - -- - - 105
9.1. Rm values in solutions containing bicarbonate ions ........................................... 105
9.2. Formation of ketoacids in solutions containing bicarbonate ion ......................... 108
9.3. Formation of AOC in solutions containing bicarbonate ions .............................. 113
9.4. Formation of THMS and HAAS in solutions containing bicarbonate ions .......... 1 16

10. EFFECT OF HYDROGEN PEROXIDE ............................................................. 119
10.1. RC1 values for hydrogen peroxide ..................................................................... 1 19
10.2. Formation of ketoacids for hydrogen peroxide .................................................. 123
10.3. Formation of AOC in the hydrogen peroxide/ozone treatment ......................... 125
10.4. Formation of THMs and HAAS for hydrogen peroxide/ozone treatment .......... 127

11. EFFECT OF PHOSPHATE ............ -_ - ...... .......................................... 130
11.1. RCT values in solutions containing phosphate ions ............................................ 130
11.2. Formation of ketoacids in solutions containing phosphate ................................ 133
11.3 Formation of AOC in solutions containing phosphate ions ................................ 135
11.4. Formation of THMs and HAAS in solutions containing phosphate ions ........... 137

12. COMPARISON OF WATER CHARACTERISTICS ........................................ 140

13. CONCLUSIONS ..................................................................................................... 147

14. REFERENCES ........................................................................................................ 151

APPENDICES...... ......... - - - ............................................................ 164

viii

LIST OF TABLES

Table 3.1. Rate Constants of Ozone and OH radical. ...................................................... 32
Table 3.2. Free Energy of Ozonation Additives. .............................................................. 34
Table 4.1. Characteristics of Lake Erie and Lake Lansing water. .................................... 50

Table 6.1. Comparison of RCT values and half-life of OH radical in Lake Erie water and
Lake Lansing water. .................................................................................................. 73
Table 7.1. Activation energy and frequency factor of bicarbonate and ozone ................. 84
Table 9.1. Concentrations of carbonate ion including added bicarbonate in Lake Erie. and
Lake Lansing ........................................................................................................... 1 10
Table 10.1. Rate constant for the H20-O3 and Aqueous HzOz-O3 Systems (Staehelin and
Hoigne’, ES& T. 1982) .............................................................................................. 120

Table 12.1. Characteristics of US. surface waters ......................................................... 144

ix

LIST OF FIGURES

Figure 3.1. Decomposition of ozone in waters. (Masten and Davies. 1994) .................... 24
Figure 4.1. Schematic diagram of batch type ozone reactor. ............................................ 46
Figure 4.2. Schematic diagram of continuous semi-batch ozone reactor. ........................ 48
Figure 4.3..Procedure for generating diazomethane. ........................................................ 54
Figure 4.4. Procedure for ketoacids analysis ................................................................... 55
Figure 4.5. Procedure for HAAS analysis. ........................................................................ 57

Figure 5.1. Effect of ozone dose on Rc-r values (TOC: 2.0 mg/L for Lake Erie and 9.5
mg/L for Lake Lansing ). .......................................................................................... 61
Figure 5.2. Formation of ketoacids for ozone dosages (TOC: 2.0 mg/L in Lake Erie, 9.5
mg/L in Lake Lansing, HRT: 12.5 min, Temperature: 25 OC) .................................. 64
Figure 5.3. Formation of AOC for ozone dosages (TOC: 2.0 mg/L in Lake Eric. 9.5 mg/L
in Lake Lansing, HRT: 12.5 min, Temperature: 25 OC). .......................................... 67
Figure 5.4. Formation of THMs on ozone dosages (HRT: 12.5 min. Temperature: 25 0C).
................................................................................................................................... 69
Figure 5.5. Formation of HAAS for ozone dosages (HRT: 12.5 min, Temperature: 25 OC).
................................................................................................................................... 70
Figure 6.1. The effect of HRTS on Ozone, OH radical concentration and Rm values in
Lake Erie water (Ozone dosage: 1mg O3/mg C, Temperature: 12.5). ...................... 74
Figure 6.2. The effect of HRTS on Ozone, OH radical concentration and RCT values in

Lake Lansing water (Ozone dosage: 1 mg O3/mg C, Temperature 25 OC). .............. 75

Figure 6.3. Formation of ketoacids for HRTS (Ozone dosage: 1 mg O3/mg C,
Temperature: 25 0C) .................................................................................................. 77
Figure 6.4. Formation of AOC for HRTS (Ozone dosage: 1 mg O3/mg C. Temperature: 25
°C). ............................................................................................................................ 79
Figure 6.5. Formation of THMs for HRTS (Ozone dosage: 1 mg Og/mg C, Temperature:
25 oC). ....................................................................................................................... 81
Figure 6.6. Formation of HAAS for HRTS (Ozone dosage: 1 mg O3/mg C, Temperature:
25 0C). ....................................................................................................................... 82
Figure 7.1. The effect of temperature on RCT values (Ozone dosage: 1 mg O3/mg C). 86
Figure 7.2. Formation of ketoacids for temperature (Ozone dosage: 1 mg O3/mg C, HRT:
12.5 min). .................................................................................................................. 88
Figure 7.3. Formation of AOC for temperature (Ozone dosage: 1 mg O3/mg C, HRT:
12.5). ......................................................................................................................... 90
Figure 7.4. Formation of THMs for temperature (Ozone dosage: 1 mg Og/mg C, HRT:
12.5). ......................................................................................................................... 91
Figure 7.5 Formation of HAAS for temperature (Ozone dosage: 1 mg O3/mg C, HRT:
12.5). ......................................................................................................................... 92
Figure 8.1. The effect of pH on RC1 values (Ozone dosage: 1 mg O3/mg C, Temperature:
25 oC) ......................................................................................................................... 95
Figure 8.2. Formation of ketoacids for pH (Ozone dosage: 1 mg O3/mg C, HRT: 12.5
min, Temperature: 25 0C). ........................................................................................ 97
Figure 8.3. Decomposition of ozone in the raw water and buffered water (Lake Erie

water, Ozone dosage: 2mg/L, pH 8, Temperature 25 0C) ......................................... 98

xi

Figure 8.4. Formation of AOC for pH (Ozone dosage: 1 mg O3/mg C, HRT: 12.5 min,
Temperature: 25 0C) ................................................................................................ 100
Figure 8.5. Formation of THMs for pH (Ozone dosage: 1 mg O3/mg C, HRT: 12.5 min,
Temperature: 25 °C) ................................................................................................ 103
Figure 9.1. The effect of bicarbonate on RCT values (Ozone dosage: 1mg O3/mg C,
Temperature: 25 °C). .............................................................................................. 107
Figure 9.2. Formation of ketoacids on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT:
12.5 min, Temperature: 25 °C). .............................................................................. 111
Figure 9.3. Formation of ketoacids on bicarbonate (Ozone dosage: 1mg Og/mg C, HRT:
12.5 min, Temperature: 25 °C). .............................................................................. 112
Figure 9.4. The effect of bicarbonate on the formation of AOC (Ozone dosage: 1mg
Og/mg C, HRT: 12.5 min, Temperature: 25 °C). .................................................... 114
Figure 9.5. Formation of AOC on bicarboante (Ozone dosage: 1mg O3/mg C, HRT: 12.5
min, Temperature: 25 °C). ...................................................................................... 115
Figure 9.6. Formation of THMS on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT: 12.5
min, Temperature: 25 °C). ...................................................................................... 1 17
Figure 9.7. Formation of HAAS on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT: 12.5
min, Temperature: 25 °C). ...................................................................................... 118
Figure 10.1. Diagram of ozone decomposition by hydrogen peroxide ions (Langlais et al.,
1990) ...................................................................................................................... 120
Figure 10.2. The effect of hydrogen peroxide on RCT values (Ozone dosage: 1 mg O3/mg

C, Temperature: 25 0C). .......................................................................................... 12]

xii

Figure 10.3. Formation of ketoacids for hydrogen peroxide (Ozone dosage: 1 mg O3/mg
C, HRT: 12.5 min, Temperature: 25 0C) ................................................................. 124
Figure 10.4. Formation of AOC for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,
HRT: 12.5 min, TemperatUre: 25 °C). .................................................................... 126
Figure 10.5. Formation of THMs for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,
HRT: 12.5 min, Temperature: 25 °C). .................................................................... 128
Figure 10. 6. Formation of HAAS for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,
HRT: 12.5 min, Temperature: 25 °C). .................................................................... 129
Figure 11.1. Schematic drawing of the proposed set of typical reactions, resulting in the
conversion of OH- to 02' radicals due to the action of phosphate ions as promoters.
(Hoigné and Bader, 1985) ....................................................................................... 131
Figure 11.2. The effect of phosphate on RCT values (Ozone dosage: 1mg O3/mg C,
Temperature: 25 °C). .............................................................................................. 132
Figure 11.3. The effect of phosphate on the formation of ketoacids (Ozone dosage: 1mg
O3/mg C, HRT: 12.5 min, Temperature: 25 °C). .................................................... 134
Figure 11.4. The effect of phosphate on the formation of AOC (Ozone dosage: 1mg
O3/mg C, HRT: 12.5 min, Temperature: 25 °C). .................................................... 136
Figure 11.5. Effect of phosphate on the formation of THMs (Ozone dosage: 1mg O3/mg
C, HRT: 12.5 min, Temperature: 25 0C) ................................................................. 138
Figure 11.6. Effect of phosphate on the formation of HAAS (Ozone dosage: 1mg O3/mg
C, HRT: 12.5 min, Temperature: 25 °C) ................................................................. 139
Figure 12.1. Correlation between DOC and £21)ch in US. surface water (Dashed line: 95

% confidence interval). ........................................................................................... 145

xiii

Figure 12.2. Correlation between ketoacids and AOC in Lake Erie and Lake Lansing

water ........................................................................................................................ l 46

xiv

NOM

AOC

DBPs

HRT

EPA

ICR

SDWA

THMS

THMFP

HAAS

MCL

BDOC

TOC

GAC

UV254

FBR

FBT

AWWA

ABBREVIATION

Natural Organic Matter
Assimilable Organic Carbon
Disinfection by-products
Hydraulic Retention Time
Environmental Protection Agency
Information Collection Rule
Safety Drinking Water Act
Trihalomethanes

Trihalomethane Formation Potential
Haloacetic Acids

Maximum Contaminant Levels
Biodegradable Organic Carbon
Total Organic Carbon

Granular Activated Carbon

UV 254nm absorbance

Fluidized Bed Reactor

Fluidized Bed Treatment

American Water Works Association

XV

1. OBJECTIVE

The major objective of this project is to investigate the effect of the reactions of
molecular ozone and the OH radical on the biodegradability of NOM (Natural Organic

Matter) upon ozonation.

The partic'ular objectives are,

1) To determine the effect of molecular and radical reaction pathways on the formation
of AOC (Assimilable Organic Carbon).

2) To investigate the effect of operational parameters (e.g. temperature, pH, HRT
(Hydraulic Retention Time), dissolved ozone concentration) on the reaction pathways
of ozone, in terms of the formation of AOC.

3) To evaluate the use of inorganic additives to control the pathways of the reaction of
ozone with NOM and the formation of AOC.

4) To evaluate the formation of DBPs (Disinfection By-products) in waters in which the
formation of AOC (and subsequent remOval during biological treatment is

optimized).

2. SIGNIFICANCE

In United States most drinking water systems (98%) use chlorination for
disinfection (Irvine Ranch Water District, 1999). While the use of chlorine has
successfully prevented outbreaks of many water-borne diseases, chlorine does have
severe drawbacks. The chlorination of natural organic matter results in the formation of
trihalomethanes (THMS) and other disinfection by-products (DBPS). The presence of
these DBPS in treated water causes much concern because of their long-term health
effects. For example, chloroform, the most prevalent THM, is a human carcinogen (Jolley
et al., 1985). There is also concern that DBPS result in a high incidence of spontaneous
abortions (Waller et al., 1998).

Concerns related to the health risks associated with DBPS have brought about
stringent regulation for those DBPS. In 1994, the EPA (Environmental Protection
Agency) proposed a Stage 1 D/DBP (Disinfectants/Disinfection by-products) Rule and
Stage 2 D/DBP Rule in accordance with a specified schedule. According to the
Information Collection Rule (ICR), modification can be made when new available
information regarding toxicology is present. With the promulgation of the Stage 1 D/DBP
rule, the MCL (Maximum Contaminant Levels) for total THMS were lowered from 100
ug/L to 80 ul/L and that for haloacetic acids (HAA) was set the MCL at 60 ug/L. The
Stage 2 D/DBP Rule must be promulgated by May 2002 based on the US. SDWA
(Safety Drinking Water Act). Although MCLs designated in the Stage 2 D/DBP Rule are
tentative (40 ug/L of total THMS, 30 ug/L of HAA), it is likely that the MCL will

continue to be reduced in accordance with the Information Collection Rule (ICR).

 

Therefore, the water treatment plants face technical problems to meet these regulations.
Clearly new technologies must be developed to better and more cost-effectively solve
these problems.

As a result of the concern over DBPS and in order to comply with increasingly the
more stringent regulations concerning DBPS, many plants in the United States have
ceased using prechlorination and are now using ozone. Although the use of ozone results
in low concentrations of THMS in the treated water (Lykins et al., 1994; Speitel et al.,
1993; Shukairy and Summers, 1992), ozonation can results in other problems, such as an
increase in the formation of easily Biodegradable Organic Carbon (BDOC) which can
enhance the regrowth of bacteria in the distribution system. Moreover, ozonation results
in the formation of haloacetic acids and haloketo acids upon the post chlorination of
ozonated water (Kuivinen and Johnsson, 1999; Najm et al., 1994; Speitel et al., 1993).

One method to control regrowth problems and to avoid the formation of
precursors of chlorinated oxygenated species (such as chloroacetic acid) is to use
biological treatment after ozonation and before chlorination. The advantage of using
biological treatment is that any BDOC produced can be consumed by the microorganisms
prior to chlorination and distribution. The goal of this biological treatment is then to
convert all BDOC and AOC to C02, H20 and minerals.

Although BDOC formation increases with increasing ozone dose (Paode et al.,
1997; Volk et al., 1993), ozonation is rather expensive compared to the other
conventional processes such as coagulation, sand filtration, and biological process
(Yavich, 1998; Qasim et al., 1992). As such, reducing the applied ozone dosage as much

as possible provides cost benefits.

For the last decade, There has been a research on the development of a combined
ozonation/biological fluidized bed treatment process for the control of DBP precursors in
drinking water. For example, the removal of THM precursors by combined ozonation and
fluidized bed treatment (F BT), with no pretreatment, was comparable to that achieved at
the Ann Arbor Water Treatment Plant (Ann Arbor, MI), which includes lime softening,
coagulation, and GAC adsorption (Masten, 2000; Ann Arbor Sanitary Survey, 1999;
Yavich et al., 1997). Moreover, the ozonation/FBT pilot system operated using source
water (Huron River water) that had a higher organic loading than the one entering the
Ann Arbor Plant, which consist of 80% Huron River water and 20% ground water. The
cost analysis indicated that the ozonation/FBT process was more cost effective than the
conventional processes similar to those employed at the Ann Arbor Water Treatment
Plant (Masten, 2000). Because this novel system has a small footprint and generates
virtually no sludge, the ozonation/FBT system can be very attractive for small water
utilities and larger facilities, where space for retrofitting or expansion is limited.

In order to operate ozonation/biological process effectively in terms of BDOC and
AOC minimization, ozonation must produce as much BDOC/AOC as possible so that the
BDOC/AOC produced from ozonation can be completely removed in the subsequent
biological process. Thus, if the ozone reaction pathways with NOM are controlled for the
formation of BDOC, the applied ozone dosage can be reduced and a biologically stable
water can still be produced. However, unlike in European countries, in the US. the
combination of ozonation and biological treatment is unpopular due to fears associated
with microorganisms present in the biological system. Nevertheless, in order to resolve

the BDOC problem, some US. water treatment plants have modified their conventional

process by trying to enhance biological growth in the sand or GAC (Granular Activated
Carbon) filtration systems. Unfortunately, sand filtration does not effectively remove
BDOC because of the frequently required back washing of the filters using chlorinated
water. This problem is exacerbated by presence of microorganisms in the filter since they
result in more rapidly increasing head loss, resulting in more frequent backwashing. As a
result, it is difficult to maintain sufficient quantities of microorganisms to obtain efficient
degradation of BDOC in the filters. Thus, for the water treatment plants that can not
accomplish their biological treatment effectively after ozonation, they need to minimize
the formation of BDOC during ozonation. The way to minimize BDOC formation is to
control the ozonation process toward minimizing the formation of BDOC by ozone. The
control of ozone decomposition pathways in water treatment, whether biological
processes are used or not, can not only prevent microbial regrowth in the distribution

system which eventually deteriorate the water quality but also save costs.

3. BACKGROUND

3.1 . Chlorination

A. Usual practices in US.

Disinfection is the most important step in the drinking water treatment process to
protect human from pathogens. For the past eight decades in United States, chlorine, well
established as a disinfectant, has successfully reduced the number of incidences of water-
borne diseases. For this reason, chlorination in United States drinking water has been
used widely (98 %).

Chlorine can be dosed at one or both of two different steps during the drinking
water treatment processes. Prechlorination is applied for the removal, inactivation or both
of pathogenic microorganisms from water source. Postchlorine is used to maintain
residual chlorine to prevent microbial regrowth in the distribution system. According to
the Stage 1 D/DBP rule, the MRDL (Maximum Disinfection Residual Level) of chlorine
is 4 mg/L. However, a minimum residual chlorine must also be maintained to hinder the
growth of microorganisms.

In order to maintain required residual chlorine concentrations as well as disinfect
pathogens during the water treatment process, chlorine dosages are controlled in the
prechlorination and postchlorination steps. Chlorine decomposes or is consumed after
prechlorination. Chlorine is also consumed through other unit processes such as sand
filtration and GAC (Granular Activated Carbon). The chlorine dosage at post-

chlorination is determined based upon the dose necessary to maintain the concentration of

residual chlorine in the distribution system as required by the Safe Drinking Water Act.
According to the questionnaire developed by the AWWA (American Water Works
Association) Disinfection Committee, 96 % of 350 water utilities in US. dose the
chlorine with 15 mg/L as the maximum (White, 1999). However, such high chlorine
dosages have considerable side effects in terms of the water quality pertaining to DBPS.

Even at much lower chlorine dosages the effects can be significant.

B. NOM

Natural organic matter (NOM) is a complex matrix of organic matter present in
natural water. Humic substances, which comprise 40-60% of the dissolved NOM having
a high molecular weight, and polyelectrolytic macromolecules, can be divided
operationally into hydrophilic and hydrophobic substances (Manahan, 1991).
Hydrophobic substances, called a humic fraction, consist of humic acids and fulvic acids.
Hydrophilic substances, called a non-humic fraction, consist of hydrophilic acids,
proteins, amino acids, and carbohydrates.

Humic substances are biologically recalcitrant materials formed during the
decomposition of vegetation occurring as deposits in soil, marsh sediments, peat, coal,
lignite, or in almost any location where large quantities of vegetation have decayed
(Manahan, 1993). Humic acids usually have a molecular weight of 2000-5000 daltons or
greater and aquatic fulvic acid usually has a molecular weight of less than 2000 daltons
(Thurman, 1985). The phenolic content of humic acid is greater than that of fulvic acid,

and there are more color centers on the humic acid molecule (Collins et al., 1986).

Non-humic substances are defined as compounds belonging to known classes of
biochemistry, such as amino acids, carbohydrates, fats, waxes, resins, organic acids.
(Stevenson, 1994). Non-humic substances have greater portions of biodegradable organic
carbon (BDOC) than humic substances (Yavich, 1998, Paode, 1994).

The major problem in water treatment caused by chlorination is that humic
substances can react during chlorination to produce trihalomethanes (THMS) and other
undesirable halogenated organics such as haloacetic acids, haloacetronitriles,
haloalcohols, and haloketones (e.g., see Richardson et al., 1999; Ichihashi et al., 1999;
Stevenson, 1994; Singer, 1994; Boyce and Homlg 1983; Christman et al., 1983).

Chloroform, dichlorobromomethane, dibromochloromethane, and bromoform are
well established as the predominant THMS formed during chlorination (McGuire and
Meadow, 1988). Some factors that affect the formation of THMS are the nature of humic
substances, chlorine dose, chlorine incubation time, bromide ion concentration,
temperature, and pH (Trussell and Umphres, 1978).

The other DBPS that have been reported to form upon chlorination of drinking
waters are haloacetic acids (HAAS), carbon tetrachloride, and dichloromethane (Kuivinen
and Johnson, 1999; Najm et al. 1994; Speitel et al. 1993). HAAS (monochloroacetic acid,
dichloroacetic acid, trichloroacetic acid, monobromoacetic acid, and dibromoacetic acid)
are often formed when chlorination follows ozonation of waters containing NOM

(Reckhow and Singer, 1990; Uden and Miller, 1983).

3.2. Ozonation

In the United States the first ozone plant used for the disinfection was located at
Philadelphia (PA) and installed in 1900-1905 (Rice, 1999). By 1979, a few water
treatment plants had installed primary ozonation for taste and odor control as well as for
color removal. Since the SWDA amendment of 1986, ozonation of drinking waters has
become more popular in the US. As of 1997, more than 200 ozone plants are being
operated for drinking water treatment in United States.

However, ozonation also produces by-products in the presence of NOM causing
biological instability. Therefore, postchlorination is used to suppress biological regrowth.
However, this results in additional production of DBPS as well as suppression of
biological activity. As a result, the stringent D/DBP rule can be different to meet.

Unlike those in the United States, European water supplies commonly apply
biological processes in the drinking water process. As an example, Ziirich water supply in
Switzerland and Mtihleim process in Germany use slow filtration and GAC as a means of
biological stabilization after ozonation. The use of biological process results not only in
biological stabilization but also reducing the chlorine requirement, thereby, minimizing
production of chlorination DBPS. Based on these successes, in the US. during the last
decade biological water treatment has risen as a candidate process to meet both DBP rule
and biological instability. Unlike in Europe, in the US. rapid sand filters are used widely
and have been retrofitted biological degradation. However, rapid sand filtration often
fails to exert biological treatability due to frequent clogging and to backwashing with

chlorinated water.

3.3. Ozone by-products

While ozonation can replace prechlorination and reduce postchlorination
disinfectant dosages, ozone still produces by-products upon reaction with NOM.
Richardson et al. (1999) and Gracia et al. (1996) identified 110 different organic
compounds, mainly carboxylic acids, aromatics, hydrocarbons, aldehydes, ketones, and
furan-carboxylic acids, which formed from the reaction of ozone with humic substances.
The by-products of the reaction of ozonation with fulvic acid were also identified by
Benoit et al. (1993). The aliphatic compounds rich in oxygen were found at high ozone
dosages whereas benzene polycarboxylic acids and polyhydroxy benzene polycarboxylic
acids were identified in low ozone concentration.

Glaze et al. (1989) have developed a method to measure carbonyl compounds
including aldehydes in drinking water. The most predominant aldehyde formed upon the
ozonation of drinking water was formaldehyde (Schmidt et al., 1999; Carlson and Amy,
1998; Miltner et al., 1992; and Zhou et al., 1992). The formation of aldehydes is affected
by alkalinity, ozone dosage, and DOC (Paode et al., 1997). Aldehydes are proportionally
produced with DOC concentration and ozone dosage (Siddiqui et al., 1997). The reaction
mechanism for aldehyde formation during ozonation was first proposed by Glaze et al.
(1989). Molecular ozone attacks unsaturated carbon-carbon bonds forming epoxide
intermediates, which subsequently decompose to form aldehydes. Waters having a high
alkalinity (high bicarbonate concentration) yield low concentrations of aldehydes due to
OH radical consumption by bicarbonate and carbonate (Paode et al., 1997). The
scavenging of OH radicals results in a greater fraction of aldehydes. This may be

explained that the ozone molecule reacts with NOM since the presence

10

carbonate/bicarbonate favors reactions with molecular ozone. Schecheter and Singer
(1995) also observed that ozonation at higher pH values produces lower concentration of
aldehydes, most likely due to greater scavenging of OH radicals.

Three ketoacids, glyoxylic acid, pyruvic acid and ketomalonic acids also were
identified in the ozonated drinking water as ozonation by-products. The concentrations of
those ketoacids were observed to be much higher than that of aldehydes (Griffini et al.,
1999; Carlson and Amy, 1998; Xie and Reckhow, 1992). The concentrations of the
ketoacids increased with increasing ozone dose (Griffini et al., 1999; Xie and Reckhow,
1992). The ketoacids are also easily biodegradable organic compounds. The formation of
aldehydes and ketoacids has been correlated with the formation of BDOC (Griffini et al.,
1999). This result suggests that the aldehydes and ketoacids would be appropriate

indicators of biologically stability.

A. BDOC formation

Ozone oxidizes NOM to more polar and biodegradable compounds (De Laat et
al., 1991; Rice,1985). Volk et al. (1997) found that ozonation of NOM resulted in an
average composition of BDOC of 75 % humic substance, 13 % carbohydrate, and 2 %
amino acids. Cipparone et al. (1997) showed that BDOC formation was maximized at an
ozone dose of 1 mg O3/mg TOC. This result is consistent with other research results
(Siddiqui et al., 1997; Volk etal., 1993). Although the BDOC formation was increased at
higher ozone concentration (>1 mg), the effect of ozone concentration is less pronounced
(Paode et al., 1997; Volk et al., 1993). BDOC formation is more affected by the ozone

dose than contact time. Most BDOC formation occurred in the first two minutes of

11

ozonation (Volk et al., 1993). Therefore, a short contact time associated with a high
ozone dose is more effective at converting NOM to BDOC than is treatment at a long
contact time with low ozone dose. BDOC formation also was maximized where the UV
absorbance achieved maximum removal efficiency of ozonated water. Based on these
studies involving the formation of ozonation by-products such as aldehydes and
ketoacids, the ozone molecule yields more ozonation by-products which are

biodegradable organic carbon than do radical reactions.

B. BDOC problems
BDOC formation can result in problems for water utilities since as BDOC
provides a good carbon source for bacteria, their number can increase with distance away
from the point of disinfection, as the disinfectant residual is reduced. This increase has
been termed regrowth (Edge and Finch, 1987). Not only may regrowth result in the
proliferation of potentially pathogenic bacteria (Burke et al., 1984; Wadowsky et al.,
1982) but their presence may cause tastes, odor, color, and pipe corrosion problems

(Levy et al., 1986).

C. AOC as a parameter to measure regrowth
Biodegradable or assimilable organic carbon is defined as either (1) the fraction of
dissolved organic carbon (DOC) that can be used by the bacteria for growth and cell
maintenance or (2) the degree to which microbial growth is stimulated by DOC (Langlais
et al., 1991). Two different methods to determine biodegradability in drinking water have

been developed. Van der Kooji et al. (1982) proposed the AOC (Assimilable Organic

12

Carbon) method, which measures assimilable organic carbon by measuring the growth of
a pure bacterial culture using the standard method plate count. Bacteria used for AOC
method are Pseudomonas fluorescens P17 and Spirillum NOX. The plate count is
converted into units of sodium acetate by referring to a calibration curve measured. The
AOC accounted for by the P17 organism utilized includes many of the easily
biodegradable compounds: amino acids, carboxylic acids, hydrocarboxylic acids,
alcohols, and carbohydrates (Van der Kooij, 1985). However, this P17 organism can not
utilize certain major ozonation by-products such as oxalate, formate, and glyoxalate. In
order to compensate this lack of capability, Spirillum NOX can be used for utilizing those
compounds which can not be utilized by P17 organism.

Van der Kooij (1992) investigated the relationship between heterotrophic bacteria
regrowh and AOC in 20 water supplies. The results showed that there exists a significant
relationship between AOC and heterotrophic bacteria. However, DOC removal during
controlled inoculation was shown to be unrelated to heterotrophic bacteria regrowth.
Recent research showed that there appear to be a poor correlation between bacterial
numbers and AOC concentrations (Miettinen et al., 1997 ; Gibbs et al., 1993,). This might
be due to the low sensitivity of AOC measurement method. Also, another reason is that
the bacterial growth is also affected by the physical conditions: temperature, dissolved
oxygen, and turbidity (Power and Nagy, 1999).

The other method used to assess BDOC has been developed by Servais et al.
(1987) with the objective of measuring DOC concentration at the beginning and end of a
specific incubation period. In order to minimize the incubation time, the modified method

was developed by Joret et al. (1988). Using this method, the sample is recycled through a

13

sand column inoculated with biomass, and DOC is measured daily until the concentration
of DOC no longer changes with time.

The methods of BDOC and AOC (P17) were compared by Frias et al. (1995) and
Kaplan et al. (1994). These results show that the AOC method gives substantially lower
values than that obtained using BDOC and a linear relationship between BDOC and
AOC. This is not surprising since BDOC is a portion of DOC that can be mineralized by
heterotrophic microorganisms, whereas AOC is the more restrictive portion that can be
converted to cell biomass. The conceptual differences might explain the difference of
BDOC and AOC values. However, Siddiqui et al. (1997) also shows as ozone dose was
increased, BDOC and AOC followed a similar trend.

Zoungrana and Prévost (1998) reported that the determination of BOC is affected
by bicarbonate concentration. The addition of bicarbonate into samples before ozonation
or alter ozonation resulted in lower BDOC than that determined without additional
bicarbonate. The higher the concentration of bicarbonate added (25-100 mg/L as CaCO3),
the lower the BDOC obtained. This is thought to be due either to suppression of the
ability of the bacterial biomass to degrade the BDOC or a modification of BDOC
biodegradability. Thus, it might be worthy to take into account the bicarbonate effect
when the bicarbonate is added into the water to be tested.

Miettinen et al. (1999) modified the AOC method by adding inorganic nutrients to
the samples. This was done because inorganic nutrients, especially phosphorus, are often
the limiting factor for microbial growth in treated drinking water. The modified AOC
method, named AOCpotemial, was compared to the AOCmvc method developed by Van der

Kooij et al. (1982). The AOCpmemia; has a positive correlation with DOC whereas

14

AOC...“vc has poor correlation. As such, the AOCpotcmial method would be more reliable
method than AOCnmiW in demonstrating potential biological activity in treated drinking
water.

Consequently, the parameter that has the best capability to measure bacterial
regrowth or growth of coliforrns is the AOC measurement. On the other hand, the
parameter that best suited to measuring DBP formation potential or chlorine demand
through a biological treatment process is DOC measurement with biomass inoculation

(BDOC) (Huck, 1990).

15

3.4. BDOC control
A. BDOC/AOC formation by molecular ozone.

The formation of BDOC and AOC was affected by some conditions during
ozonation. The total AOC was negatively correlated with alkalinity concentrationduring
ozonation whereas BDOC was independent of alkalinity (Paode et al., 1997). Also, there
exists only a limited amount of data discussing the effect of pH on BDOC or AOC
formation during ozonation. In waters having pHs raging from 6 to 9, pH was shown to
have no significant effect of BDOC formation during ozonation (Siddiqui et al, 1997).
The effect of ozone contact time on the formation of BDOC has been observed (Volk et
al., 1993). At ozone dose of 1.5 mg O3/mg C, most of BDOC was formed in the first 5
minutes, then no significant increase of BDOC was observed upon longer contact times.

The results indicate the different ozonation conditions may influence the formation of

BDOC/AOC.

B. BDOC/AOC formation by ozone/hydrogen peroxide

The use of hydrogen peroxide combined with ozone (Peroxone) has been
investigated as a means of ozone decomposition and increasing the concentration of OH
radical (Glaze et al., 1987; Duguet et al., 1985; Brunet et al., 1984). The optimal
stoichiometric ratio of hydrogen peroxide to ozone in pure water has been found to be
0.3-0.5 (w/W). However, natural water sources have different characteristics and
concentrations of NOM. As such, the optimum [O3]/[H202] ratio varies with water
characteristics. As the [O3]/[H202] ratio is increased, the yield of OH radicals also

increases. However, the OH radical rapidly reacts with hydrogen peroxide and

16

hydroperoxide (HOz') so that the overall OH radical concentration decreases due to OH
radical consumption by H202. As result, overdosing hydrogen peroxide may not only
decrease the reaction efficiency but also increase chemical costs. As such, it is important
to determine the optimum [O3]/[H202] ratio for the each water sources in terms of BDOC
formation.

Ozone combined with hydrogen peroxide (ozone dosage: 3 and 6 mg O3/L;
[H202]/[O3]: O — 1.0) has been observed to increase BDOC and AOC formation.
However, at relatively high doses of hydrogen peroxide, AOC and BDOC formation
decreased. The formation of BDOC and AOC by ozone/peroxide process has been
compared with that formed during the reaction with ozone alone. (Speitel and Wanielista
1999; Miettinen, 1998; Siddiqui et al., 1997; Volk et al., 1997; Hacker et al., 1994; Volk
et al., 1993). Using ozone alone, applied ozone dosages up to 4 mg/L (TOC: 2.3-3.7
mg/L) increased the AOC concentration. The use of the ozone/peroxide, which would
result in OH radical reactions predomination, resulted in lower AOC concentrations than
did ozone alone. At low peroxide dosages (HzOz/O3: 0.05), the concentration of AOC
increased (Hacker, 1994). Several researchers (Siddiqui et al. (1997). Volk et al. (1997),
and Volk et al. (1993)) observed that the ozone/hydrogen peroxide process yields lower
BDOC values than does ozone alone. On the contrary, Miettinen et al. (1998) found the
opposite to be true. However, the overall TOC that could not be removed by ozone alone
was removed during ozone/hydrogen peroxide treatment.

Speitel and Wanielista (1999) compared ozone, ozone/peroxide and peroxide/UV
system for BDOC formation in Lake Austin water and Lake Houston water. Ozonation

and Peroxide/UV treatment enhanced BDOC formation, whereas ozone/peroxide

17

decreased BDOC in both water sources. The ozone doses and peroxide doses applied for
Lake Austin water were 1 — 2 mg 03/mg TOC and 0.2 —— 1.3 mg H202/mg 03,
respectively. The lower dosages were applied for Lake Houston water (ozone dosages:
0.5-1 mg 03/mgT0C and peroxide dosages: 0.4 -1.3mg H202/mg 03) to see if the higher
ozone dosages were masking the hydroxyl radical effects. However, after 03/H202
treatment, the BDOC formation decreased only slightly with increased oxidant dosage
and did not differ significantly from ozone alone. At larger ozone doses the BDOC
concentration was observed to have decreased. In order to investigate the effect of OH
radicals, it is thought that the reaction mechanisms and reaction rate of the oxidants (0;,
H202, 0H-) have to be determined for the each tested NOM. Moreover, at low
concentrations of either ozone or hydrogen peroxide compared with the NOM
concentration, the radicals might be consumed by the ozone decomposition, even though
the concentration of 0H radicals is enhanced. This is especially true if the rate of reaction
of 0H radical with NOM is slower than that of 0H radicals with ozone or ozone

decomposition intermediates.

C. BDOC/AOC formation by hydrogen peroxide/U V
The hydrogen peroxide/UV process is another method of generating 0H radical.
The main pathway of this process involving UV radiation energy is to cleave a peroxide
molecule into two hydroxyl radicals (Legrini et a1. 1993).
Speitel and Wanielista (1999) also have investigated the Peroxide/UV process to
determine its effect on BDOC formation. At irradiation time of less than 16 minutes (UV

dosage: 3 W-min/umol TOC), the formation of BDOC increased with increasing UV

18

irradiation time. However, at longer irradiation times both the BDOC and TOC
concentration decreased consistently. Apparently, long irradiation times have a negative
effect for the BDOC formation.

Although the hydrogen peroxide/UV process formed BDOC, the pathways of the
formation include not only 0H radical reactions but also photosensitized reactions as a
result of UV irradiation. As such, the direct effect of UV on the humic substance
structure makes it more difficult to distinguish reactions involving ozone molecule alone
from radical species reactions involving 0H radicals. The photosensitized reaction1 may
result in the production of 0H radicals and other radical species from sensitized humic
substance as well as resulting in change of the humic substance structures (Gao and Zepp,

1998).

D. Gamma ray/electron beam irradiation
Processes using gamma ray or electron beams also generate 0H radicals. These
reactions do not use ozone. Recently several researchers have developed such processes
(Siddiqui et al., 1996; Cooper et al., 1992; and Gehringer et al., 1985). Usually, 60Co
radioisotope or an electron beam accelerator is used for an irradiation source. Irradiation
of the solution produces some electronically excited state and free radicals. The products

that have been formed in aqueous solution are:

 

' The humic substance is a well known chromophore. In other words, the humic substance
structure absorbs light in the UV and visible range. This absorption of light can result in the

. . . . . . ]
extinction of the humic substance molecules, allowmg It to transfer Its excess energy to another

molecule in the water (Schwarzenbach, RP. 1992). This reaction is called photosensitization.

19

H20 -/\/\/\--—> [2.7] 0H0 + [2.6] 6,, + [0.6] H- + [0.7] H202 + [2.6] 1130* + [0.45] H2

(Bielski and Gebicke, 1970) Eq. 3.1.

The values in parenthesis stand for the G value (Eq. 3.1), which is the number of radicals,
excited states, ions formed or lost in a system absorbing 100 eV energy. Those species
formed by irradiation can react to remove contaminants. The irradiation process generates
additional radical species reaction beyond the 0H radical. The presence of the radicals
makes it more difficult to investigate the 0H radical reaction.

The effect of y-irradiation/ozone on NOM was investigated (Arai et al. 1986). y-
irradiation with ozone cab be used to enhance 0H radical concentration. However, there
is insufficient information concerning the reactions between hydrated electrons, hydroxyl
radicals, hydrogen atom and NOM. Karpel et al. (1997) conducted y-irradiation/ozone
experiments using NOM to investigate BDOC formation (y-ray dose: 390 Gy, TOC: 2.5
mg/L, NaHC03: 4.5 mM, Ca(N03)2: 0.4 mM). The results Show that y-irradiation with
ozone and ozone/peroxide could enhance BDOC formation compared with ozone alone.
The NOM tested surprisingly did not increase BDOC concentration at ozone dose 1
mg/mg C, and 23 % of BDOC was increased at ozone dose 2.5 mg/mg C. This may
indicate that the NOM tested probably has low proportion of reactive sites with the ozone
molecule. Thus, increasing the 0H radical concentration relative to ozone/peroxide/UV
processes can further oxidize NOM to BDOC. Although Karpel et al.(1997) found that

ozone/peroxide and y-irradiation/ozone which resulted in similar BDOC formation

efficiencies, it is difficult to compare those two processes. y-irradiation uses relatively

20

more energy than other processes such as ozone, peroxide, and the mechanism of y-
irradiation involves not only 0H radicals but also the hydrated electron. Thus, both
reduction reaction and oxidation reactions occur in this process. Experiments with y-
irradiation alone support this hypothesis since y-irradiation alone also could enhance the

BDOC formation which could not be enhanced by ozone alone.

21

3.5. Ozone pathway

As mentioned in previous discussions the various processes used in water
treatment to oxidize NOM result in significantly different pathways involving NOM to
form BDOC. As such, it may be important to understand and control the pathway(s) of

ozone decomposition to control BDOC formation.

A. Ozone decomposition reaction

The pathways of the ozone decomposition can be classified in three steps:
initiation, promotion, and inhibition (Staehelin and Hoigné, 1985). As shown in Fig. 1.,
some compounds, such as hydroxide ion, hydroperoxide ion (H02'), and primary
alcohols, can initiate the decomposition of molecular ozone to superoxide ion (02')
(Staehelin and Hoigné 1985). Hydroxide ions play a fundamental role in initiating the
ozone decomposition. These ions initiate ozone decomposition by reacting with ozone,
producing superoxide anions (02'). Ozone then reacts with the superoxide anion to form
ozonide radical ions and oxygen. The protonated ozonide ion rapidly decomposes to
produce oxygen and hydroxyl radicals. The hydroxyl radicals decompose ozone further,
producing more superoxide ions, which participate in the cycle reaction. (Masten and
Davis, 1994; Bailey, 1978).

As mentioned previously, there are three steps involved in 0H radical formation
during the ozone decomposition: Initiation, promotion, and scavenging reaction. The
commencement of free radical reactions during ozone decomposition to form a
superoxide ion from the ozone molecule is called the initiation step. The promotion

reaction serves to regenerate the superoxide ion (02') from hydroxyl radical. Typical

22

 

promoters include organic compounds with aryl groups and primary alcohols. The
scavenging reaction consumes 0H radicals without producing superoxide ions.
Carbonates, bicarbonates and phosphate ions may serve as promoters. Thus, if these steps
could be controlled properly, 0H radical formation from the ozone molecule could also

be controlled and BDOC/AOC formation could be optimized.

23

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The dark reaction that occur in "pure" water are shown in blue. The additional dark reactions

that may occur when reactive solutes are present are shown in red. D is a compound that can
react directly with ozone; I is a compound that, when it reacts with ozone via an electron transfer
reaction, produces the ozonide radical ion (i.e., it is an initiator of ozone decomposition); RH is a
promotor of ozone decomposition; and S is an 'OH scavenger. Propagation of the chain reaction
may occur by the reaction of RH with -OH to form an organoperoxide radical, R00 -. Termination
of the chain reaction may occur via reaction of S with -OH to form a secondary radical, (D, which
dose not participate in the chain reaction. Additional reaction that occur in O 3/H202 and/or 03/UV
systems are shown in yellow.

Figure 3.1. Decomposition of ozone in waters. (Masten and Davies, 1994)

24

 

B. Factors affecting ozone decomposition

Ozone decomposition may be controlled by regulating inhibition, promotion, and
inhibition reactions. Yuteri and Gurol (1988) found that the ozone decomposition kinetics
were significantly influenced by such factors such as pH, TOC, and alkalinity as well as
background NOM characteristics. The retardation of ozone decomposition by altering
these factors will permit an increase in concentration of molecular ozone compared to
that of the radical species. In contrast, the acceleration of the ozone decomposition would

lead to the opposite situation.

B.l. pH
The hydroxide ion is a well known initiator of ozone decomposition. As an
initiator, the high concentration of hydroxide ions at high pH accelerates ozone
decomposition. The initiation of ozone decomposition in distilled water is first order in
ozone and hydroxide ion concentration. The reaction rate of ozone with hydroxide was
observed to be 70:7 M'ls'l (Stahelin and Hoigné, 1982). The log values for the ozone
decomposition rate were shown to be linear with pH (Stahelin and Hoigné, 1982;

Sullivan and Roth, 1979).

8.2. Alkalinity
Staehelin and Hoigné (1982) found that carbonate and bicarbonate ions have the
ability to stabilize molecular ozone. This has been confirmed by many other researchers

(Yuteri and Gurol, 1988; Reckhow et al., 1986; Fomi et al., 1982). However, as

25

 

bicarbonate and carbonate are scavengers of 0H radicals their presence, which is
ubiquitous in most natural waters having a pH>7, will result in relatively low 0H radical
concentrations in these waters. This is especially true at higher pH values where the
bicarbonate ion dissociates into carbonate ion to a greater extent (pKa = 10.3) (Snoeyink
and Jenkins, 1980). Moreover, the reaction rate of bicarbonate with CH radical is greater
than that of carbonate. As such, bicarbonate and carbonate ions may serve to counter
some of the effects of 0H radicals, thereby stabilizing molecular ozone.
Bicarbonate/Carbonate ions have also been found to result in more effective
scavenging than do phosphate ions (See Table 3.1). The bicarbonate ion inhibits the
radical chain reaction, hence increasing the concentration of molecular ozone. As such,
the presence of radical scavengers could result in the predominance of reactions
involving the ozone molecule and preferential reaction of the ozone molecule with the
aromaticity of fulvic acid, a component of humic substance. Although bicarbonate and
carbonate ions are effective at 0H radical scavenging in the ozonation, they produce
secondary oxidants, the bicarbonate radical and carbonate radical, which can oxidize
some inorganic and organic compounds. Therefore, more careful attention of the

carbonate/bicarbonate radical reactions that occur during ozonation needs to be made.

8.3. NOM
NOM also acts as a promoter and initiator of ozone decomposition (Elovitz et a1.
1999; Staehelin and Hoigné, 1985). Westerhoff et al. (1996) attempted to simulate ozone
decomposition mathematically. Without considering the promotion and initiation role of

NOM on the ozone, the simulation results differ significantly difference from

26

 

experimental data. The simulation results Show a first order reaction rate during entire
reaction period and slower decomposition rate than that of experimental results.
However, when the promotion and initiation role of NOM was considered, the ozone
decomposition rate results in two stages, with rapid rate decomposition at the early stage
and followed by slower decomposition rate.

Ozone decomposition in the presence of NOM and t-butylalcohol, also an 0H
radical scavenger, has been investigated (Xiong and Legube, 1991). The results indicate
that NOM plays both initiation and promotion roles in ozone decomposition. F ulvic acid
did not participate in the radical reaction involved in ozone decomposition under very
acidic conditions (pH = 2); whereas it becomes a promoter of the ozone decomposition
with increasing pH due to increasing degree of dissociation of the acidic groups. The
fraction of fulvic acid reacting via the radical reaction increases proportionally with
increased initial [03]/[FA] ratio. Consequently, the effect of NOM, as both a promoter
and inhibitor, on ozone decomposition reactions should be considered when investigating
the ozone — 0H radical reaction.

Recently, Elovitz et al. (1999) investigated the rate of promotion and inhibition of
ozone by NOM using y-irradiation. This experiment also was conducted to determine the
rate of hydroxyl radical scavenging by eleven NOM isolates. y-irradiation results show
that hydroxyl radical scavenging rates were relatively equal for all NOM isolates (1.3 i
0.1 x 104 L/mg/s). The 0H radical scavenging rate could be divided by the rate of
promotion (superoxide production) and inhibition (non-superoxide production). The rate

of promotion and inhibition were also relatively equal for all NOM isolates (3.7 i 0.3 x

104 L/mg/s and 9.3 :t 0.7 x 104 L/mg/s, respectively). This result indicates that the NOM

27

 

predominantly behaves as direct consumer of ozone and promoter of ozone

decomposition.

8.4. Temperature
Temperature also can affect the kinetics of ozone decomposition. Sotelo et al.
(1987) showed that the ozone decomposition rate increased with increasing pH and
temperature. The ozone initiation rate was negligible at pH below 3 at various
temperatures (10 - 40 °C).
The elementary reaction rate increases as temperature is increased. Based on the
elementary reaction of ozone decomposition, the 0H radical concentration at steady state

condition can be expressed as following equation.

 

2(k() .0”. [03 ])T [OH— 1.“
[OH]...- = ' Eq. 3.2.
(kHFU‘ ,()H )T [HCO3]
(kn )T = (kNlMllh )T ([03 ]s.\' )T + (kA’().’ll.()H )T ([()H]\\ )7' Eq' 3"3

The reaction rate that initiates the ozone decomposition and produces 0H radicals
(numerator in this equation) will increase as temperature increases. The 0H radical
scavenger reaction (denominator of the equation) will also increase with increasing
temperature. Moreover, the rate of reaction of ozone molecule reaction with NOM also

changes. Unfortunately, the activation energy for each reaction is not established.

28

 

Thereby, it is difficult to establish the extent to which the 0H radical concentration will

increase or decrease with changing temperature.

8.5. Hydrogen peroxide

As previously described, hydrogen peroxide can act as an initiator of molecular
ozone decomposition. The addition of hydrogen peroxide in ozonation treatment can
enhance 0H radical concentrations because the ozone molecule is rapidly decomposed,
producing 0H radicals (Glaze et al., 1987; Duguet et al. 1985; Brunet et al., 1984).

The efficiency of 0H radical production from hydrogen peroxide depends on pH.
Hydrogen peroxide partially dissociates into hydroperoxide ion (H02’) (pKa = 11.6).
While the hydroperoxide ion is highly reactive with ozone, thus initiating ozone
decomposition (k = 106 M'ls'l), the reaction of hydrogen peroxide with ozone is very
slow. As such, the ozone decomposition by hydrogen peroxide increases with increasing
pH since hydroperoxide ion is predominant over pH 11.6. However, competition
reactions occur at high pH since hydroxide ions are more effective at initiating molecular
ozone than are hydroperoxide ions, and 0H radicals are also consumed by bicarbonate
and carbonate ions.

The effect of hydrogen peroxide on the oxidation of NOM by ozone was
investigated by Tuhkanen et a1. (1994) and Duget et al. (1986). The influence of different
hydrogen peroxide concentrations on the ozonation of NOM were investigated by Duget
et al. (1986). The use of hydrogen peroxide (1 mg/L) with ozone (3 mg/L) increased the
extent of oxidation of NOM, especially, increasing the removal efficiency of UV

absorbance (254 nm) in the early reaction times. However, at higher ozone concentrations

29

_ n...» f'dr. m.

 

(> 1 mg/L), the removal rate of UV absorption did not increase. The results indicate that
there exists, for each natural water, an optimum hydrogen peroxide concentration where
by efficiency is maximized. In other words, the 0H radicals produced from
ozone/peroxide have to be used for NOM reaction as much as possible, thereby avoiding
the consumption of reactive radicals reacting with hydrogen peroxide. Since the optimum
dosage of hydrogen peroxide would also depend on the background organic matrix in the
water, it is necessary to determine optimum ozone and peroxide doses in each water

source.

8.6. Organics

Many compounds are able to initiate, promote, or inhibit the chain-reaction
processes involving ozone (Staehelin and Hoigné 1985). Some organic chemicals have
ability to initiate the decomposition of molecular ozone. The initiators include glyoxylic
acid, formic acid and humic substances (Langlais et al., 1991). Several organic
chemicals, including aryl groups, formic acid, glyoxalic acid, and primary alcohols, have
been found to act as promoters (Staehelin and Hoigne 1985).

The common scavengers of 0H radicals include bicarbonate and carbonate ions,
alkyl groups, tertiary butyl alcohols (Staehelin and Hoigné 1985). Guittonneau et al.
(1992) investigated the kinetics of the reaction of ozone with NOM in the presence of
formic acid and t-butylalcohol. Formic acid acts as a promoter and t-butyl alcohol inhibits
ozone decomposition (See Table 3.1). The ozonation of fulvic acid in the presence of

bicarbonate was shown to be more effective for the removal of aromaticity. Fulvic acid

30

 

consists of two parts: a nulceophilic aromatic moiety and a saturated aliphatic moiety. It
can be expected that molecular ozone mostly reacts with the nucleophilic centers of
fulvic acid, while the 0H radical attacks aliphatic portion since 0H radical is non-
selective. This is consistent with the findings of Xing and Legube (1991). They found
that when the radical scavengers, bicarbonate and tert-butyl alcohol, were added to ozone
reaction on the fulvic acid, the aromaticity from fulvic acid was effectively removed
(Xiong and Legube, 1991).

The dissociation of organic acids is a function of their pKa value. For example,
the pKa value of acetic acid is 4.7. Although acetic acid predominates at pH values less
than 4.7, acetate, the dissociated form of acetic acid, is predominant at pH values above
4.7. As shown in Table l, the degree of ionization of the acid significantly affects the
reaction rate constants of the acid with OH- and ozone. For bicarbonate (pka = 10.3)
(Snoeyink and Jenkins, 1980) and phenol (pka = 9.9) (Hoigné and Bader, 1983b), the
reaction rate of 0H radical with these compounds and their dissociated form also Show
significant difference. Staehelin and Hoigné (1983b) also found that the formate ion

reacts rapidly with ozone while the formic acid is practically inert.

31

Table 3.1. Rate Constants of Ozone and 0H radical.

 

 

 

 

 

 

 

 

Solute kO3 (Mis'i) kOH (MT-[3-1) Solute ko3 (M'is'f) kOH (M'is'i)
Acetic Acid 3x105, 1.6x107‘ t-butyl alcohol 0.003T 3.1 x 10‘)I
Acetate ion 3x10'~“ 8.5x107: Bicarbonate - 8.5.4081
F orrnic acid 5* 1.3x 10XI Carbonate - 3.9x 108:
Formate ion 100* 8.5x 107: l—Propanol 0.3 71 2.8x 10":

Methanol ~0.024“ 9.7x 10““ Phenol 1.3x103 ‘ 6.6x 10":
Ethanol 0.37T 1.9x109: Phosphate ion <2.4x104 ' <1,0x10"1

 

 

 

 

 

 

l Hoigné and Bader (1983a)

‘ Hoigné and Bader (1983b)

1 Bouxton(1988)

32

 

3.6. Biodegradation of additives for ozonation

The biodegradability of the additives, which play a role in controlling ozone
decomposition, has to be taken into account. Although the additives exert their role
during the ozonation, there may be residual additives and/or their by-products after
ozonation.

As shown in Table 2., the redox potentials (AGO) of the candidates for ozonation
additives appear to be fairly high. This is one of the indicators that the organic chemicals
can be relatively easily biodegraded.

The ozonation of organic additives produce reaction by-products during the
ozonation. Dalouche (1981) determined the by-products of alcohols to be aldehydes,
ketones, acids, and esters, all of which are easily biodegraded. Thus, it is expected that
the residual additives and their ozonation by-products also can assist to increase the

biodegradability of ozonated water.

33

Table 3.2. Free Energy of Ozonation Additives.

 

 

 

 

 

Organics AGU (kJ/elec. Mole)
Acetate -27.40
Formate -39. 19
Ethanol -31 . 18

Methanol -36.84

 

 

 

 

Ref) McCarty, PL. 1969.

34

4. RESEARCH APPROACH

The aim of the study is to investigate the ozonation pathways which optimize the
formation of easily biodegradable organic compounds. The pathways by which ozone
reacts with NOM are significantly affected by the following parameters: NOM
characteristics, initial [03]/[N0M] concentration ratio, pH, reaction temperature, reaction
time, additives concentration.

The effect of operational parameters and the presence and concentration of
additives on the concentrations of molecular ozone and the 0H radical were determined
in Phase I experiments. The influence of various operational parameters on
biodegradability formation were determined in Phase 11 experiments. The effect of

additives on biodegradability formation was investigated in Phase III experiments.

Phase I :

Determination of [Ogl/JOH'] concentration ratios in various conditions.

 

The experiments were conducted as following conditions. The effect of the initial
ratio of [03]/[NOM] on the concentration ratios of [03]/[0H-] also was investigated. At
low initial ratios of [03]/[N0M] reactions involving molecular ozone should
predominate. In contrast, high initial ratios of [031/[N0M] should lead to radical
reactions unless NOM were a scavenger of 0H radical. This hypothesis was tested by
monitoring [03]/[0H-] ratios at different initial ozone concentrations. Each NOM tested
will have significantly different characteristics for ozonation since all NOM originated

from different waters by different mechanism. Although similar experiments have been

35

conducted by other researchers, it is necessary for us to conduct the experiment to
compare our results with those already published.

The effect of pH (2-10) on the concentrations of [03] and [0H-] was investigated.
At low pH, the ozone molecule predominates due to low concentrations of the hydroxide
ion. 0n the other hand, at high pH the ozone molecule is rapidly decomposed by
hydroxide ion so that the radical reaction is predominant. Although the range of the pH in
water treatment plants varies only from 6 to 8, the extremely low or high pH values are
necessary to observe the reaction pathways. Based upon the literature results, the
experiments have been conducted over a narrow range of pH (6-9). These results can not
explain which reaction is predominantly involved in NOM oxidation since both the
molecular ozone and 0H radical reactions compete at these pH ranges.

As pH changes, the concentration of 0H radical scavengers, bicarbonate and
carbonate, both also change. To compare the effect of pH, the scavenging capacity of the
solution was maintained constant.

The effect of temperature on the ratio of [03] and [0H-] was studied. The water
temperature in water treatment plants varies significantly with their locations and
seasons. To cover this range, the temperature used in our experiments was varied from 5
to 35 °C. The temperature may influence the rates of both ozone decomposition and
radical consumption by scavengers. From this study, we expected to be elucidate which
of these reactions (molecular ozone vs. radical) is more significantly influenced by
temperatures. If the ozone decomposition rate is influenced greater than radical
consumption by scavenger on the temperatures, the concentration of 0H radical would

increase with increasing temperature.

36

Several additives were used to investigate their capability to control the reaction
pathway (either ozone molecule reaction or 0H radical reaction). In the presence of the
additive that play either initiator or promoter roles, the 0H radical concentration is
expected to predominate. 0n the other hand, in the presence of the additive that plays an
inhibitor role, the molecular ozone concentration will be predominant. The concentration
of the additives also was varied (umol to mmol), respectively, to compare the efficiency
of the additives since the stoichiometry ratio of ozone to additive is important parameter

in terms of 0H radical production.

Phase II :

Effect of reaction pathways on the formation of biodegradability.

 

Using Phase I results, the ozonation step can be controlled toward reactions
involving either the ozone molecule or 0H radical by changing operational parameters
including temperature, pH, HRT, and applied ozone concentration.

Ozonation was accomplished under conditions that allow us to compare AOC
formation by reactions involving the ozone molecule and the 0H radical. Based upon
literature results, it is expected that the AOC may be increased with increased ozone
molecule reaction rather than 0H radical reaction. For the AOC analysis, the pH was
adjusted to 7 after the sample was taken since a significantly low or high pH may inhibit
the bacterial growth during AOC analysis.

The pH range tested was same as done in Phase I. Since at low pH, reactions with
molecular ozone will predominate, AOC formation should be enhanced in waters having

a low pH.

37

Except in experiments specifically conducted to study temperature effects, the
temperature of the water was maintained at 25 °C. In those cases, the temperature was
varied from 5 to 35 °C. In order to utilize same ozone dosage at different temperatures,
the ozone mass transfer coefficient also was determined at each of the temperatures used.

It was expected that the concentration of AOC would increase with decreased
hydraulic retention time (HRT) since it was expected that the molecular ozone reaction
would predominate at short HRTS. Changes in the HRT, over the range from 5 min to 30
min, were investigated. The ozone dosage was held constant as HRT was varied.

The effect of applied ozone dose was investigated at a single HRT. The results
conducted in Phase I were useful to explain these experiment results. It was expected that
high [03]/[N0M] ratios would decrease AOC formation due to OH radical prevalent
condition.

A analyses included the following parameters: TOC, residual ozone
concentration, UV254, AOC, ketoacids (glyoxylic acid, pyruvic acid and ketomalonic

acids), SDS THMS, and SDS HAAS.

Phase III :

Evaluation of additives on the ozone reaction for the formation of biodegradability.

 

The results obtained from the Phase I gave information regarding the
effectiveness of the additives in terms of controlling the ozone pathways. The
effectiveness of additives in controlling biodegradability formation during the ozonation

was compared. The initiators (hydroxide ion) and promoters (hydrogen peroxide,

38

phosphate ion) were expected to decrease AOC formation due to the prevalent 0H
radical reaction. 0n the other hand, the 0H radical inhibitors (bicarbonate) were expected
to increase AOC formation due to prevalent molecular ozone reaction. Although there are
various additives that could have been used, the inorganic additives were used for this
study to avoid problems associated with the influence of organic compounds on AOC and
TOC analysis. The residual organic additives affect TOC measurement so that TOC value
increased due to carbon in residual additives. The organic additives and their ozonation
by-products may influence the AOC values. Thus, use of inorganic additives is necessary
to avoid analytical errors.

In addition, some experiments were conducted prior to beginning Phase III
experiments. The inorganic additives in DI water were ozonated and then AOC
concentrations were determined, respectively. An additive such as peroxide, bicarbonate,
and phosphate may inhibit AOC concentration even though it does not produce such
organic ozonation by-products. In order to determine the inhibition of inorganic
additives on the formation of AOC, the AOC for the raw water containing various

concentrations of additives was measured, respectively.

39

4.1 . Experimental methods
Phase I .

Elovitz and Guten (1997) have developed a method to determine the ratios of
[03]/[0H-] concentrations during ozonation using PCBA (p-chlorobenzoic acid) as an .
0H radical probe compound. Since PCBA has a negligibly low reactivity with molecular
ozone but reacts rapidly with OH- (ko3, pch S 0.15 M'l 3", k0”, pCBA = 5x109 M'ls"), the

PCBA decomposition in the ozonation can be expressed by Eq. 4.1.

d[PCBA] ___

dt _k()H-.PCBA [PCBAIIOH'] Eq. 4.1.

The equation can be rearranged and integrated to equation 4.2.

[PCBA]

n[PCBA_—]—0}:_ kou- PCBA leH' ldt Eq. 4.2.

1n{

Elovitz and Guten (1997) defined a term, RCT, which describes for the ratio of

[OH] and [03]. (Eq. 4.3)

j[0H]dt

_— E .4.3.
[[03160 q

RCT =

40

Substitution of Eq. 4.3 to Eq. 4.2 yields Eq. 4.4.

[PCBA],

ln{——
[PCBA]0

} = _k0Ht.PCBA Rct jl03 ldt Eq. 4.4.

When the PCBA concentrations are plotted with respect to ozonation time, the
slope represents k0”. pCBA .RCT value. Since the k0“. pCBA is well known to be 5x109 M'
‘3'1 (Elovitz and Guten, 1999), the RC]. value can be obtained from the slope.

Elovitz and Guten (1999) showed that RCT value also equals to the ratio of [0H-]

and [03] at steady state conditions where RCT is constant. (Eq. 4.5.)

_ j[0H]dt _ [021.]
I[03]dt [03]

 

RCT Eq. 4.5.

Thus, the ratio of [03] and [0H] concentrations during the ozone reaction can be
obtained from RCT values. This result also gives the change of [03] and [OH-] with
respect to reaction times since the concentrations of ozone and PCBA were
simultaneously monitored during the ozone reaction time.

Using RCT values obtained from experimental results, it not only was the ratio of
[03]/[0H-] for the various operational conditions during ozonation, but also was the role
of additives and operational parameters in controlling ozone pathways. In addition, the
effect of additive concentrations in controlling ozone decomposition pathways could be

observed.

41

The results obtained in the Phase I can also be used to calculate the rate constants
for the reaction of 0H radical with NOM (kOHNOM) based upon the methodology
developed by Haag and Yao (1993).

The rate constant for the reaction of the 0H radical with NOM was determined by
using the Q concept developed by Haag and Hoigné (1985). The Q value is defined as the
amount of ozone that is consumed in removing one log unit (37 %) of an 0H radical
probe compound. The Q value can be determined from the slope of a plot of In

{[PCBA]/[PCBA]0} vs. dosed initial ozone concentrations (Eq. 4.6). such that:

 

_ln [PCBA]
[PCBA ]O
= E .4.6.
PCBA [03 ] added q

If PCBA is present at a low concentration relative to that of NOM in water and the 0H

radical is consumed only by NOM, the assumption can be made as in Eq.4.7.:

RQH,pCBA[PCBA] <<<< kQH,NOM[NOM] Eq. 4.7.

Then, the concentration of very short lived 0H radicals at steady-state can be determined

by equation 4.8:

 

 

A[03] x 1

OH- =—
[ ]SS 77 At 2k0H,N0M[N0M]

Eq. 4.8.

42

where n is the stoichiometric yield of 0H radicals from ozone decomposition. The value
of 0.67 for 11 value is usually selected based on the based-catalyzed 0H radical yield

without promotion (Westerhoff et al. 1999).

The 0H radical concentration can be applied to the PCBA degradation in Eq. 4.9.

_ d[PCBA]

dt = kOH’PCBA [PCBAHOHJSS : kexperimem [PCBA] Eq. 4.9.

Thus, the 0H radical concentration at steady-state can be described as Eq. 4.10.

k .
1011],, =-k—P—’"—-i Eq. 4.10.
0H,PCBA

Integration of Eq. 4.9 and then substitution of Eq. 4.8 to Eq. 4.9 yields Eq. 4.1 1.

 

_ 1n [PCBA] : k0H,PCBA

t
' ' d 0 = k . , xt
[PCBA]O 77 ZkOH,N0M [NOM] j l 3] experlmcm Eq. 4.11.

0

The ozone concentration consumed at the time when ln[PCBA]/[PCBA]0 equals to l
(PCBA is degraded 37 %) represents the Q value. Thus, the Eq. 4.1 1 can be rewritten as

Eq. 4.12.:

43

Q _ [CO/twat!
77' PCBA _ _k—-— Eq. 4.12.
OHJ’CBA

The determination of Q, and n values for the raw waters tested can be used to
determine the 0H radical rate constant with NOM. This could be used to understand the
pathways for the ozonation of NOM in various water sources. When the concentration of
either ozone or promotor is much lower than that of NOM and the rate of reaction of 0H
radical with NOM is slower than that of 0H radicals with ozone or ozone decomposition
intermediates, the radicals may be consumed by molecular ozone, even though the
concentration of 0H radicals is enhanced. The k011,N()M values were used to elucidate the
extent to which the 0H radicals were involved in the reaction of NOM during ozonation.
Consequently, the rate constant of 0H radical with NOM is a good indicator of the extent
to which molecular ozone and 0H radicals are involved in the reactions with NOM.

The schematic of the bench scale system is shown in Figure 4.1. The saturated
ozone solution was prepared in phosphate buffered DI. water (1 L glass bottle) providing
gaseous ozone into the solution. A syringe type (headspace-free) batch reactor (250 mE)
was used for ozonation to prevent volatile loss of ozone from the solution. Prior to
ozonation, the water samples were pre-filtered through a 0.45 pm membrane filter to
remove any particulate matter. The stock solution of PCBA (20 ppm) was prepared in
DI. water The solutions tested were prepared by adding stock solution of PCBA (ppb
level), additives and adjusting the pH and buffer capacity with phosphate, as required.
When the desired concentration of ozone was achieved, an aliquot of the ozone stock

solution was transferred immediately into the reactor. A magnetic stirrer at the bottom of

44

the reactor was used to completely mix the solution. Samples to determine the PCBA
concentration were taken at appropriate time intervals and ozone concentration was
monitored continuously by the indigo method (Standard Method, 1995). In order to
minimize detention time of the sample in the tubing from the reactor to the point where
indigo solution is mixed with sample, 0.03 inch I.D. Teflon® tubing was employed and
the length of the tubing was reduced as much as possible (approximately 10 cm). In order
to study ozonation at different temperatures (5-35 °C), the reactor was maintained at a
constant temperature during the reaction by providing cooling-heating water to the
outside of the reactor. The temperature of the solution was monitored by a thermometer.
Afier the ozone is completely decomposed, the sample also was taken to measure the

residual additives and/or PCBA concentration.

45

 

 

 

 

 

.................

 

 

 

 

 

 

Indigo Solution 1

Spiral Mixer

HPLC Gradient Mixer

 

Spectrophtometer

 

 

 

 

Figure 4.1. Schematic diagram of batch type ozone reactor.

46

Phase II .

A continuous semi-batch reactor (water-jacked glass reactor, 250mL) was used
for ozonation experiments in which the formation of AOC, ozonation by-products
(ketoacids), and DBPs under various conditions were conducted (See Figure 4.2). pH was
adjusted as required. The water was continuously fed into the bottom of the reactor with a
peristaltic pump. Another peristaltic pump also was provided at the top of the reactor to
withdraw the ozone reactor effluent to maintain a constant liquid level in the reactor.
Gaseous ozone made from pure oxygen was immediately introduced into the reactor. The
ozone was dispersed in the reactor using a fritted glass diffuser. For improved gas
transfer efficiency and solution mixing, a magnetic bar was used to stir the solution at the
bottom of reactor. Once the conditions in the reactor reached steady-state, the
concentrations of ozone in the influent and effluent of gas streams were monitored
spectrophotometrically to calculate transferred ozone concentration. Thereafter, the
samples for the analysis of AOC, ozonation by-products (glyoxylic acid, pyruvic acid and
ketomalonic acids) and SDS THMS, and SDS HAAS were taken. The residual ozone was
quenched by sparging the solution with a high purity helium gas (99.999 %). In addition,
samples also were analyzed for residual aqueous ozone concentration using indigo
method before quenching the residual ozone by purging Helium gas. pH and UV254 nm

absorbance were also monitored in the effluents of ozonation.

47

 

 

Ozone

Generator

 

 

 

 

 

Oxygen

 

 

lnfluent

Flow

Controller

Vent

l
l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T UV Spectrophotomter
Ozone Offgas
1
g l -—+~———>—
0 O O ;E
o o o o -
o O O
o , 0
° . 0 ° Effluent

 

 

 

 

Figure 4.2. Schematic diagram of continuous semi-batch ozone reactor.

48

Phase III .

The same reactor as that used in Phase 11 experiments was used. Appropriate
concentrations of additive, such as hydrogen peroxide and bicarbonate, were
continuously be added with influent water. The parameters for analysis include residual
additive concentration, TOC, residual ozone concentration, UV254, AOC, ketoacids

(glyoxylic acid, pyruvic acid and ketomalonic acids), SDS THMS. and SDS HAAS.

49

4.2. Water sources and materials

The water sources tested in this study are followings:

Lake Erie (Monroe, MI) and Lake Lansing (Haslett, MI) water. Lake Erie water samples
were taken from Monroe Water Treatment Plant intake.

The TOC values of the Lake Erie and Lake Lansing water are around 2.0 and 9.5
ppm, respectively. Thus, the experimental results could obviously show the effect of
biodegradability on ozonation over the Wide range of TOC. The compounds, used as 0H
radical scavengers, were bicarbonate and carbonate ion. The compounds, added as
initiators/promoters were phosphate ion, and hydrogen peroxide.

The characteristics of water quality on the Lake Erie and Lake Lansing are

presented in Table 4.1.

Table 4.]. Characteristics of Lake Erie and Lake Lansing water.

 

Lake Erie Lake Lansing

 

 

 

TOC 2.0 9.5
pH 7.8 8.0
Alkalinity (mg/L as CaC03) 90 140

 

 

 

 

 

4.3. Analytical method

50

Gas-Phase ozone analysis: The concentration of generated gaseous ozone was measured
at UV 254 nm with a UV/V is spectrophotometer (Milton roy Genesis-5). An extinction
coefficient of 3000 M‘lcm'l at 254nm (Hoigné et. a1, 1988) was used for calculating

ozone concentrations.

Dissolved ozone analysis: Dissolved ozone was measured with the indigo method
(Standard Method, 1995) or an amperimetric ozone analyzer (Orbisphere Laboratories,

Geneva, Switzerland).

Residual additives:

- Hydrogen Peroxide: Low concentration of hydrogen peroxide was measured
using DPD method (Hoingé, 1988), using peroxidase catalyzed oxidation of N,N-

diethyl-p-phenylendiamine (DPD).

- Bicarbonate/Carbonate : Bicarbonate and carbonate ion concentration were

determined by acid titration according to Standard Method (1995)

Natural Organic Matter (NOM): Due to the complexity of natural organic matter, the
carbon concentration of the NOM commonly represents its concentration. In order to
measure the organic carbon concentration, the samples were filtered by 0.45 pm glass-

fiber filter, and then acidified with phosphoric acid (85 %) to lower pH to less than 4. The

51

solution was purged with high purity helium gas (99.999%) for 5 minutes to avoid the
presence of inorganic carbon in the sample. The samples were injected into TOC analyzer

(OI Analytical Model 1010).

UV absorbance: UV absorbance was measured at a wavelength of 254 nm with a Milton

Roy Genesis-5 spectrophotometer using an 1 cm quartz cell.

PCBA (p—chlorobenzoic acid): PCBA was analyzed by HPLC (Perkin Elmer series 200)
with a UV—vis detector (Waters 2487) at 234 nm. Acetonitrile and D.l. Water mixture

(60:40 wt %) were used for the HPLC eluent with flow rate of 1.0 ml/min.

Assimilable Organic Carbon (AOC): The procedure for the AOC method is described
elsewhere (Background section and van der Kooji et al., 1982). A sample was heated at
60 °C for 30 min to inactivate bacteria originally present in the water. The pure cultures,
Pseudomonas fluorescens P17 and Spirilium NOX, were then inoculated into sample.
The plate count was determined after 7 days of incubation (20 oC). The yield coefficient
of Pseudomonas fluorescens P17 and Spirilium NOX was established as 4.1 x 106 CFU
P-17/ug acetate-C, and 1.2 x 107 CFU-NOX/ug acetate-C (Standard Method, 1992). For
quality control, the yields were determined in the study to be 4.4 x 106 CFU P-17/ug

acetate—C, and 1.6 x 107 CFU-NOX/ug acetate-C. Even though the yields were different

than literature values, the literature values were used in this study since the difference

was insignificant.

52

Ketoacids: Glyoxylic acid, pyruvic acid and ketomalOnic acids were monitored as
ozonation by-products. Those compounds were determined by the PFBHA-diazomethane
double derivatization procedure followed by GC/ECD equipped with DB-S column
(J&W Scientific, Folsom, CA) (Xie and Reckhow, 1992). The diazomethane was
generated in the laboratory for every analysis. The generation method for diazomethane

and a detailed procedure for ketoacids analysis are presented in Figure 4.3 and Figure 4.4.

53

 

 

Place 33 mg MNNG (1-methyl-3-nitrosoguanidine) in
inside of tube along with 0.125 mL Water

 

 

V

 

 

Place approximately 0.75 mL MTBE in Outside of tube

 

 

 

 

 

Inject 0.15 ml of 5 N NaOH (gradually)

 

 

l

Maintain in the ice batch for 45 min

 

 

 

 

 

Take Diazomethane from the outside tube

 

 

 

Figure 4.3..Procedure for generating diazomethane.

54

 

20 ml sample

 

 

 

 

 

 

 

‘ 1 ml 6mg/L PFBHA

i

 

 

 

 

Keep 45°C, 1.75 h

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20°C, 15 min
‘ 1 ml Conc. H2804 ml
% 4 ml MBTE with 1.8 (1 ,3-DBP)
V
Shake 3 min

 

 

l

 

 

Transfer lmL extract to
autosampler vial
Keep 7 min at — 10 0C

 

 

 

 

‘_

0.25 ml diazomethane

 

 

 

 

V

 

 

Keep at 4 ‘C for 15 min
Followed 20 0C for 15 min

 

 

 

 

 

 

10 mg silica gel

 

 

 

Analyze in GC-ECD

 

 

Figure 4.4. Procedure for ketoacids analysis

55

SDS THMS : The procedures for Simulated Distribution System THMs were employed.
The samples were dosed a chlorine concentration which would allow residual
concentration in the range of 0.5 to 2 mg/L after 48 hours incubation at room temperature
(Standard Method 2350). The sample was prepared according to THMFP method in
Standard Methods (1995). Extracted samples were analyzed by gas chromatography (DB-
624 column, J&W Scientific, Folsom, CA) equipped with ECD. The oven temperature
was programmed from 50 °C increasing to 150 °C with a rate of 10 °C/min. the carrier
flow was 12.0 mL/min. The injector temperature and detector temperature were 275 and

350 °C, respectively

SDS HAAS : Samples were prepared according to the HAA method (Standard Methods
,1995). A Perkin Elmer Autosystem gas chromatograph (DB-1 column, J&W Scientific,
Folsom, CA) equipped with an electron capture detector was used in these analyses. The
oven temperature was programmed to hold 20 minute at 35 °C increasing to 250 °C with
a rate of 10 °C/min. the carrier flow was 1.0 mL/min. The injector temperature and
detector temperature were 160 and 350 °C, respectively. Detail procedure for HAAS is

shown in Figure 4.5.

56

 

 

15 ml sample

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

¢ Na2803 for dechlorination
‘ 1.5g CuSO4

¢ 6 g Na2804

‘ 2 ml MTBE with (DFBP)

 

 

v

 

 

Extract 15 min

 

 

l

 

 

Transfer lml extract to
autosampler vial
Keep at — 10 ° C for 7 min

 

 

 

 

0.25 ml diazomethane

 

 

 

 

 

Keep at 4° C for 15 min
And at 20° C for 15 min

 

 

 

 

 

 

10 mg silica gel

 

 

 

Analyze using GC-ECD

 

 

Figure 4.5. Procedure for HAAs analysis.

57

Other analyses: pH was measured by pH meter. The temperature of solution was

measured by thermometer.

58

5. EFFECT OF OZONE DOSAGE

5.1 RCT values for ozone dosages.

In order to investigate the effect of ozonation pathways on the formation of both
DBPS and biodegradable compounds, it is necessary to determine the ratio of ozone/OH
radical concentrations, which can be used to confirm the importance of the two
mechanisms. The experiments were conducted by using the method developed by Elovitz
and von Guten (1997) to determine [O3]/[0H-] concentration ratio using PCBA (p-
chlorobenzoic acid) as 0H radical probe. They developed a term, RCT value, i.e., the
ratio of [OH-]/[O3].

The RCT values were determined in two different water sources: those from Lake
Lansing and Lake Erie (at the Monroe Water Treatment Plant intake). Each water source
was tested to evaluate the RCT value. As shown in Figure 5.1, the RCT value increases as
ozone dosage increased. In other word, the production of 0H radical increases with
increasing ozone dosage. This is due to the high concentration of ozone which auto-
decomposes to produce the 0H radical. Another reason for increasing 0H radicals as
increasing ozone concentration is due to the differences of NOM characteristics and/or
background water characteristics of the water sources. Since NOM plays a promotion and
inhibition role for ozone decomposition and inorganics present in water can scavenge OH
radicals, these roles affect for the production of OH radical during ozonation.

In order to compare the effect of NOM on the 0H radical and ozone reactions, the

RCT values were also determined at the same alkalinity (140 mg/L as C aCO3) to exclude

59

the effect of background OH radical scavenger (Figure 5.1). To accomplish this, the
alkalinity of Lake Erie water was increased to that of Lake Lansing water. First,
comparing the RCT values of Lake Erie water at different alkalinities indicates that the
increases in the RCT values with ozone dosage in low alkalinity (90 mg/L as CaCO3) was
greater than that in waters of high alkalinity (140 mg/L as C8CO3) (See Figure 5.1).
Thus, it can be concluded that the reaction of NOM with OH radical is rather slow
compared to the rate of reaction with bicarbonate ion and, at the high ozone dosage, the
OH radical produced from ozone decomposition reacts predominantly with bicarbonate
ions.

The alkalinity of Lake Erie water was increased so that RCT values could be
compared for Lake Erie and Lake Lansing waters at the same alkalinity (140 mg/L as
CaCO3). A higher OH radical ratio was found in the Lake Erie water than in Lake
Lansing water and no statistically significant differences (at a 95 % confidence interval)
in RCT values of Lake Lansing water are observed as the concentration of ozone
increased. The result indicates that the NOM present in Lake Erie water is a more
effective promoter of OH radicals than is that in Lake Lansing water. Thus, not only low
alkalinity but also this phenomenon results in the high Rct values measured in Lake Erie
water. Consequently, it appears that the NOM of Lake Erie reacts with OH radicals
more slowly than does the NOM present in Lake Lansing water. This suggests that the
extent to which OH radicals react with OH radical scavengers increases as ozone dosage
increases and that the NOM of Lake Erie water has the capability to promote a greater

extent of OH radical formation than does that of Lake Lansing.

60

Rct

2. 50E-07

 

  
    

 

 

 

 

 

l TLake Erie (Alk. 90 mg/L) 1
Z'OOE-WTT I Lake En'e(Alk. 140 mg/L) . - 1,, ,

l ' 1 1
1.50507 .7 __‘ LakeLans'nQABmgfl—l]
1005-07 -_ _ i

,_ , I I

5.00E-08 . .5 .4 g, -1. TT: 4—
0.00E+OO _ .

Ozone dosage (mglL)

Figure 5.1. Effect of ozone dose on RCT values (TOC: 2.0 mg/L for Lake Erie and 9.5

mg/L for Lake Lansing ).

61

5.2. Formation of ketoacids with ozone dosage

Low molecular compounds form during the reaction with ozone and NOM.
Richardson et al. (1999) and Gracia et al. (1996) identified 110 different organic
compounds, mainly carboxylic acids, aromatics, hydrocarbons, aldehydes, ketones, and
furan-carboxylic acids, that formed from the reaction of ozone with humic substances.
The concentrations of ketoacids, as a portion of the ozonation by-products, were observed
to be much higher than that of aldehydes (Griffini et al., 1999; Carlson and Amy, 1998;
Xie and Reckhow, 1992). The concentrations of the ketoacids (gloxalic acid, pyruvic acid
and ketomaloinc acid) increased with increasing ozone dose (Griffini et al., 1999; Xie
and Reckhow, 1992). The formation of aldehydes and ketoacids, easily biodegradable
compounds, has been correlated with the formation of BDOC (Griffini et al., 1999). This
result implies that the aldehydes and ketoacids would be appropriate indicators of
biologically stability. It is believed that the formation of ketoacids is favorable when the
direct ozone reaction predominates (Griffini et al., 1999).

Based on the experimental results (Figure 5.2), the formation of ketoacids
significantly increased during the ozonation of both water sources. The increase in the
concentration of ketoacids with increasing ozone concentration was greater in Lake
Lansing water which has a higher NOM concentration than Lake Eire water. Thus, it
appears that the formation of ketoacids is also influenced by NOM concentration.
Another factor affecting the formation of ketoacids was that the molecular ozone reaction
in Lake Lansing was more important than that observed in Lake Erie water. According to

the results from RCT experiments (Figure 5.1), the NOM of Lake Lansing water promoted

62

OH radical less effectively than did the NOM in Lake Erie water. Moreover, the reaction
of the OH radical with NOM in Lake Lansing water was restricted due to high alkalinity.
The formation of ketoacids in Lake Lansing water leveled off at high ozone dosages
while the formation of ketoacids in Lake Erie water continued to increase gradually even
at high concentrations of ozone. This is thought to be the result of the NOM from Lake
Lansing reacting more quickly at the early stages of ozonation (than does the NOM from
Lake Erie water). After the initial reactivity of NOM, that which remains is essentially
non-reactive with molecular ozone. 0n the other hand, the NOM from Lake Erie water,
as it appears from the RCT experiments, appears to react slowly and more continuously

with ozone. These characteristics of NOMs favor the formation of ketoacids.

63

 

300

250 -

200 -

150 -

Hg/L ketoacids

 

  
 

100 -
0 Lake Erie

0 Lake Lansing

 

 

 

50‘

 

 

 

O I T I I I I I

0 1 2 3 4 5 6 7
Ozone dosage (mg 03/mg C)

Figure 5.2. Formation of ketoacids for ozone dosages (TOC: 2.0 mg/L in Lake Erie, 9.5

mg/L in Lake Lansing, HRT: 12.5 min, Temperature: 25 0C).

64

5.3. Formation of AOC for ozone dosages.

AOC (Assimilable Organic Carbon) was measured by a method developed by

Van der Kooji et al. (1982) and modified by adding inorganic nutrients, i.e., nitrate,
phosphate, to the samples. These inorganic nutrients were added since they, especially
phosphorus, are often the limiting factor for microbial growth in treated drinking water
(Miettinen et al., 1999).
As described in the previous section, it is expected that AOC will increase with
increasing prevalence of the molecular ozone reaction over the OH radical reaction. The
formation of AOC during ozonation might be explained that molecular ozone produces
hydrophilic compounds by the reaction with aromatic ring structures. Masten et al. (2000)
found that the reaction of molecular ozone reaction with DEET(N,N-diethy-m-
toluamide), which has an aromatic ring, produced several hydrophilic as by-products.
Yao et a1. (1998) found that the reaction mechanism of pyrene with molecular ozone
attacks the double bond, causing ring cleavage of pyrene at the 4,5 (or 9,10) position.
This results in formation of various carboxylic acids and aldehydes.

The AOC determination is based on culturing bacteria that utilize groups of
Specific compounds such as carboxylic acids or carbohydrates (van der Kooji, 1985).
NOX, one of bacteria used for determination of AOC is capable of oxidizing methyl
pyruvate, acetic acid, formic acid, DL-lactic acid, a-hydroxybutyric acid, a-ketobutyric
acid, propionic acid, bromosuccinic acid and D-alanine. Of these compounds, formic

acid, a-hydroxybutyric acid and a-ketobutyric acid are not utilized by P17. It was

65

concluded that NOX is especially adept in the utilization of carboxylic acids (Charnock
and KjOnnO, 2000).

As shown in Figure 5.3, the formation of AOC increased as ozone dosage
increased. The formation of AOC in Lake Lansing water leveled off at high ozone
dosages (approximately 1 mg ozone/mg C) whereas the formation of AOC in Lake Erie
water gradually increased even at high concentrations of ozone. The formation of AOC
follows the same trends as that observed with ketoacids. Thus, it also can be concluded
that the AOC formation predominates when the molecular ozone reaction is preferred

over OH radical reactions.

66

1800

 

 

1600 - O

1400 ~

1200 ‘

 

Hg/L AOC

1000 1 0 Lake Erie
0 Lake Lansing

 

 

 

800 -
600 -
400 -

200 . °

 

 

 

Ozone dosage (mg O3,mg C)

Figure 5.3. Formation of AOC for ozone dosages (TOC: 2.0 mg/L in Lake Erie, 9.5 mg/L

in Lake Lansing, HRT: 12.5 min, Temperature: 25 0C).

67

5.4. Formation of THM and HAA for ozone dosages.

THMS (chloroform, dichlorobromoform, dibromochlroform, bromoform) and
HAAS (monochloroacetic acid, dichloroacetic acid, trichloroacetic acid,
monobromoacetic acid, and dibromoacetic acid) were measured as representative
disinfection by-products.

The effect of ozone dosage on THM formation is shown in Figure 5.4. The
formation of THMS decreased slightly as ozone dosage increased. Although the
concentration of THM decreased with increasing [OH-]/[O3] ratio, due to the reactivity of
molecular ozone with NOM, especially with aromatic rings, (Yao et al., 1998), the
reaction rate of NOM with molecular ozone increased as the ozone dosage increased.
After the initial reactivity of NOM, that which remains is essential non-reactive with
molecular ozone. NOM would, then, react with the OH radical. It is thought by Reckow
et al. (1986) that the reaction of chlorine with NOM is less reactive than molecular ozone
reaction with NOM. Thus, based on these and the RCT experiments. THM formation
appears to prevail when the molecular ozone reaction predominates (over the reaction of
NOM with OH radicals).

Reckow et al. (1986) explains the formation pathway of DBP precursors on
ozonation. Since both ozone and aqueous chlorine can act as strong electrophiles in such
systems (Decoret et al., 1984; Morris, 1975), molecular ozone would destroy nucleophilic
centers otherwise attacked by chlorine (e.g. highly activated aromatics such as dihydroxy
benzenes). In contrast, OH radicals are more non-selective than molecular ozone and

chlorine. Hence, it presumed that molecular ozone alters the moieties in the humic

68

molecule which react with chlorine to produce THMS to a far greater extent than does the
OH radical (Reckow et al 1986).

 

 

 

 

 

250 1
200 w A- ' z, ,
I
H
150 . L
5", .Lake Erie
:1 .
100 i ,. Lake Lansing
50 I4 4 h h _
. ‘9
0
0 1 2 3 4 5 6
Ozone dosage (mg Oslmg C)

Figure 5.4. Formation of THMS on ozone dosages (HRT: 12.5 min, Temperature: 25 0C)

The formation of HAAS also follows similar trends compared with THM

formation. (See Figure 5.5) This result indicates that THM and HAA formation follows

similar reaction mechanisms involving chlorine and NOM.

69

 

100 - ,, -L. .+ _ . irate Erie

I Lake Lansing ’

 

 

 

 

0 1 2 3 4 5 6
Ozone dosage (mg O3Img C)

Figure 5.5. Formation of HAAS for ozone dosages (HRT: 12.5 min, Temperature: 25 0C).

70

6. Effect of Hydraulic Retention Time (HRT)

6.1. Effect of HRT on RCT values.

In order to investigate the effect of the ratio of [O3]/[OH-] on the formation of
ketoacids, AOC, and DBPS formation potential, the concentrations of OH radicals in each
water sources were investigated in a CSTR (since the experiments for the formation of
ketoacids, AOC, and DBPS was conducted in CSTR) based on the RC]- values obtained in
batch reactor. The effect of HRTS using an ozone dosage 1mg O3/mg C was studied.
Based on the RCT values obtained in the batch studies (Figure 5.1), the OH radical
concentrations during the ozone decomposition were determined (See Table 6.1). The
calculated concentration of 0H radical and the measured concentration of ozone was
used for determine RCT value with respect to time. The values of RCT were 8.06 x10“8 i
7.73x10'll and 4.8 x 10'8 i 3.61x10'l4 in Lake Erie water and Lake Lansing Water,
respectively (See Table 6.1.). Comparing to the values presented in Table 6.1, it is
apparent that the concentrations of OH radical were greater in Lake Erie water than that
in Lake Lansing water. This supports previous results obtained when considering the
effect of ozone dosage; that is, the NOM of in Lake Erie water has a greater capability to
promote OH radical formation than does that of Lake Lansing water.

As shown in Figure 6.1 and 6.2, although the concentration of OH radicals and
molecular ozone decrease with ozonation time, the ratio of [OH-]/[O3] was almost
constant during the ozonation process due to the significantly high concentrations of

ozone compared to that of OH radical (a factor of 6 to 7).

71

Although the changes in the RCT values with varying experimental conditions were
small, the half-life of the OH radical is increased by a factor of approximately three
between the maximum and minimum values during ozonation (See Table 6.1). Thus, it
appears that, such small changes in the concentration and half-life of the OH'radical are
critical for the ozonation reaction. However, as expressed in Eq. 6.1, if the value of
k03[03][NOM] is greater than k0H[OH][NOM]. The increased concentration of OH

radical at high HRTS insignificantly affects the ozonation reaction involving NOM.

k03[03]-[N0M]2k0H[0H]-[N0M] Eq.6.1.

72

Table 6.1. Comparison of RCT values and half-life of OH radical in Lake Erie water and

Lake Lansing water.

 

(Sec):

60
120

180

[M]

0 4.2E-05
"“me-

05

05

i

l

T 20813-]

1.7413“- 3

05

 

f
240

 

360

480 3

 

300 11315-05?

11.1213!

9s

05

9.51E-‘f
42M .

1.4913- ‘5‘

l
l

vLfi Lake Erie
Time oione “ [OH] x RCTx10'7 Half-life of

1012

, [M]

0

2.00

1.67

1.40

1.20 l

1.04

0.90

A_.-

0.76

0.65

0.57

0.43

(Sec)

0 [ O
8.06 13.45E-07M
18.06 l4.14E-07
18.06 497113-03
13.66 3.771367
8.08 +6.64E—07
8.06 “ ‘ 7.69E-07
8.06 .9.04E-07
8.06 l{6512.262
8.06 . 1.20.3061

[ L

l

8.06 1.60E-06

8 0H radical

Lake Lansing

“Ozone ¥ [OH]x RCTxltTg‘Half-lire of

[M] 1012

_ , [M]
tom—04 0
8.91E-05 4.28
7.391305 7 3.55
6.15E-05 ._- 2.95
54713-05. 2.62
4731305 I 2.27
4.03E-05' 1.96
35313-05. 1.69
3.04505 I n 1.46

 

0

0H radical

- (Sec)

1

0

4.80 4.28E-05

47—

4.80
4.80
4.80
4.80

4.80

4.80

4.80

]

1

l
I

*0

3.55E-05
2.95E-05
—.62E-05
2.27E-05

1.96E-05

1 .46E-05

 

 

According to previous research (Westerhoff et al. 1999), the reaction rate of OH

radical with NOM was determined as 3.6 x 108 L (mol C)" 3'1 (assuming a mass of 12 g

C per mole C) which corresponds to 6.0x104 3", and 3.0x105 s'l for Lake Erie water and

Lake Lansing water, respectively. The concentration of OH radicals was approximately

73

10'12 M in both Lake Erie water and Lake Lansing water. In contrast, the concentration of
ozone was 10'3 M. Thus, since the reaction rate of ozone with NOM is greater than 10'3
s", such a small changes in RCT values during ozonation time does not affect the reaction

rate of ozonation with NOM.

 

4.50E-05 . - v A- _ . 9.00E-08
4005-05 I . 8.00E-08
3.50505 1 [+Ozone . 7.00E-08
3.00E-05 . +[OH]x10E6 . 6.00E-08

g 2.50E-05 . . 5.00E-08 m

‘5’ 2.00E-05 ‘ . 4.00E-08 3

0 1.50505 l 3.00E-08
1.00E-05 1 ._ 2.00E-08
5.00E-06\E \1.00E-08
0.00E+00\ . . . \0.00E+00

012345678910

Time (min)

Figure 6.1. The effect of HRTS on Ozone, OH radical concentration and RCT values in

Lake Erie water (Ozone dosage: 1mg O3/mg C, Temperature: 12.5).

74

1.20E-04

1 .00E-04

8.00E-05

6.00E-05

Conc. [M]

4.00E-05

2.00E-05

0.00E+00 ‘

 

\
N

 

 

 

 

6.00E-08
t ‘ 5.00E-08
A—f——A——A
LI— Ozone 4.00E-08
l-o—OH radical x 10E7
*+RCT
_ 3.00E-08
2.00E-08
1.00E-08
\
x
4 0.00E+00
3 4 5 6 7 8
Time (min)

:1

0
.q

Figure 6.2. The effect of HRTS on Ozone, OH radical concentration and RCT values in

Lake Lansing water (Ozone dosage: 1 mg 03/mg C, Temperature 25 0C).

75

6.2. Formation of ketoacids for HRTS

As shown in Figure 6.1 and 6.2, the ratios of OH radical concentration to
molecular ozone concentration (RCT) were constant for the HRTS studied. However, as
shown in Figure 6.3, the formation of ketoacids in both water sources increased
unexpectedly as the HRT in CSTR increased. It is thought that the formation of ketoacids
depends on reaction time as well as RCT value because the reaction of NOM with ozone
to form ketoacids would be influenced by not only the concentration of ozone but also
reaction time.

Although the RCT values are constant with HRTS, the concentration of OH radical
follows the trends for ozone concentration. The rate of formation of ketoacids in Lake
Erie water was suppressed at short retention times due to high concentration of OH
radical or the destruction of ketoacids by the OH radical. On the contrary, in Lake
Lansing water, the concentration of ketoacids increased linearly with HRT increased.
Thus, the high concentration of the OH radical present at short HRTS may play a role in
suppressing the formation of ketoacids.

Consequently, high concentrations of 0H radical at short HRTS results in a
suppression of the formation of ketoacids in Lake Erie water. The reason that the
concentration of ketocids in Lake Erie presents comparative low level even at low
concentration of 0H (at long retention time) is the low concentration of NOM in Lake

Eric.

76

[lg/L Ketoacids

 

  
  
 

400 - i

350 i A

300

250 *- * -5 -- ~~ ]. Lake Erie

200 - 2 Lake Lansing ; _,

 

 

 

150
O
100 - -
50 z
0 . . , ,
0 5 10 15 20 25 30
HRT (min)

Figure 6.3. Formation of ketoacids for HRTS (Ozone dosage: 1 mg O3/mg C,

Temperature: 25 0C).

77

6.3 Formation of AOC for HRTS.

Based upon consistency of the RCT values with HRT, the concentration of AOC
was expected to be constant with HRTS. As shown in Figure 6.4, the changes in AOC
concentration with HRTS inversely followed trend in OH radical concentration,
suggesting that the formation of AOC dependent on the RCT values rather than the HRTS.
The steady state AOC concentrations can be compared obtained in studying the effect of
AOC with that obtained when investigating the effect of ozone dosage. In Lake Erie
water, AOC concentrations of 400 [Lg/L could be achieved at HRT of 5 minute. However,
at ozone dosage, less than 1 mg O3/mg C, concentrations of approximately 200 ug/L
were produced. In Lake Lansing water, AOC concentrations of 1600 [lg/L could be
achieved at HRT of 8 minute. However, at ozone dosage, less than 2 mg O3/mg C,

concentrations of approximately 1600 ug/L were produced.

78

2000

 

1800 *

 

1600 -

1400 -

 

. Lake Erie

1200 - 0 Lake Lansing

 

 

 

ligIL AOC

1000 r
800 r
500 E
400 r

200 *

 

 

HRT (min)

Figure 6.4. Formation of AOC for HRTS (Ozone dosage: 1 mg O3/mg C. Temperature: 25

0C).

79

6.4 Formation of THMS and HAAS for HRTS.

The formation of THMS and HAAS was investigated for HRTS (See Figure 6.5
and 6.6). As HRT was increased, the concentrations of THMS decreased in Lake Lansing
water while the concentration of THM in Lake Erie water remained essentially constant
with increasing HRT. The changes in concentration of HAAS with HRT were statistically
insignificant at a 95 % confidence interval. It appears that the formation of THM was
independent of HRTS only in Lake Erie water. This indicates that the formation of THM
precursors is influenced by the ratio of the concentration of OH radical and molecular
ozone rather than reaction time. However, the formation of THM (except in Lake
Lansing) and HAA was significantly influenced by RCT ratios ([OH-]/[O3]) which were

constant during the retention times.

80

250

 

l . lake Erie i

200 W+Lake Lansing l

 
 

 

 

 

 

,, 150
s

I

I'-

=1

3’ 100 . 1 L L

O O
50 l A 2 . _ F
0 , .
0 5 10 15 20 25 30

HRT (min)

Figure 6.5. Formation of THMS for HRTS (Ozone dosage: 1 mg O3/mg C, Temperature:

25 °C).

81

 

 

 

 

 

 

 

 

 

 

250 ,
l . Lake Erie [
200 [h , g J+Lake Lansing'~ 7
+ + H
a 150 , 2 ,, 1 ,1
3
d
g 100 A
50 z 1 L
4’47
0‘ - f o
0 5 10 15 20 25 30
HRT (min)

Figure 6.6. Formation of HAAs for HRTS (Ozone dosage: 1 mg 03/mg C, Temperature:

25 °C).

82

7. EFFECT OF TEMPERATURE

7.1. RCT values for temperatures

According to the Arrhenius law, elementary reaction rate increased with
increasing temperature. Sotelo et al. (1987) showed that the ozone decomposition rate
increased with increasing pH and temperature. The ozone initiation rate was negligible at
pH below 3 at temperatures ranging from 10 - 40 °C.

Based on the elementary reaction of ozone decomposition, the OH radical

concentration at steady state condition can be expressed by following equation (Eq. 7.1).

2(k,,,,,-10.l).10H‘l.. + (k6,,nol6103lth0M]
(kllC03.0H )T [HC03 ] + (klN'OALO/l )T [NOM]

 

[OH]... = Eq. 7.1.

The reaction rates that describe the initiation of ozone decomposition and the
production of OH radicals are found in numerator in this equation will increase as
temperature increases. The OH radical scavenger reaction rate (denominator of the
equation) will also increase. Moreover, the rate of reaction of molecular ozone reaction
with NOM also increases with increasing temperature.

When the Eq. 7.1 is divided by [03], it can be expressed as RCT term as Eq. 7.2.

83

[0H]SS _ 2(k .. )rlOH‘lSS +(k03,.~'().w)T[N0Ml

03 ,OH

[03] _ (kl/CO3.OII)T[HCO3]+(kh'ouxullrlNOMI

 

Eq. 7.2.

The activation energy and frequency factor of the reaction of bicarbonate with the OH
radical was determined by Buxton et a1 (1988). Although the activation energy for the
reaction of ozone with OH' was determined by Hewes and Davison (1971), the frequency
factor was not determined. However, since the rate constant of ozone with OH' at 20 0C
was determined (Stahelin and Hoigné, 1982), the reaction rate of ozone with hydroxide
ion at various temperatures and and the frequency factor can be calculated (See Table 7.1.
and Appendix). Since the parameters other than the rate constant for both ozone with
NOM and OH radical with NOM have been determined, the RCT value can be predicted if

the values of k03,N0M and kNOMpu at the different temperatures are known.

Table 7.1. Activation energy and frequency factor of bicarbonate and ozone

 

 

 

 

 

 

 

Activation energy (kJ mole'l) log A
BicarbonateT 21.2 12.8
Ozone” 70.5 72. 1

 

Reference: TBuxton et al. 1988,

i‘lHewes and Davison, 1971

84

Based upon the literature values, the RCT value was calculated at the temperatures
at the various temperatures and compared to the RCT values obtained in this study. Due to
two unknown parameters (ko3.NOM. kNOMQH), the calculation was attempted with the
equation which is eliminated k 033002,] term. Calculation of the rate constant of NOM with
OH radical always resulted in negative values, (See Appendix) which increased as
temperature increased. Thus, it appears that the role of NOM as a promoter of molecular
ozone decomposition is more significant for OH radical production rather than is the role
of NOM as an OH radical scavenger. Comparing the activation energy for the reactions
occurring during ozonation in the water sources, temperature had the greatest effect on
the rate constant of NOM with OH radical and on that for NOM with ozone were the
most significantly increased as temperature increased (See Appendix).As a result, the OH
radical production rate can be calculated if the characteristics (rate of promotion and
inhibition) of the reaction of NOM with ozone and OH radical can be determined
experimentally.

The experiments for RCT were conducted in the batch reactor at temperature
ranging from 5 to 35 0C (See Figure 7.1). Each experiment was maintained with constant

temperature.

85

1 .40E-07

 

 
 
 
 
 
  

 

 

 

l 775* ,
1205-07 9 Lake Erie ,
f I Lake Lansing ,
1.00E-07
8.00E-08 #4 -
g: y = 3E-09x + 215-09
6.00E-08
400508 Y = 2E-09x + 1E-08
2.00E-08 e
0.00E+00 ,

Temperature (°C)

Figure 7.1. The effect of temperature on RCT values (Ozone dosage: 1 mg 03/mg C).

During ozonation, the OH radical concentration increased as temperature
increased due to the acceleration of ozone decomposition with temperature (See
Figurel8). This result indicates that the denominator term of Eq. 7.2 does not increase as
much as the numerator of Eq. 7.2 increases. As shown in Figure 7.1, the RCT values for
both water sources increased linearly as temperature increased. The increase of RCT value
of Lake Erie was greater than that of Lake Lansing. It is thought that either the activation
energy of the OH radical with NOM of Lake Erie is less than that of Lake Lansing or that
the promotion of the OH radical by NOM of Lake Lansing is less than that of Lake Erie.

7.2. Formation of ketoacids for temperatures.

86

Based on the RCT values obtained from the batch reactor, experiments to determine
the formation of ketoacids, AOC and DBPS were conducted in a CSTR at temperatures
ranging from 5 to 35 °C at a fixed ozone dosage of 1 mg O3/mg C and a HRT of 12.5
minute.

As shown in Figure 7.2, the formation of ketoacids appears to increase as temperature
increased. Comparing the formation of ketoacids with RCT values, the formation of
ketoacids increased as RCT values increased (See Figure 7.1 and Figure 7.2). The increase
of the concentration of ketoacids follows the same order of magnitude increase with the
increase of RCT values. This result contradicts the hypothesis that AOC is formed
preferentially by molecular ozone over the OH radicals, the results show different trend.

The result indicates that the ozone molecule dose not always drive the formation of

ketoacids.

87

119/L Ketoaclds

 

  

 

 

 

 

 

350 ~
300 i__ T T T -
250 _. [oLake Erie l 77
I Lake Lansing]

200 -~-~~ 4‘“ Th— ~—~~ ,
150 —~-~ e ,
100 A 1-, z. e e ~ e e

50 T T T T T T ;i—-—/

F 4—
0 a .
0 10 20 30 40

Temperature (°C)

Figure 7.2. Formation of ketoacids for temperature (Ozone dosage: 1 mg O3/mg C, HRT:

12.5 min).

88

7.3. Formation of AOC for temperatures.

The concentration of AOC did not change as temperature increases as shown in
Figure 7.3. Moreover, the formation of AOC in Lake Lansing decreased as temperature
increased within the range 20 —35 0C in statistically at 95 % confidence interval. The
formation of AOC in Lake Lansing water was unrelated to RCT values. Although at high
temperatures the rapid decomposition ozone results in the predominance of OH radical
reactions, the complexity of ozone reactions with NOM results in less significant changes
than expected.

As temperatures increased, ozone decomposition is accelerated so that higher
concentrations of OH radical are expected, and were observed (Figure 7.1.). However,
the reaction rates of the OH radical with both NOM and OH radical scavengers are also
increased. Therefore, there may be competition reactions of OH radical with either NOM
or OH radical scavengers. Regardless of their concentrations, each reaction obviously
affects the overall ozone reaction since the specific increases in the reaction rate constant

with respect to temperature is different.

89

 

— +Monroe
300 « LL, , —~ [fl—Lake Lansing, ,

lug/L AOC
8
O
O

 

 

, ,1 I v 2 !

0 1 0 20 30 40

Temperature (°C)

Figure 7.3. Formation of AOC for temperature (Ozone dosage: 1 mg O3/mg C, HRT:

12.5).

90

7.4. Formation of THM and HAA for temperatures.

The effects of temperature on the formation of THMS and HAAS are presented in
Figures 7.4 and 7.5. The formation of THM and HAAS were independent of RCT values,
which increased as temperature increased. The reason is that as temperature changes,
even though RCT values change, the rate constant of molecular ozone with NOM and of
OH radical with NOM also change. Hence, the reaction activity of NOM with molecular
ozone and OH radical is difficult to compare at the each temperature. It may be possible
to explain the formation of AOC at different temperature with same reason. The results

also Show that THM and HAAS can not be controlled by changing the water temperature.

200
180.
160 - - — »
140 4 -4 .4 a, 4 — 4 4 e e 4
120 ,, e

100. - - - , - f _ l—oT—LfiakeErie l
80 z - z 5 , , _ , ,, I+LakeLansing 2

 

 

uglL THMs

 

40--. ’T + , ,, % T’
20—— ———————————— ———

 

 

 

0 10 20 30 40

Temperature (°C)

Figure 7.4. Formation of THMS for temperature (Ozone dosage: 1 mg O3/mg C, HRT:

12.5).

91

ng/L HAAS

180
160

140 .
120 ,
100 -
80 ,

60
40
20

 

 

1+ LakeErie I

il+Lake Lansing ’

 

 

 

 

20 30
Temperature (°C)

40

Figure 7.5 Formation of HAAS for temperature (Ozone dosage: 1 mg 03/mg C, HRT:

12.5).

92

8. EFFECT OF pH

8.1. RCT values for pH

Varying the pH affects the ozone decomposition rate not only because the
hydroxide ion concentration changes but also because the OH radical scavenging rate is
altered since the concentrations of bicarbonate, carbonate ions and other radical species
(ionic scavenging) are pH dependent. As such, the OH radical scavenging rate changes
with pH even if the alkalinity is held constant. In order to maintain the same OH radical
scavenging rate at the different pHs (from 2 to 10) to investigate the effect of pH on
ozonation, the OH radical scavenging rate was calculated considering the alkalinity of
raw water samples, the pH and calibrated by adding bicarbonate and phosphate ions,
which are well characterized OH radical scavengers at each pH (Eq. 8.1). The pH was fist
controlled within the range from 2 to 10 by adding sodium hydroxide and hydrochloric
acid. The OH radical scavenging species were, then, added to water samples to achieve
the desired OH radical scavenging rate. Based on the Eq. 8.1, the total scavenger
capacities were set as 2.34 x 105 s", 3.38 x 105 s'1 in Lake Erie and Lake Lansing water,

respectively.

93

{Scavenging Capacity} = kH2C03 ,OH [H2C03 l +km'03 —,()Il [HC03_ l + 12.03210” [C0321
+ kll3P04 .OH [H3P04 ] 1' kH2P04 -.0n [H2P04-]
kHPO42T ,0" [HP042- ] + kP043‘ .011 [P0431
Eq. 8.1.

Where, kH2C03,0H = 1.0 MJSJ, l([[C03-,()][ = 8.5 X10.6 MJSJ, kC032-,()][ = 3.9 X 10‘8 MJSJ,
kH3PO4,0H = 2-7 X 10-6 M'IS'I. k112PO4-,OH = 2-0 X 10'4 M'lS-l.

KHPO42-,OH : 1.5)(10'5 M-lS-I, 1904330“ = 1.0 X 10-7 M-1S4.

The RCT values were investigated at different pH values in solutions having the
same scavenging capacity. As shown in Figure 8.1, The RCT values increased as pH
increased in both water sources. However, the RCT values in these results were low
compared to the results obtained under uncontrolled (no added scavenging capacity)
conditions. This shows that the production of the OH radical is restricted and OH radicals
are consumed by the scavenging species added.

The larger RCT values in Lake Erie compared to Lake Lansing is due to the lower
scavenging capacity (2.34 x 105 s") in Lake Erie compared to that in Lake Lansing (3.38
x 105 s"). In order to compare the OH radical scavenging rate by scavenging species and
NOM, the rate of OH radical reacting with NOM also calculated (Eq. 8.2). The rate
constant of the OH radical and the molecular weight of NOM were used obtained from
previous research (kNOM.on z 3.6 x 108 mol C‘1 s", 12 g C per mol C) (Westerhoff et al.,
1999). In both Lake Erie and Lake Lansing waters, the ratios of OH radicals reacting
with NOM and scavengers were calculated to be 0.77, 1.28, respectively. The results

indicate that scavengers were able to compete with NOM for OH radicals. Therefore. the

94

effect of NOM on the ozone reactions is less that what would otherwise be expected,

especially in Lake Erie water.

[Scavenging Capacity] > kN()M,0”[NOM]

(kNOMpH z 3.6 x 108 mol C" s", 12 g C per mol C) (Westerhoff et al., 1999) Eq. 8.2.

As pH increased, the formation of the OH radical increased (See Figure 8.1). This
is not unexpected since the ozone decomposition rate at low pH is significantly lower
than that at high pH, due to the restricted initiation step of ozone decomposition at low

pH.

 

1.80E-08 _ _f g - - -
1.60E-08 e +4 - 1, * -l—O—Lake Erie

1.40E-08 - - e - a- 1—I—1ake Lansing

1.20E-08
1.00E-08
8.00E-09 ~
6.00E-09 —
4.00E-09 ~
2.00E-09

0.00E+00

RCT

 

 

 

 

Figure 8.1. The effect of pH on RCT values (Ozone dosage: 1 mg O3/mg C, Temperature:

25 °C).

95

8.2. Formation of ketoacids for pH

As shown in Figure 8.2, the concentration of ketoacids decreased as pH increased
for both water sources. This result also indicates that the formation of ketoacids occurs
preferentially by a molecular ozone reaction as compared to an 0H radical mediated
reaction. This is consistent with the original hypothesis. An increase in pH resulted in a
greater in the concentration of ketoacids in Lake Lansing water as compared to that in
Lake Erie. The lower reaction rate of OH radical with NOM from Lake Erie water as
compared to that with Lake Lansing may be the cause of this phenomenon.

Comparing the concentrations of ketoacids produced in the uncontrolled
conditions (no additional buffer), it was observed that the concentration of ketoacids in
buffered water was greater (See Figure 5.2 and 8.3). This was due to the greater
concentration of ozone that is maintained in waters having a greater scavenging capacity
since the reaction of the scavenging species with 0H radical consumes OH radical
without producing superoxide anion (02-), thereby slowing down the decomposition of
ozone (Langlais et al. 1991). The decomposition of ozone in the buffered and unbuffered
water is presented in the Figure 8.3. Consequently, the concentration of ketoacids was

increased at the high pH under the presence of molecular ozone dominant.

96

  
  

- + Lake-Erie ‘

l + Lake Lansing

TT 'TT' 1

 

 

Ketocaids [pg/L]

Figure 8.2. Formation of ketoacids for pH (Ozone dosage: 1 mg 03/mg C, HRT: 12.5

min, Temperature: 25 0C).

97

-0.2

-0.4

In [Cal/[Oslo

-1.2

-1.4

Figure 8.3. Decomposition of ozone in the raw water and buffered water (Lake Erie

-0.5 -
-0.8 ~ -

 

T

 

  
   

 

. concise

 

l

unbuffered, fl

 

 

water, Ozone dosage: 2mg/L, pH 8, Temperature 25 0C).

98

8.3. Formation of AOC for pH

Prior to the inoculation of bacteria into the samples for AOC determination, the
pH of samples was adjusted to neutral (pH=7) using sodium hydroxide and/or
hydrochloric acid. In order to minimize the volume of sodium hydroxide and
hydrochloric acid, the concentration of the solutions used as 4 M. The formation of AOC
is expected to decrease as pH increased since the reaction of molecular ozone with NOM
is more favorable than is the OH radical for the formation of AOC. The concentrations of
AOC in both Lake Erie and Lake Lansing water decreased at high pH. The
concentrations of AOC produced in Lake Lansing were higher than that in Lake Erie
water. This was due to not only higher concentration of NOM but also the low

concentrations of OH radical in Lake Lansing (See Figure 8.4).

99

119/L A00

2000 -
1800 i ,. -277 - m

 

 

._
1600 i 0 GT \
1400l - - . i r .
1200 i , - A-
1000 . -
+Lake Erie
800 . .
l —l— Lake Lansing
6001- e —- . ti, .
400 i - ~ ,
200 e ,
o . - i 1 L
0 2 4 6 8 10 12

Figure 8.4. Formation of AOC for pH (Ozone dosage: 1 mg Og/mg C, HRT: 12.5 min,

Temperature: 25 0C).

100

8.4. Formation of THMS and HAAS for pH

The formation of THMS and HAAS following ozonation at pH from 2 to 10 is
shown in Figures 8.5 and 8.6. Prior to adding chlorine to the ozonated samples, the pH
values of the samples were adjusted to neutral (pH=7) using sodium hydroxide and/or
hydrochloric acid. After that, the phosphate buffer was added into the samples as usual
chlorination procedure.

The concentrations of THMS in both water sources decreased as pH increased.
This can be explained by the presence of higher concentrations of molecular ozone at
lower pH. The concentration of THMS was greatest at the lowest. This conflicts with our
hypothesis that the predominance of the molecular ozone reaction would result in a
decrease the formation of DBPS. This result may also be due to the significant role the
scavengers play a significant role in the ozone reaction. The buffering was controlled
with bicarbonate at high pH and phosphate at low pH since the bicarbonate ion present as
H2CO3 at low pH (pH <6.3) has little buffering capacity. The bicarbonate and carbonate
ion produce the secondary oxidants, the bicarbonate radical and carbonate radical (HCO3-
/C03°'), which can oxidize some inorganic and organic compounds. Although the fate of
carbonate radicals in drinking water is not entirely determined, carbonate radicals have
been found to react through self-reaction (Eq. 7.1) or with NOM (Eq. 7.2) (Acero and

von Gunten, 2000).

C03“ + C03" —. C02 + C042' k=2x107 M's" Eq. 7.1.

CO.‘ + NOM —> C032' + Products k=leO7 M's" Eq. 7.2.

101

Therefore, at the high concentration of bicarbonate and carbonate that exist at the higher
pH values, the concentration of carbonate radicals would also increase, significantly
affecting the reaction of ozone with NOM.

Although the concentration of HAAS in Lake Erie did not change with pH, the
concentration of HAAS in Lake Lansing decreased as pH increased. These trends are

similar to those for the THMS

102

 

140

  
  

 

 

 

120 ST S * ”*****—**TT’*******T 0 Lake Erie l
100 “T T 1 .Lake Lansing \-
a .0 __ L---
E
=1 60 A z
2 40*
20 t - g
0 I r
0 2 4 6 8 10

12

Figure 8.5. Formation of THMS for pH (Ozone dosage: 1 mg O3/mg C, HRT: 12.5 min,

Temperature: 25 °C).

103

 

ugILHAAs
88888888

 

 

 

 

 

12

Figure 8.6. Formation of HAAS for pH (Ozone dosage: 1 mg O3/mg C, HRT: 12.5 min,

Temperature: 25 °C).

104

9. EEFFECT OF BICARBONATE

9.1. RCT values in solutions containing bicarbonate ions

Scavenging reactions consume OH radicals without producing superoxide ions.
Although there are various additives that could have been used as scavengers, nontoxic
inorganic additives were used for this study so as to avoid problems associated with the
influence of the organic scavenging compounds on AOC and TOC analysis. Since
residual organic additives will affect the measurement of TOC measurement, these
compounds were avoided. The organic additives and their ozonation by-products may
also influence the formation of AOC, providing another reason for using inorganic
scavengers.

Staehelin and Hoigné (1982) found that carbonate and bicarbonate ions have the
ability to stabilize molecular ozone. This has been confirmed by many other researchers
(Yuteri and Gurol, 1988; Reckhow et al., 1986; Fomi et al., 1982). The presence of the
OH radicals, bicarbonate and carbonate ions, which are ubiquitous in most natural waters,
will result in relatively low OH radical concentration in these waters. This is especially
true when the bicarbonate ion is dissociated into carbonate ion at pH > 10.3 (pKa = 10.3)
(Snoeyink and Jenkins, 1980). The reaction rate of carbonate with OH radical is greater
than that of bicarbonate. As such, bicarbonate and carbonate ions may serve to consume
OH radicals, thereby stabilizing molecular ozone. The bicarbonate ion inhibits the radical
chain reaction, hence increasing the concentration of molecular ozone. As such, the

presence of radical scavengers could result in the predominance of reactions involving

105

the ozone molecule, resulting in preferential reactions involving the ozone molecule and
the aromatic moieties of fulvic acid, a component of humic substance.

According to previous research conducted on the effect of alkalinity on ozonation
(alkalinity ranging from 0 to 1.5 mM) in Lake Zfirich water (Elovitz et al. 2000), RCT
values were seen to decreased five-fold when the alkalinity was increased from 0 to 1.5
meq/L. The concentration range of bicarbonate applied in our study was to 10 mM. As
shown in Figure 9.1, the RCT values in Lake Erie decreased as the concentration of
bicarbonate increased while the RCT values in Lake Lansing increased as bicarbonate
increased. However, the RCT values in Lake Lansing water remained relatively constant
at bicarbonate concentrations greater than 4 mM. The addition of bicarbonate (to 10 mM)
resulted in an increase in the pH values in Lake Erie water from 7.8 to 8.2, and from 8.0
to 8.9 in Lake Lansing water. The pH increase is expected to lead to higher
concentrations of 0H radicals due to an increase in the ozone decomposition rate.
However, the change of RCT values in Lake Erie was 75 % and in Lake Lansing was 37
% comparing at bicarbonate concentration of 0 and 10 mM. Increase in the pH led to
higher concentration of 0H radicals. As a result, the impact of the pH which resulted in
an increase in the concentration of OH radicals and the addition of bicarbonate which

decreased the concentration of OH radicals counteracted one another.

106

1.40E-07 i - -7- i . ,-

 

1.20E-07 .A -— -A ,,7 if. , 7, 3
1005-07 . ——.. _.. -_- ___- -_ _
J I. ‘
8.00E-08 --- 4+ 4 2-- -_ . z ,
é LI -——I” 7‘? ,7, 77
6.005-08 f - - - l 0 Lake Erie '

   

' :7 Like Lansing !

 
 

4.005-08 .
2.00E-08~1 -

0,005+00 .— -- _. -- - . _. __
0 2 4 6 8 10 12

Bicarbonate (mM)

Figure 9.1. The effect of bicarbonate on RCT values (Ozone dosage: 1mg O3/mg C,

Temperature: 25 0C).

107

9.2. Formation of ketoacids in solutions containing bicarbonate ion

The effect of bicarbonate addition on the concentration of ketoacids is shown in
Figure 9.2. While increasing the bicarbonate concentration in Lake Erie water had no
effect on the concentration of ketoacids, the ketoacids concentration in Lake Lansing did
decrease with increasing bicarbonate concentration in Lake Lansing water. Although
bicarbonate and carbonate ions are effective OH radical scavengers, they produce the
secondary oxidants, the bicarbonate radical and carbonate radical (HCO3- /CO3-'), which
can oxidize some inorganic and organic compounds. Although the fate of carbonate
radicals in drinking water is not entirely determined, carbonate radicals have been found
to react through self-reaction (Eq. 9.1) or with NOM (Eq. 9.2) (Acero and von Gunten,

2000)

coy + coy —» (:02 + co.” k=2><107 M‘s" Eq. 9.1.

coy + NOM —> CO32' + Products k=1x107 M's" Eq. 9.2.

Therefore, as the concentration of bicarbonate increases, the concentration of carbonate
radicals would also increase, significantly affecting the reaction of ozone with NOM.
This may cause a decrease in the concentrations of ketoacids at the high concentrations of
bicarbonate since the radical species can directly attack NOM to inhibit the formation of
ketoacids.

The results regarding the effect of bicarbonate on the formation of ketoacids have

been plotted along with the concentration of carbonate ion. (See Figure 9.3).

108

Considering the amount of bicarbonate added and pH in water, the concentration of
carbonate ion in water was calculated (Snoeyink and Jenkins, 1980). The total alkalinity,

including added bicarbonate, can be expressed as Eq. 9.3

 

K
[Total alkalinity] =C “-03 (a, + 2612) + [HPZ] — [H +] Eq. 9.3.

The ionization fractions obtained from the exact equations for diprotic acid are

[H’T ___ [H503]

 

 

 

a :
0 E CT,A Eq. 9.4.
[H+]CKa,l [HC03—]
a, = E = ___—C Eq. 9.5.
m
CKaJ CKa.2 [C032-]
0‘2 = = —— Eq. 9.6.

E C m

Where E = [H’L]2 + [H+]CKa.l +CKa,1CKa.2

For bicarbonate and carbonate ions, K3,] = 10'“, Kal =10'l0‘3.

Using the Eq. 9.3 through 9.6 and constants (KM, K32), C1103 (concentration of
total inorganic carbon) can be obtained from Eq. 9.3. Thus, the concentration of

bicarbonate ion and carbonate ion can be calculated based on the Eq. 9.7 and 9.8.

[HCOfl = CT.C03 ~ 0t: Eq. 9.7.

109

[C032] 2 C1103 . (12 Eq. 9.8.

The calculated values of carbonated concentration in Lake Erie and Lake Lansing are

presented in Table 9.1.

Table 9.1. Concentrations of carbonate ion including added bicarbonate in Lake Erie, and

 

 

 

 

 

 

 

 

 

Lake Lansing.
Added amount pH
[C032] in Lake
of bicarbonate (Lake Erie and [CO32'] in Lake Eric (M)
Lansing (M)
ion (mM) Lake Lansing)
O 7.8, 8.0 5.692E-06 1.403E-05
1 7.9, 8.1 885415-06 190415-05
2 8.0, 8.3 1.201E-05 3.028E-05
4 8.0, 8.6 2.309E-05 8.800E-05
6 8.1, 8.7 3.105E-05 1.394E-O4
8 8.2, 8.7 4.91 lE-OS 2.712E—04
10 8.2, 8.9 5.914E-05 5.095E-04

 

 

 

 

 

The concentrations of ketoacids on the effect of bicarbonate were plotted according to the

concentration of carbonate ion in Figure 9.3.

110

 

250 ~

‘ +Lake Erie

200 [I . I f
,. . a Lake Lansing

150 ~ -

100 - -- - -

PgIL Ketoacids

 

 

 

 

Bicarbonate [mM]

Figure 9.2. Formation of ketoacids on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT:

12.5 min, Temperature: 25 oC).

As shown in Figure 9.2, the concentration of ketoacids in Lake Lansing increased
until approximately 6.0 x 10'5 M of carbonate concentration. The concentrations of
ketoacids were decreased at the carbonate ion concentration greater than approximately
6.0 x 10'5 M. (See Figure 9.2.) Therefore, due to increases of the carbonate radical
concentration, the formation of ketoacids seems to be restricted otherwise increased by

prevalent molecular ozone concentration.
Comparing the results presented in Figure 9.2 with those in Figure 9.3, the

concentration of ketoacids in Lake Erie did not change significantly while the

111

concentration of Lake Lansing changed over the bicarbonate range from 0 to 10 mM
(Figure 9.2). However, when considering the concentration of carbonate ion as shown in
Figure 9.3, it is apparent that the carbonate ion concentration in Lake Erie water is
significantly smaller than that in Lake Lansing water, even though the addition of
bicarbonate was equivalent (0 to 10 mM of bicarbonate) in both waters. Thus, the effect

of carbonate radical would be expected to be less significant in Lake Erie water.

 

 

   

 

 

 

 

250 j

200 L L z - - - -
m l
13 l
'8 150 :- . . -
3 l #l I
2.6 s
—l 100 A E ’ " '5 5 ’5' ’ ,:_m;-;: T_ ’_._
E: i 1+Lake Erie

50 A - - - - -- ¥I+LakeLansing~

o P i . . . .
0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006
[0032‘]

Figure 9.3. Formation of ketoacids on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT:

12.5 min, Temperature: 25 0C).

112

9.3. Formation of AOC in solutions containing bicarbonate ions

The addition of bicarbonate was expected to increase AOC formation since as an
OH radical scavenger its presence would result in molecular ozone reaction
predominating. As shown in Figure 9.1, although the RCT values measured in Lake Erie
water decreased as the concentration of bicarbonate increased, the concentration of AOC
was essentially unchanged. This is consistent with the previous results, that is, the rate of
reaction of the OH radical with NOM from Lake Erie is slow. In Lake Lansing water, the
concentration of AOC increased as bicarbonate concentration increased (within the range
from 0 to 4 mM). At bicarbonate concentrations greater than 4 mM, the concentration of
AOC decreased with increasing bicarbonate concentration. This may be due to the
reaction of radical species (bicarbonate radicals and carbonate radicals) produced as
secondary oxidants. These radical species also seem to affect the concentration of AOC
differently that the OH radical does. The relationship between the concentration of AOC
and the concentration of carbonate ion is shown in Figure 9.5. As was shown in Figure
9.3 (with ketoacids), the concentration of AOC also decreased as the concentration of
carbonate ion, increased above approximately 6.0 x 10'5 M (See Figure 9.4). The radical
reaction involving carbonate and bicarbonate radicals seems to effectively react with
NOM. This indicates that higher concentrations of carbonate ion during ozonation will
promote carbonate radicals, increasing the reactivity with NOM that otherwise might

react with OH radicals.

113

 

Peg/L AOC
r— { ‘
m
x-
CD
III l
a.

1000 , if * fiJ-Ij-Lake Lansing. ,- l

 

 

Bicarbonate [mM]

Figure 9.4. The effect of bicarbonate on the formation of AOC (Ozone dosage: 1mg

O3/mg C, HRT: 12.5 min, Temperature: 25 0C).

114

2500

 

 

 

 

2000 .
0 3
(31500 ~» ;
< l
.1 : z. , ,
31000 g * i—o—Lake Erie ' *
E ; I Lake Lansing
500 E a ‘4 ‘5 * z: '
|W
O . § .

 

0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006
[003’]

Figure 9.5. Formation of AOC on bicarboante (Ozone dosage: 1mg O3/mg C, HRT: 12.5

min, Temperature: 25 0C).

115

9.4. Formation of THMS and HAAs in solutions containing bicarbonate ions

The effects of bicarbonate on the formation of THMS and HAAS are shown in
Figures 9.6 and 9.7. The addition of 1 mM bicarbonate to Lake Lansing water resulted in
a decrease in the THMS concentration approximately from 260 ug/L to 190 ug/L. No
further decreases were observed even at a bicarbonate concentration of 10 mM. The
concentration of THMS in Lake Erie water did not change (with in a 95 % confidential
interval) with increase in the concentration of bicarbonate ranging from O to 10 mM. This
result indicates that small amounts of bicarbonate are effective at reducing the formation
of THMS in Lake Lansing water, but that the effective of bicarbonate is maximized at
some bicarbonate concentration of 4 mM (in this case). Further work would need to be a
accomplished to determine if this concentration is consistent. Due to the slow reaction
rate of the OH radical with NOM in Lake Erie, bicarbonate is less effective for the
control of the formation of THMS and HAAs in this water source. A similar trend for the
decrease in the HAA concentration with increasing bicarbonate concentration is in Figure
9.7. This result is consistent with the findings of Reckhow et al. (1986). They observed
that comparing to no addition of bicarbonate, the concentration of THMS and HAAS
decreased from lmM of bicarbonate and no significant changes until 10 mM of
bicarbonate in extracted fulvic acid from Black Lake near Elizabethtown, NC.
Consequently, both Reckhow et al (1986) and our results supports the hypothesis that the
precursors of THMS and HAAS are preferentially formed by the reaction of OH radical

with NOM.

116

 

300

 

 

 

 

 

 

 

 

 

250 - 0 Lake Erie
0 Lake Lansing
200 —
0
en 0
E
I- 150 ~ 0
<'
c:
:L
100 -
0 i l j i 1
0 2 4 6 8 10

Bicarbonate [mM]

Figure 9.6. Formation of THMS on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT: 12.5

min, Temperature: 25 °C).

117

12

180

 

 

 

 

 

 

 

 

 

 

160 5 0 Lake Erie
C O 0 Lake Lansing
140 - O
O
a:
i 100 —
I
§’, 80 -
:1.
60 —
4O -
O l f i I i
0 2 4 6 8 10 12

Bicarbonate [mM]

Figure 9.7. Formation of HAAS on bicarbonate (Ozone dosage: 1mg O3/mg C, HRT: 12.5

min, Temperature: 25 °C).

118

10. EFFECT OF HYDROGEN PEROXIDE

10.1. RCT values for hydrogen peroxide

The combination of hydrogen peroxide with ozonation can enhance OH radical
production because the decomposition of the ozone molecule is accelerated by
superoxide ion (HOz'), producing OH radicals (Masten et al., 1997; Glaze et al., 1987;
Duguet et al. 1985; Brunet et al., 1984, Staehelin and Hoigné, 1982). However,
competition reactions occur at high pH since hydroxide ions are more effective at
initiating molecular ozone than is hydrogen peroxide (See Figure 10.1 and Table 10.1)
and at high pH OH radicals are also consumed by bicarbonate and carbonate ions. (Eq.
10.1 and 10.2) The OH radical concentration is also influenced by the composition of
water sources and types and the concentrations of OH radical scavengers. Also, OH
radical concentrations will increase with increasing hydrogen peroxide up to a certain
dosage. Beyond this dosage, the concentration of OH radical decreases due to the

recombination reaction of OH radicals. (Eq. l0.3)

119

Table 10.1. Rate constant for the HzO-O3 and Aqueous HzOz-O3 Systems (Staehelin and

Hoigné, ES& T. 1982).

 

H20 kn: < 5 x 10'6 s" kmmo. = < 4 x 10'4 M‘s" 1.010.-.-"k‘a5d = 210 i 20 M" s"
H202 k0,..202 = < 10'2 M"s"a ko3,H02-'"“asd= (5.5 a: 1.0) x 10" M‘s"
mm: 5.5 x 106[Hzoz].,.10”""“ s'”’

3For Ionic strength I: 0.05 M. bTaube, H and Bray, W.C., J. Am. Chem. Soc. 1940

 

on + HCO3' —» coy +H20 8.5 x 108 L mol" s" Eq. 10.1. (Buxton et al., 1988)

 

011 +003“2 ——) orr + my 3.9 x 109 L mol" s" Eq.]0.2. (Buxton et al., 1988)
-OH + -OH —9 H202 2k=1.1 x 10'0 L mol" 5" Eq.10.3. (Buxton et al., 1988)
05 § H202
m2
9 11
H05
03 03
03
H, H04
0
H03 2
\- OH
02

Figure 10.1. Diagram of ozone decomposition by hydrogen peroxide ions (Langlais et al.,

1990).

120

 

4.00E-07

 

 

 

3.505-07 m—e . , . .
3005-07 .
2.505-07 *9 , ~ - ,~ '
0:5 2005-07 e .
- , 9"! , / 7 ,2- _
150E 07 ,5 /' 9 Lake Erie
x l
1-005'07 3’ I/ z“ ’ " Tl r- Lake Lansing '
5.00E-08 ri/ * r " * *
0.00E+00 ~ .
0 0.2 0.4 0.6 0.8 1 1.2
[Hzozlllozl

Figure 10.2. The effect of hydrogen peroxide on RCT values (Ozone dosage: 1 mg O3/mg

C, Temperature: 25 °C).

The peroxide concentrations applied in this study were varied from 0 to 1.0
[HzOz]/[O3] (wt/wt) with a fixed ozone dosage (1 mg O3/mg C). As shown in Figure 10.2,
over the range of [HzOz]/[O3] ratio of 0.3- 0.5, the RCT values increased as peroxide
concentration increased. However, the RCT values decreased at the high peroxide dosages
greater than an [HzOz]/[O3] ratio of 0.5. The results indicate that the production of OH
radical can be maximized at a certain dosage (0.3 - 0.5) and higher concentrations of

H202 actually have a negative effect on the production of OH radicals. This is due to OH

121

 

radical scavenging effects at high concentrations of H202. These results are consistent
with that of previous researcher (Glaze et al., 1987; Duguet et al. 1985).

At a particular ozone dosage and [H202]/[O3] ratio, greater OH radical
concentrations were produced in Lake Erie water as compared to that in Lake Lansing
water. It is thought that the OH radicals in Lake Lansing water may be consumed by high
alkalinity in Lake Lansing water. Also, due to the high concentration of NOM in Lake
Lansing, the NOM plays a greater role as scavenger than as a promoter of OH radicals.
As stated previously in the section on the effect of ozone dosage, the properties of the

NOM in Lake Erie make it effective at promoting OH radical from ozone.

122

10.2. Formation of ketoacids for hydrogen peroxide

As peroxide concentration is increased, the production rate of OH radical
increased up to a certain dosage (See Figure 10.2). However, the formation of ketoacids
decreased as peroxide concentration increased (Figure 10.3). This result shows that the
reaction of molecular ozone with NOM preferentially produces ketoacids over the OH
radical reaction. Comparing the formation of ketoacids in Lake Lansing water and Lake
Erie water, the OH radical is more effective at reducing the ketoacids concentration in
Lake Lansing water. However, decreases both about 50 % increase in concentration of
OH radical more than 7 fold. This result shows good agreement with the previous
conclusion made in the section on the effect of ozone dosage that the reaction of OH

radical with NOM in Lake Erie is slower than that with Lake Lansing NOM.

123

 

14o ~ ~ , ; —+?-—Lake Eta
120222.? *—~*- f” L,“ .IeLakeLansing l ,2

HgIL Ketoacids
8
o
l
I
l
l
I
l

 

 

 

 

 

0 0.2 0.4 0.6 0.8 1 1 .2
[Ham/[03] (wt/wt)

Figure 10.3. Formation of ketoacids for hydrogen peroxide (Ozone dosage: 1 mg O3/mg

C, HRT: 12.5 min, Temperature: 25 0C).

124

10.3. Formation of AOC in the hydrogen peroxide/ozone treatment

The formation of AOC during ozone/hydrogen peroxide treatment is expected to
be hindered by the reactions involving OH radicals, which are produced from the reaction
of ozone with hydrogen peroxide. Therefore, it is predicted that low concentrations of
AOC will result during ozone/hydrogen peroxide treatment as compared to that resulting
during ozonation.

The concentration of AOC decreased as the concentration of hydrogen peroxide
increased, following a similar trend as that observed with the formation of ketoacids (See
Figures 10.3 and 10.4). This result also indicates that the formation of AOC is also
related to the reaction of molecular ozone with NOM. The concentration of AOC in Lake
Lansing at a peroxide ratio of 0.01 [H202]/[O3] (wt/wt) resulted in a higher value of AOC
as compared to the AOC concentration produced at other [H202]/[O3] ratios. Although
the experimental error ranges are small as shown in Figure 10.4. the datum apprears to be

an artifact.

125

Hg AOC/L

Figure 10.4. Formation of AOC for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,

3000

 

 

 

 

 

2500 ., I

2000 f“? ,.' Lake Erie ' _
if ‘ I— Lake Lansmg

1500 -2 a A

1000 ~ f

i‘ I r

500 a . ~ , '1” r ‘

0 0.2 0.4 0.6 0.8 1
[Hzozmoal (MM)

HRT: 12.5 min, Temperature: 25 °C).

126

1.2

10.4. Formation of THMS and HAAS for hydrogen peroxide/ozone treatment

The relationship between the concentrations of THMS and HAAS and the
peroxide/ozone ratio (0 - 1.0 [H202]/[O3] wt/wt) also was determined (Figures 10.5 and
10.6). As shown in Figure 10.5, the concentration of THMS decreased as the
concentration of peroxide was increased. Although there was no significant difference (at
95 % confidential interval) in the concentrations of THMS in Lake Eric at different
[H2O2]/[03] ratios, the concentration of THMS in Lake Lansing decreased as the
concentration of peroxide increased. These results also indicate that increased OH radical
concentrations compared to ozone only may further oxidized NOM thatwould react with
chlorine in Lake Lansing water. Reckhow et al. (1986) explained the formation pathway
of DBP precursors on ozonation. Since both ozone and aqueous chlorine can act as strong
electrophiles in such systems (Decoret et al., 1984; Morris, 1975), molecular ozone
would destroy nucleophilic centers otherwise attacked by chlorine (e.g. highly activated
aromatics such as dihydroxy benzenes). In contrast, OH radicals are more non-selective
than are molecular ozone and chlorine. Hence, it presumed that molecular ozone alters
moieties in the humic molecule which react with chlorine to produce THMs to a far
greater extent than does the OH radical (Reckow et al 1986). The results excluding that

for THMS in Lake Lansing show good agreement with this proposed mechanism.

127

250

 

 
  
 

200
;+ Lake Erie
1 I Lake Lansing I

 

 

 

 

g 150 e.
I
I'-
_l
7! 100 -- ._-. A. f - . - L L
50 ,7, L7 ,_. _ j
\
I
0 , . f .
0 0.2 0.4 0.6 0.8 1 1.2

[Hzozlllos] (Wt/Wt)

Figure 10.5. Formation of THMs for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,

HRT: 12.5 min, Temperature: 25 0C).

The formation of HAAS is shown in Figure 10.6. The formation of HAAS follows
the trend for OH radical cocentration. As the concentration of OH radical increases, the
formation of HAAS also increases. As the concentration of OH radical decreases at the
high ratio of [H2O2/O3], the formation of HAAS also decreases in Lake Lansing water and
the increases in HAAS concentrations were less significant in Lake Erie water since the
reaction rate of the OH radical with NOM in Lake Erie water is rather slow. The
concentration of THM and HAA were maximized at a [H2O2/O3] (wt/wt) of 0.1 even
though the concentration of OH radical was maximized at 0.3 [H2O2/03] (wt/wt).

According to Acero and von Guten, (2000), carbonate radical produced by OH radical

128

with bicarbonate exert to promote ozone decomposition. The carbonate radical
recombination takes place to produce CO2 and C042' (k = 2x107 M‘ls'l). Thus, at
peroxide/ozone ratios greater than 0.3, these reactions may result in low concentrations of
molecular ozone and OH radical available react to NOM. The lack of variability in THMS
and HAAS concentration resulting in treated Lake Erie water seems to indicate that the
reaction of NOM with OH radical is too slow for these factors to influence the formation

of DBPS.

450
400 u 7 v , , e
350
300
250
200
150
100
50

 

     

I +* kLake Efie
1+ Lake Lansing l

119/L HAAS

 

 

 

 

0 0.2 0.4 0.6 0.8 1 1.2
[1120211103] (wt/wt)

Figure 10. 6. Formation of HAAS for hydrogen peroxide (Ozone dosage: 1 mg O3/mg C,

HRT: 12.5 min, Temperature: 25 0C).

129

11. EFFECT OF PHOSPHATE

11.1. RCT values in solutions containing phosphate ions

In order to more thoroughly investigate the influence of OH radical reactions on
the formation of AOC, THMS and other by-products during the ozonation of NOM,
phosphate (range from O to 10 mM) was used in this study. The phosphate ion is well
documented as a promoter of OH radicals during the decomposition of ozone (Hoigné
and Bader, 1985). Therefore, it was expected that the RCT values would increase as the
concentration of phosphate increased since the phosphate ion generates the superoxide
anion (02') from oxygen. In the presence of the phosphate ion and organic alcohols, the
superoxide anion is formed by the following mechanism (See Figure 11.1). Superoxide
(02') can react with 03' to produce the 0H radical. At very high phosphate
concentrations, (e.g., in buffer solutions) the phosphate ions can act as an inhibitor of the

OH radical (Hoigné and Bader, 1985).

130

OH OH- H po

0 -

H,Po,- l2 92

A U A : \j
\A

w’ RCHOH ———-> X

 

 

 

 

H Po-
HPO42' HPO,‘ 2 4

Figure 11.1. Schematic drawing of the proposed set of typical reactions, resulting in the
conversion of OH- to 02' radicals due to the action of phosphate ions as promoters.

(Hoigné and Bader, 1985)

The effect of phosphate concentration on RCT values is shown in Figure 11.2. As
expected, the RCT values increased with increasing phosphate concentration in both Lake
Erie and Lake Lansing water. Also as expected, the RCT value decreased at high
concentrations of phosphate (10 mM). At the lower concentrations of phosphate, the
phosphate ion acted as a promoter, increasing the OH radical concentration while at high
concentrations the inhibition reactions predominated, resulting in a decrease in the
concentration of OH radicals. The result substantiates the well-established theory by

Hoigné and Bader (1985).

131

RCT

1.20E-07 . ,, f- _ f

1.00E-07. _ z,

   
  
 
  

l
8.00E-08 l
I 7 I
6.00E-08 _ L, _.. ” ___ ,/ _. i
l l
4'00E-08 7 7 7' I i—o—lfakeiErie 7 I
i l J Lake Lansing I
2.00E-08 A . e 2 2 L L, L L ,2. : 7,.
0.00E+00 . 2 -- _.l - . _._.. --__ .2 , ., . L. ., ,
0 2 4 6 8 10 12

Phosphate [mM]

Figure 11.2. The effect of phosphate on RCT values (Ozone dosage: 1mg 03/mg C,

Temperature: 25 °C).

132

l 1.2. Formation of ketoacids in solutions containing phosphate

The concentration of ketoacids formed in the presence of phosphate is shown in
Figure 11.3. Based upon our hypothesis that the predominance of molecular ozone
reactions would increase the formation of ketoacids, it was expected that the
concentration of ketoacids would decrease as the concentration of phosphate increased
since the concentration of 0H radical is increased due to the promotion reaction of
phosphate.

As observed in Figure 11.3, the concentration of ketoacids increased at low
concentrations of phosphate ion (<2mM). However, the concentration of ketoacids
decreased as the concentration of phosphate increased above this threshold value. The
result confirms the hypothesis that the formation of ketoacids is hampered under
conditions where the OH radical reaction predominates. It also substantiates the
hypothesis that promoters and scavengers play a role greater than just reacting with ozone
and 0H radicals. It appears that as with bicarbonate, the relationship between RCT and
by-product concentration is much more complex than originally hypothesized. As with
bicarbonate/carbonate radicals, the phosphate radicals may react with the NOM present in

solution.

133

300

 

 

 

 

 

250 .22 . ___
/l

«n 200 2. . 2
g // + Lake Lansing
8 l . .Lake Erie l
g 150 222 , .2 2 2 2 - ,2 9 ,
X
d
g? 100

50

0 2 .
0 2 4 6 8 10 12
Phosphate [mM]

Figure 11.3. The effect of phosphate on the formation of ketoacids (Ozone dosage: 1mg

O3/mg C, HRT: 12.5 min, Temperature: 25 0C).

134

11.3 Formation of AOC in solutions containing phosphate ions

Since the formation of AOC occurs by reactions involving molecular ozone, the
concentrations of AOC are expected to decrease as the concentration of phosphate
increases, since 0H radical reactions will predominate at high concentration of
phosphate. As shown in Figure 11.4, the concentration of AOC decreased as the
concentration of phosphate was increased in both Lake Erie and Lake Lansing water (See
Figure 11.4). This result also confirms the hypothesis that the reaction of molecular
ozone with NOM is the predominant source of AOC rather than OH radical reaction. The
addition of phosphate (IOmM) before ozonation resulted in 77% and 22% lower

concentrations of AOC in Lake Erie and Lake Lansing than that of the raw water.

135

2500 . 7 ,2 '—0— Lake Erie ‘
' .+ Lake Lansing '

119/L AOC

 

 

12

Phosphate [mM]

Figure 11.4. The effect of phosphate on the formation of AOC (Ozone dosage: 1mg

O3/mg C, HRT: 12.5 min, Temperature: 25 0C).

 

136

11.4. Formation of THMS and HAAS in solutions containing phosphate ions

Since phosphate is a promoter of 0H radicals, the formation of THMS and HAAS in
the presence of phosphate ions is expected to increase due to the increased prevalence of
the 0H radical reaction. The results obtained for THMS and HAAS in both Lake Erie and
Lake Lansing water are consistent with that observed with AOC and ketoacids, (See
Figure 11.5 and Figure 11.6) and the hypotheses put forth in this work. In general, the
concentration of HAAS increased as the concentration of phosphate increased. This result
is consistent with the hypothesis that the formation of THMs and HAAS occurs,
predominately, as a result of the reaction of OH radicals with NOM. However, as with
RCT and AOCs, the concentrations of THMS and HAAS leveled off at high phosphate
concentrations. This result further supports the hypothesis that at high concentrations, the
phosphate ion acts as an inhibitor, thereby decreasing the concentration of 0H radicals
and the importance of the 0H radical reaction compared with the molecular ozone

reaction.

137

300 ,

 

  
 

250 2 — -. 2 _ - «2.

200- 2, f . Lake Erie

1+ Lake Lansing

rig/L THMs
a;
o

 

 

Phosphate [mM]

Figure 11.5. Effect of phosphate on the formation of THMs (Ozone dosage: 1mg O3/mg

C, HRT: 12.5 min, Temperature: 25 °C).

138

350

 

 

 

 

 

 

 

300 _

250 2 I 0 Lake Erie' |
m ~+ Lake Lansing
5200- , _~. ,, ,2 ~
:1:
:1 150- ,. _ __ __
a
:1.

100 2 2 2 , 2 2

50 ,, v 2 2 2, —2fi2 2

N
0 WM 2 .
0 2 4 6 8 10 12

Phosphate [mM]

Figure 11.6. Effect of phosphate on the formation of HAAS (Ozone dosage: 1mg O3/mg

C, HRT: 12.5 min, Temperature: 25 0C).

139

12. COMPARISON OF WATER CHARACTERISTICS

The results obtained in this study were used to calculate the rate constants for the
reaction of 0H radical with NOM (k0H,N0M) based upon the methodology developed by
Haag and Yao (1993). The rate constant for the 0H radical reaction with NOM was
determined by using the Q concept developed by Haag and Hoigné (1985). The Q value
is defined as the amount of ozone that is necessary to degrade one log unit removal (37
%) of an OH radical probe compound. The Q value was determined from the slope of a

plot of In {[PCBA]/[PCBA]O} vs. dosed initial ozone concentrations, such that:

 

_ ln [PCBA ]
[PCBA 10
PCBA [ 03 ]added Eq. 12.1

 

If PCBA is present at a low aqueous concentration relative to that of NOM and the OH

radical is consumed only by NOM, the assumption can be made:

kon.PCBA[PCBA] <<<< kon.NOM[NOM] Eq. 12.2

Then, the steady-state concentration of the very short lived 0H radicals can be

determined by Eq. 12.3:

140

A[03] x 1
Al ZkOH,NOM[N0M]

 

[011155 = -77 Eq. 12.3.

where 11 is stoichiometric yield of OH radicals from ozone decomposition. A value of

0.67 is usually selected for 7] based on the based-catalyzed OH radical yield without
promotion (Westerhoff et al. 1999). This 0H radical concentration can be applied to the

PCBA degradation equation (Eq. 12.4), such that:

dPCBA
— l d, ]=k0H,pCB.[PCBA1[0H-i.. =k

 

[PCBA] Eq. 12.4.

exp erimenl

Thus, the OH radical concentration at steady-state can be described as Eq. 12.5.

k exp eriment

 

lOHlss = Eq. 12.5.

k0H,PCBA

Integration of equation 12.4 and then substitution of Eq. 12.3 into Eq. 12.4 yields

 

Eq. 12.6.
PCBA k ’
_ 1n-[_—__]_ = 17 . OHJ’CBA ' J'd[03] = kexperiment X t Eq-126
[PCBA]. ZkOH.N0M[N0M1

0

141

When ln[PCBA]/[PCBA]O =1 (PCBA is degraded 37 %), the amount of ozone consumed

until this time represents the Q value. Thus, Eq. 12.6 can be rewritten as:

k
OH. NO.”
77 ° 9mm = ___—k Eq. 12.7.
CH .PC BA

The determination of Q and n values for the raw water tested can give the rate
constant for the reaction of 0H radical with NOM. These values can be used in the
analysis to understand the pathways for the ozonation of NOM in various water sources.
When the concentration of either ozone or promoter is much less than that of NOM and
the rate of reaction of 0H radical with NOM is slower than that of OH radicals with
ozone or ozone decomposition intermediates, the OH radicals might be consumed during
ozone decomposition, even though the formation of the 0H radical is enhanced.

The values for DOC, alkalinity, QPCBA, RCT, and rate constant of 0H radical with
NOM obtained in this study were compared to that obtained in other US. surface waters
(Table 12.1). The ratios of [0H°]/[O3] were quite constant, ranging from 0.76 to 1.56 for
all waters except Lake Lansing water (value of 0.06) and Eno River water (Durham, NC)
(value of 2.66). As such, the ratio of [0H-]/[O3] appears to be independent of the DOC
concentration.

The relationship between QPCBA and DOC is plotted in Figure 12.1. The results
indicate that (2mm is linearly correlated with the DOC. The equation describing this
relationship is (0.519 2L. 0.04) DOC - 0.286 when data for Lake Erie and Lake Lansing

water are included. This resulted in a lower slope compared to the equation (QPCBA =

142

(0.573: 0.05)-DOC - 0.42; r2: 0.989) obtained by Haag and Yao (1993). However, using
the F test (See Appendix, Park J .H. 2000, Anderson R.L. 1987), yielded F values for the
results of 0.25 and the F value based on the degree of freedoms (n) is 8.18 for a
confidence interval of 99%. Therefore, the slope obtained by us and Haag and Yao
(1993) are not statistically different at a 99 % confidential interval. These results are
important since they allow for the estimation of ozone consumption in US. surface
waters, along with estimating 0H radical concentrations based upon DOC concentrations.

The rate constant for the reaction of the 0H radical with NOM in Lake Lansing
water is quite low compared to that obtained using Lake Erie water. This result contrasts
with the previous results that indicated that the rate constant for the reaction of OH
radical with NOM in Lake Erie water (in Chapter 5) is greater than that obtained using
Lake Lansing water. In order to determine the rate of reaction for the 0H radical with
NOM, a value of 0.67 was used for the stoichiometric yield, n, of 0H radicals from
ozone decomposition. This value does not consider the promotion or inhibition roles of
NOM (Westerhoff et a1. 1999). As such, the difference in the rate constants obtained
using this calculation and that obtained from experiments indicates the importance of
promotion and inhibition reactions. As expected, this difference also indicates that the

influence of these reactions varies with water source.

143

Table 12.1. Characteristics of US. surface waters

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

104kouooc
DOC Alkalinity ope... 107
Water sample L-s/(m
mg/L meq/L mg/L Ozone [0H]/[03]
DOC)
Hetch Hetchy Res.
Yosemite Park, CA 1.4 0.0065 0.7 1.56 3.49
*Lake Erie,
Monroe, MI 2.0 1.8 1.02 1.13 3.56
L.Crystal Springs Res.
Hillsborough, CA 2.1 0.39 0.68 1.26 2.26
San Andreas Res.
Millbrae,C A 2.4 0.41 0.72 1.11 2.09
San Antonio Res,
Sunol, CA 2.6 0.6 0.84 0.81 2.25
P°t°ma° 13:“ Region: 3.3 0.68 1.34 1.09 2.83
Calaveras Res.
Milpitas, CA 3.7 1.25 2.12 0.76 4.00
Eno River Durham, NC 4.8 0.19 2.18 2.66 3.17
a“ Hemessey Napa’ 5.6 1.33 2.77 1.05 3.45
*Lake Lansing,
Haslett, MI 9.5 2.5 3.59 0.06 2.64
Virginia Key
Wastewater Miami, FL 14.1 2.13 7.62 1.08 3.77
. 1.14 :i:
Averages. 0.63 3.05 :t 0.66
Aldrich Humic Acid 4.8 0 2.75 4.0

 

 

 

 

 

 

*Obtained from this research

'l'Reference: Haag and Yao (1993)

144

 

 

10

 

 

 

r’ =0.9530751114 / ’o
< 6 7
m
0
D.
C:

4 2

2 _

0

0 16

 

DOC (mg/L)

Figure 12.1. Correlation between DOC and (2;)ch in US. surface water (Dashed line: 95

% confidence interval).

It was attempted to correlate ketoacids and AOC concentrations on formed upon
ozonation of Lake Erie and Lake Lansing water. Since the ketoacids as biodegradable

compounds, the ketoacids would be portion of AOC. However, as shown in Figure 12.2,

145

 

no correlation between ketoacids and AOC was found. This result indicates that the
portion of ketoacids concentration in AOC is significantly smaller than expected. Not
only the variability of the biodegradable organic carbon formed upon ozonation but also
small portion of ketoacids in AOC would result in insignificant correlation (r2 = 0.499)

between ketoacids and AOC.

 

 

 

 

 

 

 

400
Lake Erie Lake Lansi
350 2,2,_2_______2_.22ng ”2
{ V‘ V
300 2 2 2 2 2 2 2 , .
3
a 250 ..
3 0
g 200 . 7. 2 2 o 2
E 150 . 2° ° ° ’ °
3 0 9
Q 100 °— — e—firj' —2 22_2._ 2 2 2 9
50 3,2“; __°__ _ _ 2 ,
v4? ‘ °
0 o o e .
0 500 1000 1500 2000 2500

AOC (11 911-)

Figure 12.2. Correlation between ketoacids and AOC in Lake Erie and Lake Lansing

water.

146

 

13. CONCLUSIONS

1. The RCT values obtained experimentally allowed us to not only determine the ratio of
OH radical to molecular ozone concentration at the various ozonation conditions but
also indicated the reaction rate of OH radical with NOMs tested. The NOM of Lake
Erie reacts with CH radicals more slowly than does the NOM present in Lake
Lansing and the NOM of Lake Erie has a capability to promote more OH radical

formation than does that of Lake Lansing.

2. The OH radical concentration was calculated based on RCT values. During the
ozonation time, the RCT values were constant even though the ozone decomposes
continuously. The formation of DBPS on ozonation was influenced by the ratio of the

concentration of OH radical and molecular ozone rather than reaction time.

3. The RCT values at various temperatures (ranges from 10 to 40 °C) linearly increased
as temperature increased. However, due to the changes in the reaction rate of 0H
radical and molecular ozone with NOM on the temperature, the formation of

ketoacids, AOC and DBPS are independent of RCT values.

4. Ozonation at various pH (2 to 10) significantly changed the RCT values, the RCT did
not affect the reaction of ozonation with NOM due to high concentration of

scavenging capacity of the OH radical. The secondary radical species produced from

147

 

OH radical reaction with scavenging species also affected the reaction of NOM with

020116.

. Bicarbonate ion inhibited the reaction of the OH radical with NOM. However,

carbonate radicals produced from the reaction of OH radical with carbonate ion

resulted in oxidation of NOM otherwise oxidized by OH radicals.

OH radical concentrations increased with increasing hydrogen peroxide up to the
ratios [H202/O3] of 0.3 to 0.5 [Wt/Wt] in both water sources. The higher the
concentration of OH radical formed, the lower the concentration of ketoacids, AOC

and DBPS produced.

. Phosphate ion plays a promoter role in ozonation increasing OH radical. Ozonation in
the presence of phosphate ion (0 to 10 mM) the concentration of AOC resulted in
77% and 22% lower concentration of AOC in Lake Erie and Lake Lansing than that

of the raw water.

. The (2;:ch obtained from Lake Erie and Lake Lansing was compared with that of
other surface water in US. A correlation between DOC and QPCBA for surface water
including Lake Erie and Lake Lansing was predicted. The results using Lake Erie and
Lake Lansing water were not statistically different at a 99 % confidential interval
included within a confidence interval of 99% obtained by us and Haag and Yao

(1993). The rate constants of 0H radical with NOMs were determined in Lake Erie

148

10.

11.

and Lake Lansing, 3.56 x 10" and 2.64 x 10“4 L's/(M of DOC), respectively. Since
the calculation excluded the promotion role of NOM on ozonation, it showed
difference from the RCT value obtained in this study. The calculation of the rate
constant including promotion role of NOM on ozonation is recommended for future

research.

The concentration of ketoacids and AOC were not correlated since the ketoacids has

smaller potion of AOC thereby, giving more variability than expected.

The formation of AOC indicates that the reaction of molecular ozone with NOM is
the predominant source of AOC rather than OH radical reaction and that the
formation of DBPS decreased under the molecular ozone predominant condition.
However, the NOM characteristics and background characteristics affecting ozone

decomposition also need to be considered for predicting the formation of AOC and

DBPs

These results would important for water treatment industry to not only control the
concentration of AOC and DBPs but also understand ozone reaction with NOM on

ozonation.

149

 

 

12. The followings are recommended for a future research.

a The rate constants of NOM with OH radical and ozone considering the promotion
effect is needed to predict the concentration of AOC and DBPs.

o More water sources in US. need to be tested to confirm the correlation obtained in
this study.

0 Prior to application of controlling AOC and DBPs for the fields, the cost estimation
needs to be calculated.

0 The computer modeling work including the mechanisms suggested in this research
need to be tested by the field operators if the results would be incorporated with the

results in the fields.

150

 

14. REFERENCES

Anderson R.L., 1987. Practical Statistics for Analytical Chemists, Van Nostrand
Reinhold. New York. New York.

Arai H., Arai M., and Sakumoto A., 1986. Exhaustive Degradation of Humic Acid
in Water by Simultaneous Application of Radiation and Ozone, Wat. Res. ,
20(7):885-891.

Acero J., and von Gunten, 2000. Influence of Carbonate on the Ozone/Hydrogen
Peroxide-Based Advanced Oxidation Process for Drinking Water Treatment,
Ozone Sci. & Engrg., 22:305-328.

Bader H., Sturzenegger V., and Hoigné J., 1988. Photometric Method for the
Determination of Low Concentrations of Hydrogen peroxide by the Peroxidase
Catalyzed Oxidation of N,N,-diethyl-p-phenylenediamine (DPD), Wat. Res,
22(9):]109-1115.

Bailey PS, 1978. Ozonation in Organic Chemistry. Academic Press, Inc. New
York., pp.45-82.

Benoit F. M., Helleur R., Malaiyandi M., Ramaswamy S., and Williams D. T.,
1993. Soil Fulvic Acid Degradation by Ozonation in Aqueous Medium, Ozone
Sci. & Engrg., 15:19-38.

Bielski B.H.J., and Gebicke J.M., 1970. Species in Irradiated Oxygenated Water,
Adv. Radiation Chemistry, 22177-279.

Bouxton G.V., Greenstock C.L., Helman W.P., and Ross AB, 1988. Critical
Review of Rate Constants for Reactions of Hydrated electron, Hydrogen
Atoms and Hydroxyl Radicals (OH-/O'-) in Aqueous Solution, .1. Phys. Chem.
Ref Data., 17:518-759.

Boyce S. D., and James H. F., 1983. Reaction Pathways of Trihalomethane
Formation from the Halogenation of Dihydroxyaromatic Model Compounds for
Humic acid, Environ. Sci. Technol., 17(4):202-211.

151

 

Brunet R., Bourbigot M. M., and Dore M., 1984. Oxidation of Organic
Compounds through the Combination Ozone-Hydrogen Peroxide, Ozone
Sci. & Engrg., 6:163-183.

Burke V., Robinson J., Gracey M., Peterson D., Partridge K., 1984. Isolation of
Aeromonas-Hydrophila from a Metropolitan Water-Supply — Seasonal
Correlation with Clinical Isolates, App]. Envir. Microbiol., 48(2):361-366.

Carlson K.H., and Amy G.L., 1998. BOM Removal during Biofiltration, J. A W WA.,
90(12):42-52.

Chamock C., and KjOnnO 0., 2000. Assimilable Organic Carbon and Biodegradable
Dissolved Organic Carbon in Norwegian Raw and Drinking Waters, Wat. Res.
34(10):2629-2642.

Christman, R. F., Norwood D.L., Millington D.S., Johnson J.D., Stevens A.A.,
1983. Identity and Yields of Major Halogenated Products of Aquatic Fulvic acid
Chlorination, Environ. Sci. T echnol., 17(10): 625-628.

Cipparone L.A, Diehl A, and Speitel G. E., 1997. Ozonation and BDOC
Removal: Effect on Water Quality, J. A WWA., 89(2):84-97.

Cooper W.J., Nickelesen M.G., Meacham D.E., Cadavid E.M., Waite TD,
and Kurucz ON. 1992. High Energy Electron Beam Irradiation: An Innovative

Process for the Treatment of Aqueous Based Organic Hazardous Wastes, J.
Environ. Sci. Health, A27(1):219-244.

Collins, M. R., Amy, Gary L. Steelink, Comellus. 1986. Molecular weight
distribution, Carboxylic acidity, and Humic substances Content of Aquatic
organic matter: Implications for Removal during Water Treatment, Environ.

Sci. Technol., 20(10):1028-1032.

Dalouche A., 1981. Degradation de Composes Organiques Malodorants par
Traitement Biologique et Ozonation. These Rennes, No. 696, France.

Decoret C., Royer J., Legube B., and Dore M., 1984. Experimental and Theorectical
Study of the Mechanism of the Initial Attack of Ozone on Some Aromatics in

152

 

Aqueous Medium, Envir. Technol. Lett. 52207-218.

De Laat J ., Dore’ M., and Mallevialle J ., 1991. Effects of Preozonation on the
Adsorbability and the Biodegradability of Aquatic Humic Substances and

on the Performance of Granular Activated Carbon Filters, Wat. Res.,
25(2):151-164.

Duguet J .P.,. Brodard E., Roustan M., 1986. An Automated procedure
for monitoring the Effectiveness of Ozonation Process, Ozone Sic. & Engrg.,
8:321-338.

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 Sci. & Engrg., 7:241-258.

Edge 1.0, and Finch RE, 1987. Observations on Bacterial after Growth in
Water Distribution Systems: Implications for Disinfection Strategies, Wat.
Environ. Manag., 1:104-110.

Elovitz M., von Guten, and Kaiser, H., 2000. Hydroxyl radical/Ozone Ratios during
Ozonation process. 11. The Effect of Temperature, pH, Alkalinity, and DOM
properties, Ozone Sci. & Engrg., 22:123-150.

Elovitz M., and von Gunten U., 1999. The Effect of DNOM Properties on the
Kinetics of Ozone Decomposition and Hydroxyl Radical Scavenging, 14th
Ozone World Congress, Dearbom, Michigan, pp. 95—101.

Elovitz M., and von Gunten U., 1997. Influence of Temperature, pH, and DOM
Source on Hydroxyl radical/Ozone Ratios during Ozonation of Natural
Waters, 13th Ozone World Congress, Kyoto, Japan, pp. 475-480.

Forni L., Bahnemann D, Hart E.J., 1982. Mechanism of the Hydroxide Ion
Initiated Decomposition of Ozone in Aqueous Solution, J. Phys. Chem, 86:255-
259.

Frais J., Ribas F., and Lucena F., 1995. Comparison of Methods for the

153

 

Measurement of Biodegradable Organic Carbon and Assimilable Organic Carbon in
Water, Wat. Res., 29(12):2785-2788.

Gao H., and Zepp R. G., 1998. Factors Influencing Photoreactions
of Dissolved Organic Matter in a Coastal River of the Southeastern United
States, Environ. Sci. Technol., 32(19):2940-2946.

Gehringer P., Proksch E., and Szinovatz W., 1985. Radiation-induced
Degradation of Trichloroethylene and Tetrachloroethylene in Drinking
Water, Int. J. Appl. Radiat. Isot., 36:313- 314.

Gibbs R.A, Scutt J.E., and Croll B.T., 1993. Assimilable Organic Carbon
Concentrations and Bacterial Numbers in a Water Distribution system, Water
Science and Technology, 27(3-4): 159-166.

Glaze W.H., Koga M., and Cancilla D., 1989. Ozonation Byproducts. 2.
Improvement of an Aqueous-Phase Derivatization Method for the

Detection of Formaldehyde and Other Carbonyl Compounds Formed by
the Ozone of Drinking Water, Eviron. Sci. & Technol., 23(7):838-847.

Glaze W. H., Kang, J .W., and Chapin, H. D.,1987. The Chemistry of
Water Treatment Processes Involving Ozone, Hydrogen Peroxide and
Ultraviolet Radiation, Ozone Sci. & Engrg., 9:335-352.

Gracia R.A., and Ovelleiro J. L., 1996. Study of the Catalytic Ozonation of

Humic substances in Water and their Ozonation Byproducts, Ozone. Sci. &
Engrg., 18:195-208.

Griffini 0., 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
Sci. & Engrg. 21:79-98.

Guittonneau S., Glaze W.H., Duget J .P., and Walbe 0., 1992. Characterization
of Natural Water for Potential to Oxidize Organic Pollutants with Ozone, Ozone
Sci. & Engrg., 14:185-196.

154

Haag W.R., and Yao C.C.D., 1993. Ozonation of US. Drinking Water Sources:
HO- Concentration and Oxidation-competition Values. 1 1th ozone world
congress, Sanfracisco, CA. USA. pp.S-17-119 — 126.

Haag, W.R. and Hoigné. J., 1985. Photo-Sensitized Oxidation in Natural Water
via OH Radicals, Chemosphere, 14(1 1/ 12). 1659-1671.

Hacker P.A., Paszko-Kolva 0, Stewart M.H., 1994. Production and Removal of
Assimilable Organic Carbon under Pilot-plant Conditions through the Use of
Ozone and Peroxone, Ozone Sci. & Engrg., 16:197-212.

Hoigné J., and Bader H., 1983b. Rate Constants of Reactions of Ozone with
Organic and Inorganic Compounds- 11, Water. Res., l7(2):185-194.

Hoigné J ., and Bader H., 1983a. Rate Constants of Reactions of Ozone with
Organic and Inorganic Compounds- 1, Water. Res., 17(2):]73-183.

Huck P.M., 1990. Measurement of Biodegradable Organic Matter and Bacterial
Growth Potential in Drinking Water, J. A WWA., 82(7):78-88

Ichhashi K., Teranishi K, and Ichimura A., 1999. Bromianted Trihalomethane
Formation in Halogenation of Humic acid in the Coexistence of Hydrochlorate
and Hypobromite ions, Wat. Res. 33(2): 477-483.

Irvine Ranch Water District, 1999. http://www.irwd.con1/\I\I"atchcr\'icc/r111p.htm.

 

Jolley R.L., Bull R.J., Davis W.P., Katz S., Roberts M.H., and Jacobs V.,
1985. Water Chlorination: Chemistry, Environmental Effect and Health Effects,
Vol. 5. Chelsea, MI. Lewis Publishers. 145-152.

Joret J.C., Lévi Y., Dupin T., and Gilbert M., 1988. Rapid Method for
Estimating Bioeliminable Organic Carbonc in Water, AWWA Annual
Conference, Orlando-Florida, pp.1715-1725.

Kaplan L.A., Reasoner DJ, and Rice E.W., 1994. A survey of BOM in US
Drinking Waters, J. A WWA., 86(2):]21-132.

155

 

Karpel V., Leitner N., Gehringer P., Eschweiler H. Legube B., and Dore M.,
1997. Oxidation of Natural Organic Matter in Aqueous Solution by Processes

Involving Ozone and/or Ionizing Radiation, 13th Ozone World Congress, Kyoto,
Japan, pp.535-540.

Kuivinen J. and Johnsson H., 1999. Determination of Trihalomethanes and
Some Chlorinated Solvents in Drinking Water by Headspace Technique with
Capillary Column Gas-Chromatography, Wat. Res., 33(5):1201-1208.

Langlais B., Reckhow D.A., and Brink DR, 1990. Ozonation in Water
Treatment; Application and Engineering, Lewis Publisher, Chelsea, MI. pp.
11-13.

Legrini O., Oliveros E., and Braun A.M., 1993. Photochemical Processes for
Water Treatment, Chem. Rev., 93:671-698.

Levy R.V., Hart FL, and Cheetham RD, 1986. Occurrence and Public Health
Significance of Invertebrates in Drinking Water, J.A W WA., 78(9):105-1 10.

Lykins B.W., Koffskey W.E., and Patterson KS, 1994. Alternative Disinfectants
for Drinking Water Treatment, J. Environmental Engineering. 120(4):745-
758.

McCarty P.L., 1969. Energetics and Bacterial Growth, 5th Rudolf Research
Conference, Rutgers, the State University, New Brunswick, NJ. 1-38.

McGuire M.J., and Meadow KG, 1988. AWWARF Trihalomethane Survey,
J.A WWA., 80(1):61-68.

Manahan S. E., 1993. Fundamental of Environmental Chemistry, Lewis
Publisher, Chelsea, MI. pp. 388.

Manahan S. E., 1991. Environmental Chemistry, Lewis Publisher, Chelsea, MI.
pp. 68.

156

 

Masten S.J., Tian M., Upham BL, and Trosko J .13., 2000. Effect of Selected Pesticides
and their Ozonation By-products on Gap Junctional Intercellular Communication
Using Rat Liver, Chemosphere, In Publish.

Masten S.J., Galbraith M.J., and Davies S., 1997. Oxidation of 1,3,5-Trichlorobenzene
Using Advanced Oxidation Processes, Ozone Sci. & Engrg., 18:535-547.

Masten 8.1., and Davis 8., 1994. The Use of Ozonation to Degrade Organic
Contaminants in Wastewater, Environ. Sci. Technol., 28(4):180A-185A.

Miettinen I, T., Vartianinen T, and martikainen P.J., 1999. Determination of

Assimilable Organic Carbon in Humus-Rich Drinking Waters, Wat. Res.,
33(10):2277-2282.

Miettinen I.T., Vartiainen T., Nissinen T., Tuhkanen T., and Martikainen P.,
1998. Microbial Growth in Drinking Waters Treated with Ozone,
Ozone/Hydrogen Peroxide or Chlorine, Ozone Sci. & Engrg., 20:303-315.

Miettinen I. T., Vartiainen T., and Martikainen P.J., 1997. Microbial Growth and
Assimilable Organic Carbon in Finnish Drinking Waters, Water Science
and Technology, 35(11-12):301-306.

Miltner R. J ., Shukairy H. M., and Summers R. S., 1992. Disinfection By-product
Formation and Control by Ozonation and Biotreatment, J.A WWA.,
84(11):53-62.

Morris J. C., 1975. Formation of Halogenated Organics by Chlorination of Water
Supplies. US. EPA 600/1-75-002, Washington , D.C.

Najm I.N., Patania N.L., Jacangelo 1.0., and Krasner S.W., 1994. Evaluating
Surrogates for Disinfection By-products, J.A W WA., 86(6):98-106.

Paode R. D., Amy G. L., Krasner S. W., Summers R. S., and Rice E.W., 1997.
Predicting the formation of aldehydes and BOM, J. A WWA., 89(6):79-93.

Paode R., 1994. Characterization of Natural Organic Matter and its Relationship

157

 

to Treatability by Ozone, University of Colorado, Dissertation.

Park J .H., 2000, Bioavailavility of Sorbed Organic Contaminants, Michigan State
University, Dissertation.

Power K.N. and Nagy L.A., 1999. Relationship between Bacterial Regrowth and

Some Physical and Chemical Parameters within Sydney’s Drinking Water
Distribution System, Wat. Res., 33(3):741-750.

Qasim S.R., Lim S.W., Motley E.M., and Heung KG, 1992. Estimating Costs
for Treatment Plant Construction, J. A WWA., 84(8):56-62.

Reckhow D.A., and Singer, P.C., 1990. Chlorination By-products in Drinking
Water: From Formation Potentials to Finished Water Concentrations,
J.A WWA., 82(4): 173-180.

Reckhow D.A., Legube 8., Singer P.C., 1986. The Ozonation of Organic Halide
Precursors: Effect of Bicarbonate, Water. Res., 20:987-998.

Rice R.G., 1999. Ozone in the United States of America - State-of-the-art, Ozone
Sci. & Engrg., 21(2): 99-118.

Rice R.G., 1985, Ozone in Drinking Water Treatment-Evolution and Present
Status in Safe Drinking Water, Lewis publishers, Chelesa, MI. pp. 123-166.

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 GR, and Majetich G.,

1999. Identification of New Ozone Disinfection Byproducts in Drinking
Water, Environ. Sci. & Techno]. 33(19)3368-3377.

Schechter D.S., Singer P.C., 1995. Formation of Aldehydes during Ozonation,
Ozone Sci. & Engrg., 17:53-69.

Schmidt W., Wricke B. Bdhme U., and Baumgardt W., 1999. Formation and
Behaviour of Bioavailable Ozonation - And Disinfection By-products in Drinking
Water Treatment, 14th Ozone World Congress, Dearbom, Michigan, pp.367-383.

158

 

Schwarzenbach R.P., Gschwend P.M., and Imboden D.M., 1992. Environmental
Organic Chemistry. John Wiley & Sons, INC., pp.443-445.

Servais P., Billen G, Hascoet M.C., 1987. Determination of the Biodegradable
Fraction of Dissolved Organic-matter in Waters, Wat. Res., 21: (4) 445-
450.

Shukairy H.M., and Summers RS, 1992. The Impact of Preozonation and
Biodegradation on Disinfection By-product Formation, Wat. Res., 26(9):1217-
1227.

Siddiqui M.S., Amy G.L., and Murphy B. D., 1997. Ozone Enhanced Removal
of Natural Organic Matter from Drinking Water Sources, Wat. Res.,
31(12):3098-3106.

Siddiqui M.S., Amy G.L., Cooper W.L., Kurucz C.N., Waite TD, and
Nickelesen M.G., 1996. Bromate ion removal by HEEB Irradiation, J. A W WA.,
88(10):90-101.

Singer Philip C., 1994. Control of Disinfection By-products in Drinking Water,
J. Environmental Engineering, 120(4):727-744.

Snoeyink V.L., and Jenkins D., 1980. Water Chemistry. John Wiley & Sons, r
pp.157-158.

Sotelo L. J ., Beltran F.J., Benitez F. J., Beltranheredia J ., 1987. Ozone
Decomposition in Water: Kinetic Study, Ind. Eng. Chem. Res., 26. 39-43.

 

Speitel G.E., and Wanielista M.M., 1999. Advanced Oxidation and
Biodegradation Processes for the Destruction of TOC and DBP Precursors.
American Water Works Association Research Foundation and American
Water Works Association. pp. 33-53.

Speitel G.E., Symons J.M., Diehl A.C., Sorensen H.W., Cipparone L.A., 1993.

159

Effect of Ozone Dosage and Subsequent Biodegradation on Removal of DBP
Precursors, .1. AW WA., 85(5):86—95.

Standard Methods for the Examination of Water and Wastewater. 18'h ed. 1992.
Greenberg, A.E., Clesceri, LS, and Easton, A.D., Eds, APHA, AWWA,
WEF.

Staehelin J. and Hoigné J ., 1985. Decomposition of Ozone in Water in the
Presence of Organic Solutes Acting as Promotors and Inhibitors of
Radical Chain Reactions, Environ. Sci. & T echnol., 19(12):120-126.

Staehelin J ., and Hoigné J., 1982. Decomposition of Ozone in Water: Rate of
Initiation by Hydroxide Ion and Hydrogen Peroxide, Environ. Sci. &
T echno]., 16(10):676-681.

Stevenson F.J., 1994. Humus Chemistry: Genesis, Composition, Reactions,
John Wiley & sons, Inc., 2“d edition. pp.48-53.

Sulivan D. E., and Roth J. A., 1979. Kinetics of Ozone Self-Decomposition in
Aqueous Solution, AICHE symposium series, 76:197, 142-149.

Thurman E.M., 1985. Humic substances in ground water. Humic substance in
Soil, Sediment and Water. G.R. Aiken, D.M. Mcknight, R.B. Wershaw, and
P. MacCarthy, Eds. New York: John Wiley and Sons. pp. 87-103.

Trussell RR. and Umphres M.D., 1978. The Formation of Trihalomethanes, J.
AWWA., 70(11):604-612.

Tuhkanen T.A., Kainulainen T.K., Vartiainen T.K., Kalliokoski P.J., 1994. The
Effect of Preozonation, Ozone/Hydrogen Peroxide Treatment, and Nanofiltration
on the Removal of Organic Matter from Drinking Water, Ozone. Sci. & Engrg.,

16:367-383.

Uden P.C., and Miller J.W., 1983. Chlorinated Acids and Chloral in Drinking
Warer, J. AW WA., 75(10): 524-527.

160

 

Van der Kooij D., 1992. Assimilable Organic Carbon as Indicator of bacterial
Regrowth, J.A W WA ., 84(2):57-65.

Van der Kooij D., and Hijnen W.A.M., and kruithof 1.0, 1985. The effects of
Ozonation, Biological Filtration, and Distribution on the Concentration of

Easily Assimilable Organic Carbon in Drinking Water, Ozone Sci. & Engr.,
11(3):297-311.

Van Der Kooij D., Hijnen W.A.M., 1984. Substrate Utilization of an Oxalate-
Consuming Spilium Species in Relation to Its Growth in Ozonated Water,
Appl. Environ. Microbiol., 47:551-559.

Van der Kooij D., Visser A., Hijnen W.A.M., 1982. Determining the
Concentration of Easily Assimilable Organic Carbon in Drinking Water,
J.A WWA., 74(10):540-545.

Volk C.J., Volk OB, and Kaplan L.A., 1997. Chemical Composition of
Biodegradable Dissolved Organic Matter in Streamwater, Limno]. Oceanogr.,
42(1):39-44.

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, Wat. Res., 31 :32650-
656.

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 Sci. & Engrg., 15:389-404.

Volk C., Roche P., Renner C., Paillard H., and Joret J .C., 1993. Effects of
Ozone-Hydrogen Peroxide Combination on the Formation of Biodegradable
Dissolved Organic Carbon, Ozone Sci. & Engrg., 15:405-418.

Wadowsky R.M., Yee R.B., Mezmar L., Wing E.J., and Dowling J .N., 1982. Hot
Water-Systems as Sources of legioneIIa-pneumophila in Hospital and
Non-Hospital Plumbing Fixtures, App]. Envir. Microbiol., 43(5): 1104—1110.

161

 

Waller K., Swan S.H., Delorenze G., and Hopkins 8., 1998. Trihalomethanes
in Drinking Water and Spontaneous Abortion, Epidemiology, 9(2): 1 34-140.

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, Wat. Res., 33(10):2265-2276.

Westerhoff P., Song R., Amy G., and Minear R., 1996. Application of Ozone
Decomposition Models, Ozone Sci. & Engrg., 19:55-73.

White G. C., 1999. Handbook of Chlorination and Alternative Disinfections, 4'h
ed. John Wiley&Sons, Inc. 423-424.

Xie Y., and Reckhow D.A., 1992. Formation of Ketoacids in Ozonated Drinking
Water, Ozone Sci. & Engrg., 14:269-275.

Xiong F., and Legube B., 1991. Enhancement of Radical Chain Reactions of
Ozone in Water in the Presence of an Aquatic Fulvic acid, Ozone Sci. &
Engrg.,13z349-363.

Yao J .J ., Huang 2., and Masten S.J., 1998. The Ozonation of Benz[A]anthracene:
Pathway and Product Identification, Wat. Res. 32(11):3235-3244.

Yao J .J ., Huang 2., and Masten S.J., 1998. The Ozonation of Pyrene: Pathway and
Product Identification, Wat. Res. 32(10):3001-3012.

Yavich A.A., 1998. The use of ozonation and biological fluidized bed treatment
for the control of THM precursors in Drinking water. Ph.D. Dissertation.
Michigan State University.

Yurteri C., and Gurol M.D., 1988. Ozone Consumption in Natural Waters: Effects
of Background Organic Matter, pH and Carbonate Species, Ozone Sci. &

Engrg., 10:277-290.

Zhou X., Reckhow D.A., and Tobiason J. E., 1992. Formation and Removal of
Aldehyde in Drinking Water Treatment Processes. Water Quality

162

 

Technology Conference Proceedings. Toronto. Ontario. pp.291-315.

Zoungrana C., Dejardins R., and Pre'vost M., 1998. Effect of Sodium
Bicarbonate Recarbonation on the Biodegradable Dissolved Organic

Carbon Measured by Fixed Biomass, Environmental Technology, 19:193-
201.

163

 

APPENDICES

164

Calculation of rate constant of knomon

 

Reaction rate of OH radical with
bicarbonate

 

 

 

 

 

 

 

 

 

 

 

 

 

Temperature InA kHCOSDH
278 12.8 0.009172359 12.79082764 358910.2517
288 0.008853874 12.791 14613 359024.5773
298 0.008556764 12.79144324 359131 .2629
308 0. 008278947 12.79172105 359231 .0496
Reaction rate of ozone with hydroxide Determination of A for O3OH’
koao“- 0.971434034
278 69.89062148 A= 72.05841834
288 6996478778
298 70.03404744
308 70.09887174
Denominator
Temperature RCT Calculated ko3,011- 2*k*OH
5 2.65E—08 2.47E-08 69.89062148 1 .39781 E-05
15 4.42E-08 4.43E-08 69.96478778 1 .3993E-05
25 8.61 E-08 8.61 E-08 70.03404744 1 .40068E-05
35 1 .27E-07 1 .27E-07 70.09887174 1 .40198E-05
kHC03,0H kNOM.OH Numerator
3.5891E+05 74758017 1 202 5.66E+02
3.59021~:+05 ("f-183890414493 3.16E+02
3.5913E+05 -.-:-2799017.37.871. 1.63E+02
3.5923E+05_..-3112688290438 1.101=.+02
RCTX108 Saleulatedxl 0 Substract
26.5000000 24.6955249 3.2561304
44.2000000 44.2651054 0.0042387
86.1000000 86.1001 841 0.0000000

127.0000000 127.0004653 ..

99909902

    

-4,.,;

  

 

165

 

 

F test2

A variance ratio test used to decide whether two independent estimates of variance can
reasonably be accepted as being two estimates of the variance of a single normally

distributed universe.

F225:
5.3

Where

S A2 = the larger variance estimate

832 = the smaller variance estimate

F test for a group of sets3

2 (RSSIZ — RSS1 — RSS2)/(df12 — df, — dfz)
RSS9 /df12

F

 

where

RSS = 2(yi-Ayi)2: Residual sum of square
SST= 2(yi-'yi)2: Total sum of square

yi = datum point

Ayi = predicted point by equation

'yi= mean of data

RSS1= Residual sum of square at first set
RSS2 = Residual sum of square at second set

RS812: Residual sum of square at combined set

 

2 Anderson R.L., 1987

3 Park J.H., 2000

166