;

 

,u, < 44mg.»

”393‘?!

O:§\Ill§l,f‘l

 

m

0

u a

L 9.c:

s .v..;x1'.

4 .57.!5; I.w

9 VI... liabrl

1!. .1}

'11.. . I

x

c.0‘1m‘0-1iflfl‘in‘

gnunnurfl-t

I! In

«qu5n

w

«you

‘3'". Ina much,“- .

.1 .. , . V , .Pvia.

, E. .Ifi‘r?‘ . , . ‘ ‘ . ‘ , 91.3.2.2

 

{rt-gm a
90°?

This is to certify that the
dissertation entitled

THE USE OF OZONATION AND CATALYTIC OZONATION
COMBINED WITH ULTRAFILTRATION FOR THE CONTROL
OF NATURAL ORGANIC MATTER (NOM) AND
DISINFECTION BY-PRODUCTS (DBPS)

IN DRINKING WATER

presented by
BHAVANA SUSHILKUMAR KARNIK

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Civil & Environmental Engineering

x5014») aha/ZEN

Major yofessor’s Signature

CLAW JO, MOL
/ Date

 

MSU is an Afiirrnative ActiorVEquaI Opportunity Institution

‘__ _ ___—_._ , w_— —— .__.—.__... ‘

LIBRARY
Michigan State
w University

 

 

 

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

 

 

DATE DUE

DATE DUE

DATE DUE

 

W5 C 7 2008

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/C|RC/DateDue.indd-p.1

 

THE USE OF OZONATION AND CATALYTIC OZONATION COMBINED WITH
ULTRAFILTRATION FOR THE CONTROL OF NATURAL ORGANIC MATTER
(NOM) AND DISINFECTION BY-PRODUCTS (DBPS) IN
DRINKING WATER

By

Bhavana Sushilkumar Kamik

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

2006

ABSTRACT

THE USE OF OZONATION AND CATALYTIC OZONATION COMBINED WITH
ULTRAFILTRATION FOR THE CONTROL OF NATURAL ORGANIC MATTER
(NOM) AND DISINFECTION BY-PRODUCTS (DBPS) IN
DRINKING WATER
By
BHAVANA SUSHILKUMAR KARNIK
Commercially available titania membranes, with a molecular weight cut-off of 15, 5, 1
kD were used in a ozonation/membrane system that was fed with water from Lake
Lansing. The effects of ozonation on permeate flux recovery and membrane fouling was
investigated. In addition the effects of ozonation/membrane filtration hybrid process on
the removal of the natural organic matter (N OM) and the formation of disinfection by-
products (DBPs) were monitored. The commercial membrane (CéRAM Inside, Tami
North America, St. Laurent, Quebec, Canada) was coated with iron oxide nanoparticles
(4-6 nm in diameter) using a layer-by-layer technique and sintered in air for 30 minutes.
The number of coatings was varied from 20 layers to 60 layers and sintering temperatures
of 500 and 900 °C were used. Surface characterization was carried out using electron
microscopy techniques and atomic force microscopy, to study the changes in structure
and surface morphology of the membranes. The removal and survival of bacteria in the
process was also evaluated using fluorescence microscopy and microbial assays. Finally
the surface catalytic reaction was investigated to propose the mechanism responsible for

the improved performance of the hybrid process in terms of degradation of NOM and the

control of disinfection by-products.

The permeate flux through a titania coated ceramic membrane was significantly
affected by ozonation and the pH of the feed water in the system; a minimum threshold
ozone concentration (2.5 g/m3) could achieve complete recovery of permeate flux after
fouling. Ozonation/ filtration resulted in the formation of partially oxidized compounds
from NOM that were less reactive with chlorine, decreasing the concentration of
simulated distribution system total trihalomethanes (SDS TTHMs) and simulated
distribution system halo acetic acids (SDS HAAs) by up to 80% and 65%, respectively.
With catalyst coated membranes, the concentration of dissolved organic carbon was
reduced by >85% and the concentrations of SDS TTHMs and SDS HAAs decreased by
up to 90% and 85%, respectively. With the coated membrane, the concentrations of
aldehydes, ketones, and ketoacids in the permeate were reduced by >50% as compared to
that obtained with the uncoated membranes, thus reducing the risk of potential regrowth
of bacteria in the distribution system. Furthermore, with the hybrid process, greater than 7
log removal of bacteria was achieved. Surface characterization showed that optimum
water quality was achieved at 40 layers, which corresponds to a surface coating
morphology consisting of a uniform, coarse-grained structure with open, nano-sized
interconnected pores. A mechanism is proposed showing strong surface catalytic effect,
where ozone first adsorbs on the membrane surface then decomposes resulting in the
formation of hydroxyl radicals that degrade NOM and other organic compounds and thus
improved the performance of the ozone-membrane filtration hybrid process. Thus we can
meet the regulatory requirements for DBPs using a 5 RD membrane, 40 layers, sintered at

900 °C and a gaseous ozone concentration of 2.5 g/m3.

Copyright by
BHAVANA SUSHILKUMAR KARNIK

2006

To my parents, for their love, motivation, support and sacrifice.

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to Dr. Susan J. Masten for giving me
an opportunity to work and successfully complete this research project. She has been
extremely co-operative and supportive during the course of this project which made me
resolVe harder to give my best to this research. She generously extended her technical
knowledge at every instance when I was required to make a critical decision. It is her
compassionate attitude and reassurance that have been tremendous motivating factors to
succeed in this research and accomplish the goals of this work.

I want to acknowledge Dr. Melissa J. Baumann for her incessant encouragement
and overwhelming support throughout this research. I want to especially thank her for
introducing me to the fascinating world of ceramics and materials science. Her
sympathetic and coaxing attitude persuades you to excel in every task you undertake.

I extend my gratefulness to my committee members Dr. Simon H. Davies and Dr.
Vladimir Tarabara who have been very supportive during this research and at the same
time giving valuable suggestions and advice.

I would like to thank the US Environmental Protection Agency (U S EPA) Science
To Achieve Results (STAR) Program (Grant No.RD830090801) for financial support of

this work.

Acknowledgements are due to Ms. Ewa Danielewicz and Dr. Alicia Pastor for
enduring support, cheerful and comforting attitude that helped me sustain through the
best and worst times of this research.

I also acknowledge Mr. Yan-lyang Pan and Mr. Robert Pcionek for their technical
support in the laboratory. I thank undergraduate students Dan J aglowski, Nikhil Theyyuni
and David Jackson for their assistance with the laboratory work. I thank all my friends
and colleagues for their encouragement and valuable suggestions.

Finally my parents and Milind Prabhu, my core supports who have been
extremely co-operative and accommodating in all my decisions and always helping me
maintain a balanced frame of mind. It is their persistent back-up that has helped me

revitalize every time when I was weakest.

vii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. xii
LIST OF FIGURES ................................................................................. xiii
ABBREVIATIONS .............................................................................. xxii
CHAPTER ONE

INTRODUCTION

1.1 Significance ...................................................................................... 1

1.2 Objectives ........................................................................................ 3
1.3 Hypotheses ....................................................................................... 4

1.4 References ....................................................................................... 6

CHAPTER TWO

LITERATURE REVIEW

2.1 Membrane Filtration ........................................................................... 9

2.2 Ozonation ........................................................................................ 12
2.3 Catalytic Ozonation ............................................................................ 14
2.4 Ozonation-Membrane Filtration ............................................................... 16
2.5 References ...................................................................................... 17
CHAPTER THREE

EFFECTS OF OZONATION ON THE PERMEATE FLUX OF
NANOCRYSTALLINE CERAMIC MENIBRANES

3.1 Abstract ......................................................................................... 25
3.2 Introduction .................................................................................... 25
3.3 Materials and Methods
3.3.1 Membrane Filtration System .......................................................... 27
3.3.2 Water Source ............................................................................. 28
3.3.3 Gas-phase Ozone Analysis ............................................................ 30
3.3.4 Membrane cleaning and preparation ................................................ 30

viii

3.3.5 Experiments

3.3.5.1 Effect of ozone on the permeate flux ........................................ 30
3.3.5.2 Effect of securing on the permeate flux ..................................... 31
3.3.5.3 Effect of the feed water pH on the permeate flux .......................... 32
3.4 Results and Discussion
3.4.1 Effect of ozone on the permeate flux ................................................. 32
3.4.2 Effect of scouring on the permeate flux of the nanofiltration ceramic
membrane ................................................................................ 35
3.4.3 Effect of the feed water pH on the permeate flux ................................... 38
3.5 Conclusions ..................................................................................... 42
3.6 Acknowledgements ............................................................................ 43
3.7 References ....................................................................................... 44
CHAPTER FOUR

THE EFFECTS OF COMBINED OZONATION AND FILTRATION ON
DISINFECTION BY-PRODUCT FORMATION

4.1 Abstract ......................................................................................... 48
4.2 Introduction .................................................................................... 49
4.3 Materials and Methods
4.3.1 Ozonation/ Membrane Filtration ...................................................... 51
4.3.2 Water Source ............................................................................. 55
4.3.3 Membrane cleaning and preparation .................................................. 55
4.3.4 Analytical Methods ................................................................ ' ..... 56

4.4 Results and Discussion
4.4.1 Effect of ozonation, ultrafiltration, and ozonation-ultrafiltration on

water quality ..................... ‘ ........................................................ 59

4.4.2 Effect of membrane MWCO on the water quality .................................. 62
4.4.3 The effect of gaseous ozone concentration on water quality ...................... 68
4.4.4 Effect of pH on the water quality ..................................................... 71

4.5 Conclusions .................................................................................... 76
4.6 Acknowledgements ............................................................................ 76
4.7 References ...................................................................................... 77

CHAPTER FIVE

FABRICATION OF CATALYTIC MEMBRANES FOR THE TREATMENT
OF DRINKING WATER USING COMBINED OZONATION

NAN OFILTRATION
5.1 Abstract ......................................................................................... 83
5.2 Introduction .................................................................................... 84

ix

5.3 Materials and Methods

5.3.1 Membrane Preparation and Characterization ........................................ 87

5.3.2 Ozonation/ Membrane Filtration ...................................................... 89

5.3.3 Water Source ............................................................................ 92

5.3.4 Analytical Methods .................................................................... 92

5.3.5 Ozone-membrane filtration experiments ............................................ 94
5.4 Results and Discussion ....................................................................... 95
5.5 Supplemental Information .................................................................. 102
5.6 Acknowledgements ... ....................................................................... 103
5.7 References .................................................................................... 104
CHAPTER SIX

AFM AND SEM CHARACTERIZATION OF IRON OXIDE COATED
CERAMIC MEMBRANES

6.1 Abstract ....................................................................................... 109
6.2 Introduction ................................................................................... 110
6.3 Materials and Methods

6.3.1 Membrane Preparation ................................................................ 113

6.3.2 Characterization of Membranes ..................................................... 114
6.4 Results and Discussion

6.4.1 AFM imaging .......................................................................... 1 17

6.4.2 SEM imaging .......................................................................... 125
6.5 Conclusions ................................................................................... 13 1
6.6 Acknowledgements .......................................................................... 132
6.7 References ..................................................................................... 134
CHAPTER SEVEN

TEM CHARACTERIZATION OF IRON OXIDE COATED CERAMIC
MEMBRANES

7.1 Abstract ......................................... . ............................................... 137
7.2 Introduction .................................................................................... 138
7.3 Materials and Methods

7.3.1 Membrane Preparation ................................................................ 140

7.3.2 Characterization of Membranes .......... ' ............................................ 143
7.4 Result and Discussion ............................................... . ............................ 145
7.5 Conclusions .................................................................................... 145
7.6 Limitations-Problems ........................................................................ 155
7.7 Acknowledgements .......................................................................... 146
7.8 References .......................... ' ........................................................... 157

CHAPTER EIGHT

REMOVAL AND SURVIVAL OF ESCHERICHIA COLI AFTER
TREATIVIENT USING OZONATION-ULTRAFILTRATION WITH
IRON OXIDE COATED MEMBRANES

8.1 Abstract ........................................................................................ 160
8.2 Introduction .................................................................................... 161
8.3 Materials and Methods
8.3.1 Water Source ........................................................................... 164
8.3.2 Experimental Methods ............................................................... 165
8.3.3 Analytical Methods
8.3.3.1 Bacterial analysis by fluorescence microscopy .......................... 169
8.3.3.2 Regrth potential of heterotrophic bacteria ............................ 171
8.3.3.3 Microscopic observation of the membrane surface ...................... 171
8.4 Results and Discussion
8.4.1 Bacterial analysis by fluorescence microscopy ................................... 172
8.4.2 Regrowth potential of heterotrophic bacteria ..................................... 177
8.4.3 Microscopic observation of the membrane surface .............................. 179
8.5 Acknowledgements .......................................................................... 1 8 1
8.6 References .................................................................................... 182
CHAPTER NINE

USE OF SALICYLIC ACID AS A MODEL COMPOUND TO INVESTIGATE
HYDROXY L RADICAL REACTION TN OZONATION-MEMBRANE
FILTRATION HYBRID PROCESS

9.1 Abstract ........................................................................................ 186
9.2 Introduction .................................................................................... 187
9.3 Material and methods

9.3.1 Ozone-Membrane Filtration Experiment ............................................ 189

9.3.2 Quantification of salicylic acid and its byproducts ................................ 193

9.3.3 Confirmation of the identity of byproducts using GC-MS ....................... 193
9.4 Results and discussions ...................................................................... 194
9.5 Acknowledgements ........................................................................... 207
9.6 References ..................................................................................... 208
CHAPTER TEN

CONCLUSION AND RECOMIVIENDATIONS

10.1 Conclusions .................................................................................. 212
10.2 Recommendations for Future Research .................................................. 214
APPENDIX A

xi

Table 3.1

Table 3.2

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 5.1

Table 5.2

Table 7.1

Table 8.1

Table 8.2

Table 8.3

Table 8.4

Table 9.1

Table 9.2

Table 9.3

LIST OF TABLES

Typical Characteristics of Lake Lansing Water (Haslett, MI) ............... 28
Operating Conditions for the Ozone—Membrane Filtration System ....... 31
Typical Characteristics of Lake Lansing Water (Haslett, MI) ............. 53
Operating Conditions for the Ozone—Membrane Filtration System ....... 55
Initial concentrations of filtered raw water from Lake Lansing ............ 60

Comparison of performance parameters Ozonation alone, UF alone,

and combined Ozonation/UP ................................................... 61
Typical Characteristics of Lake Lansing Water (Haslett, MI) ............... 90
Operating Conditions for the Ozone—Membrane Filtration System ........95
Summary of surface area and average pore size ............................. 154
Typical Characteristics of Lake Lansing Water (Haslett, MI). ............ 164

Operating Conditions for the Ozone—Membrane Filtration System ...... 166
Live-Dead bioassays using fluorescence spectroscopy ..................... 174
Reduction in Bacterial Counts Using Heterotrophic plate count method 177
Operating Conditions for the Ozone—Membrane Filtration System ...... 190

Total molar concentration (M) of the by-product in individual
treatment processes .............................................................. 199

Dissolved ozone concentration in the permeates of Lake Lansing water
for different treatment processes 204

xii

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

LIST OF FIGURES

Schematic Representation of the Ozone-Membrane Filtration System . 29

Permeate Flux Recovery Pattern with Ozonation
(Operating Conditions: Table 2, Initial Specific flux 153I/m2h-bar,
pH 8.2, Gaseous ozone concentration 12.5 g/m3) ......................... 33

Effect of Ozone or Oxygen on Permeate Flux Recovery
(Experimental setup: Fig.3.]. Operating Conditions: Table 3. Initial
Specific flux 153L/m2h-bar, pH 8.2) .................................... 34

Effect of Continuous Ozone or Oxygen on Permeate Flux Recovery
(Experimental setup: Fig. 3.1. Operating Conditions: Table 3.2.
Continuous Ozonation (Ozone gas concentration 1.5 g/m3) and
Continuous Oxygenation, Initial Specific flux 153L/m2h-bar, pH 8.2) . 36

Effect of Ozone on Permeate Flux Recovery under

different conditions

(Experimental setup: Fig 3.1. Operating Conditions: Table 3.2.
Continuous ozonation/No Scouring:

Ozone concentrations of 12, 5, 2.5, 1.5 g/m3 used

afier12 hours operation, Pulsed Ozonation: Pulse of ozone lasting
1 minute at frequency of very 5 minutes,

Ozone concentration 12, 5, 2.5, 1.5 g/m3 afier12 hours operation
and Scouting-increase distance between Y-line mixer and
membrane module from 10 cm to 50 cm,

Ozone dose12.5, 5, 2.5, 1.5 g/m3 after12 hours operation) ............... 37

Effect of pH on Permeate Flux Recovery

(Experimental setup: F ig.3.1 Operating Conditions: Table.3.2,
Continuous Ozonation, Ozone gas concentration: 1.5 g/m3,

pH: natural pH, 7.0, 6.0 and 4.0) ............................................. 39

Dissolved Ozone Concentration in Feed water

(Experimental setup: Fig.3.1 Operating Conditions: Table3.2,
Continuous Ozonation, Ozone gas concentration: 1.5 g/m3’,

pH: natural pH, 7.0, 6.0 and 4.0) .............................................. 4O

xiii

Figure 3.8

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Relationship between Steady State Flux and Steady State

Ozone Concentration ........................................................

41

Schematic Representation of the Ozone-Membrane Filtration System . 52

Effect of molecular weight cut-off on Permeate 1
(First 400 mL of the sample)
(Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.
Ozone Dose. 2. 5 g/m3, Membrane Size: 15,5 and 1 kD.

All values are average of triplicates within experiments and
duplicate experiments.

The values have a maximum std. deviation of5 %)

Effect of molecular weight cut-off on Permeate 2
(Latter 1000 mL of the sample)
(Experimental setup: Fig. 4. 1 Operating Conditions. Table 4.1.
Ozone Dose: 2. 5 g/m3, Membrane Size: 15,5 and 1 kD.

All values are average of triplicates within experiments and
duplicate experiments.

The values have a maximum std. deviation of 5 %) ....................

Effect of molecular weight cut-off on Ozonation By-products
(Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.
Ozone Dose. 2. 5 g/m3, Membrane Size: 15, 5 and 1 kD.

All values are average of triplicates within experiments and
duplicate experiments.

The values have a maximum std. deviation of 5 %) ....................

Effect of gaseous ozone concentration on Permeate 1
(First 400 mL of the sample)
(Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.

Ozone concentration: 1. 5, 2. 5 and 10 g/m3 Membrane Size: 15 kD.

All values are average of triplicates within experiments and
duplicate experiments.

The values have a maximum std. deviation of 5 %) ....................

Effect of gaseous ozone concentration on Permeate 2

(Latter 1000 mL of the sample)

(Experimental setup. Fig. 4. 1 Operating Conditions. Table 4.1.

Ozone Dose: 1. 5, 2. 5 and 10 g/m3 Membrane Size: 15 kD.
All values are average of triplicates within experiments and

duplicate experiments.

The values have a maximum std. deviation of 5 %) ....................

xiv

.64

64

68

69

69

Figure 4.7

Figure 4.8a

Figure 4.8b

Figure 4.9a

Figure 4.9b

Figure 4.10a

Effect of gaseous ozone concentration on Ozonation By-products
(Experimental setup: Fig. 4.1 Operating Conditions: Table 4.1.
Ozone concentration: 1. 5, 2. 5 and 10 g/m3, Membrane Size: 15 kD.
All values are average of triplicates within experiments and
duplicate experiments.
The values have a maximum std. deviation of 5 %) ....................... 71

Effect of pH on Permeate 1 of 15 kD molecular weight cut off
membrane
(Experimental setup: Fig. 4. 1 Operating Conditions. Table 4.1.
Ozone Dose: 2. 5 g/m3, Membrane Size: 15 kD, pH 7 and 8. 2.
All values are average of triplicates within experiments and
duplicate experiments.
The values have a maximum std. deviation of 5 %) ....................... 72

Effect of pH on Permeate 2 of 15 kD molecular weight cut off
membrane
(Experimental setup: Fig. 4. 1 Operating Conditions. Table 4.1.
Ozone Dose: 2. 5 g/m3, Membrane Size: 15 kD, pH 7 and 8. 2.

All values are average of triplicates within experiments and

duplicate experiments.
The values have a maximum std. deviation of 5 %) ........................ 72

Effect of pH on Permeate l of 5 RD molecular weight cutoff
membrane

(Experimental setup. 3Fig. 4. 1 Operating Conditions: Table 4.1.

Ozone Dose: 2. 5 g/m3, Membrane Size: 5 kD, pH 7 and 8. 2.

All values are average of triplicates within experiments and

duplicate experiments.
The values have a maximum std. deviation of 5 %) ...................... 73

Effect of pH on Permeate 2 of 5 RD molecular weight cut off membrane
(Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.
Ozone Dose. 2. 5 g/ms, Membrane Size: 5 kD, pH 7 and 8. 2.
All values are average of triplicates within experiments and
duplicate experiments.
The values have a maximum std. deviation of 5 %) ....................... 73

Effect of pH on Ozonation By-products in Permeate 1 of

15 kD & 5 kD molecular weight cut off membrane

Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.

Ozone Dose: 2. 5 g/m3, Membrane Size: 5 and 15 kD, pH 7 and 8. 2.
All values are average of triplicates within experiments and

duplicate experiments.

XV

Figure 4.10b

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

The values have a maximum std. deviation of 5 %) ...................... 74

Effect of pH on Ozonation By—products in Permeate of

15 kD & 5 kD molecular weight cut off membrane

(Experimental setup: Fig.4.1 Operating Conditions: Table 4.1.

Ozone Dose: 2.5 g/m3, Membrane Size: 5 and 15 kD, pH 7 and 8.2.

‘All values are average of triplicates within experiments and

duplicate experiments.

The values have a maximum std. deviation of 5 %) ...................... 74

TEM characterization of the iron oxide particles

(average size 4-6nm)

(TEM: JEOL 100CX, Accelerating voltage: 100 kV,

Imaging system: Analysis, Digital imaging: Megaview HI) 88

Schematic Representation of the Ozone-Membrane Filtration System 91

Permeate flux for different membrane coating modifications
(Experimental setup: Fig. 5 .2. Operating Conditions: Table 5.1.
Membrane Size: 5 or 15 kD.

'All values are average of triplicates within experiments and
duplicate experiments.

The values have a maximum std. deviation of 5%.

For the coated membranes the first number in the legend corresponds
to MWCO of the membrane, the second number is the number of
coatings and the third number is the sintering temperature.

For example, 15-20-500 is a membrane with 15 kD MWCO

coated 20 times with the catalyst and sintered at 500 °C.

All values are average of triplicates within experiments) ................ 95

Water quality results for two different sintering temperatures
(Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1.
Membrane Size: 15 kD, Coating: 20 or 40 coatings,

Sintering temperature: 500 °C or 900 °C.

All values are average of triplicates within experiments.

Explanation of the legend is given in the caption of Fig.5.3) ............ 96

Concentrations of ozonation by-products in the permeate for two
different sintering temperatures

(Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1.
Membrane Size: 15 kD, Coating: 20 or 40 coatings,

Sintering temperature: 500 °C or 900 °C.

All values are average of triplicates within experiments.

Explanation of the legend is given in the caption of F ig.5.3) ............ 97

xvi

Figure 5.6

Figure 5.7

Figure 6.1

Figure 6.2a

Figure 6.2b

Figure 6.2c

Figure 6.3

Figure 6.4a

Figure 6.4b

Figure 6.4c

Figure 6.5

Water quality results for two different sintering temperatures
(Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1.
Membrane Size: 5 kD, Coating: 20 or 40 coatings,

Sintering temperature: 500 °C or 900 °C.

All values are average of triplicates within experiments.

Explanation of the legend is described in the caption of F ig.5.3) ........97

Concentrations of ozonation by-products in the permeate for two
different sintering temperatures

(Experimental setup: Fig.5.2. Operating Conditions: Table 5.1.
Membrane Size: 5 kD, Coating: 20 or 40 coatings,

Sintering temperature: 500 °C or 900 °C.

All values are average of triplicates within experiments.

Explanation of the legend is described in the caption of Fig.5.3) ........98

Schematic Representation of sample preparation for SEM and
AF M imaging ............................. L ................................... 1 15

AFM images of the AZT ceramic membranes: 5kD MWCO AZT
membrane uncoated .......................................................... 1 18

AFM images of the AZT ceramic membranes: 5kD AZT
membrane uncoated, sintered at 900 °C for 30 minutes ................. 118

AF M images of the AZT ceramic membranes: 5kD AZT membrane
with 40 coatings of iron oxide, sintered at 900 °C for 30 minutes ......119

Height data for AFM images of membranes with different
number of coatings ........................................................... 119

AFM image of an AZT ceramic membrane with iron oxide coating:

5kD AZT membrane with 20 coatings of iron oxide, sintered at
900 °C for 30 minutes ......................................................... 120

AF M image of an AZT ceramic membrane with iron oxide coating:
5kD AZT membrane with 30 coatings of iron oxide, sintered at
900 °C for 30 minutes ......................................................... 121

AFM image of an AZT ceramic membrane with iron oxide coating:
5kD AZT membrane with 40 coatings of iron oxide, sintered at
900 °C for 30 minutes ........................................................ 122

Roughness data for AF M images of membranes with different
number of coatings ........................................................... 123

xvii

Figure 6.6

Figure 6.7

Figure 6.8

Figure 6.9

Figure 6.10

Figure 6.11

Figure 7.1
Figure 7.2

Figure 7.3a

Figure 7.3b

Figure 7.4a

SEM images of an AZT ceramic membrane: a) 5kD MWCO

AZT membrane uncoated, b) 5kD AZT membrane uncoated,

sintered at 900 °C for 30 minutes and c — i) 5kD AZT membrane

with 20, 30, 40, 45 coatings, respectively, of iron-oxide, sintered at

900 °C for 30 minutes ........................................................ 124

Grain size measurements of AZT ceramic membrane:
Membranes: 5kD MWCO uncoated, 5kD with 20, 30, 40 coatings
of iron oxide sintered at 900 °C for 30 minutes ........................... 126

Average pore size measurements for the AZT membranes:

5kD MWCO uncoated, 5kD sintered at 900 °C for 30 minutes,

5kD with 20, 30, 40, 45 coatings of iron oxide sintered at

900 °C for 30 minutes and 5kD with 40 coatings of iron oxide
unsintered ...................................................................... 126

EDS mapping of the membranes: a) EDS mapping of uncoated
membrane and b) EDS mapping of membrane with 40 coatings
and sintered at 900 °C for 30 minutes ...................................... 128

Relative Fe concentration from EDS scans. The graph represents
relative Fe concentrations measured as Fe counts in the EDS scans 129

Water Quality data for the permeate after combined ozonation
membrane treatment process.

(The data for water quality parameters for a 5kD AZT membrane
uncoated, unsintered, with 20, 30, 40, 45 coatings of iron-oxide,

sintered at 900 °C for 30 minutes.

The insert in the graph is the plot of dissolved organic

carbon concentrations for the same membranes) ......................... 130

Schematic Representation of sample preparation for TEM imaging 142
TEM image of Fe203 nanoparticles processed from sol suspension ...143
TEM cross section of the micro porous AZTC membrane supplied
from the manufacturer. The insert is electron diffraction pattern of

the T102 coating on the AZT C membrane ................................. 146

TEM cross section of micro porous AZTC membrane with 40 coatings
of iron oxide sintered at 900 °C for 30 minutes ........................... 146

T EM cross section and Electron diffraction patterns of a 5kD

AZTC membrane with 20 coatings of iron oxide,
sintered at 900 °C for 30 minutes ........................................... 147

xviii

Figure 7.4b

Figure 7.4c

Figure 7.4d

Figure 7.5

Figure 7.6

Figure 8.1

Figure 8.2

Figure 8.3

TEM cross section and Electron diffraction patterns of a 5kD
AZTC membrane with 30 coatings of iron oxide,
sintered at 900 °C for 30 minutes ........................................... 147

TEM cross section and Electron diffraction patterns of a 5kD
AZTC membrane with 45 coatings of iron oxide,
sintered at 900 °C for 30 minutes ........................................... 148

TEM cross section and Electron diffraction patterns of a 5kD
AZTC membrane with 60 coatings of iron oxide,
sintered at 900 °C for 30 minutes ........................................... 148

Water Quality data fer the permeate after combined ozonation
membrane treatment process.

(The data for water quality parameters for a 5kD AZT membrane
uncoated, unsintered (0 nm coating thickness), with 20, 30, 40, 45

and 60 coatings of iron-oxide, sintered at 900 °C for 30 minutes

with respectively, 20, 30, 50, 55, 57 nm coating thickness.

The insert in the graph is the plot of dissolved organic

carbon concentrations for the same membranes) ......................... 151

X-ray diffraction patterns of AZTC membrane

a) 5kD MWCO AZTC membrane uncoated, b) 5kD AZTC

membrane uncoated, sintered at 900°C for 30 minutes

0) 5kD AZTC membrane with 40 coatings of iron-oxide unsintered

and d-g) 5kD AZTC membrane with 20, 30, 40 or 45 coatings

of iron oxide, sintered at 900 °C for 30 minutes ........................... 152

Schematic Representation of the Ozone-Membrane Filtration System 167

Fluorescence images indicating bacteria presence in the permeate

after different treatments

(Experimental setup: see Fig. 8.1 Operating Conditions: see

Table 8.2. Membrane Size: 5 kDa, Gaseous ozone concentration

2.5 g/m3, catalyst coated membrane: coated with 40 times with

iron oxide nanoparticles, and then sintered in air at 900 °C) ............ 173

Percent of live — dead bacteria in the permeate after different treatments.
(Experimental setup: see Fig. 8.1 Operating Conditions: see

Table 8.2. Membrane Size: 5 kDa, Gaseous ozone concentration

2.5 g/m3, catalyst coated membrane: coated with 40 times with

iron oxide nanoparticles, and then sintered in air at 900 °C.

The results were determined using fluorescence spectroscopy after
staining with molecular probes.

All values are the average of triplicates within experiments ............ 175

xix

Figure 8.4

Figure 8.5

Figure 9.1

Figure 9.2

Figure 9.3

Figure 9.4

Figure 9.5

’ Assimilated Organic Carbon (AOC) concentration after different

treatments for the permeate and reject streams.

(Experimental setup: Fig. 8.1 Operating Conditions: Table 8.2.
Membrane Size: 5 kDa, Gaseous ozone concentration 2.5 g/m3,

catalyst coated membrane: coated with 40 times with

iron oxide nanoparticles, and then sintered in air at 900 °C.

‘All values are the average of triplicates within experiments) .......... 178

SEM images of membrane surface before and after treatment

(SEM JEOL 6400V, accelerating voltage 15 kV, a-c) SEM images

of an uncoated-unsintered membrane from the manufacturer,

d-t) SEM images of a membrane coated with 40 times with

iron oxide nanoparticles, then sintered in air at 900 °C.

Pretreatment and post-treatment refers to samples before and

after ozonation-membrane filtration hybrid process treatment) ......... l 80

Schematic Representation of the Ozone-Membrane Filtration System 191

Disappearance of salicylic acid at pH 2.5-3 with different

treatment processes

(Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1
Membrane filtration (MF), Ozonation (OZ),

Ozone-uncoated unsintered Membrane filtration (MF+OZ)) ............ 195

Disappearance of salicylic acid at natural pH with different treatment
processes ‘

(Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1,

pH 7.0-8.1. Membrane filtration (MF), Ozonation (OZ),
Ozone-uncoated unsintered Membrane filtration (MF+OZ),

Iron oxide catalyst coated membrane filtration combined with

ozonation (Catalyst coated MF+OZ)) ....................................... 196

Formation of oxidation byproducts at pH 2.5-3 with different

treatment processes

(Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1
Membrane filtration (MP), Ozonation (OZ), Ozone-uncoated

unsintered Membrane filtration (MF+OZ)) ............................... 198

Formation of byproducts at pH 8 with different treatment process
(Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1,
pH 7 0-81. Membrane filtration (MF), Ozonation (OZ),
Ozone-Membrane filtration (MF+OZ), Iron oxide catalyst

coated membrane filtration combined with ozonation

(Catalyst coated MF+OZ)) ................................................... 200

XX

Figure 9.6

Figure 9.7

Figure 9.8

Total ion chromatograrns (TIC) for Salicylic acid (SA),

2, 3- dihydroxy benzoic acid (2, 3-DI-IBA), and

2, 5- dihydroxy benzoic acid (2, 5-DHBA) derivatized by

BSTFA+ TMCS. a) TIC obtained from the sample before treatment,

b) TIC obtained from the permeate from ozone-membrane

filtration process and c) TIC obtained from the permeate from
ozone-catalyst coated membrane filtration process ....................... 202

Mass spectra (MS) for compounds derivatized by BSTFA+ TMCS.
a) Mass Spectra for Salicylic acid, b) Mass spectra for 2,3-DHBA
and c) Mass spectra for 2,5-DHBA .......................................... 203

Schematic of the suggested mechanism during ozonation

membrane filtration hybrid process.

A: is TiOz in case of uncoated/unsintered membranes and is

TiOz-Fe203 in case of iron oxide coated sintered membrane ............ 206

2, 3-DHBA
2, 5-DI-IBA
AOC
AOPs
BDOC
DBPs

DOC

NOM

RO

SA

SDS HAAs
SDS THMs
THMs

TOC

ABBREVIATIONS

2, 3-Dihydroxy benzoic acid

2, 5-Dihydroxy benzoic acid
Assimilable organic carbon
Advanced oxidation processes
Biodegradable organic carbon
Disinfection by-products

Dissolved organic carbon

Haloacetic acids

Humic substances

Microfiltration

Molecular weight cutoff
Nanofiltration

Non-humic substances

Natural organic matter

Reverse Osmosis

Salicylic acid

Simulated distribution system HAAs
Simulated distribution system THMs
Trihalomethanes

Total organic carbon

xxii

UF Ultrafiltration

UV-254 Ultraviolet absorbance measured at 254 nm

xxiii

CHAPTER ONE

INTRODUCTION

1.1 SIGNIFICANCE

There is an increasing interest in the application of ozone and membrane filtration by
water utilities in order to meet the requirements of the Surface Water Treatment Rule
(SWTR), the Disinfectant and Disinfection By—Product Rule (D/DBPR), and the Long
Term 1 Enhanced Surface Water Treatment Rule (LTlESWTR) (USA EPA).
Collectively, the SWTR, D/DBPR, and LTlESWTR place stringent treatment
requirements on systems using surface water as a source. Membrane processes are
considered one of the best available technologies (BAT) for meeting the Stage 2 D/DBP
requirements (Arora et al., 1997). Indeed the development of better membranes and the
characterization of membrane surfaces along with an increased understanding of
membrane fouling make membrane filtration a effective tool in meeting regulatory
requirements ( Lin et al., 2001; Carroll et a1. 2000; Clark et al., 1998). In addition, for
systems using conventional and direct filtration, the Filter Backwash Recycling Rule
(FBRR) complements the surface water treatment rules by reducing the potential for
microbial pathogens, particularly, Cryptosporidium oocysts, to pass through the filters
into the finished water. These rules reduce the risks of the presence of microbial
pathogens in the finished water, thereby providing additional protection to consumers.

Ozone is an effective alternative disinfectant and a more powerful oxidant than

chlorine, which is widely used as a disinfectant in the United States. Ozone is capable of

decreasing the numbers of microorganisms (Lee and Deininger, 2000) and the
concentration of DBP precursors (Y avich, 1998; Cipparone et al., 1997). It also reacts
with organic substances and increases their biodegradability (Leiknes et a1, 2005;
Takeuchi et a1. 1997). The use of ozonation in water treatment processes decreases the
formation of THMs and halo acetic acids (HAAs) (Zhang et a1, 2001; Richardson et al.,
1999). The reactions that occur during ozonation produce by—products that are of
particular concern due to their mutagenicity and carcinogenicity (Bull and McCabe,
1984). These by—products are easily biodegradable and can serve as substrates for
microbial regrowth in the distribution system. These by-products can be easily be
removed by biofiltration (Y avich and Masten, 2003; Griffini et a1, 1999). In addition,
catalytic ozonation would reduce the concentration of ozonation by-products, thereby
generating more biologically stable water.

In addition catalytic ozonation is a promising technology for water treatment as it can
result in the effective degradation of NOM and other organic compounds that serve as
DBPs precursors (Legube and Karpel Vel Leitner, 1999). Based on extensive research
involving various ozonation methods for drinking water treatment, catalytic ozonation
has been determined to be the best alternative for the oxidation of ozone by—products to
carbon dioxide, and the reduction in the chlorine demand (V olk et al., 1997; Alleman et
al., 1993). Ceramic membranes are ozone resistant and when combined with ozonation,
generate very high and stable permeate fluxes without membrane damage (Schlichter et
al., 2004; Kim et al., 2001; Kim et al., 1999). Also, it has been found when metal or metal
oxide catalysts are used as the support for ceramic membranes the degradation of NOM

by ozone results (Ernst et al., 2004). Ozone in the presence of different metal oxide

catalysts, including manganese oxide, titania, alumina, and zirconia, degrades refractory
compounds, including saturated carboxylic acids, phenols, aromatic hydrocarbons, dyes,
humic substances and herbicides (Beltran et al., 2003a, 2003b; Ni and Chen, 2001;
Radhakrishnan and Oyama, 2001; Gracia et al., 1996, 2000a, 2000b; Legube and Karpel
Vel Leitner, 1999). Thus, catalyst coated ceramic membranes will enhance degradation of
NOM and improve the quality of the treated water. Also, these membranes are expected
to result in further deceases in the concentration of DBPs than achievable with ozone.
This process further facilitates the inactivation and complete removal of pathogens and
microorganisms. Thus, the proposed ozone-membrane filtration process combines the
advantages of ozonation and membrane filtration to produce water meeting required
regulatory standards. In addition, the optimum ozone doses evaluated will reduce
membrane fouling and maintain stable permeate fluxes without membrane damage.
1.2 OBJECTIVES

The objective of this research is to determine the feasibility of using a combined
ozonation and membrane filtration system to control disinfection by—products (DBPS)
precursors in drinking water treatment process. Commercially available ceramic
membranes and ceramic membranes coated with a nano-crystalline catalyst like iron
oxide colloidal particles, that decompose ozone will be used in the study.

In particular this study will try to attain the following four objectives:
1. To determine the effect of ozonation on the permeate flux of the nano-crystalline

ceramic membranes.
2. To investigate the effects of combined ozonation and membrane filtration on

disinfection by—product formation in drinking water treatment systems.

3.

To develop ceramic membranes coated with catalytic metal oxide nanoparticles and
to study the effect of catalytic ozonation and membrane filtration on the formation of
disinfection by—products.

To study the removal and survival of bacteria with treatment using the catalyzed

ozonation—membrane filtration hybrid process.

1.3 HYPOTHESES

The following hypotheses of this proposed ozonation-membrane filtration will be tested:

1.

The combined ozonation-membrane filtration can maintain stable permeate flux and
extend the operation period.

The reaction of OH radicals generated at the surface with natural organic matter
(NOM) sorbed on the ceramic membrane surface will reduce membrane fouling and
will result in better water quality.

The molecular ozone and -OH radical reactions in the bulk water and on the ceramic
membrane surface will cause a reduction in the UV-254 absorbance, dissolved
organic carbon and result in the conversion of the humic substances to non-humic
substances.

The ozonation will further reduce the chlorine demand and result in a decrease in the

disinfection by-product formation.

. Conventional sol-gel and sintering methods can produce coated membranes with the

desired molecular weight cut-off (MWCO), permeability and favorable catalytic
properties.
The use of a metal oxide, such as iron oxide, for coating of ceramic membranes will

enhance ozone decomposition as a result of changes in the surface morphology,

which enhance the formation of the 'OH radicals and other reactive species at the
membrane surface, thereby further improving the water quality and degrading NOM
and reducing the formation of DBPs.

. The metal oxide further facilitates the sorption or catalytic oxidation of ozonation by-
products such as aldehydes, ketones, ketoacids on the catalytic coated surfaces, thus
the concentration of these compounds in the treated water is very low. This results in
water that is biologically stable water with low concentrations of biodegradable
organic carbon.

. The catalytic ozonation-nanofiltration will produce biologically stable water, which
will further result in the reduced regrowth potential of bacteria; also the iron-oxide
coating traps the bacteria by sorption or filtration. As such, the combined effect of

these processes will produce water free of bacteria.

1.4. REFERENCES

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ O3-H202 and O3/Ti02) for the oxidation of natural organic matter
in water” Ozone: Science and Engineering, 15(5), 419-432.

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

Beltran, F. J ., Rivas, F. J., and Montero-de-Espinosa, R. (2003a) "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts.1. Homogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14): 3210-3217.

Beltran, F. J ., Rivas, F. J., and Montero-de-Espinosa, R. (2003b)"Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts. 2. Heterogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research. 42(14): 3218-3224.

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

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

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

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

Ernst M., Lurot F., and Schrotter J .C. (2004) "Catalytic ozonation of refractory organic

model compounds in aqueous solution by aluminum oxide" Applied Catalysis B:
Environmental, 47(1), 15-25.

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

Gracia, R., Cortes, S., Sarasa, J., Orrnad, P., and Ovelleiro, J. L. (2000a)"Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Onnad, P., and Ovelleiro, J. L. (2000b) "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

Griffini, O., Bao, M. L., Barbieri, K., Burrini, D., Santianni, D. and Pantani, F. (1999)
"Formation and removal of biodegradable ozonation by-products during ozonation-

biofiltration treatment: Pilot-scale evaluation" Ozone: Science and Engineering,
21(1), 79-98

Kim, JD. and Somiya, I., (2001) "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

Kim, J. 0., Somiya, I. and Fujii, S. (1999) "Fouling control of ceramic membrane in
organic acid fennenter by intermittent ozonation", In the Proceedings of the 14‘”
Ozone World Congress, Dearborn, MI, 131-143.

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

Leiknes, T., Lazarova, M. and Odegaard, H. (2005) "Development of hybrid ozonation
boifilm membrane filtration process for the production of drinking water" Water
Science and Technology, 51 (6-7), 241-248. '

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

Ni, CH; Chen, J .N. (2001) 'IHeterogeneous catalytic ozonation of 2-chlorophenol
aqueous solution with alumina as a catalyst", Water Science and Technology, 43 (2),
213-220.

Radhakrishnan, R., Oyama, ST. (2001) "Ozone Decomposition over manganese oxide
supported on ZrOz and TiOzz A kinetic study using in situ Laser Raman
Spectroscopy”, Journal of Catalysis, 199(2), 282-290.

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

Schlichter, B., Mavrov, V. and Chmiel H. (2004) "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production fiom
surface water" Desalination, 168, 307-317.

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

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

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

Volk, C., Roche, P., Joret, J.C., and Paillard, H. (1997) "Comparison of the effect of
ozone, ozone-hydrogen peroxide system and catalytic ozone on the biodegradable
organic matter of a fulvic acid solution" Water Research, 31(3), 650-656.

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

Yavich, AA. and Masten, SJ. (2003) "Use of ozonation and FBT to control THM
precursors" Journal of American Water Works Association, 95(4), 159-171.

Zhang, X. Z., Ni, Y. and van Heiningen, A. (2001) "Effect of temperature on the
kinetics of pulp ozonation" Journal of Pulp and Paper Science 27(8), 279-283.

CHAPTER TWO

LITERATURE REVIEW

Present and future regulations concerning drinking water quality necessitates the
use of new techniques over conventional water treatment processes, including
prechlorination, coagulation/ flocculation, sedimentation, rapid sand filtration,
disinfection and biological treatment, to meet the increasing stringent regulations of
Surface Water Treatment Rule (SWTR), Interim Enhanced Surface Water Treatment
Rule (IESWTR) and Disinfectants/Disinfectant By Product Rule (D/DBPR) which have
together introduced tighter limits on turbidity and particle removal requirements, with
additional microorganism removal and limits on disinfectants and disinfections by-
product fonned in the process.

2.1 MEMBRANE FILTRATION

Membrane filtration is considered to be one of the best available technologies to
meet these regulations (Arora et al., 1997). Membrane filtration i) serves as anabsolute
barrier to suspended particulate matter and pathogens; ii) provides treatment and
disinfection without supplemental chemical addition beyond ozone; iii) provides
consistent filtered water quality irrespective of the feed water; iv) results in the formation
of minimal amounts of chemical sludge and, therefore, minimizes residual disposal costs;
v) is simple, reliable, and can be easily automated; and vi) offers a compact modular
construction (Cleveland 1999; Nakatsuka et al., 1996; Rachwal 1995). Pressure driven

membrane processes, including microfiltration (MF), ultrafiltration (UF), nanofiltration

(NF) and reverse osmosis (R0), are among the most promising new techniques for the
treatment of water and waster water treatment processes. '(Bruggen et al., 2003;
Cleveland, 1999). UF and MF can be used in combination with traditional treatment or as
a partial replacement of traditional methods including prechlorination,
coagulation/flocculation, sedimentation, rapid sand filtration, disinfection and biological
treatment to meet the increasing stringent regulations. MF can be used for the removal of
colloidal particles, microorganisms and other particulate material, so it can serve as a
pretreatment to UP or in combination with traditional treatment methods (Mavrov et al.,
1998). UP allows for the efficient removal of suspended particles and colloids, turbidity,
algae, bacteria, parasites and viruses for clarification and disinfection purposes, but must
be preceded by effective particulate removal using coagulation/sedimentation or MF. The
current trend in water treatment is to use UF rather than MF as a single treatment option
or to use UF in combination with other treatment processes because UP is more efficient
at removing viruses, and also are less prone to fouling by fine particles than MF (Guigui,
C. et al., 2002). Researchers found that the combination of membrane filtration with other
treatment options, such as coagulation, adsorption, oxidation, and biological treatment,
enhances the efficacy of the treatment system, produces higher quality water, and
improves system performance as compared to the operation of membrane filtration alone
(Freeman et al., 2002; Lai et al., 2002; Westrell et al., 2002; Odegaard et al., 1986).
Several researchers demonstrated the efficacy of membrane filtration for effective
removal of natural organic matter (NOM) and disinfection by—product (DBPS) precursors
(Jacangelo 1995; Lainé et al., 1993). Thompson (2001) demonstrated the use of UP

membranes as an effective barrier to microorganisms and organic matter in highly turbid

10

surface water sources with minimal pre-treatment and pre-filtration requirements.
Siddiqui et a1. (2000) found that NF was very effective at removing DBPs from low
turbidity surface water when compared to MF and UF. Similar results were reported by
Mijatovié and collegues (2004) who showed that NF should be used, rather than UP, for
the removal of NOM from lake water or to improve the quality of conventionally treated
water because most of the organic molecules had relatively low molecular weight and
cannot be separated by UF alone. Karakulski et a1. (2001) reported that over 80% of total
trihalomethanes (TTHMs) were retained on the RO membranes. Some of the researchers
found that the efficiency of organic matter removal was around 99% of the total organic
carbon (TOC) on the ceramic membrane and 90% of the TOC on the polymeric
membrane used in these studies (Klomfas and Konieczny, 2004; Konieczny et al., 2002;
Adham et al., 1993, 1991). The UF hybrid process achieved the removal of DBP
precursors and dissolved organic materials, humic materials, fulvic acids and hence
reduced the formation of DBPs (Ngo et al., 2000; Jacangelo et al., 1995). UP is the
preferred option for NOM and DBP removal due to its fairly low pressure and high
permeate flux in comparison to NF and R0 (Bruggen et al., 2003).

Until the recent decade, it was not feasible to use membrane technology for the
production of drinking water because of economic constraints and problems related to
fouling (Cho et al., 2002; Lee et al., 2001). One of the major challenges associated with
the operation of membrane filtration plants is the decrease in the permeate flux due to
membrane fouling (Siedel and Elimelech, 2002; US EPA, 2001; Crozes et al., 1997). The
deposition of NOM on the filter surface is a primary cause of membrane fouling (Fan et

al., 2001; Lee et al., 2001; Nilson and DiGiano, 1996; Ravindran et al., 1993). Fouling

11

not only reduces the efficiency of the membrane, but the characteristics of the foulants
also control the rejection of other substances by the membrane (Schafer et al., 2000). In
drinking water natural organic matter is a major reason for membrane fouling (Fan et al.,
2001; Lee et al., 2001; Carroll et al., 2000; Nilson and Digiano let al., 1997). Fouling can
be controlled either by preventing the particles from reaching the membrane and resulting
in cake formation or by flushing them out using pretreatment options (Milisic, 1986).
Another option for the removal of organic compounds is the addition of powdered
activated carbon (PAC) which reduces irreversible fouling regardless of membrane
hydrophobicity (Lainé et al., 1993).

Several researchers have showed that the permeability rates of ceramic
membranes are superior to the permeability rates of polymeric membranes and are more
effective for the treatment of textile waste water, alkaline and acidic solutions (Weber et
al., 2003). Lee and Cho (2004) compared ceramic membranes with polymeric membranes
for the removal of NOM and found that a ti ght-ultrafiltration ceramic membrane showed
the same potential as a similar nanofiltration polymeric membrane in terms of reducing
the formation of haloacetic acid formation, and was comparable. for the removal of NOM
to polymeric membranes with similar molecular weight cut-off (MWCO). The higher
chemical, mechanical and thermal stability of the ceramic membranes over polymeric
membranes is increasing their popularity for filtration applications (Benfer et a1. 2004).
2.2 OZONATION I
Ozone (03) is a strong and powerful oxidant that preferentially oxidizes electron-rich
moieties containing carbon-carbon double bonds and aromatic alcohols (Bablon et al.,

1991). It oxidizes the naturally occurring organic matter (NOM) that is believed to be

12

largely responsible for the fouling of membranes (Y avich et al., 2004). The reaction with
ozone can be either direct reaction with molecular ozone or indirect one with the
hydroxyl radicals formed by the decomposition of ozone in water at elevated pH (Bablon
et al., 1991). The decomposition reactions are catalyzed by hydroxide ions and NOM,
where OH radicals reacts to dissolved organic compound and accelerate decomposition
of ozone GBablon et al., 1991; Staehelin and Hoigné, 1985). The reaction between OH'
ions and ozone lead to the formation of super-oxide anion radical 02'- and hydroperoxyl
radical HOz'. Reaction between ozone and super oxide anion radical results in the
formation of the ozonide anion radical (03"), which decompOses immediately, giving
rise to the 'OH radical (Acero et al., 1999).

The overall reaction being,
303 + OH' + H+ —>2'OH + 402

The bicarbonate and carbonate play an important role as scavengers of 'OH radicals in
the natural systems. The products of reaction between 'OH radical and carbonate or
bicarbonate ions are passive carbonate or bicarbonate radicals which do not interact
further with ozone or organic compounds (Acero et al., 1999).

The 'OH radical reactions can bring about mineralization of the natural organic
water, reducing the concentration of TOC and hence that of the DBPs precursors. It is a
well known fact that the addition of a strong oxidant such as ozone or hydrogen peroxide
(H202) to water containing humic substances can bring about decolorization of that water
and the degradation of organic matter. The concentration of UV absorbing organic
compounds decreased following ozonation. The dissolved organic carbon content was

also reduced following ozonation (Schlichter et al., 2004; Hashino et al., 2000; Koechling

l3

et al., 1996; Shukairy et al., 1994; Amy et al., 1988; Odegaard et al., 1986). Also
replacing chlorination with ozonation can significantly reduce the concentration of some
of the DBP precursors and the overall DBP-formation potential (Guay, 2005; Schlichter
et al., 2004; Chang et al., 2002; Hu et al., 1999). The permeate flux was also seen to
increase with increase in ozone concentration (Schlichter et al., 2004; Park, 2002). Ma
(1999) showed that the pre-ozonation of the water reduced the formation of the DBPs as
a result of the conversion of the hydrophobic groups contained in the humic acid to
hydrophilic ones.

The ozonation of water containing organic compounds and humic substances
results in formation of ozonation by-products including fOrmaldehyde, aldehydes,
ketones, ketoacids and carboxylic acids (Gracia et al., 1996; Glaze et al., 1989). These
biodegradable compounds can cause problem of microbial regrowth in the distribution
network. Some of the aldehydes like formaldehyde, acetaldehyde are of particular
concern afier being identified as carcinogenic and mutagenic in nature (Richardson,
1998)

2.3 CATALYTIC OZONATION

Another way to accelerate ozonation reaction is use of catalysts. Several metal
oxides and ions have been studied and significant decomposition of targeted compound is
reported. The dissolved ozone is adsorbed on the catalyst surface, where it decomposes to
form 'OH radicals. The refractory character of some of the organic pollutants, such as
dyes, phenols, chlorinated hydrocarbons, pesticides and herbicides, humic substances and
the Iirnited oxidation of the by-products formed, necessitates a more thorough oxidation

by the generation of hydroxyl radicals, using advanced oxidation processes.

14

The use of catalytic ozone resulted in significant decrease in the concentrations of
TOC and COD (Acero et al., 1999; Legube et al., 1999; Paillard et al., 1991). Several
studies showed that catalytic ozonation results in the complete degradation of such
compounds as salicylic acid, which are only very slowly oxidized by molecular ozone
(Legube et al., 1999; Paillard et al., 1991). In the presence of different metal oxide
catalysts, such as iron oxide, manganese oxide, titania, alumina and zirconia, ozone
degrades organic compounds, including saturated carboxylic acids, phenols, aromatic
hydrocarbons, dyes, humic substances and herbicides (Trapido et al., 2005; Ernst at al.,
2004, Beltran et al., 2003a, 2003b; Ni and Chen 2001; Radhakrishnan and Oyarna 2001;
Gracia et al., 2000a, 2000b, 1996). Masten and Davies (1997) reported that the presence
of reactive soil surfaces catalyzed the decomposition of ozone and contaminants sorbed
on the soil. Paillard et al. (1991) reported that TiOz-catalyzed ozonation was more
efficient than ozone alone for the degradation of humic acid. Pines and Reckhow (2003)
showed that ozonation using metal oxides, such as titanium dioxide, cobalt oxide, nickel
oxide, copper oxide and mixed metal oxide of these compounds, was very effective for
the destruction of recalcitrant micropollutants in water as a result of hydroxyl radical
reactions. Beltran et a1. (2005) demonstrated that the use of catalytic ozonation with
homogeneous Fe (III) and heterogeneous (FezO3/A1203) led to total mineralization of
oxalic acid. Similarly Park et al. (2004) showed removal of para-chlorobenzoic acid using
iron type catalyst in combination with ozonation. Huang et al. (2005) showed that the
combined use of ozone and ZnO catalyst generated hydroxyl radicals during the

oxidation process, leading to effective degradation of trichlorophenol in water.

15

2.4 OZONATION-MEMBRANE FILTRATION

Several researchers have attempted to combine ozone with polymeric membranes
with little success. As ozone is a powerful oxidant, polymerictmembranes are prone to
destruction by the ozone (Hashino et al., 2000; Shanbag et al., 1998; Castro and Zander
1995; Shen et al., 1995). Ceramic membranes are ozone resistant and have been shown to
maintain stable permeate fluxes without membrane damage (Schlichter et al., 2004; Kim
et al., 2001, 1999; Allemane, 1993). When attempts were made to combine ceramic
membranes with ozone filtration, researchers found this hybrid process to be more
efficient in eliminating DBPs and other organic contaminants in raw water, without
membrane damage (Schlichter, 2004; Shioyama et al., 2001). Sawada et al. (2001)
observed that ozonation of the water prior to filtration, resulted in the decomposition of
organic matter by the residual ozone in the membrane filtrate, making the organic matter
easily detachable, resulting in a suppression of fouling (Sawada et al., 2001).

The oxidation/membrane filtration leads to an elimination of mineral compounds,
color, turbidity and suspended solids, bad tastes and odors, in addition to the degradation
of natural organic matter, the degradation of toxic micropollutants, the elimination of
trihalomethane precursors, an increase in the biostability of water and the deactivation of
microorganisms (Camel et al., 1998). The ozonation of water containing humic
substances results in the formation of hydroxyl, carbonyl, and carboxyl groups, and
aliphatic and alicyclic ketones (Glaze et al., 1989). Ozonation / membrane filtration can
effectively remove these biodegradable compounds generating biostable water and also

controlling NOM induced fouling (Leiknes et al., 2005).

16

2.5 REFERENCES

Acero, IL. and Gunten, U.V. (1999). "Influence of carbonate on the ozone/hydrogen
peroxide basedadvanced oxidation process for drinking water treatment" Ozone:
Science and Engineering, 22, 305-328.

Adham S.S., Snoeyink,V.L., Clark, M.M., and Bersillon, IL. (1991). "Predicting and
verifying organics removal by PAC in an UF system" Journal of American Water
Works Association, 83 (12), 81-89.

Adham S.S., Snoeyink,V.L., Clark, M.M., and Ansehne, C. (1993). "Predicting and
verifying TOC removal by PAC in pilot-scale UF systems" Journal of American
Water Works Association, 83 (12), 58-68.

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ 03-H202 and 03/T102) for the oxidation of natural organic matter
in water” Ozone: Science and Engineering, 15(5), 419-432.

Amy, G.L., Kuo, C.L., Sierka, RA. (1988). "Ozonation of humic substances: Effects on
molecular weight distributions of organic carbon and trihalomethane formation
potential" Ozone- Science and Engineering, 10, 39-54.

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

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

Beltran, F. J ., Rivas, F. J., and Montero-de-Espinosa, R. (2003a). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts.1. Homogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14), 3210-3217.

Beltran, F. J ., Rivas, F. J ., and Montero-de-Espinosa, R. (2003b). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts. 2. Heterogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research. 42(14), 3218-3224.

Beltran, F. J ., Rivas, F. J ., and Montero-de-Espinosa, R. (2005). "Iron type catalysts for
the ozonation of oxalic acid in water" Water Research 39, 3553-3564.

Benfer, S., Arki and Tomandl, G. (2004). "Ceramic membranes for filtration
applications-Preapration and Characterization" Advanced Engineering Materials,
6(7), 495- 500.

17

Bruggen, B.V., Vandecasteele, C., Gestel, T.V., Doyen, W. and Leysen, R. (2003). "A
review of Pressure-driven membrane processes in waste water treatment and drinking
water production" Environmental Progress, 22(1), 46-56.

Camel, V. and Bermond, A. (1998). "The use of ozone and associated processes in
drinking water treatment" Water Research, 32 (11), 3208-3222.

Carroll, T., King, 8., Gray, S.R., Bolto, BA. and Booker, NA. (2000). "The fouling of
microfiltration membranes by NOM after coagulation treatment" Water Research, 34
(1 1), 2861-2868. '

Castro K. and Zander AK. (1990). "Membrane air-stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

Chang, C.N., Ma, Y.S. and Zing, RF. (2002). "Reducing the formation of disinfection
by- products by pre-ozonation" Chemosphere, 46, 21-30.

Cho, J ., Sohn, J ., Choi, H., Kim, [S and Amy,G. (2002). "Effects of Molecular Weight
Cut off, f/k ratio; and hydrophobic interactions on natural organic matter rejection
and fouling in membranes" Journal of Water Supply Research and T echnology-Aqua,
51, 109-123.

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

Crozes, G.F., Jacangelo, J.G., Anselme, C. and Laine, JIM. (1997). "Impact of
ultrafiltration operating conditions on membrane irreversible fouling" Journal of
Membrane Science. 124(1), 63-76.

Ernst M., Lurot F ., and Schrotter J.C. (2004). "Catalytic ozonation of refractory organic
model compounds in aqueous solution by aluminum oxide" Applied Catalysis B:
Environmental, 47(1), 15-25.

Fan, L.H., Harris, J .L., Roddick, F.A., Booker, NA. (2001). "Influence of characteristics
of natural organic matter on the fouling of microfiltration membranes" Water
Research, 35, 4455-4463.

Freeman, S., Leitner, G.F., Crook, J., and Vernon, W. (2002). "A Clear Advantage"
Water Environment and Technology, 16-21.

Glaze, W.H., Koga, M., Cancilla, D., Wang, K., Mcguire, M.J., Sun, L.A., Davis, M.K.,
Tate, C.H., Aeita, EM. (1989). "Evaluation of ozonation by-products from two
California surface waters" Journal of American Water Works Association, 81(8), 66-
73.

18

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

Gracia, R., Cortes, S., Sarasa, J., Onnad, P., and Ovelleiro, J. L. (2000a). "Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Onnad, P., and Ovelleiro, J. L. (2000b). "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

Guay, C., Rodriguez, M., and Sérodes, J. (2005). " Using ozonation and chlorarnination
to reduce the formation of trihalomethanes and haloacetic acids in drinking water"
Desalination, 176, 229-240.

Guigui, C., J., Rouch, L. Durand-Borlier, V.B., and Aptel., P. (2002). "Impact of
Coagulation conditions on the inline coagulation /UF process for drinking water
production" Desalination, 147 (1-3), 95-100.

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

Hu, Y.U., Wang, Z.S., NG, W.J., and ONG, S.L. (1999). "Disinfection by-products in
water produced by ozonation and chlorination" Environmental Monitoring and
Assessment, 59, 81-93.

Huang, W.H., Fanf, GO, and Wang, CC. (2005). "A nanometer-ZnO catalyt to enhance
the ozonation of 2,4,6- trichlorophenol in water" Colloids and Surfaces A: Physico
chemical and Engineering Aspects, 260, 45-51.

Jacangelo, J.G., Lainé, J.M., Cummings, E.W., and Adham, SS. (1995). "UP
pretreatment for removing DBP precursors" Journal of American Water Works
Association, 87(3), 100-112.

Karakulski, K., Gryta, M., and Morawski, AW. (2001). "Pilot plant studies on the
removal of trihalomethanes by composite reverse osmosis membranes" Desalination,
140 (3), 227-234.

Kim, J .O. and Somiya, I., (2001). "Effective combination of microfiltration and

intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

19

Kim, J. O., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14‘”
Ozone World Congress, Dearborn, MI, 131-143.

Klomfas, G. and Konieczny,K., (2004). "Fouling phenomenon in unit and hybrid
processes for potable water treatment" Desalination, 163, 31.1 -322.

Koechling, M.T., Shukairy, H.M., and Summers, RS. (1996). "Effect of ozonation and
biotreatrnent on molecual rsize and hydrophilic fractions of natural organic water" In

Water Disinfection and Natural Organic Matter, Characterization and Control. ACS
Symposium 649, 196-210.

Koneiczny, K. and Klomfas, G. (2002). "Using activated carbon to improve natural water
treatment by porous membranes" Desalination, 147, 109-116.

Lai, W., Yeh, H.H., Tseng, I.C., Lin, T.F., Chen, J.J. and Wang, GT. (2002).
"Conventional versus advanced treatment for eutrophic source water" Journal of
American Water Works Association, 94(12), 96-108.

Lainé, J .M., Jacangelo, J.G., Cummings, E.W., Carns, K.E., and Mallevialle, J. (1993).
"Influence of bromide on low pressure membrane filtration for controlling DBPs in
surface waters" Journal of American Water Works Association, 85(6), 87-99.

Lee, H., Amy, G., Yoon, J.W., Moon, SH, and Kim, LS. (2001). "Cleaning Strategies
for flux recovery of an ultrafiltration membrane fouled by natural organic matter"
Water Research, 35, 3301-3308.

Lee, S. and Cho, J. (2004). "Comparison of ceramic and polymeric membranes for
natural organic matter (N OM) removal" Desalination, 160, 223-232.

Legube, B. and Karpel Vel Leitner, V. (1999). "Catalytic Ozonation: A promising
advanced oxidation technology for water treatrnen " Catalysis Today, 53, 61-72.

Leiknes, T., Lazarova, M. and Ddegaard, H. (2005). "Development of hybrid ozonation
boifilm membrane filtration process for the production of drinking water" Water
Science and Technology, 51 (6-7), 241-248.

Ma, Y.S. (1999). "Reaction mechanisms for DBPs reaction in hunric acid ozonation"
Ozone: Science and Engineering, 26, 153-164.

Masten SJ. and Davies S.H.R. (1997). "Efficacy of in-situ ozonation for the remediation
of PAH contaminated soil" Journal of Contaminant Hydrology, 28 (4), 327-335.

Mavrov, V.H., Chmiel,H., Kluth, J., Meier, J., Heinrich, P.A., Backes, K. and ster, P.

(1998). "Comparative study of different MF and UP membranes for drinking water
production" Desalination, 117 (1-3), 189-196. '

20

Mijatovié, 1., Matosié, M., Hajduk Cemeha, B., Bratulié, D. (2004). "Removal of natural
organic matter by ultreafiltration and nanofiltration for drinking water production"
Desalination, 169, 223-230.

Milisic, V. (1986). "Antifouling techniques in cross-flow filtration microfiltration"
Filtration and Separation, 23 (6), 347-349.

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

Ni, CH; Chen, J .N. (2001). "Heterogeneous catalytic ozonation of 2-chlorophenol
aqueous solution with alumina as a catalyst" Water Science and Technology, 43 (2),
213-220.

Nilson, J.A. and Di Giano, RA. (1997). "Influence of NOM composition on
nanofiltration" Journal of the American Water Works Association, 88(5), 53-66.

Ngo, H.H., Vigneswaran, S., Kim, SH, Bidkar, A., and Moon, H. (2000).
"Microfiltration-adsorption hybrid system in organics removal from water" Water
Science and Technology, 41 (10), 51-57.

Odegaard, H., Brattebo, H., Eikebrokk, B., and Thorsen, T. (1986). "Advanced
technigues for the removal of humic substances in potable water" Water Supply, 4,
129-158.

Park, Y., (1995). "Effect of ozonation for reducing membrane-fouling in the UP
membrane" Desalination, 147, 43-48.

Park, J .S., Choi, H., and Cho, J. (2004). "Kinetic decomposition of ozone and para-
chlorobenzoic acid (p-CBA) during catalytic ozonation" Water Research, 38, 2284-
2291.

Paillard, H., Dore, M. and Borbigout, M.M. (1991). "Prospects concerning applications
of catalytic ozonation in drinking water treatment" In the proceedings of the 10th
Ozone world Congress (IOA), Monaco, 313-329.

Pines, D.S., and Reckhow, D.A., (2003). "Solid Phase Catalytic Ozonation Process for
the Destruction of a Model Pollutant" Ozone: Science and Engineering, 25(1), 25-39.

Rachwal, A.J. (1995). "Water treatment for public supply in the 1990’s-A role for
membrane technology?" Water Resources Management under drought or water
shortage conditions; In the proceedings of the EWRA 95 Symposium, Nicosia,
Cyprus, 14-18 March 1995, 153-159.

21

Radhakrishnan, R., Oyama, ST. (2001). "Ozone Decomposition over manganese oxide
supported on ZrOz and T102: A kinetic study using in situ Laser Raman
Spectroscopy‘ ’, Journal of Catalysis, 199(2), 282-290.

Ravindran, V., Badriyha, RN. and Pirbazari, M. (1993). "Crossflow membrane filtration
for the removal of natural organic matter" In Proceedings of the Membrane
Technology Conference, Baltimore, MD, 587-599.

Richardson, S. D. (1998). Drinking water disinfection by-products. In Encyclopedia of
Environmental Analysis and Remediation, RA. Meyer, Ed. New York, John Wiley &
Sons, Inc.

Sawada,S., Sumida, I., and Matsumoto, K. (2001). "Membrane filtration of surface water
in the presence of ozone" Water Science and Technology, 1 (5-6), 141-150.

Schafer, A.I., Fane, AC. and Waite, TD. (2000). "Fouling effects on rejection in the
membrane filtration of natural waters", Desalination, 131, 215-224.

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking . water production from
surface water" Desalination, 168, 307-317.

Siddiqui, M., Amy, 6., Ryan, J., and Odem, W. (2000). "Membranes for the control of
natural organic matter from surface waters" Water Research, 34 (13), 3355-3370.

Seidel, A. and Elimelech, M. (2002). "Coupling between chemical and physical
interactions in natural organic matter (NOM), fouling of nanofiltration membranes:
Implications for fouling control" Journal of Membrane Science. 203(1-2), 245-255.

Shanbhag P.V., Guha AK, and Sirkar K.K. (1998). "Membrane-based ozonation of
organic compounds Membrane-based ozonation of organic compounds" Industrial
and Engineering Chemistry Research, 37 (11), 4388 -4398.

Shen Z.S., Semmens M.J., and Collins AC. (1990). "A novel-approach to ozone water
mass-transfer using hollow fiber reactors" Environmental Technology, 1990, 11, 597-
608. .

Shioyama, M., Kawanishi, T., Yokoyama, S., Nuno, M., and Yamamoto, T. (2001).
"Development of advanced ceramic membrane filtration system combined with
ozonation and powdered activated carbon treatment" Water Science and Technology,

1 (5-6), 91-96.
Shukairy, H.M., Miltner, R.J., and Summers, RS. (1994). "Bromide’s Effect on DBP

formation, speciation, and control: Part 1, Ozonation" Journal of American Water
Works Association, 86(6), 72-87.

22

Staehelin, J. and Hoigné, J. (1985). "Decomposition of ozone in water in the presence of
organic solutes acting as promoters and inhibitors of radical chain reactions"
Environmental Science and Technology, 19(12), 1206-1213.

Thompson MA. (2001). "Membrane filtration of high turbidity water sources" Water
Science and Technology: Water Supply, 1 (5-6), 325-330.

Trapido, M., Veressinina, R., Munter, R., and Kallas, J. (2005). "Catlytic ozonation of m-
Dinitrobenzene" Ozone: Science and Engineering, 27, 359-363.

US EPA (2001). "Low-pressure membrane filtration for pathogen removal: Application,
implementation and regulatory issues" Office of Water. Washington, DC. 815-C-01-
001.

Weber, R., Chmiel, H., and Mavrov, V. (2003). "Characteristics and application of new
ceramic nanofiltration membranes" Desalination, 157, 113-125.

Westrell, T., Bergstedt, O. Heinicke, G., and Karrman, E. (2002). "A systems analysis
comparing drinking water systems-central physical-chemical treatment and local
membrane filtration" Water Science and Technology: Water Supply, 2(2), 11-18.

Yavich, A.A., Lee, K.H., Chen, K.C., Pape, L. and Masten, SJ. (2004). "Evaluation of
biodegradability of NOM after ozonation" Water Research 38(12), 2839-2846.

23

CHAPTER THREE

Kamik, Bhavana.S., Davies, Simon.H., Chen, Kuang.C., Jaglowski, Dan.R., Baumann,
Melissa.J. and Masten, Susan.J., (2005). Effects of ozonation and pH on the permeate
flux of nanocrystalline ceramic membranes. Water Research, 39(4): 728-734.

24

CHAPTER THREE

EFFECTS OF OZONATION ON THE PERMEATE FLUX OF
NAN OCRYSTALLINE CERAMIC MEMBRANES.

3.1 ABSTRACT
Titania membranes, with a molecular weight cut-off of 15 kD were used in a
ozonation/membrane system that was fed with water from Lake Lansing, which had been
pre-filtered through a 0.45 pm glass fiber filter. The application of ozone gas prior to
filtration resulted in significant decreases in membrane fouling. The effects of ozonation
could not be explained by physical scouring of the filter cake. Decreases in the pH
resulted in a concomitant increase in the dissolved ozone concentration in the feed water
and in an improvement in permeate flux recovery. Increasing the ozone concentration
beyond a threshold value had no beneficial effect on permeate flux recovery. Ozone
decomposition, resulting in the formation of ’OH or other radicals at the membrane
surface, is thought to result in the decomposition of organic foulants at the membrane
surface and reduce the extent of membrane fouling.
Keywords: ceramic membranes; membrane filtration; ozonation; permeate flux; fouling.
3.2 INTRODUCTION

Membrane filtration is an effective method to remove particles, microorganisms and
organic matter from drinking waters. Compared to conventional treatment methods,

membrane processes i) can provide higher quality water, ii) minimize disinfectant

25

demand, iii) are more compact, iv) provide easier operational control and less
maintenance, and v) generate less sludge (Cleveland, 1999; Nakatsuka et al., 1996).

One of the major challenges associated with the operation of membrane filtration
plants is an increase in the operational cost as a result of the decrease in the permeate flux-
due to membrane fouling (Siedel and Elimelech, 2002; US EPA, 2001; Crozes et al.,
1997). It is important to enhance membrane performance, so that these systems can
become more affordable and reliable.

As ozone has been employed for the intermittent cleaning of membranes in chemical
processing (Kim et al., 1999), it could also be used for membrane cleaning in drinking
water treatment applications. Ozone is a powerful oxidant that preferentially oxidizes
electron-rich moieties containing carbon-carbon double bonds and aromatic alcohols
(Bablon et al., 1991). It oxidizes the naturally occurring organic matter (NOM) that is
believed to be largely responsible for the fouling of membranes (Y avich et al., 2004).
However, few researchers have investigated the combination of ozonation and membrane
processes. This is most likely because polymeric membranes, commonly used in the
water industry, are prone to destruction by ozone (Hashino et al., 2000; Shanbhag et al.,
1998; Castro et al., 1995; Shen et al., 1990). Ceramic membranes, which are ozone-
resistant, in combination with ozonation achieved a high permeate flux without
membrane damage (Schlichter et al., 2004; Chen, 2003; Kim and'Somiya, 2001; Bablon
et al., 1991). In addition, intermittent ozonation is effective in preventing membrane
fouling caused by particle accumulation on the membrane surface (Kim and Somiya,
2001; Verberk et al., 2001; Klijn et al., 2000; Laborie et al., 1998; Laborie et al., 1997;

Cui and Wright, 1994; Bablon et al., 1991; Moulin et al., 1991). Most commercial

26

ceramic membranes are fabricated from metal oxides, such as titania, alumina and
zirconia. These metal oxides are known to promote ozone decomposition and the
formation Of -OH or other radicals (Gracia et al., 2000a; Gracia eta1., 2000b; Legube and
Vel Leitner, 1999; Gracia et al., 1996). Therefore, the reaction of ozone on the
membrane surface is likely to facilitate the removal of organic foulants.

In this work we have considered the effects of ozone gas concentration and pH on the
permeate flux using continuous and pulsed ozonation with membrane filtration. The
effect of scouring on the permeate flux was also studied.

3.3 MATERIALS AND METHODS
3. 3.1 Membrane Filtration System

The ozone-membrane filtration system used in these experiments is illustrated in
Figure 3.1. Tubular ceramic membranes (Clover-leaf (three channels) design, CéRAM
Inside, Tami North America, St. Laurent, Quebec, Canada) with a molecular weight cut-
off 15 kD were used. Teflon tubing, stainless steel joints and stainless steel valves were
. used throughout the system. Other components included: 3.5-liter and 1.5-liter water-
jacked glass reservoirs made of Pyrex glass, and a simple Y inline mixer (Ozone Service,
Burton, BC, Canada). The external diameter of each titania membrane was 10 mm and
the active length was 25 cm. The filter had a total filtering area of 41.2 cmz. The
membranes can be operated in the pH range, from 0-14.

Pure oxygen gas from a gas cylinder was dried using a molecular sieve trap, and then
fed to the ozone generator (Model OZZPCS, Ozotech Inc., Yreka, Calif). The gaseous
ozone concentration was controlled by varying the voltage applied to the ozone

generator. The excess ozone gas was vented by passing the gas through a 2% potassium

27

iodide (KI) solution to destroy any residual ozone gas. A constant temperature was
maintained during the experiment using the external refiigerating and heating bath
circulator (NESLAB RTE-10, Thermo Electron Corp). The transmembrane pressure for
the system was fixed at a value of approximately 0.17 bar to prevent degasification in the
3.5 L reservoir tank.

3. 3.2 Water Source

Experiments were carried out using samples taken from Lake Lansing (Haslett, MI).
Lake Lansing is a borderline eutrophic lake. The typical characteristics of the Lake
Lansing waters are shown in Table 3.1. The samples were collected at the boat ramp at
the Lake Lansing Park-South, Haslett, Michigan in five-gallon carboys and stored at 4°C.
The maximum storage period was seven days. Water samples were pre-filtered through a
0.45-um mixed cellulose ester (Millipore-HA) filters before testing.

Table 3.1
Typical Characteristics of Lake Lansing Water (Haslett, MI) '

Parameters . Lake
TOC 8.6 to 11.6

7 .7 to 8.6

145 to 157
UV-254 abs. 0.160 to 0.180
SDS 240

SDS HAA 75

BDOC 1.0 to 4.1
Nitrate 0.44

Total 0.06
Hardness as 190 to 198

 

“All data reported is obtained Lake Lansing Watershed Advisory Committee Report(1998)
except for SDS T HM and SDS HAA, which were measured as part of this study

b SDS THM and SDS HAA were measured using Standard Method 5710 and USEPA
Method 552.2 respectively.

28

Waste gas

  

 

  

 

 

 

 

 

 

 

 

 

 

Moisture
Adsorbent sPectl'OPI'IOtometer
Ozone
generator “'—
Oxygen ‘ l |
Cylinder I
. Kl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11— < :) A
l><l o o
r\ M Membrane
—® 1
v *
Water I \ h J
Tank 6) Flowmeter
Cb (FD Pressure gauge

 

 

 

 

 

 

g Pump

0 X Valve

 

 

 

 

 

Figure 3.1 Schematic Representation of the Ozone-Membrane Filtration System

29

3.3.3 Gas-phase Ozone Analysis

The concentration of ozone in gas phase was measured at a wavelength of 254 nm
with a Milton Roy Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) using
a 2-mm quartz flow-through cell. An extinction coefficient of 3000 M'lcm'1 at 254 nm
(Hoigné, 198 8) was used to calculate the ozone concentration.
3. 3.4 Membrane cleaning and preparation

Prior to each experiment the membrane was thoroughly cleaned by soaking the
membrane in a sodium hydroxide solution (15 g/L) at 85°C for 30 minutes; following
this, the membrane was rinsed with distilled deionized (DDI) water. Then the membrane
was soaked in a nitric acid solution (0.1M) at 50°C for another 30 minutes followed by
thorough rinsing with DDI water. Finally the membrane was steam sterilized at 121°C for
30 minutes. The effectiveness of the cleaning procedure was verified by measuring the
permeate flux through the membranes using DDI water to ensure that the initial
membrane flux was the same in all experiments.
3. 3.5 Experiments
3. 3. 5.1 Efl'ect of ozone on the permeate flux

For the experiments investigating permeate flux recovery, a centrifugal pump
(Model 4RH12, Dayton Electric Mfg. Co., Niles, Illinois) was used to circulate water
from the 3.5-liter water reservoir through the membrane module. A magnetic stirrer at
the bottom of the reservoir was used to mix the water. The permeate flux was monitored
by measuring the volume of the permeate in a graduated cylinder over specified time
intervals. A peristaltic pump (Masterflex Model 7520-35, Cole-Parmer Co., Chicago,

Illinois) was used to pump the water from the 1.5-liter water tank into the 3.5-liter water

30

reservoir to maintain the constant volume of water in the larger reservoir. The operating
conditions used are listed in Table 3.2. The system was operated without ozone injection
until the permeate flux decreased to 60% of the initial flux. At this time, ozone gas was

inj ectedthrough the Y inline mixer into the feed water stream of the membrane module.

 

 

 

 

 

Table 3.2
Operating Conditions for the Ozone—Membrane Filtration System
Water source Lake Lansing (Haslett, MI)
Water flow rate 2.75 LPM
Water temperature 20°C
Ozone gas flow rate 100 mL/min
TMP 0.17 bar

 

 

 

 

The experiment was repeated for different ozone gas concentrations (1, 1.5, 2.5, 5
and 12.5 g/m3). Also, the experiment was repeated with pulsed ozonation, wherein the
ozone gas was introduced for period of one minute every five minutes.

3. 3. 5.2 Efi'ect of scouring on the permeate flux

The recovery in the permeate flux after ozonation could result fiom the reaction of ozone
with the NOM in the filter cake/gel layer or it could be due to the dislodging of the
cake/ gel as a result of the turbulence caused by the ozone/oxygen bubbles. To determine
the effect of turbulence induced by gas bubbles, the distance between inline Y-mixer and
the membrane module was increased from 10 cm to 50 cm and the decline in permeate
flux was studied. The operating conditions were maintained as in Table 2. Experiments
were conducted using gaseous ozone concentrations of 12.5, 5.0, 2.5, and 1.5 g/m3. An

experiment was also conducted using pure oxygen.

31

3.3.5.3 Effect of the feed water pH on the permeate flux

The decomposition of ozone in solution to form -OH radicals is highly pH dependent
(Masten and Davies, 1994). In order to study the effect of pH on permeate flux recovery,
the natural pH of the Lake Lansing water was reduced using concentrated hydrochloric
acid and the permeate flux recovery was investigated using a gaseous ozone
concentration of 1.5 g/m3, the minimum ozone concentration shown to achieve permeate
flux recovery in previous experiments. The other operating conditions are listed in Table
2. The permeate flux was measured at pH values of 4.0, 6.0, 7.0 and at the natural pH of
the feed water. The system was operated with continuous ozonation.
3.4 RESULTS AND DISCUSSION
3. 4. I Efi’ect of ozone on the permeate flux

The trends in the permeate flux during 12 hours of filtration are shown in Figure 3.2.
As a result of membrane fouling, the permeate flux decreased to about 60 % of its initial
value within 12 hours of operation. After this time, ozonation was begun and continued
for 2 hours. When ozone was applied to the system, the permeate flux recovered to
approximately 95% of the initial permeate flux within a period of 30 to 40 minutes.
Figure 3.3 shows the effect of ozone dosage on fouling. The permeate flux recovered to
100% following ozonation at an ozone concentration of 12.5 g/m3 and 5 g/m3. The
reduction in the ozone concentration from 12.5 to 5.0 g/m3 had no significant difference
in the final permeate flux (p> 0.05). However, when the ozone gas concentration was
further reduced to 2.5 g/m3 there was a significant decrease in permeate flux to about
90% of the initial value (p< 0.05). No permeate flux recovery was observed at an ozone

concentration of 1.0 g/m3. When the permeate flux recovery experiment was repeated for

32

an ozone concentration of 1.5 g/m3, differences in the permeate flux recovery were
statistically insignificant from that obtained at an ozone concentration of 2.5 g/m3 (p>
0.05). Thus, a minimum threshold ozone concentration is required to obtain a permeate
recovery; the greater the ozone concentration, the greater the permeate flux. However,
beyond a certain ozone gas concentration, further increases in ozone concentration do not

affect the permeate flux (as the flux is essentially equivalent to the permeate flux

achieved with pure water).

 

110 I I I

Permeate Flux (%)

 

 

 

 

 

<——g—-— W/O Ozone >§< W/ Ozone ->
so i i n” i i i 4
0 2 4 6 8 10 12 14 16

Time, (hours)

Figure 3.2 Permeate Flux Recovery Pattern with Ozonation.
(Operating Conditions: Table 3. 2, Initial Specific flux l53L/m2h-bar, pH 8. 2,
Gaseous ozone concentration 12.5 g/m’).

33

Permeate Flux PM

110 I l

 

   
   
  
 

5 FE

 

 

 

 

03-1 2.5 glm3
03-5 glm3

9... 03-25 glm3
\ \ 345$ Boa-1.5 glm3

 

 

80 T _
«KM Oxygen
70 - '1 3 a
:llll'-I' [1111] [:1 103'1 9"“
‘-i
60 T _
I” W/O 03 or 02 >< W/ 03 or 02 +
50 l J l l
0 4 8 ' 12 16

Time (hours)

Figure 3.3 Effect of Ozone or Oxygen on Permeate Flux Recovery
(Experimental setup: Fig. 3.1. Operating Conditions: Table 3.2).
Initial Specific flux I53L/m’h-bar, pH 8. 2.)

34

When the experiment was performed by introducing pure oxygen at Operating
conditions in Table 3.2, after 12 hours of operation, the permeate flux was only 72% of
the initial value (see Figure 3.3); also the extent of permeate flux did not improve
significantly under continuous oxygenation, resulting in a final recovery of only 76%.
The results shown in Figure 3.4, which compare continuous oxygenation and continuous
ozonation, suggest that the permeate recovery is largely due to the reaction of NOM with
ozone and/or 'OH (or other) radicals, presumably due to a reduction in the thickness
and/or resistance of the filter cake formed during filtration. An experiment was conducted
to compare the results under continuous and pulsed ozonation. As shown in Figure 3.5,
when ozone gas was injected into the system for two hours in pulses lasting one minute at
an interval of five minutes and when ozone was injected continuously for 2 hours, the
results were similar. This shows that continuous rather than intermittent ozonation could
possibly be used in full-scale operations. This could result in considerable savings in
capital and operating costs.

3. 4.2 Efl'ect of scouring on the permeate flux of the nanofiltration ceramic membrane

I Increasing the distance between the Y inline mixer and the membrane module did
not result in any measurable change in the steady state permeate flux. As shown in Figure
3.5, there is no significant difference in the permeate flux when the in-line mixer was
placed 50 cm from the membrane module (no securing) and when it was placed 10 cm
from the membrane module (scouring). Increasing the distance between the mixer and
the inlet would have reduced the
turbulence effects at the membrane interface. As the permeate flux was not affected by

this change, it appears that the flux recovery is largely due to the reactions of ozone

35

and/or -0H radicals with the NOM, rather than the dislodging of the filter cake due to the

increased turbulence created by gas bubbles in the feed water.

 

 

 

 

 

 

1‘
fl
4': _
“S
E
:I s e e H
g 130 — Bax Continuous Ozonation -
E
o
E
o ‘. _
a 120 T ‘.
‘0 ‘EEL ------ aim- - as ------ w-muaa
Continuous Oxygenation
110 T _
100 L l
0 5 10 15

Time, (hours)

Figure 3.4 Effect of Continuous Ozone or Oxygen on Permeate Flux Recovery
(Experimental setup: Fig. 3.1. Operating Conditions: Table 3. 2.

Continuous Ozonation (Ozone gas concentration 1.5 g/m’) and Continuous Oxygenation, Initial
Specific flux 153L/m2h-bar, pH 8.2.)

36

 

—0—- Continuous Ozonation! No scouring
X Pulse Ozonation

 

 

 

 

 

 

 

 

...... Securing

160 I T ’

155 — —
"g 150 — —
a
5‘ 145 — __
E
O
E
i 140 — ~
or

135 — -

130 1

0 5 10 15

Ozone Dose, (glm’)

Figure 3.5 Effect of Ozone on Permeate Flux Recovery under different
conditions

(Experimental setup: Fig. 3.1. Operating Conditions: Table 3.2. Continuous ozonation/No
Securing: ozone concentrations of 12, 5, 2. 5, 1.5 g/m’ used afterIZ hours operation, Pulsed
Ozonation: Pulse of ozone lasting 1 minute at frequency of very 5 minutes, ozone concentration 12,
5, 2. 5, 1.5 g/rn3 afierIZ hours operation and Scouring-increase distance between Y-line mixer and

37

3. 4.3 Eflect of the feed water pH on the permeate flux
The effects of pH of the feed water on permeate flux recovery are shown in Figure 6.
Unlike experiments discussed earlier, in these experiments, the filter was not fouled
before ozonation was commenced. At the natural pH of the water (pH 7.9) no dissolved
ozone was detected in the feed water, although ozone was supplied at a concentration of
1.5 g/m3. Over the first eight hours of operation, the permeate flux decreased to
approximately 79% of the initial flow (Figure 3.6). Afier 8 hours of operation, the
permeate flux began to recover and reached a steady state value that was 84% of the
initial flux. Similar results occurred at pH 7 .0; again the flux decreased in first eight
hours of operation to approximately 80% of the initial flux. As observed when the pH
was 7.9, the specific flux began to recover after 8 hours of ozonation. In this case, it was
observed that the dissolved ozone concentration also began to increase after 8 hours.
When the pH of the feed water was further decreased to 4.0 and 6.0, the permeate
flux decreased initially, although the recovery occurred earlier (6 hours for pH 6.0 and 4
hours for pH 4.0). The steady state specific flux also increased with decreasing pH. The
extent of recovery was statistically significant when the natural pH (7.9) was reduced to a

pH of 4.0 (p<0.05).

38

Specific Fqu,(LIm’h-bar)

200

190 -

180

170

160

150

140

130

120

 

     
 

pH7

 

 

 

 

s e a
_ Natural pH 7.9 T
l J l l
0 4 8 1 2 16
Time,(hours)

Figure 3.6 Effect of pH on Permeate Flux Recovery
(Experimental setup: Fig. 3.1 Operating Conditions: Table. 3. 2, Continuous Ozonation,
Ozone gas concentration: 1.5 g/m’, pH: natural pH, 7. 0, 6.0 and 4. 0)

39

Dissolved Ozone Concentration, (mg/L)

 

 

   

 

0.16 l l l l
0-14 - ’1 --------- I ----- I -1 pH 4‘
0.12 - ,’ ‘
0.1 - ,’ ‘
0.08 — ,’ ‘
__,-.‘r i if I 9““
0.06 — , “
ni'.-.. . pH7
0.04 -— "
0.02 -/ I ‘

: Natural pH 7.9
o A

A n
V V

 

 

V

0 4 8 12 16
Time, (hours)

Figure 3.7 Dissolved Ozone Concentration in Feed water
(Experimental setup: Fig. 3.1 Operating Conditions: Table3. 2, Continuous Ozonation,
Ozone gas concentration: 1.5 g/m”, pH: natural pH, 7. 0, 6.0 and 4. 0)

40

 

185 l I l I I I

180

175

170

165 '

 

Steady State Flux ,(LIm 2h-bar)

 

 

 

160 I I I H l I
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Steady State Ozone Concentration, (mg/L)

Figure 3.8 Relationship between Steady State Flux and Steady State Ozone
Concentration

As shown in Figure 3.7, the dissolved ozone concentration profiles are related to the
specific flux profiles. Changes in the steady state ozone concentrations and increases in
the specific flux occur at similar time. Higher dissolved ozone concentrations resulted in
higher steady state fluxes, and at pH 4.0, when the dissolved ozone concentration was
0.14 mg/L, the steady state specific flux was nearly 95% of the initial flux. The
relationship between specific flux and dissolved. ozone concentration is seen in Figure
3.8. A minimum ozone concentration of 0.05 mg/L in the recirculation feed water is
required to obtain steady permeate fluxes that are >95% of the initial value. Thus, it

appears that membrane fouling can be effectively reduced and stable fluxes can be

41

maintained throughout the operation if the concentration of dissolved ozone is greater
than 0.05 mg/L in the recirculation water.

We hypothesize that the increase in dissolved ozone concentration results in an
increase in the concentration of the -OH or other radical species at the membrane surface.
These very reactive species decompose the organic foulants present at the surface and
bring about the degradation of the cake/ gel layer, thereby reducing the membrane fouling
and increasing the permeate flux.

3.5 CONCLUSIONS

The permeate flux through a titania coated ceramic membrane was significantly
affected by ozonation and the pH of the feed water in the system. Increase in the ozone
gas concentration resulted in a concomitant increase in the permeate flux. Also, beyond a
particular ozone gas concentration (5 g/m3), further increases in gaseous ozone
concentration had no effect on the level of permeate flux recovery, thus a minimum
threshold concentration could achieve complete recovery. The permeate flux recovery in
these experiments was found to be due to the reaction of ozone with potential foulants
and not the increased turbulence created at the membrane surface by gas bubbles. The pH
of the feed water affected the permeate flux. As expected, the dissolved ozone
concentration increased with decreasing pH. The improved permeate flux recovery at
lower pH may be due to the higher concentration of dissolved ozone that is present under
these conditions, resulting in a greater extent to which the ozone reacts with the NOM

species that are responsible for fouling.

42

3.6 ACKNOWLEDGEMENTS

The authors would like to thank the US Environmental Protection Agency (US EPA)
Science To Achieve Results (STAR) Program (Grant No. RD830090801) for financial
support of this work. We would also like to thank Dr. Vladimir Tarabara for reviewing

this manuscript and for his helpful comments.

43

 

3.7 REFERENCES

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

Castro, K. and Zander, AK (1995). "Membrane air-Stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

Chen, KC. (2003). "Ozonation, ultrafiltration, and biofiltration for the control of NOM
and DBP in Drinking Water", Ph.D. dissertation, Michigan State University, East
Lansing, MI 48824.

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

Crozes, G.F., Jacangelo, J.G., Ansehne, C. and Laine, J.M. (1997). "Impact of
ultrafiltration operating conditions on membrane irreversible fouling" Journal of
Membrane Science, 124(1), 63-76. I

Cui, Z.F. and Wright, K.I.T. (1994). "Gas-Liquid 2-phase cross-flow ultrafiltration of
BSA and dextran solutions" Journal of Membrane Science. 90(1-2), 183-189.

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

Gracia, R., Cortes, S., Sarasa, J ., Onnad, P., and Ovelleiro, J. L. (2000a). "Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Onnad, P., and Ovelleiro, J. L. (2000b). "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

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

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

Kim, J. 0., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14m
Ozone World Congress, Dearborn, MI, 131-143.

44

r—

Kim, J .0. and Somiya, I., (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1), 7-15.

Klijn, R.B., van der Meer, W.G.J., Vriezen, H., and van Ekkendonk, EH]. (2000).
"Surface water treatment with zenon microfiltration membranes: minimisation of
energy and chemical use" In the Proceedings of the Membranes in Drinking and
Industrial Water Production, L'Aquila, Italy, 557-563.

Laborie, S., Cabassud, C., Durand-Bourlier, L. and Laine, J.M. (1997). "Flux
enhancement by a continuous tangential gas flow in ultrafiltration hollow fibres for
drinking water production: Effects of slug flow on cake structure" Filtration &
Separation 34(8), 887-891. ,

Laborie, S., Cabassud, C., Durand-Bourlier, L. and Laine, J .M. (1998). "Fouling control

by air sparging inside hollow fibre membranes - Effects on energy consumption"
Desalination 118(1-3), 189-196.

Lake Lansing Watershed Advisory Committee. (1998). Progress Report, Ingham County
Drain Commissioner’s Office, Mason, Michigan.

Legube, B. and Karpel Vel Leitner, N. (1999). "Catalytic ozonation: A promising
advanced oxidation technology for water treatment" Catalysis Today, 53, 61-72.

Masten, SJ. and Davies, S.H.R. (1994). "The use of ozonation to degrade organic
contaminants in wastewaters" Environmental Science & Technology, 28(4), A180 —
A1 85.

Moulin, C., Bourbigot, M.M., Taizi-Pain, A. and Faivre, M. (1991). "Potarrilisation of
surface waters by crossflow ultra- and microfiltration on mineral membranes: Interest

of ozone" In the Proceedings of the Membrane Technologies in the Water Industry,
Orlando, Florida, 729-737.

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

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

Shanbhag P.V., Guha A.K., and Sirkar K.K. (1998). "Membrane-based ozonation of

organic compounds Membrane-based ozonation of organic compounds" Industrial
and Engineering Chemistry Research, 37 (11), 4388 -4398.

45

l—4

CHAPTER FOUR

Karnik, Bhavana 8.; Davies, Simon H.; Baumann, Melissa I; Masten, Susan J. (2005).
The effects of combined ozonation and filtration on disinfection by-product formation.
Water Research, 39(13): 2839-2850.

47

CHAPTER FOUR

THE EFFECTS OF COMBINED OZONATION AND FILTRATION
ON DISINFECTION BY-PRODUCT FORMATION

4.1 ABSTRACT

The effects of combined ozonation and membrane filtration on the removal of the natural
organic matter (NOM) and the formation of disinfection by-products (DBPS) were
investigated. Ozonation/filtration resulted a reduction of up to 50% in the dissolved
organic carbon (DOC) concentration. Furthermore, humic substances were converted to
non-humic substances, with changes in the humic and non-humic substance
concentrations of up to -50% and +20%, respectively. Ozonation/filtration resulted in the
formation of partially oxidized compounds from NOM that were less reactive with
chlorine, decreasing the concentration of simulated distribution system total
trihalomethanes (SDS TTHMs) and simulated distribution system halo acetic acids (SDS
HAAs) by up to 80% and 65%, respectively. Reducing the MWCO of the membranes
resulted in reductions in the concentrations of SDS TTHMs and SDS HAAs. Using a
membrane with a 5kD MWCO, the minimum gaseous ozone concentration required to
bring about effective NOM degradation and meet regulatory requirements for chlorinated
DBPs was 2.5 g/m3.

Keywords: ceramic membranes, nanofiltration, ultrafiltration, ozonation, disinfection by-

products (DBPs), water quality, natural organic matter (NOM).

48

4.2 INTRODUCTION

Natural organic matter (NOM) is composed of a heterogeneous mixture of organic
compounds that can be of human origin or the result of natural processes. NOM can be
broadly divided into two fractions: humic substances, which are composed of fulvic and
humic acids, and non-humic substances, which include carbohydrates, lipids, and amino
acids.

In water treatment systems, the presence of NOM is a cause of concern because of its
reaction with disinfectants. Chlorination of drinking water results in the formation of
disinfection by-products (DBPS), such as trihalomethanes (THMs), some of which are
known carcinogens (Morris et al., 1992; Mughal, 1992; Kool et al., 1985). While humic
substances have been recognized as the primary precursors of chlorination byproducts
(Ichihashi et al., 1999; Manahan, 1993; Reckhow et al., 1990; Collins et al., 1986), non-
humic substances also result in the formation of many regulated or pctentially regulated
DBPs. The non-humic fraction of the NOM is generally more biodegradable and, as such,
supports bacterial regrowth in water distribution systems (Y avich, 1998; Mogren et al.,
1990)

The use of ozonation in water treatment processes results in a decrease in the
formation of THMs and halo acetic acids (HAAs) upon subsequent chlorination (Zhang
et al., 2001; Richardson etal., 1999). Increases in ozone dosages result in a concomitant
decrease in the concentrations of THMs and HAAs formed from subsequent chlorination
(Lee, 2001; Cipparone, 1997; Amy et al., 1988). Ozonation results in the formation of
more polar compounds and an increase in the biodegradability of the chemicals found in

the water as compared to that generated with chlorination (Koechling et al., 1996; Owen

49

et al., 1995; Amy et al., 1992). The reactions that occur during ozonation produce by-
products, including aldehydes (formaldehyde, glyoxal, and methylglyoxal), ketones,
gyloxylic acid, and pyruvic acid (Paode et al., 1997; Weinberg and Glaze, 1996). Some
of these by—products are of particular concern due to their mutagenicity and
carcinogenicity (Bull and McCabe, 1984). Also, as they are easily biodegradable, they
can serve as substrates for microbial regrowth in the distribution system. The ozonation
by-products can be easily removed by biofiltration (Y avich and Masten, 2003; Griffini et
al., 1999).

Membrane filtration is an effective method to remove particles, microorganisms and
organic matter from drinking waters. Compared with conventional treatment methods,
membrane processes i) can provide higher quality water, ii) minimize disinfectant
demand, iii) are more compact, iv) provide easier operational control and less
maintenance, and v) generate less sludge (US EPA, 2001; Cleveland, 1999; Nakatsuka et
al., 1996).

One of the major challenges associated with the operation of membrane filtration
plants is the decrease in the permeate flux due to membrane fouling (Siedel and
Elimelech, 2002; US EPA, 2001; Crozes et al., 1997). The deposition of NOM on the
filter surface is a primary cause of membrane fouling (Lee et al., 2001; Nilson and
DiGiano, 1996; Ravindran et al., 1993). Fouling not only reduces the efficiency of the
membrane, but the characteristics of the foulants also control the rejection of other
substances by the membrane (Schafer et al., 2000). The application of ozonation prior to
membrane filtration reduces membrane fouling and enhances permeate flux (Karnik et

al., 2005; Schlichter et al., 2004, 2003; Hashino et al., 2000; Hyung et al., 2000; Kim et

50

al., 1999). The use of ozonation in combination with membrane processes has not been
extensively investigated; however, the limited research in this area has shown that
ceramic membranes in combination with ozonation achieved a high permeate flux
without membrane damage (Schlichter et al., 2004; Chen, 2003; Kim et al., 1999; Kim
and Somiya, 2001; Moulin et al., 1991; Bablon et al., 1991).

In this study, we have investigated the quality of water after combined ozonation-
membrane filtration. The permeate collected was used to determine the effect of
treatment on the UV absorbance measured at 254 nm (UV -254), DOC, humic substances
(HS), and non-hurrric substances (non-HS). The concentrations of SDS TTI-IMs, SDS
HAAs, aldehydes, ketones, and ketoacids were also evaluated. The effect of gaseous
ozone concentration on the water quality of the permeate was investigated.

4.3 MATERIALS AND METHODS
4. 3. I . Ozonation/ Membrane Filtration

A schematic representation of the ozonation/membrane system is shown in Figure
4.1. Tubular ceramic membranes (Clover-leaf design (containing three channels),
CéRAM Inside, Tami North America, St. Laurent, Quebec, Canada) with molecular
weight cut-offs of 15, 5, and 1 kD were used. The external diameter of each titania
membrane was 10 mm and the active membrane length was 25 cm. The membrane had a
total filtering area of 41.2 cmz. A stainless steel filter holder, Teflon® tubing and
stainless steel or Teflon® joints and valves were used throughout the system. Other
components included: 3.5-liter and 1.5-liter water-jacked glass reservoirs made of Pyrex

glass, and a simple Y inline mixer (Ozone Service, Burton, BC, Canada). Ozone gas

51

was added into the water stream through the simple Y inline mixer just before entering

the membrane module.

     
    
 

 

 

 

 

 

 

 

 

 

 

 

 

Waste gas
MOISture Spectrophotometer
Adsorbent
Ozone
generator 1 [—7
Oxygen ‘ I
C linder '
Y .‘ KI ‘

 

 

Membrane

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water
Tank ® Flowmeter
® Pressure gauge

 

 

 

 

 

 

 

Figure 4.1 Schematic Representation of the Ozone-Membrane Filtration System

52

To generate ozone, pure oxygen gas (99.999%) from a pressurized cylinder was dried
using a molecular sieve trap, and then fed to the ozone generator (Model OZZPCS,
Ozotech Inc, Yreka, Calif). Varying the voltage applied to the ozone generator
controlled the gaseous ozone concentration. The excess gas was vented after passing the
gas through a 2% potassium iodide (KI) solution to destroy any residual ozone gas. The
water level in the 3.5-liter reservoir was maintained at a constant level during the
experiments using a peristaltic pump (Masterflex Model 7520-35, Cole-Parmer Co.,
Chicago, Illinois) to transfer water from the 1.5-liter reservoir into the 3.5-liter reservoir.
A constant water temperature of 20°C was maintained using a recirculating water bath.
The operating conditions used are given in Table 4.1. The gaseous ozone concentration

was 2.5 g/m3, unless otherwise stated.

 

 

 

 

Table 4.1 . _
Operating Conditions for the Ozone -Membrane Filtration System
Water recirculation rate 2.75 LPM
Water temperature 20°C
Ozone gas flow rate 100 mL/min
T MP 0.2 bar

 

 

 

 

The experiments were performed with membrane cross flow velocity of 0.6 m/s; the
flow was turbulent with Reynolds number in the range of approximately 6000. Previous
studies in our laboratory considered such important factors, as gas flow rate, water flow
rate, and the characteristics of the source water, which influence the ozone transfer
efficiency (Chen, 2003). The ozonation/membrane system used in this study can achieve

high ozone mass transfer, and thus, requires a lower ozone dose, gas flow rate, and water

53

flow rate than comparable systems (Chen, 2003). The volumetric mass transfer
coefficient for ozone in the experimental setup was determined to be 0.138 rnin'1 (Chen,
2003)

Ceramic membranes with molecular weight cut-offs of 15, 5, and 1 kD were used.
The specific flux for 15, 5 and 1 kD membranes were 60, 20 and 8 L/mz-bar,
respectively. The permeate flux recovery trends are discussed in detail in our earlier work
(Karnik et al., 2005). The conductivity remained practically unchanged for the duration
of the experiment (< 0.01 uS/cm).

Permeate samples were collected in bottles covered with Parafilm® and stored in an
ice-bath for the duration of the experiment. The first 400 mL of permeate collected was
labeled as P1 and latter 1000 ml as P2. P1 and P2 samples were collected to study the
effect of ozone contact time on the water quality.

Samples of the pre-filtered raw water (FRW), P1, P2 and from the 3.5-liter water tank
reservoir (WT) were analyzed for UV-254 absorbance, dissolved organic carbon (DOC),
humic substances and non-humic substances, chlorine residual, SDS total
trihalomethanes (SDS TTHMs), SDS halo acetic acids (SDS HAAs), aldehydes, ketones
and ketoacids. The effect of gaseous ozone concentration on water quality was
investigated using a membrane with a MWCO of 15 kD. The gaseous ozone
concentration was varied between 1.5 and 10 g/m3.

To study the effect of pH on the process, the pH of Lake Lansing water (initial pH
8.2) sample was adjusted to 7.0 by the addition of hydrochloric acid (concentrated, ACS

reagent grade). We chose a pH of 7 because earlier studies revealed no difference in the

54

’penneate flux recovery at pH 6 and pH 7 (Karnik et al., 2005). The membranes used in
these experiments had MWCOs of 5 and 15 kD.
4. 3.2 Water Source

Experiments were carried out on samples taken fiom Lake Lansing (Haslett, MI). The
typical characteristics of the water from Lake Lansing, a borderline eutrophic lake, are
given in Table 4.2. The samples were collected at the boat ramp at the Lake Lansing
Park-South, Haslett, MI in five-gallon polyethylene carboys and stored at 4°C. The
maximum storage period was seven days. Water samples were pre-filtered through a

0.45-um mixed cellulose ester (Millipore-HA) filter before testing.

 

 

 

 

 

 

 

 

 

 

 

Table 4.2

Typical Characteristics of Lake Lansing Water (Haslett, MI) '
Parameters Lake Lansirg
TOC(mg/Q 8.6 to 11.6

H 7.7 to 8.6

Alkalinity (mg/L as CaCO3) 145 to 157
UV-254 (abs) 0.160 to 0.180
SDS THMsb (pg/L) 24o
SDS HAAsb (pg/L) 75
BDOC (mg/L) 1.0 to 4.1
Nitrate (mg/L) 0.44
Total Phosmhate (mg/L) 0.06
Hardness (mg/L as CaC03) 190 to 198

 

 

 

 

“All data reported is obtained fi'om the Lake Lansing Watershed Advisory Committee

Report(1998) except for SDS T HMs and SDS HAAs, which were measured as part of this
study

b SDS THM and SDS HAA were measured using Standard Method 5710 and USEPA Method 552.2
respectively.

4. 3.3 Membrane cleaning and preparation
Prior to each experiment, the membrane was thoroughly cleaned using a procedure
based on that deveIOped by Xing et al. (2003). Membranes were soaked in a sodium

hydroxide solution (15 g/L) at 85°C for 30 minutes; following this, the membrane was

55

rinsed with distilled deionized (DDI) water. The membrane was then soaked in a nitric
acid solution (0.1M) at 50°C for another 30 minutes followed by thorough rinsing with
DDI water. Finally, the membrane was steam sterilized at 121°C for 30 minutes. The
effectiveness of the cleaning procedure was verified by measuring the permeate flux

through the membranes using DDI water to ensure that the initial membrane flux was the

same in all experiments.
4.3.4 Analytical Methods

Gas-phase Ozone Analysis

The absorbance of ozone in the gas phase was measured at 254 nm with a Milton Roy
Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) using a 2-mm path
length quartz flow-through cell. An extinction coefficient of 3000 M'lcm'1 (Hoigné,
1988) was used to calculate the ozone concentration.

UV-254 absorbance
The UV absorbance of the water samples was measured at a wavelength of 254 nm

with a Milton Roy Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) using

a 1 cm quartz cell.
Dissolved Organic Carbon (DOC)

DOC was analyzed using an OI Analytical Model 1010 analyzer. The TOC analyzer
uses the UV/persulfate method (Standard Method, 1998). To ensure the reliability of the
method, standards having TOC concentrations of 2.5, 5, 7, 10 mg/l (01 Analytical) were

run and samples were analyzed in triplicate. A blank Was also run with every set of

samples.

56

Humic substances and Non-humic substances

The humic substances in the samples were isolated from the water samples by
adsorption on XAD-8 resin according to Method 5510C (Standard Methods, 1998). A
100 mL sample was acidified with concentrated phosphoric acid and eluted through a 10
mm diameter (ID) x 15 cm long column at a flow rate of 2 mL/min. The effluent from the
column was collected and then analyzed for TOC, which represented the non-humic
fraction of the dissolved organic matter in the water sample. The resin-packed column
was then back eluted with 100 mL of 0.1 N sodium hydroxide at a flow rate of 2 mIJmin.
The eluent was collected and acidified with concentrated phosphoric acid to a pH less
than 4, purged with high-purity helium for 3 minutes to remove inorganic carbon, and

analyzed for TOC. The organic content of the eluent represented the concentration of

humic substances.
Chlorine residual

Chlorine residual was measured using the iodometric method, Method 4500B
(Standard Methods, 1998).

SDS total trihalomethanes (SDS TIHMs) and SDS halo acetic acids (SDS HAAs).

Water samples were dosed with a chlorine concentration that ensured a residual
chlorine concentration in the range of 0.5 to 2 mg/L after 48 hours incubation at room
temperature according to the procedures in Standard Method 2350 (Standard Methods
1998). The trihalomethane (TI-1M) compounds, chloroform (CHCI3),
bromodichloromethane (CHBrClz), dibromochloromethane (CHBr2Cl), and bromoform

.(CHBr3), were extracted from the water samples using hexane and analyzed by gas

chromatography (Method 5710, Standard Methods 1998). A Perkin Elmer Autosystem

57

gas chromatograph (Perkin Elmer Instruments, Shelton, CT) equipped with an electron
capture detector (ECD), an auto sampler, and a 30 m X 0.25 mm ID, 1 pm DB-Sms
column (J &W Scientific, Folsom, CA) was used for the analysis. The oven temperature
was ramped from 50 °C to 150 °C at a rate of 10°C/min. The flow rate of the carrier gas
(N2) was 12.0 mL/min. The injector temperature and detector temperature were 275 and
350 °C, respectively.

SDS HAAs were produced by chlorination as described above. The concentrations of
monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), dichloroacetic acid
(DCAA), bromochloroacetic acid (BCAA), trichloroacetic acid (TCAA), and
dibromoacetic acid (DBAA) were determined using US EPA Method 552.2. A Perkin
Ehner Autosystem gas chromatograph (Perkin Elmer Instruments, Shelton, CT) equipped
with an ECD, an autosampler, and a 30 m X 0.32 mm ID, 3 pm DB-l column (J&W
Scientific, Folsom, CA) was used for the analysis. The oven temperature was
programmed to hold for 15 minutes at 32°C, then increased to 75 °C at a rate of 5°C/min '
and held 5 minutes, then increased to 100 °C at a rate of 5°C/min. The carrier gas flow
(nitrogen) was 1.0 mI/min with the injector temperature and detector temperatures at 200
°C and 260 °C, respectively.

Aldehydes, Ketones and Ketoacids

USEPA Method 556 (Munch et al., 1998) was used to monitor for formaldehyde,
propionaldehyde, glyoxal, methyl glyoxal, acetone, and 2-butanone, ketomalonic acid,
pyruvicacid and gyloxylic acid. A Perkin Elmer Autosystem gas chromatograph (Perkin
Elmer Instruments, Shelton, CT) equipped with an ECD, an autosampler, and a 30 m X

0.25 mm ID, 0.5 pm DB-Sms column (J&W Scientific, Folsom, CA) was used in the

58

analysis. The oven temperature was programmed to hold for 1 minute at 50 °C, then
increased to 220 °C at a rate of 4°C/min followed by an increase to 250 °C at a rate of 20
°C/min with a 5 minute hold time. The carrier gas flow was 1.0 mL/min and the injector
temperature and detector temperatures were 180 °C and 300 °C, respectively.

4.4 RESULTS AND DISCUSSION

All data is reported as a percent decrease as compared to the concentrations present in
the raw feed water. The results found for the feed water are given in Table 4.3.

4. 4.1 Eflect of ozonation, ultrafiltration, and ozonation-ultrafiltration on water quality.

A study was conducted to compare the improvements in water quality results
achieved using ultrafiltration, ozonation, and ozonation-ultrafiltration. The apparatus
illustrated in Figure 4.1 was used for all three experiments. In the case of the ozonation
experiment, the membrane filter element was removed and the permeate collection ports
were sealed. In the ozonation experiment, samples were collected from the 3.5 L
reservoir after the same time as used for sampling in the ozonation-UP experiment. As
shown in Table 4.4, ultrafiltration was the least effective of the three processes for the
removal of DOC, HS, NHS, SDS TTHMs, and SDS HAAs. The quality of the treated
water was further improved when ozonation and ultrafiltration were combined. Not only
was the removal of UV-254, DOC, HS, NHS, SDS TTHMs, and SDS HAAs enhanced
over either of the processes used alone, but also the combined process resulted in the
production of lower concentrations of aldehydes, ketones, and ketoacids than ozonation

alone. This suggests a synergy between ozonation and membrane filtration in providing

high quality water.

59

5.35. a}: S Vufluouufi Saw—Susanne 3983‘ on: S ~20;wa 3.95 a 333.2% .835 one: u

Sun—ES 5 5: babe—S guns 5.; “REESE. 395% 6 n = .
SEEKS 5 5E hang «on» 5...: SEEKS Scuba: an: -
REESE DRESS» a 332— :cnuhnuueoo 35.0: S 32.83.. was was?» 3k

 

 

 

 

 

 

 

 

 

60

 

 

 

 

 

 

 

 

 

 

 

 

 

£6 a a: 8.0 I a: as I So as I So «to I So 3 a S 3 a S 3 a 2 She Essex
2 a Na 2 a as Z In em 2 a a Z a 3a 3 a 3m 2 a 3a «A a ”am fine moaosoxéafiuz
f « ta ..1 a Eu 2... a new as a as S. 4 new 3 a a; S. a 2» MS. a 2» fine 3: wow
R a on” 3 a cm 98 a gen 98 a 2% can a was we a one as a $2 nfi « one :53 AEEIH mom
a; a 22 Re a am one a a.” one a SN 85 “a“ 3d I am... e .o a one. a; a 83. SEE mmz
......o A S no a G :3 a ow 33 a 9m :3 a on :8 a n: 28 a n5. #5 4 2;. SEE m:
as 3.: 3.3%: S 2.2 212: $3.2 3 «2: S 3.2 S 5.2: SEE 009
was in Ed Sod a :3 83 a 4: .o 83 a E .o 83 a £3 82. 4 Ed 83 a 2.3 59.0 a :3 Es $95
a.» a 32...... an 2 8 S. E. .222. 2 2 a.» E. :23:
e. m _ e. 2 ms _ ma _ S e: _ e. m _ 9. m. 0.555%”.—
..In we «ovum .82. 2.3: no “outm— .OU>)E he tuba ,
. .. . c :25 as: E: 2.22:... . n

 

 

.wEmneA 315 SP: .33.; 3a.. 1235 .«c meet-228:8 3::—

MJ. oEfiH

...fitaomEcu 943.3 Eonahzuunou .388 as v.6
.83.... NR VS. ~n~ 3. 2.; NR 3% fix .... 982...... £8688. R5. ...»..SS. .n%§§.u\o washabzmonou m: .

.3 2% 8.3 a B. 28355 .3...
.~ .5. “fig 3......fizou 3.2ufitmmxm .~ .V 333% 3.3% 3.2%....axmu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

..o a 3: 3.. 3.8. ..c a .... ..m a 32 we a 3.8 2 a 2 3 «... 3.3 8.893:
2 a 3.. 2: a «.8 3 « we 3 a ...: 3 a as. 2 a «:8 ..N a 3. Sue 883.82.82.
.... a 3m 3‘. a S: .3 a N... E a 5.. no a 2: I. a 3” N... a N... Sue 85.. RE
m .354 m. #8. ...“ ...: Z «3. 2 3.2 ... a? m... «32 3838...... mam
E a v.8- ed «3.- q. a 3.. .... a 3..- e... a .....- e... a ..N- a... u. 3a 38281.8:de mmz
. o... a .3“ .... a do. 3. a N... 3 « 3:. 3 a. E. ..N a ...m as a “S. 3.95 m;
3 a Em .2 a ..9 o. u. n: a... a 8.8 3. « SN 8.. a m... S a 2: 3.85 won
3. a a... mm « ...: N. a 33 .... « a... N... a N3 ... a 3:. so... a E... .23 .35:
57.83....ch 53339 b: kD....ouuncuO 532.30 RD 8......» =5...— «Eunice
€030.63. o\ov m 83.50.. €28..qu $1838.89 3M... 9.82555—
ancaacnc 62.553

...... .28... RD .28... 3:330 ...... 9.82:2...— oocagctom .... ..cm_....._=.o0

v... 03.;

61

The effects of ozonation time on the removal efficiencies can be observed by comparing
the results for permeate 1 and 2. The longer ozone contact time did not result in a large
increase in the removal efficiency for UV-254 (65.2 vs. 70.1%), suggesting that most of
the. UV-254 absorbing material were degraded in the time necessary to collect the first
400 mL of sample (i.e., within 4-5 hours). On the contrary, the removal efficiencies of
DOC, SDS TTHMs, and SDS HAAs for permeate 2 were roughly twice that for permeate
1, indicating that the reaction of ozone with TTHM and HAA precursors is slower than
that for ozone with the UV-absorbing material.
4. 4.2 Eflect of membrane M WCO on the water quality

As shown in Figures 4.2 and 4.3, there is a little difference in the DOC levels in the
P1 samples for all three membranes. However, for the P2 samples, there is a statistically
significant (p<0. 05) decrease in the DOC levels when the molecular weight cutoff of the
membrane was decreased from 5 kD to l kD. One explanation for this result is that after
extended ozonation a significant fraction of the DOC is in the molecular size range from
1-5 kD. Ozonation of NOM is known to result in a decrease in the molecular weight of
the organic matter (Mellema 1998), which would then result in these compounds passing
through the coarser membranes, but not through the 1 kD membrane. An alternate
explanation is that the 1 RD membrane is a more effective catalyst for the degradation of
NOM than are the coarser membranes, presumably because the smaller pores have a
greater surface area. If this is the case, then with the 1 kD membrane prolonged

ozonation could result in the mineralization of a greater portion of the NOM to C02 and

water.

62

As shown in Figures 4.2 and 4.3, for the 5 and 15 kD membranes, the molecular weight
cut off of the membranes did not have a statistically significant effect on the UV-254 of
the P1 and P2 samples (p>0.05). Also, as the results for the P1 and P2 samples were not
very different, increasing the ozone contact did not lead to a great increase in the removal
of UV absorbing substances. Even after extensive ozonation, approximately 15% of the
UV-254 absorbance of P2 samples remains in the samples, suggesting that while most of
the UV absorbing substances react with ozone, there is a recalcitrant fraction that. does
not react with ozone.

The data presented in Table 4.4, shows that the removal’ of UV-254 was much
greater with ozonation than with ultrafiltration. This suggests that the removal of the
UV-254 absorbing compounds is predominately due to the reaction of ozone with these
substances and not due to filtration. These results are consistent with previous research
on the ozonation of Lake Lansing water. Yavich and Masten (2003) found that ozone
reacts rapidly with aromatic fraction of the NOM, resulting in a significant decrease in
UV-254 even at low ozone dosages. These workers also found that after this initial
decrease increasing the ozone contact time did not lead to a further large decrease in the
UV-254.

Only with the 1 kD membrane, did extended ozonation result in an increase in the
removal of UV-254. The lower removal of UV-254 absorbance in P1 samples with the 1
kD membrane (as compared to that with the 5 and 15 kD membranes) cannot be
explained by differences in the seasonal nature of the NOM, since replicate experiments

(for the 1 kD membrane) were conducted in May and November, yielding consistent

results.

63

 

 

 

015 kD

I5kD

D1kD

 

 

 

    

"ALI/vii.
Vivi/’3

              
      

       
  

 

3, Membrane Size:

 

   

       

    

 

 

 
        
   

m
52/ mew.” m ”annex...“

5
a

s

H. flit”...
m m //////
1 N I
D

wwwwwwwwwwwwwm

  

H8
H8

I

C““““‘OOOOO““‘A

nunuunnaaunuuunuunnz

I
1
/

  
 

 

    
   

     

;/
I
/

/

W254
W254

      
   

 

 

 

     

‘O‘OJ‘A

A
”wwwww,

DOC
DOC

 

 

 

 

Membrane Size: 15,

I

3

64

 

 

 

 

 

100.00

 

l 5, 5 and 1 ID. All values are average of triplicates within experiments and duplicate experiments.

Experimental setup: Fig. 4. 1 Operating Conditions: Table 4. 1. Ozone Dose: 2.5 g/m
The values have a maximum std deviation of 5 %

Figure 4.2 Effect of molecular weight cut-off on Permeate 1

(first 400 mL of the sample)

 

 

q q u

l _ . _ m m m m m w w
o 0 0 . . . . .
W W W 0 0 0 0 5 0 5 0 5 0
5 . 5 0 5 . 0 7 5 2 2 5
w 1 . .

7 2 9.. & 3.3.03 .x.

33.88 .\.

5 and I kD. All values are average of triplicates within experiments and duplicate experiments. The

values have a maximum std deviation of 5 %

 

Experimental setup: Fig.4. I Operating Conditions: Table 4.1.0zone Dose: 2.5 g/m

Figure 4.3 Effect of molecular weight cut-off on Permeate 2

(Latter 1000 mL of the sample)

 

 

It is possible that with the 1 RD membrane there are catalytic reactions that produce UV
absorbing compounds that pass through the membrane. It should also be noted that as the
molecular weight cutoff of the membranes is reduced, the permeate flux decreased, thus,
the ozone contact time increased, as the time required to collect equal volumes of sample
increases. While we plan to continue to investigate this phenomenon, one should note
that this result does not negate our conclusions regarding the overall effectiveness of the
hybrid process in reducing the concentrations of disinfection byproducts.

With all three membranes, ozonation/membrane filtration resulted in a reduction of
approximately 45% in the humic substance (HS) concentration in the P1 samples and an
approximately 55% reduction in the P2 samples (see Figures 4.2 and 4.3). This reduction
is, in part, due to the reaction of NOM with either ozone or '0H radicals, since an
increase in the non-humic substance (non-HS) concentration after ozonation/ filtration
was observed. The increase in non-HS concentration could only be caused by the
conversion of HS to non-HS. Filtration would not have resulted in such a conversion.
This conclusion is substantiated by data shown in Table 4.4, which show that the percent
removal of HS using ultrafiltration is 5.4 and 13.2% for P1 and P2, respectively, while
for the P1 and P2 samples, 37.3 and 50.2%, respectively, of the HS were removed by
ozonation.

The concentrations of non-HS measured in P1 samples increased by approximately
10%, while that in P2 samples increased by approximately 20% (see Figures 4.2 and 4.3),
indicating that the reaction of HS to form non-HS continued throughout the course of the
experiment. The increased reduction in the concentration of humic substance in the P2

samples compared to that in P1 samples provides further evidence of the importance of

65

oxidation reactions. If the humic substances were removed purely by filtration, the level
of removals would not likely change with ozonation time. These results are consistent
with those of Mellema (1998), who found that ozonation resulted in a significant
reduction in the concentration of humic substances with an apparent molecular weight of
3 to 7 kD (Mellema, 1998). For the UP experiment, the increase in HS removal from
5.4% to 13 % in the P1 and P2 samples, respectively, suggests that there may be some
formation of a fouling layer that results in improved removal of HS. If this is the case,
then the presence of a fouling layer did not have a detrimental effect on permeate flux.
Ozonation/filtration resulted in a significant reduction (p<0. 05) in the SDS TTHMs
and SDS HAA formed after chlorination (see Figures 4.2 and 4.3), as compared to that
removed by filtration alone (see Table 4.4). This reduction was seen in both the P1 and
P2 samples. The reduction in SDS TTHMs found in the chlorinated P2 samples increased
from 44 to 88% when the membrane pore size was decreased from 15 kD to l‘kD. The
reduction in SDS TTHNIs was significantly greater in the P2 than in the P1 samples
(p<0. 05). This decrease in the concentration of TTHM precursors with extended
ozonation, along with the data comparing removal efficiencies for the hybrid system with
ozonation alone and membrane filtration alone (see Table 4.4), is fiirther confirmation of
the importance of oxidation reactions in the removal of DBP precursors. This is
consistent with the work of Lee (2001) and Chen (2003) who showed that ozonation
resulted in a significant decrease in SDS TTHM formation after chlorination. Similar
trends were observed for SDS HAAs, although the levels of reductions were less than
that achieved for SDS TTHMs (38% compared to 68%), indicating that the precursors of

TTHMs and HAAs react at different rates with ozone and/or OH radicals, resulting in

66

different removal efficiencies. In both cases it appears that after ozonation a significant
fraction of the TTHM and HAA precursors that remain are in molecular weight range
from 1 to 15 kD.

As shown in Figure 4.4 and Table 4.4, the concentrations of aldehydes, ketones and
ketoacids increased after both ozonation and ozonation/membrane filtration and with
ozone contact time (compare P1 and P2 results). The concentrations of these species
found in the permeate after ultrafiltration was less than 10% of that afier ozonation or
ozonation/filtration, indicating the importance of ozonation in forming these chemicals.
The influence of ozone contact time on the concentrations of these chemicals is
consistent with the work of Lee (2001) who found that the concentration of ketoacids in
treated Lake Lansing water ranged from a 42 to 370 ug/L for retention times of 4 to 25
min at an ozone dose of 1 mg/mg C, and that the concentrations increased with increasing
retention time. Similar concentration ranges were reported by Chen (2003) who found
that the concentration of ketoacids in treated Lake Lansing water ranged from 40 to 1200
ug/L for an ozone dose of 2.5 mg ozone/mg C. As shown in Figure 4.4, the
concentration of aldehydes, ketones and ketoacids decreased ten-fold when the
membrane MWCO was decreased from 5 to l kD. This is quite surprising, since the
molecular weights of these compounds measured are much smaller than 1 kD and would
be expected to pass through the l kD membrane. Again these results suggest that
oxidation reactions play a significant role in determining the effectiveness of the
ozone/membrane system and that the catalytic oxidation of compounds appears to be

more effective on the 1 kD membrane than that on the 5 or 15 kD membrane.

67

 

mm um n15ko
800.00"

600.00

200.00

Concentration. ”91L
8
O
8

 

0.00

P1 AldehydosluKetoms P1 -Ketoaclds AldehydesaKetones Kate-old:

 

 

 

Figure 4.4 Effect of molecular weight cut-off on Ozonation By-products
Experimental setup: Fig. 4. 1 Operating Conditions: Table 4.1.0zone Dose: 2.5 g/m’ , Membrane Size:
15, 5 and 1 kD. °All values are average of triplicates within experiments and duplicate experiments.
The values have a maximum std deviation of 5 %

4. 4.3 The efihct of gaseous ozone concentration on water quality.

83 shown in Figures 4.5 and 4.6, with 15 kD MWCO membrane, variations in the
gaseous ozone concentration (over the range from 1.5 to 10 g/m3) had little effect on the
extent DOC removal. An explanation for this behavior is that, at the dosages used in this
experiment, only a small fraction of the DOC is mineralized (converted to C02 and
water) and that ozone simply converts larger molecules into smaller ones, which then
pass through the membrane. Chen (2003) and Mellema (1998) also found, that at ozone
dosages in the range 1 to 4 mg/mg C, little of the organic carbon was mineralized. This
was confirmed by the apparent molecular weight distribution of the organic carbon,
which increased in lower molecular weight compounds (<1000 Daltons) at ozone doses

of 2.0 and 7.0 mg/mg C (Mellema, 1998).

68

 

 

  
   

  
   
 
       
 

    

 
   
    

    

   

   

D 1.5 glm3 2.5 glm3 D 10 glm3
100.00
75.00 -
50.00 - id
7/ $5: .
g %e: :3:
. 26.00 A flat w.
:s %:x %:s
3‘ DOC HAA
45.00 -
60.00 -
45.00

 

 

 

 

 

 

Figure 4.5 Effect of gaseous ozone concentration on Permeate 1

(First 400 mL of the sample)

Experimental setup: Fig. 4. I Operating Conditions: Table 4. I.Ozone concentration: 1.5, 2.5 and 10
g/m’ Membrane Size: 15 kD. .All values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std deviation of 5 %

 

D 1.5 glm3 2.5 glm3 a 10 glm3

 

       
   

  

  
  
  

   

 

 

 

100.00
75.00 - 9225'
% a
///:x
50.00 . é'fl
O {/%"'Ii /
§ 25.00 1 4%; %
3 /”1 /,§ ’4
3 0.00 _ /m: //,//,
* DOC W254
45.00 -
-5o.oo -
45.00

 

 

 

 

 

 

Figure 4.6 Effect of gaseous ozone concentration on Permeate 2

(Latter 1000 mL of the sample)

Experimental setup: Fig. 4.1 Operating Conditions: Table 4.1. Ozone Dose: 1.5, 2.5 and 10 g/m’
Membrane Size: 15 kD. °AlI values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std deviation of 5 %

69

Increasing the gaseous ozone concentration from 1.5 to 2.5 g/m3 resulted in an
increase in the percent reduction of both UV-254 in the Pl samples, suggesting that, at
the lower ozone gas concentration, the ozone dosage was not sufficient to remove the
reactive UV—254 absorbing compounds.

As shown in Figure 4.6, the levels of SDS TTHMs in P2 were reduced by about 50%
by ozonation/membrane filtration. The levels of SDS HAAs decreased by approximately
35 to 45% compared to that in the filtered raw water. No statistically significant
decreases were observed in the concentration of SDS TTHMs when the ozone
concentration was increased from 1.5 to 10 g/m3 (p<0.05). There was a smaller
reduction in the overall levels of the SDS HAAs compared to the SDS TTHMs in the P1
and P2 samples (Figures 4.5 and 4.6).

The results given here support the previously mentioned hypothesis that the
precursors of TTHMs and HAAs have different reaction rates with ozone. Chen (2003)
also found that TTHMs and HAAs precursors had different reactiOn rates with ozone. He
observed the concentrations of SDS TTHMs decreased by approximately 40 to 45%
whereas, the concentrations of SDS HAAs decreased by around 30%. K0 et a1. (2000)
also reported that the TTHMs and HAAs produced following ozonation and chlorination
had different formation rates. The concentrations of aldehydes and ketones increased with
increasing gaseous ozone concentrations. Also, the concentrations of these compounds in
the P2 samples were greater than those in P1 samples, due to increased contact time with
ozone. The concentration of ketoacids is almost twice that of the aldehydes and ketones
(see Figure 4.7). While the highest ketoacid concentrations were observed when the

ozone gas concentration was 10 g/m3, there was no significant difference (p>0. 05) in the

70

ketoacids concentrations found at the two lower ozone gas concentrations (i.e., 1.5 and

 

 

2.5 g/m3).
CI 1.5 glm3 2.5 gIm3 B 10 glm3

1600.00

5‘» 1200.00
:3.
5

a 800.00
'2
0
2
O

0 400.00

0.00

P1 Aldehyde: 8. P1 Ketoacids P2 Aldehyde: a P2 Kntoacids
ketones ketones

 

 

 

Figure 4.7 Effect of gaseous ozone concentration on Ozonation By-products
Experimental setup: Fig. 4. I Operating Conditions: Table4. 1. Ozone concentration: 1.5, 2.5 and 10
g/m’, Membrane Size: 15 kD. .All values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std deviation of 5 %

4. 4.4 Efiéct of pH on the water quality

Membranes with molecular weight cut-offs of 5 and 15 kD were used to evaluate the
performance of the system at pH 7.0 and pH 8.2 with an ozone dose of 2.5 g/m3 (Figure
4.8a to Figure 4.10b). The pH was measured during the course of the experiment and it
did not change appreciably. The results show that decreasing the pH from 8.2 to pH 7.0
resulted in significant changes in the permeate characteristics. As shown in Figure 4.8b,
ozonation/filtration through the 15 kD membrane resulted in a reduction in the DOC
concentration of around 35% at pH 8.2 and approximately 45% at pH 7 .0 (in P2). With
the membrane having a 5 kD molecular weight cut off, the DOC removal for P2 was

approximately 46% at pH 8.2 and > 95% at pH 7.0 (Figure 4.8b).

71

 

 

a pH 7.0 I pH 8.2
100

 

75-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0
a ,
r 25 ~
Non-HS

8
o 0 q
3‘ DOC uv-254 HS F TTHM HAA

'25 ...!

-50

.75

 

 

 

 

 

Figure 4.8a Effect of pH on Permeate 1 of 15 kD molecular weight cut off

membrane

Experimental setup: Fig. 4. I Operating Conditions: Table 4.1.0zone Dose: 2.5 g/m’ , Membrane Size:
15 kD, pH 7 and 8. 2. .All values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std deviation of 5 %

 

 

0 pH 7.0 I pH 8.2
100

 

 

 

75

 

 

 

 

 

 

Non-HS

 

 

 

 

 

 

 

 

 

 

 

% Decrease

 

 

 

 

 

 

 

 

 

Figure 4.8b Effect of pH on Permeate 2 of 15 kD molecular weight cut off

membrane

Experimental setup: Fig. 4. I Operating Conditions: Table 4.1.0zone Dose: 2.5 g/m’ , Membrane Size:
I 5 kD, pH 7 and 8. 2. °All values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std deviation of 5 %

72

 

 

a pH 7.0 I pH 3.2

 

 

 

 

 

 

 

 

 

% Decrease

DOC UV-254

 

 

 

 

 

 

 

 

Figure 4.9a Effect of pH on Permeate 1 of 5 RD molecular weight cut off
membrane

Experimental setup: Fig. 4. I Operating Conditions: Table 4. I .Ozone Dose: 2.5 g/m’, Membrane Size: 5
1:0, pH 7 and 8.2.‘AII values are average of triplicates within experiments and duplicate experiments.
The values have a maximum std deviation of 5 %

 

 

13 pH 7.0 I pH 8.2

 

75-

 

 

 

50-

 

25- ——-—~

 

 

 

% Decrease

DOC UV-254 HS TTHM HAA

 

 

 

 

 

 

 

Figure 4.9bEffect of pH on Permeate 1 of 5 RD molecular weight cut off
membrane

Experimental setup: Fig. 4.1 Operating Conditions: Table 4. 1. Ozone Dose: 2.5 g/m’, Membrane Size:
5 kD, pH 7 and 8.2.3411 values are average of triplicates within experiments and duplicate
experiments. The values have a maximum std. deviation of 5 %

73

 

 

0 pH 7.0 I pH 8.2

 

§

 

 

 

 

“2' §

 

 

N
8

 

Concentration, uglL

 

 

-I

 

 

     

 

 

 

 

 

 

0,—2- . r .

5kD Aldehydes 8. 15 kD Aldehydes 8. 5 k0 Ketoacld 15 kD Ketoacld
Ketones Ketones

 

 

Figure 4.10a Effect of pH on Ozonation By-products in Permeate l of 15
kD & 5 RD molecular weight cut off membrane

Experimental setup: Fig. 4. I Operating Conditions: Table 4. I .Ozone Dose: 2.5 g/m’,
Membrane Size: 5 and I 5 kD, pH 7 and 8. 2. °All values are average of triplicates within
experiments and duplicate experiments. The values have a maximum std deviation of 5 %

 

a pH 7.0 I pH 8.2

 

§

 

é

 

 

 

§

Concentration, pglL

 

 

   

5kD Aldehydes 8 15 kD Aldehyde: 8: 5 k0 Ketoacld 15 kD Ketoacld
Ketones Ketones '

 

 

 

 

 

 

 

 

Figure 4.10b Effect of pH on Ozonation By-products in Permeate 2 of 15

kD & 5 kD molecular weight cut off membrane

Experimental setup: Fig. 4. I Operating Conditions: Table 4. I .Ozone Dose: 2.5 g/m’,
Membrane Size: 5 and 15 [(0, pH 7 and 8. 2. 'All values are average of triplicates within
experiments and duplicate experiments. The values have a maximum std deviation of 5 %

74

Similar results were observed with the 5 kD membrane and for the Pl samples (Figures
4.8a and 4.9a). Thus, DOC removal is favored at the lower pH, where ozone is more
stable, and the dissolved ozone concentrations are higher (Karnik et al., 2005).

For the 5 kD MWCO membrane, the UV-254 absorbance of the permeate was similar
at both pH 7.0 and 8.2. For the 15 kD membrane, the reduction in the UV—254
absorbance was greater at the higher pH, suggesting that an OH radical mechanism may
play a role in degrading UV absorbing substances under these conditions. Also, greater
reductions in UV-254 were seen for the 5kD membrane than for the 15 kD membrane.
However, the direct comparison of the results for the two membrane sizes is difficult,
since, due to the lower permeate flux for the 5 kD membrane, the contact time with ozone
is longer than it is for the 15 kD membrane.

Neither varying the molecular weight cut-offs of the membrane nor the pH resulted in
a statistically significant change in the reduction of HS. The greater extent of conversion
of HS to non-HS substances could be attributed to the increased residual ozone
concentration at circumneutral pH (Karnik et al., 2005).

Decreasing the pH resulted in a statistically .significant (p< 0. 05) decrease in the
concentrations of SDS TTHMs and SDS HAAs found after chlorination in the permeate
samples of both the 5 kD and the 15 kD membranes (see Figures 4.83 to 4.9b). The
higher of residual ozone concentration (Karnik et al., 2005) found at pH 7 is the likely
cause for the lower concentrations of SDS T THMs and SDS HAAs found at this pH.

For the aldehydes and ketones (shown in Figures 4.10a and 4.10b), there is a
reduction of at least 50% in the concentrations of these compounds at pH 7.0, compared

to pH 8.2. With a 5 kD MWCO membrane, there is greater reduction in the

75

concentrations of these compounds as compared to that obtained with the lSkD MWCO
membrane at pH 8.2 (see Figures 4.10a & 4.10b). If the formation of these compounds is
predominately due to a radical mechanism, then the lower concentrations of these
compounds found at pH 7 .0 may be explained by the slower formation of these
compounds at the lower pH, where the radical concentration would be expected to be
lower as ozone degradation is slower. Alternatively, it may also be explained by the
catalytic degradation of these compounds at the membrane surface. Further studies are
required to confirm the reaction mechanism. For the 15 kD MWCO membrane, a higher
concentration of ketoacids is found at pH 7.0 than at pH 8.2. For the 5 kD MWCO
membrane, the opposite is true. At this time we have no clear explanation for this

behavior.
4.5 CONCLUSIONS

Use of the combined ozonation/filtration treatment system resulted in significant
improvements in water quality compared to the filtered raw water and to that using either
ozonation or membrane filtration alone. The levels of DOC, UV absorbing compounds,
SDS TTHMs and SDS HAAs were reduced by ozonation/membrane filtration, as
compared to either ozonation or filtration alone. The concentration of aldehydes, ketones
and ketoacids after ozonation/filtration were significantly less than the concentrations of
these compounds found after ozonation (at the same ozone dosage).
4.6 ACKNOWLEDGEMENTS

The authors would like to thank the US Environmental Protection Agency (U S EPA)

Science To Achieve Results (STAR) Program (Grant No.RD830090801) for financial

support of this work.

76

4.7 REFERENCES

Amy, G.L., Kuo, C.L., Sierka, RA. (1988). "Ozonation of humic substances: Effects on

molecular weight distributions of organic carbon and trihalomethane formation
potential" Ozone- Science and Engineering, 10, 39-54.

Amy, G.L., Sierka, R.A., Bedessem, J ., Price, D. and Tan, L. (1992). "Molecular size

distributions of dissolved organic matter" Journal of the American Water Works
Association, 84(6), 67-75.

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

Bull, R. J. and McCabe, L. J. (1984). "Risk Assessment Issues in Evaluating the Health
Effects of Alternate Means of Drinking Water Disinfection. Water Chlorination:

Chemistry, Environmental Impact, and Health Effects", R. L. Jolley and et.al., Eds.,
Lewis Publishers, Chelsea, Michigan.

Chen, KC. (2003). "Ozonation, ultrafiltration, and biofiltration for the control of NOM

and DBP in Drinking Water", Ph.D. dissertation, Michigan State University, East
Lansing, MI 48824.

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

Cipparone, L. A., Diehl, A. C. and Speitel, G. E. (1997). "Ozonation and BDOC removal:

Effect on water quality" Journal of the American Water Works Association, 89(2),
84-97.

Collins, M. R., Amy, G. L. and Steelink, C. (1986). "Molecular-weight distribution,
carboxylic acidity, and humic substances content of aquatic organic-matter -

implications for removal during water-treatment" Environmental Science &
Technology 20(10), 1028-1032.

Crozes, G.F., Jacangelo, J.G., Anselme, C. and Laine, J.M. (1997). "Impact of

ultrafiltration operating conditions on membrane irreversible fouling" Journal of
Membrane Science. 124(1), 63-76.

Griffini, O., Bao, M. L., Barbieri, K., Burrini, D., Santianni, D. and Pantani, F. (1999).
"Formation and removal of biodegradable ozonation by-products during ozonation-

biofiltration treatment: Pilot-scale evaluation" Ozone Science & Engineering 21(1),
79-98.

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

77

microfiltration system for drinking water treatment" Water Science and Technology
41(10—11), 17-23.

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

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

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

Karnik, B.S., Davies, S.H., Chen, K.C., Jaglowski, D.R., Baumann, M.J. and Masten,

S.J., (2005). "Effects of ozonation and pH on the permeate flux of nanocrystalline
ceramic membranes" Water Research 39(4), 728-734.

Kim, J. O., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation" In the Proceedings of the 14‘"
Ozone World Congress, Dearborn, MI, 131-143.

Kim, J .O. and Somiya, I. (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery fiom coagulated
raw sludge" Environmental Technology, 22, 7-15.

Ko, Y.W., Abbt-Braun, G. and Frimmel, RH. (2000). "Effect of preozoation on the
formation of chlorinated disinfection by-products for river R " Acta Hydrochimica
et Hydrobiologica, 28 (5), 256-261.

Koechling, M.T., Shukairy, HM. and Summers, RS. (1996). "Effect‘of Ozonation and '
Biotreatrnent on molecular size and hydrophilic fractions of natural organic matter ” In
water Disinfection and Natural Organic Matter: Characterization and Control. ACS
Symposium Series 649, 196-210.

Kool, H. J., van Kreijl, C. F. and Hrubec, J. (1985). "Mutagenic and carcinogenic
properties of drinking water, Water Chlorination: Chemistry, Environmental Impacts,

and Health Effects" R. L. Jolley, W. A. Brungs, and R. B. Cumming, eds., Lewis
Publishers, Chelsea, Michigan.

Lee, K.H. (2001). "The effects of ozonation pathways on the formation of ketoacids and
assimilable organic carbon (AOC) in drinking water" PhD Dissertation, Michigan
State University, East Lansing, MI 48823.

78

Lee, 8., Amy, G., Cho J .W., Yoon, SH. and Kim, LS. (2001). "Cleaning strategies for

flux recovery of an ultra filtration membrane fouled by natural organic matter" Water
Research, 35, 3301-3308.

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

Mellema, J. (1998). "The use of apparent molecular weight distribution to evaluate the
transformation of natural organic matter during ozonation and biological treatment"
MS Thesis, Michigan State University, East Lansing, MI.

Mogren, E. M., Scarpino, P. and Summers, R. S. (1990). "Measurement of biodegradable
dissolved organic carbon in drinking water" In Proceedings of the American Water
Works Association Annual Conference, Cincinnati, Ohio, 573.

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

Moulin,'C., Bourbigot, M.M, Taizi-Pain, A. and Faivre, M. (1991). "Potanilisation of
surface waters by crossflow ultra- and microfiltration on mineral membranes: Interest
of ozone" In the Proceedings of the 1991 Membrane Technologies in the Water
Industry, Orlando, Florida, 729-737.

Mughal, F. H. (1992). "Chlorination of drinking water and cancer: A review" Journal of
Environmental Patholology, Toxicology & Oncology, 11(5-6), 287-292.

Munch, J. W., Munch, D. J., Winslow, S. D., Wendelken, S. C. and Pepich, B. V. (1998).
"Determination of carbonyl compounds in drinking water by
pentafluorobenzylhydroxylamine derivatization and capillary gas chromatography
with electron capture detection" USEPA, Cincinnati, Ohio.

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

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

Owen, .M., Amy, G.L., Chowdhury, Z.K., Paode, R., McCoy, G. and Viscosil, K.(1995).
"NOM characterization and treatability" Journal of American Water Works
Association, 87(1), 46-63.

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

79

Ravindran, V., Badriyha, B.N. and Pirbazari, M. (1993). "Crossflow membrane filtration
for the removal of natural organic matter" In Proceedings of the Membrane
Technology Conference, Baltimore, MD, 587-599.

Reckhow, D. A., Singer, P. C. and Malcolm, R. L. (1990). "Chlorination of humic
materials -by-product formation and chemical interpretations" Environmental Science
& Technology, 24(11), 1655-1664.

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

Schafer, A.I., Fane, AG. and Waite, TD. (2000). "Fouling effects on rejection in the
membrane filtration of natural waters", Desalination, 131, 215-224.

Schlichter, B., Mavrov, V. and Chmiel H. (2003). "Study of a hybrid process combining
ozonation and membrane filtration — filtration of model solutions" Desalination,
156, 257-265.

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

Standard Methods for Examination of Water & Wastewater (1998). L. S. Clescerl, A.E.
Greenberg, A.D. Eaton Eds., 20th edition, Publishers American Public Health
Association.

US EPA (2001). "Low-pressure membrane filtration for pathogen removal: Application,
implementation and regulatory issues" Office of Water. Washington, DC. 815-C-01-
001.

Weinberg, H. S. and Glaze, W. H. (1996). An overview of ozonation disinfection by-
products, Disinfection By-products in Water Treatment, R. A. Minear and G. L. Amy,
eds., Lewis Publisher, Boca Raton, FL., pp.165-186.

Xing, C.H.; Wen, X.H.; Qian, Y.; Sun, D.; Klose, P.S.; Zhang, X.Q. (2003). "Fouling and
cleaning of microfiltration membrane in municipal wastewater reclamation" Water
Science and Technology, 47(1), 263-270.

Yavich, A. A. (1998). "The use of ozonation and biological fluidized bed treatment for
the control of NOM in drinking water" PhD. Dissertation, Michigan State University,
East Lansing, MI -48824.

80

Yavich, AA. and Masten, SJ. (2003). "Use of ozonation and FBT to control THM
precursors" Journal of American Water Works Association, 95(4), 159-171.

Zhang, X. 2., Ni, Y. and van Heiningen, A. (2001). "Effect of temperature on the kinetics
of pulp ozonation” Journal of Pulp and Paper Science 27(8), 279-283.

81

CHAPTER FIVE

Karrrik, Bhavana.S.; Davies, Simon.H.; Baumann, Melissa. 1.; Masten, Susan. J. (2005).
Fabrication of catalytic membranes for the treatment of drinking water using combined
ozonation and ultrafiltration. Environmental Science Technology 39(19): 7656-7661.

82

CHAPTER FIVE

FABRICATION OF CATALYTIC MEMBRANES FOR THE
TREATMENT OF DRINKING WATER USING COMBINED
OZONATION NAN OFILTRATION.

5.1 ABSTRACT

The removal of disinfection by-products and their precursors was investigated using a
combined ozonation-ultrafiltration system. A commercial membrane was coated 20 or 40
times with iron oxide nanoparticles (4-6 nm in diameter). With this membrane, the
concentration of dissolved organic carbon was reduced by >85% and the concentrations
of simulated distribution system total trihalomethanes and simulated distribution system
halo acetic acids decreased by up to 90% and 85%, respectively. When the coated
membrane was used, the concentrations of aldehydes, ketones, and ketoacids in the
permeate were reduced by >50% as compared to that obtained with the uncoated
membranes. Hydroxyl or other radicals produced at the iron oxide coated membrane
surface as a result of ozone decomposition are believed to have enhanced the degradation
of the natural organic matter, thereby reducing the concentration of disinfection by-
products. While increasing the number of times the membrane was coated from 20 to 40
did not significantly reduce the concentrations of most of the parameters measured, it did
result in a significant decrease in the concentrations of ozonation by-products. Increasing
the sintering temperature from 500 0C to 900 °C also resulted in an improvement in the

removal of the ozonation by-products.

83

Keywords: Catalytic ozonation; ceramic membranes; iron oxide; Disinfection By-

products (DBPS), ozonation, nanofiltration, ultrafiltration
5.2 INTRODUCTION

The increased demand for water has lead many water utilities to use water sources
that elevated levels of natural organic matter (Bursill, 2001; Skjelkvéle et al., 2000),
making them undesirable as water sources. This increasing demand, combined with
stricter govermnent regulations, necessitates improved drinking water treatment. The
presence of natural organic matter (NOM) in the source water is a cause of concern to
health professionals and environmental engineers because the reaction of NOM with
disinfectants, such as chlorine, results in the formation of disinfection by—products
(DBPS), such as trihalomethanes (THMs) and haloacetic acids (HAAs). Because of their
toxicity (Mughal, 1993; Morris et al., 1992; Kool et al., 1985) the trihalomethanes and
haloacetic acids are regulated by the US. Environmental Protection Agency (US EPA).

In the United States there is an increasing interest in the application of both ozone
and membrane filtration for DBP and DBP precursor removal in order to meet the
requirements of the Surface Water Treatment Rule (SWTR), the Disinfectant and
Disinfectant By-products Rule (D/DBPR), and the Long Term 1 Enhanced Surface Water
Treatment Rule (LTlESWTR). Several researchers have attempted to combine ozone
with polymeric membranes with limited success, in part because the organic membranes,
which are commonly used in water and wastewater treatment applications, are prone to
destruction by ozone (Hashino et al., 2000; Shanbag et al., 1998; Castro and Zander,
1995; Shen et al., 1990). Hashino et al., (2000) studied the use of ozonation combined

with an ozone resistant polyvinylidenefluoride (PVDF) microfiltration membrane. They

84

found that ozone prevented foulants from adhering to the membrane surface, thus
decreasing membrane fouling. However, high dissolved ozone concentrations (> 1 mg/L)
were necessary to obtain high permeate fluxes and prevent membrane fouling.

Ceramic membranes are ozone resistant and when these membranes are used in
combination with ozone, stable permeate fluxes can be achieved without membrane
damage (Karnik et al., 2005a; Schlichter, 2004; Kim et al., 2001; et al., 1999; Allemane
et al., 1993). Kim and colleagues (2001) used ceramic membranes to investigate the
effect of ozone bubbling on flux recovery. The results showed that intermittent ozonation
effectively maintained high permeate fluxes and prevented membrane fouling caused by
particle accumulation on the membrane surface. Our earlier work demonstrated that
stable fluxes can be obtained with ozonation-ceramic membrane filtration. Ultrafiltration
alone did not achieve the levels of treatment obtained with combined ozonation/
membrane filtration. Ozonation-filtration resulted in a reduction of 50% in the dissolved
organic carbon (DOC) concentration. It also resulted in the formation of partially
oxidized compounds fi'om NOM that were less reactive with chlorine, decreasing the
concentration of simulated distribution system total trihalomethanes (SDS TTHMs) and
simulated distribution system halo acetic acids (SDS HAAs) by up to 80% and 65%,
respectively (Karnik 2005a;2005b).

Catalytic ozonation has been used to degrade NOM and other organic compounds
in drinking water and wastewater (Legube and Karpel Vel Leitner, 1999). In the presence
of different metal oxide catalysts, such as iron oxide, manganese oxide, titania, alumina
and zirconia, ozone degrades organic compounds, including saturated carboxylic acids,

phenols, aromatic hydrocarbons, dyes, humic substances and herbicides (Beltran et al.,

85

20033; 2003b; Ni and Chen, 2001; Radhaloishnan and Oyama 2001; Gracia et al., 2000a;
2000b; 1996; Legube and Karpel Vel Leitner, 1999). Based on extensive research
involving various ozonation methods for drinking water treatment, catalytic ozonation
has been determined to be the best alternative for the oxidation of ozone by-products to
carbon dioxide, and the reduction in the chlorine demand (V olk et al., 1997 ; Allemane at
al., 1993). Masten and Davies (1997) reported that the presence of reactive soil surfaces
catalyzed the decomposition of ozone and contaminants sorbed on the soil. Paillard et al.
(1991) documented that TiOz-catalyzed ozonation was more efficient than ozone alone
for the degradation of humic acid. Mn(II) is effective for the catalytic degradation of
carboxylic acids that do not react appreciably with molecular ozone. It is believed that
Mn(II) complexes with these carboxylic acids to form an intermediate by—product that is
more easily degraded by ozone (Andreozzi et al., 2000; 1998a; 1998b; 1992). Ma et al.,
(2000; 1999) confirmed that the degradation of compounds by ozone in the presence of
manganese follows a radical mechanism. Pure alumina, which is often used as a support
material for metal or metal oxide catalysts, was also found to be an effective catalyst for
the degradation of NOM by ozone (Ernst et a1; 2004). Pecchi and Reyes (2003) prepared
iron oxide coatings supported on TiOz and A1203 using the sol gel method. These
coatings catalyzed the degradation of phenol by ozone.

This work focuses on the fabrication of ceramic membranes with catalytic
properties using a layer-by-layer method to deposit iron oxide particles on a titania coated
membrane. We have tested the application of these membranes in a combined ozonation-

nanofiltration process to remove disinfection by—products and their precursors.

86

5.3 MATERIALS AND METHODS
5. 3.1 Membrane Preparation and Characterization

Tubular AZT (a mixture of alumina, zirconia, and titania) ceramic membranes
(Clover-leaf design (containing three channels), CéRAM Inside, TAMI North America,
St. Laurent, Quebec, Canada) with nominal molecular weight cut-offs of 15 kilodaltons
(RD) and 5 kD were used as a support for the catalytic coatings. The external diameter of
each membrane was 10 mm and the active membrane length was 8 cm. The total filtering
area of the membrane was approximately 11 cm2 and the membranes can be operated in
the pH range from 0-14. The initial permeability of the membranes was tested using DDI
water (Karnik et al., 2005a).

The colloidal particles used for coating the membranes were prepared by Sorum’s
method (Mulvaney et al., 1998 The procedure used was as follows: double deionized
water (DDI) water (450 mL) was heated until it boiled vigorously; then 50 mL of freshly
prepared 20 mM FeCl3 solution was added at a rate of approximately two drops per
second. The sol rapidly turned golden brown and finally deep red. After all the ferric
chloride solution was added, the suspension was allowed to boil for an additional 5
minutes; it was then cooled to room temperature and dialyzed, using cellulose dialysis
tubing with an average flat width of 33 mm, for 48 hours against a dilute nitric acid
solution having a pH of 3.5.

Transmission electron microscopy (T EM) characterization was performed using a
JEOL 100CX at an accelerating potential of 100 kV and magnifications ranging from
5,000X to 370,000X. The TEM protocol for the particle characterization involved

diluting the suspension with DDI water in the ratio of 1:4. Double-sided sticky tape was

87

attached to a glass slide (76.2 mm x 25.4 mm x 1 mm), leaving a small section
(approximately 2-3 mm) of the tape hanging off the long side of the slide. Masking tape
was then used to cover the portion of the double-sided tape, which rested on the glass
slide, leaving the excess double-sided sticky tape uncovered. Grids (0.25% Formvar and
carbon) were placed on the overhanging double-sided sticky tapes with light tweezer
pressure to ensure that the grids would stick. The suspension was then placed in a
dropwise manner onto the grids and the excess suspension was removed by lightly
wiping across the grid with filter paper. The grids were then air-dried in a dust free
environment until TEM analysis. Photomicrographs were collected using a Megaview 111
digital camera. The photomicrographs, which are provided in the Figure 5.1, showed that

the average particle diameter was 4 to 6 nm.

 

Figure 5.1 TEM characterization of the iron oxide particles (average size 4-6nm)
TEM: JEOL 100CX, Accelerating voltage: 100 kV, Imaging system: Analysis, Digital imaging: Mega view III

The layer-by-layer technique used to coat the membranes is based on a protocol
described by McKenzie et al., (2002) for coating doped tin oxide electrodes. for coating
doped tin oxide electrodes. The membrane was immersed in the colloidal suspension for

one minute and then rinsed with DDI water. Then, the membrane was immersed in an

88

aqueous phytic acid (40 mM) for one minute and rinsed with the DDI water. This
sequence was repeated the desired number of times (20 or 40). After coating, the
membrane was either sintered at 500 °C for 60 minutes or sintered at 900 °C for 30
minutes. These two temperatures were chosen to produce membranes on which the iron
oxide particles were attached but not fused to each other (500 °C) or completely sintered
to each other and to the membrane surface (900°C).

5. 3.2 Ozonation/ Membrane Filtration

A schematic representation of the ozonation/membrane system is shown in Figure 5.2. A
stainless steel filter holder, Teflon® tubing and stainless steel or Teflon“) joints and valves
were used throughout the system. Other components included: 3.5-liter and 1.5-liter
water-jacked glass reservoirs made of Pyrex® glass, and a simple Y inline mixer (Ozone
Service, Burton, BC, Canada). The membranes described above were used for
membrane filtration. A Teflon‘1° valve was placed in the retentate line of the membrane
system to create transmembrane pressures of 0.2 to 0.5 bars. Ozone gas was added into
the water stream through a simple Y inline mixer, just before the aqueous stream entered
the membrane module.

To generate ozone, pure oxygen gas (99.999%) fi'om a pressurized cylinder was
dried using a molecular sieve trap, and then fed to the ozone generator (Model OZ2PCS,
Ozotech Inc., Yreka, Calif). The voltage applied to the ozone generator was varied to
control the gaseous ozone concentration. The excess gas was vented to the atmosphere
after it was passed through a 2% potassium iodide (KI) solution to destroy any residual
ozone. The water in the 3.5-liter reservoir was maintained at a constant level during the

experiments using a peristaltic pump (Masterflex Model 7520-35, Cole-Partner Co.,

89

Chicago, Illinois) to transfer the water from a 1.5-liter reservoir into the 3.5-liter
reservoir. A constant water temperature of 20 °C was maintained using a recirculating
water bath. The experiments were performed with a membrane cross flow velocity of 0.6
m/s; the flow was turbulent with a Reynolds number of approximately 6000.

The operating conditions are shown in Table 5.1. The operating conditions were
determined based on the previous experiments with uncoated membranes (Karnik et al.,
2005a; 2005b; Chen, 2003). The conductivity remained practically unchanged for the
duration of the experiment. The change in conductivity was < 0.01 uS/cm.

Permeate samples were collected in bottles covered with Parafilrn® and stored in
an ice-bath throughout the duration of the experiment. The first 400 mL of permeate
collected was labeled as P1 and the latter 1000 ml as P2.

Table 5.1
Typical Characteristics of Lake Lansing Water (Haslett, MI) "

Parameters Lake
TOC 8.6 to 11.6
7.7 to 8.6
145 to 157
UV-254 abs. 0.160 to 0.180

SDS 240

SDS HAA 75
BDOC 1.0 to 4.1
Nitrate 0.44
Total 0.06
Hardness as 190 to 198
“All data reported is obtained fiom the Lake Lansing Watershed Advisory Committee
Report( I 998) except for SDS THM and SDS HAA, which were measured as part of this
study
l’SDS THM and SDS HAA were measured using Standard Method 5 7 I 0 and USEPA Method 552.2

respectively.

 

90

Waste gas

     

 

 

 

 

 

 

 

 

 

 

 

 

 

Moisture
Adsorbent Spectrophotometer
Ozone
generator ‘ r
Oxygen fl I |
Cylinder i" . ,
.Kr

 

 

    
    

Membrane

 

 

 

 

 

 

. W

\
® Flowmeter

C—D l 6’) Pressure gauge

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0| X Valve

 

Figure 5.2 Schematic Representation of the Ozone-Membrane Filtration System

91

5. 3.3 Water Source

Experiments were carried out using samples taken from Lake Lansing (Haslett, MI),
which is a borderline eutrophic lake. The typical characteristics of the water from Lake
Lansing, are given in Table 5.1. The samples were collected at the boat ramp at the Lake
Lansing Park-South, Haslett, MI in five-gallon polyethylene carboys and stored at 4°C.
The maximum storage period was seven days. Water samples were pre-filtered through a
0.45-um mixed cellulose ester filter (Millipore-HA) before testing.

5.2.4 Analytical Methods

The absorbance of ozone in the gas phase was measured at 254 nm with a Milton Roy
Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY) using a 2-mm path
length quartz flow-through cell. An extinction coefficient of 3000 M'lcm'1 (Hoigné,
1988) was used to calculate the ozone concentration.

The UV absorbance of the water samples was measured at a wavelength of 254
nm with a Milton Roy Genesis-5 spectrophotometer (Milton Roy, Inc., Rochester, NY)
using a 1 cm quartz cell.

DOC was analyzed using an 01 Analytical Model 1010 analyzer. The TOC
analyzer uses the UV/persulfate method (Standard Method, 1998). To ensure the
reliability of the method, standards having TOC concentrations of 2.5, 5, 7, 10 mg/l (01
Analytical) were run and samples were analyzed in triplicate. A blank was also run with
every set of samples.

The concentrations of humic substances in the samples were measured by
adsorption on an XAD-8 resin according to Method 5510C (Standard Methods, 1998). A

100 mL sample was acidified with concentrated phosphoric acid to a pH of 2, the

92

acidified sample was then eluted through a 10 mm diameter (ID) x 15 cm long column at
a flow rate of 2 mI/min. The effluent from the column was collected and then analyzed
for TOC, which represents the non-humic fraction of the dissolved organic matter in the
water sample. The resin-packed column was then back eluted with 100 mL of 0.1 N
sodium hydroxide at a flow rate of 2 mL/min. The eluent was collected and acidified with
concentrated phosphoric acid to a pH less than 4, purged with high-purity helium for 3
minutes to remove the inorganic carbon, and analyzed for TOC. The organic content of
the eluent represents the concentration of humic substances.

Water samples were dosed with a chlorine concentration that ensured a residual
chlorine concentration in the range of 0.5 to 2 mg/L after 48 hours incubation at room
temperature, according to the procedures in Method 2350 (Standard Methods 1998). The
THM compounds, chloroform (CHC13), bromodichloromethane (CI-IBrClz),
dibromochloromethane (CI-IBrzCl), and bromoform (CI-IBr3), were extracted from the
water samples using hexane and analyzed by gas chromatography (Method 5710,
Standard Methods 1998). A Perkin Elmer Autosystem gas chromatograph (Perkin Elmer
Instruments, Shelton, CT) equipped with an electron capture detector (ECD), an auto
sampler, and a 30 m X 0.25 mm ID, 1 um DB-Sms column (J&W Scientific, Folsom,
CA) was used for the analysis. The oven temperature was ramped from 50°C to 150 °C at
a rate of 10 °C/min. The flow rate of the carrier gas (N2) was 12.0 mI/min. The injector
temperature and detector temperature were 27 5 and 350 °C, respectively.

SDS HAAs were produced by chlorination as described above. The
concentrations of monochloroacetic acid (MCAA), monobromoacetic acid (MBAA),

dichloroacetic acid (DCAA), bromochloroacetic acid (BCAA), trichloroacetic acid

93

(TCAA), and dibromoacetic acid (DBAA) were determined using US EPA Method
552.2. A Perkin Elmer Autosystem gas chromatograph (Perkin Ehner Instruments,
Shelton, CT) equipped with an ECD, an autosampler, and a 30 m X 0.32 mm ID, 3 pm
DB-l column (J&W Scientific, Folsom, CA) was used for the analysis. The oven
temperature was programmed to hold for 15 minutes at 32°C, then increased to 75 °C at a
rate of 5 °C/min and held 5 minutes, then increased to 100 °C at a rate of 5 °C/min. The
carrier flow (nitrogen) was 1.0 mL/min with the injector temperature and detector
temperatures at 200 °C and 260 °C, respectively.

USEPA Method 556 (Munch et al., 1998) was used to monitor formaldehyde,
propionaldehyde, glyoxal, methyl glyoxal, acetone, and 2-butanone, ketomalonic acid,
pyruvic acid and glyoxylic acid. A Perkin Ehner Autosystem gas chromatograph (Perkin
Ehner Instruments, Shelton, CT) equipped with an ECD, an autosampler, and a 30 m x
0.25 mm ID, 0.5 um DB-Sms column (J&W Scientific, Folsom, CA) was used in the
analysis. The oven temperature was programmed to hold at 1 minute at 50 °C, then
increased to 220 °C at a rate of 4°C/min followed by an increase to 250 °C at a rate of 20
°C/min with a 5 minute hold time. The carrier flow was 1.0 mL/min and the injector
temperature and detector temperatures were 180 °C and 300 °C, respectively.

5. 3.5 Ozone-membranefiltration experiments
The ozone-filtration system used in these experiments is shown in Figure 5.2. The
operating conditions used are shown in Table 5.2.

Permeate samples were collected in bottles covered with parafilrn and stored in an

ice-bath throughout the duration of the experiment. The first 400 mL of permeate

collected was labeled as P1 and latter 1000 m1 as P2.

94

Table 5.2

Operating Conditions for the Ozone—Membrane Filtration System

 

 

 

 

 

 

Water recirculation rate 2,75 LPM
Water temperature 20°C

Ozone gas flow rate 100 mL/min
TMP 0.5 bar
Ozone dose 2.5 g/m3

 

 

 

5.4 RESULTS AND DISCUSSION

 

Penneablllty (LImZ-bar)

 

 

 

 

 

 

 

80
“as e fir—a 9 s 8
40
"‘1 1!: H I III F

 

 

 

 

 

Time (hours)

 

o Uneoated-15 kD
e Uncooted-s kD
A 1 5-20-500

a 5-20-500

o 1 540-500

e 540-500

x 1 5-20-000

X 5-20-000

O 1 540-000

0 540-900

 

 

Figure 5.3 Permeate flux for different membrane coating modifications.
Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1.Membrane Size: 5 or 15 kD
'All values are average of triplicates within experiments and duplicate experiments. The values have a
maximum std. deviation of 5 %. For the coated membranes the first number in the legend corresponds to
M WCO of the membrane, the second number is the number of coatings and the third number is the
sintering temperature. For example, 15-20-500 is a membrane with I 5 kD M WCO coated 20 times with
the catalyst and sintered at 500 °C. All values are average of triplicates within experiments.

95

 

 

Our earlier work showed no significant decrease in the permeate flux when using ozone
at gas phase concentrations greater than 2.5 g/m3 (Karnik et al., 2005a). Experiments
were conducted to determine the effect of the coating procedure on membrane
permeability. As shown in Figure 5.3, stable fluxes were maintained throughout the
course of each experiment. The coating of the membrane had little effect on its
permeability, suggesting that processing did not damage the integrity of the membrane

and that the resistance of the iron oxide coating is comparatively small.

Figures 5.4-5.7 compare the results obtained for the coated and uncoated
membranes. The results for the 15 kD membrane are shown in Figures 5.4 and 5.5 and

Figures 5.6 and 5.7 show the results for the 5 kD membrane.

 

 

I 15-20-500 El 15-20-900 n 1540-500
I 1540-900 1:] Uncoated

 

 

 

100

 

75

 

 

 

 

 

 

 

 

% Mnlng

 

 

 

 

 

 

 

-1 00

 

 

 

Figure 5.4 Water quality results for two different sintering temperatures.
Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1. Membrane Size: 15 kD, Coating: 20
or 40 coatings, Sintering temperature: 500 °C or 900 °C. All values are average of triplicates within
experiments. Explanation of the legend is given in the caption of Fig.5 .3

96

 

 

 

I 15-20-500 El 15-20-900 m 1540-500
I 1540-900 El Uncoated

 

 

300

—I—

 

 
   
 

 

l _

 

Hr l"‘

 

 

 

 

 

Aldehydesl-Katonea Ketoaclda

 

 

Figure 5.5 Concentrations of ozonation by-products in the permeate for two
different sintering temperatures.

Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1. Membrane Size: 15 kD, Coating: 20 or
40 coatings, Sintering temperature: 500 °C or 900 °C. All values are average of triplicates within
experiments. Explanation of the legend is given in the caption of F ig.5.3

 

 

 

I 5-20-500 El 5-20-900 I 540-500
I 540-900 El Uncoated

 

 

 

 

% Remlnlng

 

-75~ , #7 * 7* ..i—- min
-100

 

 

 

 

 

Figure 5.6 Water quality results for two different sintering temperatures.
Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1. Membrane Size: 5 [(0, Coating: 20 or
40 coatings, Sintering temperature: 500 °C or 900 °C. All values are average of triplicates within
experiments. Explanation of the legend is described in the caption of Fig.5.3

97

 

 

l 5-20-500 El 5-20-900 I 540-500
I 540-900 D Uncoated

 

 

 

300

 

 

100

 

i ll

'l;

 

 

 

 

 

 

‘W ‘

Aide hydee&Ketonea Ketoaclds

 

 

 

 

 

 

Figure 5.7 Concentrations of ozonation by-products in the permeate for two

different sintering temperatures.

Experimental setup: Fig. 5.2. Operating Conditions: Table 5.1. Membrane Size: 5 1:0, Coating: 20 or 40
coatings, Sintering temperature: 500 °C or 900 °C. All values are average of triplicates within
experiments. Explanation of the legend is described in the caption of Fig. 5 .3

As shown in Figure 5.4, the permeate fluxes are different for the 15 and 5 kD
MWCO membranes. Thus, due to the different ozone contact times, a direct comparison
of the results for the membranes with different MWCOs is impossible.

Figures 5.4 and 5.6 show that the reduction in the DOC concentration in the P2 samples
is greater for the coated membranes than for the uncoated membrane. This reduction in
DOC concentrations suggests that the iron oxide coating catalyzes the degradation of
ozone to produce radical species at the membrane surface, which degrade the NOM.
Losses due to sorption of NOM on the iron oxide coating are expected to be very small,
since the iron oxide coatings are extremely thin. Based on the observed thickness of the

coating (using TEM) the total quantity of iron oxide deposited on the membrane is

98

estimated to be less than 0.1 ug. The quantity of DOC removed is > 4 mg C. To remove
this amount of NOM via sorption, the sorptive capacity of the iron oxide would have to
be of the order of 4 x 107g/kg. This figure is too large to be reasonable, even for
nanoparticles, so we conclude that sorption to the iron oxide particles cannot explain the
enhanced NOM removal seen with the coated membranes. As with all parameters
measured, the results for NOM removal in the P1 samples follow the same trends as
observed with P2 samples. As such, only the data for P2 samples is presented in the
figures. The data for P1 samples is available in Fig 5.8 to 5. 15.

There is a little difference between the coated and uncoated membranes in the
extent to which the absorbance of the UV-254 absorbing compounds is reduced. In our
earlier work, we showed that the removal of the UV-254 absorbing compounds is due
predominately to the reaction of ozone with these substances and not due to filtration
(Karnik et al., 2005b). Together, these results suggest that the reduction in UV absorbing
compounds is due to solution phase ozonation rather than surface catalytic reactions.

Similarly, no statistically significant differences were observed in the
concentrations of the humic substances found in the permeate after combined treatment
of ozonation-membrane filtration using either the coated and uncoated membranes.
Consistent with these results for the removal of humic substances, the concentrations of
non-humic substances formed were also similar in the permeates fi'om all membranes
studied (see Figures 5.4 and 5.6). The behavior of HS and non-HS in the ozone
membrane filtration system is discussed in detail in our earlier work where we studied the
destruction of HS and formation of the non-HS during ozonation alone, membrane

filtration alone, and in the hybrid process (Karnik et al., 2005b). The concentration of HS

99

remaining in the P2 samples after ozonation/membrane filtration was less than 50% of
that in the raw water. This reduction is, in part, due to the reaction of NOM with either
ozone or OH radicals, since an increase in the non-humic substance (non-HS)
concentration after ozonation/ filtration was observed. The increase in non-HS
concentration could only be caused by the conversion of HS to non-HS. Filtration would
not have resulted in such a conversion. This conclusion is substantiated by results
presented by Karnik et al. (2005b), which show that the percent removal of HS using
ultrafiltration was 13% for P2, while 50% of the HS was removed by ozonation.

The concentrations of non-HS measured in the P2 samples increased by
approximately 20% compared to that in the P1 samples, indicating that the reaction of HS
to form non-HS continued throughout the course of the experiment. If the humic
substances were removed purely by filtration, the extent of removal would not likely have
increased with ozonation time.

Despite the results for HS, the extent to which the DBPs precursors were removed
was greater with the coated membranes than with the uncoated membrane (see Figures
5.4 and 5.6). The concentrations of TTHMs and HAAs were reduced by up to 90% and
up to 85%, respectively, with ozonation combined with an iron oxide coated 5 kD
membrane. The membrane surface coated with iron oxide appears to catalyze reactions
that lead to a reduction in DBPs and DBP precursors. For the 15 kD membranes, the
concentrations of aldehydes, ketones, and ketoacids in the permeate following treatment
using the coated membranes were less than that obtained with the uncoated membrane
(see Figure 5.5). Ozone may decompose on the active metal sites of the iron oxide

surface, resulting in increased rates of hydroxyl radical production (Ernst et al., 2004; Ma

100

et al., 2000; 1999), which in turn leads to a concomitant decrease in the concentration of
disinfection by-products and their precursors. To improve the adhesion of the coating to
the membrane, several coated membranes were sintered at 900 °C. The results for the
coated membranes treated at 500 °C and 900 °C are compared in Figures 5.3 to 5.6. A
small decrease in the concentration of ozonation by-products was found when the higher
sintering temperature was used. It appears that sintering at higher temperatures alters the
properties of the ceramic membrane surface, which further enhances its catalytic
properties.

Ongoing studies are being conducted using scanning electron microscopy (SEM)
and TEM imaging of these sintered surfaces along with chemical and phase analysis of
the membrane surface to better understand the changes that occur during sintering.

As seen in Figures 5.4 to 5.7, increasing the number of coatings of iron oxide did
not result in a significant improvement in the system performance, except for the
ozonation by-products. The lowest concentrations of aldehydes, ketones, and ketoacids
were achieved using the membrane that was coated 40 times and sintered at 900 °C. The
5 kD membrane performed better than the 15 kD membrane. A statistical analysis of the
data presented in Figures 5.4-5.7 using ANOVA indicates that at the 95% confidence
level, with the exception of the results for HAAs with a 5 RD membrane (see Figure5.6)
and the ozonation by—products with a 15 kD membrane (see Figure 5.5), there is no
statistically significant difference for the removal of NOM, DBPs or DBP precursors
using the membranes coated 20 or 40 times.

The US EPA, under the Stage 2 Disinfection/Disinfection By-Product (D/DBP)

Rule, sets standards for maximum DBP concentrations in drinking water. The maximum

101

contaminant levels for TTHMs and HAAs are 80 rig/L and 60 ug/L, respectively.
Catalytic ozonation membrane filtration met regulatory limits for both contaminants.
Using a 5 RD MWCO membrane, coated 20 times and sintered at 900°C, the
concentrations of TTHMs and HAAs after chlorination were approximately 25 to 30 ug/L
and 20 to 25 rig/L, respectively. Even better quality water was achieved using a 5 RD
MWCO membrane, coated 40 times and sintered at 900°C. Afier chlorination the
concentration of TTHMs was approximately 15 to 20 ug/L and the concentration of
HAAs was approximately 7 to 15 ug/L. These results are especially significant because
these limits are difficult to meet with poor quality waters, such as those used in this work.
Previous work has demonstrated we can meet the regulatory requirements for
DBPs using a 1 kD membrane and a gaseous ozone concentration of 2.5 g/m3 (Karnik et
al., 2005b). Comparable results could be obtained using iron oxide coated 5 kD
membranes. As the permeability of the 5 kD membrane is three times greater than that of
the 1 kD membrane, a significant decrease in the costs associated with the process can be
achieved using the coated membrane while still producing high quality water that meets
the pertinent regulatory requirements of the Stage 2 D/DBP Rule.
5.5 SUPPLEMENTAL INFORMATION
This data is included in Appendix A, showing water quality for permeate 1 (P1) of the
lSkD and 5kD MWCO is available in supplemental information Figures A1 to A4. The
effect of sintering temperatures on the permeates of 15 and 5 kD MW CO is shown in

Figures A5 to A8.

102

5.6 ACKNOWLEDGEMENTS

The authors would like to acknowledge the US Environmental Protection Agency
(US EPA) Science To Achieve Results (STAR) Program (Grant No. RD830090801) for
financial support of this work. We would also like to thank Nikhil Theyyuni for his

assistance with membrane fabrication.

103

5.7 REFERENCES

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ 03H202 and O3/Ti02) for the oxidation of natural organic matter in
water” Ozone: Science and Engineering, 15(5), 419-432.

Andreozzi, R.; Insola, A.; Caprio, V.; D’Amore, M.G. (1992). "The kinetics of Mn(II)
catalyzed ozonation of oxalic acid in aqueous solutions" Water Research, 26, 917-
921.

Andreozzi, R.; Caprio, V.; Insola, A.; Marotta, R. Tufano, V. (1998a). "The Ozonation of
Pyruvic acid in aqueous solutions catalyzed by suspended and dissolved manganese"
Water Research, 32, 1492-1496.

Andreozzi, R.; Caprio, V.; Insola, A.; Marotta, R.; Tufano, V. (1998b). "The use of
manganese dioxide as a heterogeneous catalyst for oxalic acid ozonation in aqueous
solutions" Applied Catalysis: A, 131, 75-81.

Andreozzi, R.; Marotta, R.; Sanchirico, R. (2000). "Manganese catalyzed ozonation of
glyoxlic acid in aqueous solutions" Journal of Chemical Technology &
Biotechnology, 75, 59-65.

Beltran, F. J ., Rivas, F. J., and Montero-de-Espinosa, R. (2003a). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts.l. Homogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14), 3210-3217.

Beltran, F. J., Rivas, F. J ., and Montero-de-Espinosa, R. (2003b). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts. 2. Heterogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research. 42(14), 3218-3224.

Bursill, D. (2001). "Drinking water treatment - understanding the processes and meeting
the challenges" Water Science& Technology: Water Supply, 1, 1-7.

Castro, K. and Zander, A.K. (1995). "Membrane air-stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

Chen, KC. (2003). "Ozonation, ultrafiltration, and biofiltration for the control of NOM
and DBP in Drinking Water" PhD Dissertation, Michigan State University, East
Lansing, MI.

Ernst M., Lurot F., and Schrotter J .C. (2004). "Catalytic ozonation of refractory organic
model compounds in aqueous solution by aluminum oxide" Applied Catalysis B:
Environmental, 47(1), 15-25.

Gracia, R., Aragtles, J .L., and Ovelleiro, J. L. (1996). "Study of the catalytic ozonation of

humic substances in water and their ozonation byproducts" Ozone: Science and
Engineering, 18(3), 195-208.

104

Gracia, R., Cortes, S., Sarasa, J., Ormad, P., and Ovelleiro, J. L. (2000a). "Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Ormad, P., and Ovelleiro, J. L. (2000b). "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

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

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

Karnik, B.S.; Davies, S.H., Chen, K.C.; Jaglowski, D.R.; Baumann, M.J.; Masten, S.J.
(2005a). "Effects of ozonation and pH on the permeate flux of nanocrystalline
ceramic membranes" Water Research, 39(4), 728-734.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005b). "The effects of
combined ozonation and filtration on disinfection by-product formation" Water
Research, 39(13), 2839-2850.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005c). "Fabrication of
catalytic membranes for the treatment of drinking water using combined ozonation
and ultrafiltration", Environmental Science Technology 39(19), 7656-7661.

Kim, J. O., Sorrriya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14'”
Ozone World Congress, Dearborn, MI, 131-143.

Kim, J .O. and Somiya, 1., (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

Kool, H.J.; van Kreijl, C.F.; Hrubec, J. (1985). "Mutagenic and carcinogenic properties of
drinking water" Water Chlorination: Chemistry, Environmental Impacts, and Health
Eflects, Jolley, R.L.; Brungs, W.A.; Cumming, R.B., Eds.; Lewis Publishers, Chelsea,
Michigan, 1985.

Lake Lansing Watershed Advisory Committee. Progress Report, Ingham County Drain
Commissioner’s Office, Mason, Michigan, 1998.

Legube, B. and Karpel Vel Leitner, V. (1999). "Catalytic Ozonation: A promising
advanced oxidation technology for water treatment" Catalysis Today, 53, 61-72.

105

Ma, J.; Graham, N.J.D. (1999). "Degradation of atrazine by manganese-catalyzed
ozonation-influence of humic substances" Water Research, 33(3), 785-793.

Ma, J .; Graham, N.J.D. (2000). "Degradation of atrazine by manganese-catalyzed
ozonation-influence of radical scavengers" Water Research, 34 (15), 3822-3828.

Masten, SJ. and Davies, S.H.R. (1994). "The use of ozonation to degrade organic
contaminants in wastewaters" Environmental Science & Technology, 28(4), A180 —
A1 85.

McKenzie, K.J.; Marken, F.; Hyde, M.; Compton, R.G. (2002). "Nanoporous ironoxide
membranes: Layer-by-layer deposition and electrochemical characterization of
processes within nanopores" New Journal of Chemistry, 26, 625-629.

Morris, R.D.; Audet, A.M.; Angelillo, I.F.; Chalmers, T.C.; Mosteller, F. (1992).
"Chlorination, chlorination by-products, and cancer - a metaanalysis" American
Journal of Public Health, 82, 955-963.

Munch, J .W.; Munch, D.J.; Winslow, S.D.; Wendelken, S.C.; Pepich, B.V. (1998).
"Determination of carbonyl compounds in drinking water by
pentafluorobenzylhydroxylamine derivatization and capillary gas chromatography
with electron capture detection" Method 556.1, USEPA, Cincinnati, Ohio.

Mughal, RH. (1992). "Chlorination of drinking water and cancer: A review" Journal of
Environmental Pathology, Toxicology and Oncology, 11, 287-292.

Mulvaney, P.; Cooper, R.; Grieser, F .; Meisel, D. (1998). "Charge trapping in the
reductive dissolution of colloidal suspensions of Iron (III) oxides" Langmuir, 4, 1206-
1211.

Ni, C.H.; Chen, J .N. (2001). "Heterogeneous catalytic ozonation of 2-chlorophenol

aqueous solution with alumina as a catalyst" Water Science and Technology, 43 (2),
213-220.

Paillard, H., Dore, M. and Borbigout, M.M. (1991). "Prospects concerning applications
of catalytic ozonation in drinking water treatment" In the proceedings of the 10"I
Ozone world Congress (10A), Monaco, 313-329.

Pecchi, G.; Reyes, P. (2003). "Fe supported catalysts prepared by the sol-gel method.
Characterization and evaluation in phenol abatement" Journal of Sol-Gel Science and
Technology, 26, 865-867.

Radhakrishnan, R., Oyama, ST. (2001). "Ozone Decomposition over manganese oxide

supported on ZrOz and TiOz: A kinetic study using in situ Laser Raman
Spectroscopy”, Journal of Catalysis, 199(2), 282-290.

106

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

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

Shen Z.S., Semmens M.J., and Collins AG. (1990). "A novel-approach to ozone water
mass-transfer using hollow fiber reactors" Environmental Technology, 1990, 11, 597-
608.

Skjelkvale, B.L.; Andersen, T.; Halvorsen, G.A.; Raddurn, G.G.; Heegaard, E.; Stoddard,
J .; Wright, R. (2000). "The 12-year report: Acidification of Surface waters in Europe
and North America; trends biological recovery and heavy metals" ICP-Waters report
52, 115.

Volk, C.; Roche, P.; Joret, J .C.; Paillard, H. (1997). "Comparison of the effect of ozone,

ozone-hydrogen peroxide system and catalytic ozone on the biodegradable organic
matter of a fulvic acid solution" Water Research, 31(3), 650-656.

107

CHAPTER SIX

Karnik, Bhavana 8.; Davies, Simon H.; Baumann, Melissa J .; Masten, Susan J. (2006).
AFM and SEM Characterization of Iron Oxide Coated Ceramic Membranes. Accepted
for publication in Journal of Materials Science.

108

CHAPTER SIX
AFM and SEM Characterization of

Iron Oxide Coated Ceramic Membranes

6.1 ABSTRACT

Alurnina-zirconia-titania (AZT) ceramic membranes coated with iron oxide nanoparticles
have been shown to improve water quality by significantly reducing the concentration of
disinfection by—product precursors, and in the case of membrane filtration combined with
ozonation, to reduce ozonation by-products such as aldehydes, ketones and ketoacids.
Commercially available ceramic membranes with a nominal molecular weight cut-off of
5 kilodaltons (ch) were coated 20, 30, 40 or 45 times with sol suspension processed
FezO3 nanoparticles having an average diameter of 4-6 nm. These coated membranes
were sintered in air at 900 0C for 30 minutes. The effects of sintering and coating layer
thickness on the microstructure of the ceramic membranes were characterized using
atomic force microscopy (AFM), scanning electron microscopy (SEM) and energy
dispersive x-ray spectroscopy (EDS). AFM images show a decreasing roughness after
iron oxide coating with an average surface roughness of ~ 161 nm for the uncoated and ~
130 nm for the coated membranes. SEM showed that as the coating thickness increased,
the microstructure of the coating changed fiom a fine grained (average grain size of ~ 27
nm) morphology at 20 coating layers to a coarse grained (average grain size of ~ 66 nm)

morphology at 40 coating layers with a corresponding increase in the average pore size

109

from ~ 57 nm to ~ 120 nm. Optimum water quality was achieved at 40 layers, which
corresponds to a surface coating morphology consisting of a uniform, coarse-grained
structure with open, nano-sized interconnected pores.

Keywords: ceramic membrane, nanosized, iron oxide, catalytic coating, scanning electron
microscopy, atomic force microscopy, nanofiltration

6.2 Introduction

In the United States there is an increasing interest in the application of both ozone and
membrane filtration for disinfection by-products (DBPS) and DBP precursor removal in
order to meet the requirements of the US Environmental Protection Agency’s (US EPA)
Surface Water Treatment Rule (SWTR), the Disinfectant and Disinfectant By-products
Rule (D/DBPR) and the Long Term 1 Enhanced Surface Water Treatment Rule
(LTl ESWTR). During the last decade, researchers have attempted, with limited success,
to combine ozonation and polymeric membrane filtration as a water and wastewater
treatment option (Hashino et al., 2000; Shanbhag et al., 1998; Castro and lander, 1995;
Shen et al., 1990). The increasing severity of operating parameters, including higher
temperatures and pressures, higher resistance to chemicals and overall durability, made
ceramic membranes the natural choice in spite of their much higher costs (Zuzek et al.,
2001). Potential applications for ceramic membranes include separation, purification,
catalysis and chemical sensors at high temperatures as well as use in chemically reactive
environments (Zeng et al., 1997). Ceramic membranes are ozone resistant and when used
in combination with ozone, can achieve stable permeate fluxes without membrane
damage (Karnik et al., 2005a; Schlichter et al., 2004; Kim and Somiya, 2001; Kim et al.,

1999; Allemane et al., 1993).

110

Our earlier work showed that stable fluxes could be obtained with ozonation in
combination with ceramic membrane filtration (Karnik et al., 2005a; 2005b). Catalytic
degradation of ozone at the membrane surface is thought to oxidize foulants that
accumulate at the membrane surface, thereby preventing membrane fouling. Ozonation-
filtration resulted in a reduction of 50% in the dissolved organic carbon (DOC)
concentration. It also resulted in the formation of partially oxidized compounds from
natural organic matter (NOM) that were less reactive with chlorine, decreasing the
concentration of carcinogenic compounds such as trihalomethanes and haloacetic acids.
The formation of simulated distribution system total trihalomethanes (SDS TTHMs) and
simulated distribution system halo acetic acids (SDS HAAs) were decreased by up to
80% and 65%, respectively (Karnik et al., 2005a; 2005b).

Based on extensive research involving various ozonation methods for drinking
water treatment, catalytic ozonation has been determined to be one of the best alternatives
for oxidizing NOMs and reducing the demand for chlorine, a common disinfectant used
in water purification (Volk et al., 1997; Allemane et al., 1993). In the presence of
different metal oxide catalysts, such as iron oxide, manganese oxide, titania, alumina and
zirconia, ozone degrades organic compounds, including known harmful and potential
carcinogens like saturated carboxylic acids, phenols, aromatic hydrocarbons, dyes, humic
substances and herbicides (Karnik et al., 2005c). We have developed a novel procedure
based on a layer-by layer method (McKenzie et al., 2002) for coating alumina-zirconia-
titania (AZT) nano-crystalline ceramic membranes. Iron oxide coated membranes
reduced the concentration of DOCs by >85% and the concentrations of SDS TTHMs and

SDS HAAs by up to 90% and 85%, respectively, compared to that found with untreated

111

water (Karnik et al., 2005c). Similarly, the iron oxide coated AZT membrane reduced the
concentrations of ozonation disinfection by-products (aldehydes, ketones, and ketoacids)
in the permeate by >50%, as compared to that obtained using uncoated membranes
(Karnik et al., 2005c).

Surface modification is significantly affected by surface morphology; it is an
important way to enrich the functionality of the ceramic membranes. Atomic force
microscopy (AFM), scanning electron microscopy (SEM) and energy dispersive x-ray
spectroscopy (EDS) have been used to examine ceramic membrane surfaces. Researchers
have successfully characterized the fabrication and microstructure of ceramic membranes
derived from alumoxane, ferroxane nanoparticles and A1203 - Al nano-composite
powders (Lu et al., 2005; Cortalezzi etal., 2003 ; 2002). AFM, SEM and EDS have been
used to characterize these coatings on titania membranes, composites of alumina-titania,
metal doped ceramics and similar ultra and nanofiltration membranes (Lu et al., 2005;
Siriwardane et al., 2000; Chou et al., 1999; Bee et al., 1998; Pedersen et al., 1997).

In this work, AFM, SEM and EDS were used to investigate the surface
characteristics of the iron oxide nanolayered coated AZT ceramic membranes. A
suspension of iron oxide nanoparticles was passed over the AZT membrane surface 20,
30, 40 or 45 times followed by sintering at 900 °C. The water quality analysis performed
on the permeate of different membrane coatings did not show any significant
improvement in the reduction of DBPs concentrations neither was there increased
removal of DBP precursor with the increase in the number of catalyst coatings (Karnik
2005c). However, the ozonation by-products monitored in the permeate showed a

significant reduction in concentration with increasing number of catalyst coatings from

112

20 to 40. No significant reduction in the concentrations of the ozonation by-products was
reported for 60 coating layers making 40 coatings the optimum choice in terms of water
quality performance and biological stability.

6.3 MATERIALS AND METHOD

6. 3. I Membrane Preparation

Tubular AZT (a mixture of alumina, zirconia and titania) ceramic membranes
(Clover-leaf design (containing three channels), CéRAM Inside, TAMI North America,
St. Laurent, Québec, Canada) with nominal molecular weight cut-offs of 5 kilodaltons
(kD) were used as a support for the iron oxide catalytic coatings. The external diameter of
each membrane was 10 mm and the active membrane length was 8 cm. The total filtering
area of each membrane was approximately 11 cm2 with each membrane operational from
pH 0-14. The initial permeability of the membranes was tested using distilled deionized
(DDI) water (Karnik 2005a).

A detailed description of the membrane preparation is available in our earlier
published work (Karnik 20050). The colloidal particles used for coating the membranes
were prepared using Sorum’s method (McKenzie et al., 2002). Transmission electron
microscopy (TEM) characterization showed that the average particle diameter was 4 to 6
nm. The layer-by-layer technique used to coat the membranes is based on a protocol
described by McKenzie et al. (2002) for coating doped tin oxide electrodes. The
membrane was immersed in the colloidal suspension for one minute and then rinsed with
DDI water. Then, the membrane was immersed in aqueous phytic acid (40 mM) for one
minute and rinsed with DDI water. This sequence was repeated the desired number of

times (20, 30, 40 or 45). After coating, the membrane was sintered at 900 °C for 30

113

minutes. This temperature was chosen to produce membranes on which the iron oxide
particles were completely sintered to each other and to the membrane surface. These
sintered membranes were then examined using AF M, SEM and EDS.

6. 3.2 Characterization of Membranes

To obtain images of the coated surface of the tubular ceramic membranes, the membrane
was first sliced into circular discs of 1 mm thickness using a diarnond-wafering saw.
Subsequently, these sections were cut to form small arcs of length 3 mm and width 1 mm.
These arcs were then mounted on aluminum discs for AFM and aluminum mounting
stubs for SEM using carbon adhesive tape. The schematic representation of this
procedure is shown in Figure 6.1. The samples were then imaged and data collected using
AFM, SEM and EDS.

AFM images of the uncoated and coated membranes were obtained using a
Nanoscope IV Multimode Atomic Force Microscope (Digital Instruments Inc.), in
ambient air in contact mode, which is ideal for examination of textured samples like
ceramics. Silicon nitride (Si3N4) NP triangular cantilever probes were used to image the
iron oxide coated membranes along with uncoated membranes for comparison purposes
(Digital Instruments, Veeco Metrology group, CA) with a cantilever spring constant of
0.12 N/m and a frequency of 20 kHz. The tip has a nominal radius of curvature of 20 nm
with a height 2.5-3.5 um and a side angle of 35°. Scans of 20 um X 20m were taken at a
scanning rate of ~ 0.5 Hz. Height and deflection data was taken simultaneously for the
same scan area. The 3D surface plot and roughness analysis using the height data were

performed on the images to study the surface morphology.

114

Iron oxide (Fe203)
coafing

‘ AZT Ceramic

 

membrane
Membranes from supplier Slice of membrane with
Coated by layer by layer method three channels
STEP: 2 AZT Ceramic membrane
Cut a X-section “k
lron oxide (Fe203)
coafing
Slice of membrane X-section Sample
(1 mm thickness) (3mm length x 1mm width thickness)
STEP: 3
Direction of viewing for Microscope Direction of viewing for
Microscope
1 Sample 1 Sample
Aluminum Aluminum '/
mounting disc mounting stub ’7—
AFM (Front view) SEM (Front view)

Figure 6.1 Schematic Representation of sample preparation for
SEM and AFM imaging

115

SEM images of the membranes were obtained using a JEOL 6400V scanning
electron microscope equipped with a La36 emitter operated at an accelerating potential of
15 kV at magnifications from 5,000X to 60,000X. The mounted samples were gold
coated using an Emscope SC 500 sputter coater at a rate of 7 urn/min with 20 mA current.
EDS microanalysis was performed on the samples using a Noran EDS analyzer (Noran
Instruments Inc.) at accelerating potential of 20 kV and magnifications ranging from
lOOX to 5000X. Samples were carbon coated using an EFFA Mk H carbon coater (Ernest
Fullam Inc., Latharn NY) in preparation for EDS analysis.

EDS microanalysis has a in built software module to measure the average grain
size using the line intercept method. The grain sizes are measured using the average
calculated fi'om five rnicrographs for each sample where on average; 200 grains were
measured per micrograph. ANOVA testing was performed on the grain size data within
90% confidence intervals to determine if the differences in the measurements were
significant. An identical procedure was followed for the pore size measurements,
however instead of grain size module; arbitrary distance was measured between the
grains to determine the pore size.

The natural organic matter (N OM) and the DBP precursors were monitored in
terms dissolved organic carbon concentrations (DOC) (Standard Methods, 1998). The
DBPs total trihalomethanes (TTHMs) and halo acetic acids (HAAs) were measured using
standard methods and reported elsewhere (Karnik 2005b; 2005c). Similarly, ozonation
by-products (aldehydes, ketones and ketoacids) were measured using USEPA standard

methods (Munch et al., 1998).

116

6.4 RESULTS AND DISCUSSION
6. 4.] AFM imaging

AF M analysis provided data on the surface morphology and surface roughness.
The manner in which these properties correlate with the surface porosity and filtration
performance provide insight into the structure of the filtration membrane. The surface
roughness from AFM measurements can be correlated to the grain size found using SEM.
Figure 6.2 a-c shows AFM images of a typical 5 kD uncoated AZT membrane, an
uncoated AZT membrane sintered at 900 °C and an AZT membrane coated 40 times with
the iron oxide nanoparticle suspension and sintered in air at 900 °C.

For each AFM image, the area in view represents a 20 um X 20 um square. The
features within any given sample are relatively uniform throughout the sample. With
sintering, the surface of the AZT uncoated membrane (Figure 6.2a) undergoes a gradual
transition from flat featureless regions of ~ 2.5 11m (:1: 0.2) height to more sharp surface
features of ~ 3 um (d: 0.08) height (Figure 6.2b). With coating and sintering (Figure 6.2c),
the height of these features reduces to ~ 1.5 um (i 0.1). When comparing membranes
coated for 20, 30 and 40 times followed by sintering, the respective AFM height data
(plotted in Figure 6.3) show no statistical difference between membranes coated 30 and
40 times. However, both are significantly decreased in comparison to the membranes

coated 20 times.

117

a) 5kD uncoated

 

X axis- 5.000 nmldlv
Z axis- 2000.000 nmldlv

b) 5kD uncoated- sintered (900°C)

 

15

X axis- 5.000 nmldlv
Z axls- 3000.000 nmldlv

Figure 6.2 AFM images of the AZT ceramic membranes: a) 5kD MWCO AZT
membrane uncoated, b) 5kD AZT membrane uncoated, sintered at 900 °C for
30 minutes

118

c) 5kD - 40 coatings- sintered (900°C)

  

X axis- 5.000 nmldlv
Z axis- 1500.000 nmldlv

Figure 6.2 c) AFM images of the AZT ceramic membranes: 5kD AZT
membrane with 40 coatings of iron oxide, sintered at 900 °C for 30 minutes

 

 

5.00

4.00 l

E 3.00 ~
g 2.00 -

1.00 -

       

 

 

. . , .ii *
- ».. wit, 4. r. ml" me. it ..

 

 

0.00 17

5kD-uncoated 5kD-900'C 5kD-20-OOO'C 5kD-304001: 5kD40-900'c
Membrane

 

 

 

Figure 6.3 Height data for AFM images of membranes with different
number of coatings

119

However, while the height of each feature is not significantly reduced as the number of
coats increases (Figure 6.4), there is an increase in the number of these features. This
result is substantiated by AFM roughness analysis which shows a significant decrease in
roughness from ~ 161 nm to ~ 78 nm upon sintering of the uncoated AZT membrane.
There is also a reduction in the surface roughness from ~ 161 nm for the uncoated
membrane to an average of ~ 130 nm for the coated membranes, with no statistical

difference in the rouglmess with an increase in the number of coats.

a) 5kD - 20 coatings

 

x axis- 5.000 umldiv
z axis- 1500.000 nmldlv

Figure 6.4 a) AFM image of an AZT ceramic membrane with iron oxide coating:
5kD AZT membrane with 20 coatings of iron oxide, sintered at 900 °C for 30
minutes

120

b) 5kD - 30 coatlngs

 

X axis- 5.000 nmldlv
Z axis- 1500.000 nmldlv

Figure 6.4 b) AFM image of an AZT ceramic membrane with iron oxide coating:
5kD AZT membrane with 30 coatings of iron oxide, sintered at 900 °C for 30
minutes

121

c) 5kD - 40 coatings

 

10

   

X axls- 5.000 nmldlv
Z axis- 1500.000 nmldlv

Figure 6.4 c) AFM image of an AZT ceramic membrane with iron oxide coating:
5kD AZT membrane with 40 coatings of iron oxide, sintered at 900 °C for 30
minutes.

122

For all membranes tested, the observed magnitude of the roughness, shown in
Figure 6.5, is much greater than the size of the original iron oxide particles (4-6 nm). As
shown in the SEM rrricrographs in Figure 6.6, this increase in roughness is due to
interparticle sintering between the iron oxide nanoparticles and, given the integrity of the

iron oxide nanocoating, a result of the coating sintering to the underlying porous ceramic

 

 

 

 

 

 

AZT membrane.

300.00 :
A 250.00 < ‘
E
5 200.00 ~
«0
er
3 150.00 - , l
g .
§’ 100.00 — l
°‘ I l l

50.00 -
0.00 . . , , !
5kD-uncoated 5kD-900°C 5kD-20-eOO°c 5kD-30-900°c 5kD40-900’C
Membrane

 

 

 

Figure 6.5 Roughness data for AFM images of membranes with different
number of coatings

123

a) Uncoated - Unsmtered

b) Uncoated - srntered .__ .g.

(1)30 coatings ~ sintered f) 45 coatings - smiered
. ‘r ' \ - '

I b
A

        

Eislum

Figure 6.6 SEM images of an AZT ceramic membrane: a) 5kD MWCO AZT
membrane uncoated, b) 5kD AZT membrane uncoated, sintered at 900°C for 30
minutes and c — f) 5kD AZT membrane with 20, 30, 40, 45 coatings, respectively,
of iron-oxide, sintered at 900°C for 30 minutes.

124

6. 4.2 SEM imaging

The SEM micrographs (Figure 6.6) of the coated AZT ceramic membranes exhibit a
similar coarsening behavior of the membrane surface with increases in the size and
number of surface grains with number of coats, as was found using AFM (Figure 6.4 a-c).
The nanoparticles on the surfaces have sintered together and there is an overall
coarsening of the surface (increase in the average grain size) as the number of coatings
increases from 20 to 40.

The average grain size for the membrane surfaces are plotted in Figure 6.7. After
sintering for 30 minutes, the average grain size increased from ~ 21 um (i 0.24), for the
uncoated membranes, to ~ 66 nm (:1: 20.0) for the coated membranes. Further increasing
the number of coatings from 20 to 30 and subsequently 30 to 40, resulted in a significant
increase in the average grain size from ~ 27 um (i 10) to ~ 31 nm (:1: 11) to ~ 66 nm (at
23) respectively. This particle growth is likely a result of the large driving force for
sintering posed by the high surface area of these nanosized particles, whereby
agglomerated regions of nanoparticles rapidly sinter and are separated by larger pores
(Barsoum, 2003). This finding is verified by noting that both the average grain size for
the sintered membranes coated 40 times (Figure 6.7, 5 kD-40-900 °C) and the average
pore size following sintering (Figure 6.8, 5 kD-40-9OO °C) have both increased over
those average grain and pore sizes reported for 20 and 30 coatings. This indicates that a
greater degree of agglomeration of the iron oxide particles has occurred for membranes

coated 40 times.

125

 

 

 

 

 

 

 

 

 

 

 

 

 

100
E
5 80 —
o
it
.5 6° ‘ 1
L“
(D 40 -
o
8
2 2°“ W 1 ,
< .
0 I 7 ' i, l d a
5kD-uncoated 5kD-20-900°C 5kD-30-900°C 5kD-40-900°C
Men'brane

 

 

Figure 6.7 Grain size measurements of AZT ceramic membrane:
Membranes: 5kD MWCO uncoated, 5kD with 20, 30, 40 coatings of iron
oxide sintered at 900 °C for 30 minutes.

 

 

 

Average Pore Size (nm)

 

 

"f”.a, m W W ill ..

5kD-uncoated skoeoo'c same-owe sumo-owe 51040-000’0 WC
Membrane

 

 

Figure 6.8 Average pore size measurements for the AZT membranes: 5kD
MWCO uncoated, 5kD sintered at 900 °C for 30 minutes, 5kD with 20, 30, 40,
45 coatings of iron oxide sintered at 900 °C for 30 minutes and 5kD with 40
coatings of iron oxide unsintered.

126

 

With 45 coatings, no significant increase in porosity is observed, and more
importantly, no water quality improvements were found, making 40 coatings a critical
processing parameter. So, while the average surface pore size has increased from 40 nm
(i 10) for the uncoated membrane to 120 nm (:1: 40) for the membranes coated 40 times,
the more open porosity has significantly increased the water quality while maintaining
the average pore size at the nanoscale.

SEM micrographs (Figure 6.6) show a clear evidence of uniform coverage of
coating and sintering as verified by AFM results shown in Figures 6.2 and 6.4. Also the
coarsening of the membrane surface explains the decrease in roughness value from
uncoated membranes to coated and sintered membranes.

Energy dispersive x-ray microanalysis (EDS) mapping was done for Ti, Al, Zr, 0
and Fe. EDS mapping of the uncoated AZT membrane (Figure 6.9a) showed a uniform
distribution of titania and zirconia over a porous alumina matrix. The skin of the as-
received uncoated membrane was therefore a mixture of titania and zirconia which
formed an ultrafiltration layer (ultrafiltration occurs between microfiltration (1045 m) and
nanofiltration (10'9 m)). EDS mapping of a coated and sintered membrane (Figure 6%)
confirmed the morphology and composition of the uncoated membrane with the addition
of an iron oxide layer predominantly present at the surface, with a unifonnly diffused
iron oxide presence into the membrane surface. This uniform distribution of iron oxide
into the membrane could be a result of capillary action during the coating process and/or

a result of the diffusion of iron oxide nanoparticles in the sintering process.

127

a) 5kD uncoated

15 gm

 

b) 5kD - 40 coatings- sintered (900°C)

 

Figure 6.9 EDS mapping of the membranes: a) EDS mapping of
uncoated membrane and b) EDS mapping of membrane with 40
coatings and sintered at 900 °C for 30 minutes.

128

As expected, EDS line scans for the membranes coated 20, 30 and 40 times
revealed a corresponding increase in the concentration of iron (Fe) present in the

membrane surface. The Fe concentration was proportional fi'om 20 to 30 to 40 coatings

(Figure 6.10).

 

 

 

 

 

10
14, 0W0 DSirD-ZO-OOO’C A5kD-30-000‘c 051040-0001: ill
12 0 (ii
310‘ m
B
3 a a x
O
.
... u a
I 01
44 I
0 I
2* B
5
0I ‘r 6 9 9 6
0 2M 400 000 5M 1M 1200

 

 

 

Figure 6.10 Relative Fe concentration from EDS scans.
The graph represents relative Fe concentrations measured as Fe counts in the EDS scans

The water quality data shown in Figure 6.11 gives evidence as to how the catalyst
coating has improved water quality performance in terms of reducing DBP precursors
when compared to uncoated membranes. Iron oxide coated ceramic membranes were
superior in terms of performance as compared to uncoated ceramic membranes in terms
of DBP precursors as well as DBPs as shown in Figure 6.11. DOC concentrations showed
a significant decrease for the F e203 catalyst coated ceramic membranes with increasing
Fe203 coating layers when compared to uncoated membranes. Our earlier work details
the improvement in water quality for R203 catalyst coated ceramic membranes with
different treatment processes in comparison to the uncoated ceramic membranes (Karnik

et al., 20050).

129

 

 

 

amonoome ISO-20mm IMGOID-OOOC IMMMOOC lib-45mm]

 

 

 

 

 

    
 

 

 

 

200
00”
3 5000—
34m.
150 - S
3000‘
A E20”-
3 .....-
or
C o-
f 100 . DI-olvod Organlc Carbon
:
8
I:
O
U
50 4

 

 

 

 

Tohl Trihalomethanes Halo acetic acid: NdehydealK-tonea KItoIddI

 

Figure 6.11 Water Quality data for the permeate after combined ozonation
membrane treatment process

The data for water quality parameters for a 5kD AZT membrane uncoated, unsintered, with 20, 30, 40, 45
coatings of iron-oxide, sintered at 900°C for 30 minutes. The insert in the graph is the plot of dissolved
organic carbon concentrations for the same membranes. Experiment details (Karnik et al., 2005c).

Fe203 catalyst coated ceramic membranes are a promising tool for reducing the ozonation
by-products which serve as substrates for growth of microorganisms. This reduces their
regrth potential in the permeate making the water more biologically stable and safe for
consumption. Further, the figure shows no significant changes in the water quality in
terms of measured concentration of DBPs like TTHMs and HAAs with increasing Fe203

catalyst coating layers from 20 to 40 coats. A concomitant decrease in the ozonation by-

130

 

products was seen with increasing number of Fe2O3 coating layers, however, no
significant changes in the concentrations were reported beyond 40 coating layers.

AFM characterization showed a decrease in the surface roughness of the ceramic
membrane with Fe203 coating which led to an improved effective filter separation layer
most likely comprised of nanosized iron oxide grains. SEM micrographs show a
nanoscale average pore size and uniform coverage of the coating layers, not only across
the ceramic membrane surface, but also into the membrane itself. This has likely led to
the catalytic reactions that resulted in a significant improvement of water quality in terms
of removal of disinfection by-products and ozonation by-products shown in Fig.11. The
increased Fe concentration into and away from the outer membrane surface, as measured
by EDS, further supports the explanation given for the improved water quality data for
the ceramic membranes coated with 40 layers of R203 nanoparticles.

6.5 CONCLUSIONS

Coating and sintering of AZT membranes with nanoscale iron oxide particles
resulted in significant changes in the membrane surface morphology as a result of
sintering and coarsening of the coating nanoparticles. SEM details the changes in surface
morphology of the coated membrane where the surface morphology changes from a fine
grained uniform structure at 20 coats to a coarser grain uniform structure at 40 coats.
SEM also captured the change in the average pore size, fiom the micropores in the
underlying AZT membrane to the nanopores within the iron oxide surface layer of the
coated membrane. This decrease in porosity into the nanopores regime is one possible
reason for the improved performance of these iron oxide coated ceramic membranes over

uncoated membranes. AFM and SEM data are consistent, where decreasing surface

131

roughness correlates with a coarsened average grain size on both the sintered uncoated
and coated membranes, with the smoothest and coarsest surface existing at 40 coats,
which is the Optimum in terms of water filtration (Karnik et al., 20050). It is at 40 layers
that we have the largest average pore size (although still in the nanopores range at 120
nm (:t 40)) and largest average grain size that results fi'om the greater degree of
agglomeration of the iron oxide particles during the coating process. Capillary action
during the coating process, and/or diffusion during sintering, results in the uniform
distribution of iron oxide particles throughout the membrane interior. This further
enhances water filtration because of the increased exposure to the catalytic iron oxide, not
only at the membrane surface, but into the membrane itself. Ongoing studies using SEM
of coated and unsintered membranes will determine whether capillary forces during the
coating application process are sufficient to drive the iron oxide nanoparticles into the
interior of the membrane. We have found that 40 coats of the nanosized iron oxide
particles on the underlying AZT ceramic membrane is the optimum coating in terms of
water quality performance which meets the stringent EPA regulatory requirements, while
still in the regime of nanofiltration. Future research will examine the mechanisms for the
degradation of NOM and the removal of harmful DBPs by the iron oxide coated AZT
ceramic membranes.

6.6 ACKNOWLEDGEMENTS

The authors would like to acknowledge the US Environmental Protection Agency (US
EPA) Science To Achieve Results (STAR) Program (Grant No. RD830090801) for
financial support of this work. Our thanks also go to Ms. Ewa Danielewicz, from the

Center for Advanced Microscopy for her assistance during SEM sample preparation. Dr.

132

Hazel Hosein from the Composite Materials and Structures Center is also acknowledged
for her assistance during the AFM work. We would also like to thank Mr. David Jackson

for his assistance with the membrane coating process.

133

6.7 REFERENCES

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ O3-H2O2 and 03/1“ iO2) for the oxidation of natural organic matter
in water” Ozone: Science and Engineering, 15(5), 419-432.

Barsoum, M. (2003). In. "Fundamentals of ceramics" (Institute of Physics Publishing,
2003), p.336.

Bae, D., Cheong, D., Han, K., and Choi, S. (1998). "Fabrication and microstructure of
Al2O3-TiO2 composite membranes with ultrafine pores" Ceramics International,
24(1), 25-30.

Castro, K. and Zander, A.K. (1995). "Membrane air-stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

Chou, K., Kao, K., Huang, C., and Chen, C. (1999). "Coating and characterization of
titania membrane on porous ceramic supports" Journal of Porous Materials, 6(3),
217-225.

Cortalezzi, M., Rose, J ., Barron, A., and Weisner, M. (2002). "Characteristics of
ultrafiltration ceramic membranes derived from alumoxane nanoparticles" Journal of
Membrane Science 205(1), 33-43.

Cortalezzi, M., Rose, J ., Wells, G., Bottero, J ., Barron, A., and Weisner (2003). "Ceramic
membranes derived from ferroxane nanoparticles: a new route for the fabrication of
iron oxide ultrafiltration membranes" Journal of Membrane Science, 227 (1-2), 207-
217.

Hashino, M., Mori, Y., Fujii, Y., Motoyarna, N.; Kadokawa, N., Hoshikawa, H.,
Nishijima, W., and M. Okada, M., (2000). "Pilot plant evaluation of an ozone-
microfiltration system for drinking water treatrnen " Water Science T echnologi, 41,
17-23.

Karnik, B.S., Davies, S.H., Chen, K.C., Jaglowski, D.R., Baumann, M.J. and Masten, SJ.
(2005). "Effects of ozonation and pH on the permeate flux of nanocrystalline ceramic
membranes" Water Research, 39(4), 728-734.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005b). "The effects of
combined ozonation and filtration on disinfection by—product formation" Water
Research, 39(13), 2839-2850.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, 8.1. (20050). "Fabrication of

catalytic membranes for the treatment of drinking water using combined ozonation
and ultrafiltration" Environmental Science and Technology, 39, 7656-7661.

134

Kim, J. O., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14’”
Ozone World Congress, Dearborn, MI, 131-143.

Kim, J.O. and Somiya, I., (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

Lu, H., Hu, J., Chen, C., Sun, H., Hu, X., and Yang, D. (2005). "Characterization of
Al2O3—A1 nano-composite powder prepared by a wet chemical method" Ceramics
International, 31(3), 481-485.

McKenzie, K.J., Marken, F., Hyde, M., and Compton, R.G. (2002). "Nanoporous
ironoxide membraneszLayer-by-layer deposition and electrochemical characterization
of processes within nanopores" New Journal of Chemistry, 26, 625-629.

Munch, J.W., Munch, D.J., Winslow, S.D., Wendelken, S.C., Pepich, B.V., (1998).
"Determination of carbonyl compounds in drinking water by

pentafluorobenzylhydroxylamine derivatization and capillary gas chromatography
with electron capture detection" Method 556.1, USEPA, Cincinnati, Ohio.

Pedersen, H., Tranto, J., and Haj, J .W. (1997). "Characterization of multilayer ceramic
membranes with the atomic force microscope" Key Engineering Materials, 132-136,
1707-1710.

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

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

Shen Z.S., Semmens M.J., and Collins A.G. (1990). "A novel-approach to ozone water
mass-transfer using hollow fiber reactors" Environmental Technology, 1990, 11, 597-
608.

Siriwardane, R., Poston jr., J., Fisher, E., Lee, T., Dorris, S., and Balachandran, U.
(2000). "Characterization of ceramic hydrogen separation membranes with varying
nickel concentrations "Applied Surface Science, 167(1-2), 34-50.

Standard Methods for Examination of Water & Wastewater (1998). L. S. Clesceri, A.E.

Greenberg, A.D. Eaton Eds., 20th edition, Publishers American Public Health
Association.

135

Volk, P. Roche, Joret, J. and Paillard, H (1997). "Comparison of the effect of ozone,
ozone-hydrogen peroxide system and catalytic ozone on the biodegradable organic
matter of a fulvic acid solution" Water Research, 31(3), 650—656.

Zeng, Z., Xiao, X., Gui, Z., and Li, J. (1997). "AFM study on surface morphology of
Al2O3-SiO2-TiO2 composite ceramic membranes" Journal of Membrane Science, 136,
153-160.

Zuzek, E., Catan, S., Arciprete, C., Dirnitrijewits, M., Almandoz, M.,, and Marchese, J.

(2001). "Asymmetric ceramic microporous membranes" Granular Matter, 3(1-2),
145-148.

136

CHAPTER SEVEN

TEM CHARACTERIZATION OF
IRON OXIDE COATED CERAMIC MEMBRANES

7.] ABSTRACT

Commercially available porous alumina-zirconia-titania ceramic (AZTC) membranes
having a titania surface coating were characterized using transmission electron
microscopy (TEM). TEM photomicrographs showed the as-received AZTC membrane to
be a multi-layered structure consisting of a porous almnina-zircona-titarria core having
ultrafine pore sizes, coated by an additional layer of micro porous titania. Electron
diffraction studies revealed an amorphous surface titania layer while the underlying
AZTC membrane was crystalline. The AZTC membranes were coated with iron oxide
(F0203) nanoparticles, synthesized from a colloidal sol suspension 20, 30, 40, 45 or 60
times using a layer-by-layer procedure after which the membranes were sintered in air at
900 °C for 30 minutes. Following Fe203 coating and sintering, TEM revealed relatively
uniform Fe2O3 coverage, with nanoporosity in the Fe203 layer, where the F e203 coating
thickness increased with increasing number of layers. Electron diffraction patterns
showed the Fe2O3 coating to be crystalline in nature, verified by x-ray diffraction (XRD)
results showing the structure to be or-Fe203. The average pore size of the underlying
AZTC membrane increased after the Fe2O3 coated membrane was sintered. However, no
significant increases in the average pore size were observed beyond 40 layers of Fe2O3.

Coating the membrane with iron oxide particles improved the catalytic properties of the

137

membrane when used in water treatment applications, where the maximum benefit, in
terms of water quality improvements, were found at 40 layers of F 0203.

Keywords: ceramic membrane, nanoparticles, iron oxide, catalytic coating, transmission
electron microscopy (TEM), x-ray diffraction (XRD), nanofiltration

7.2 INTRODUCTION

The use of ceramics as catalyst materials is a well accepted practice (Keane, 2003) and
has led to the development of ceramic materials that are effective catalyst supports and
catalytic agents. Recent advances in our ability to manipulate structures at the molecular
and atomic levels have further advanced the use of nanosized ceramics as catalytic
materials (Keane, 2003; Trudeau and Ying, 1996). Ceramic catalysts are used in the
production of commodity chemicals and pharmaceuticals, as well as finding increasing
application in environmental pollution control and abatement (Keane, 2003). Mixed
metal oxides have displayed promising catalytic properties in addition to improved
structural and acid-base properties (N eri et al., 2004).

Membrane filtration is an effective technique in water treatment for the removal
of particulate matter, micro-organisms and organic matter (US EPA, 2001). During recent
years, there has been increasing interest in the application of micro porous ceramic
membranes because of their chemical, mechanical and thermal stability (Puhlfier et al.,
2000). A combined water treatment process of ozone disinfection and polymeric
membrane filtration has been attempted by several researchers with very little limited
success (Hashino et al., 2000; Shanbhag et al., 1998; Castro and Zander, 1995; Shen et
al., 1990). In contrast to polymeric membranes, ceramic membranes are ozone resistant

and when these membranes are used in combination with ozone, stable permeate fluxes

138

can be achieved without membrane damage (Karnik et al., 2005a; Schlichter et al., 2004;
Kim and Somiya, 2001; Kim et al., 1999; Allemane et al., 1993).

Our earlier work showed stable permeate fluxes could be maintained using
uncoated AZTC membranes in a combined ozonation-membrane filtration process
(Karnik et al., 2005a). Further combined ozonation and membrane filtration resulted in a
decrease in the concentration of disinfection by—products (DBPS), such as total
trihalomethanes (TTHMs) and halo acetic acids (HAAs), of up to 80% and 65%,
respectively. This results from the formation of partially oxidized compounds from
naturally organic matter (NOM) that were less reactive with chlorine (Karnik et al.,
2005b). When the removal of ozonation by-products was investigated using a combined
ozonation-membrane filtration system, where now the AZTC membranes were coated
with R20; nanoparticles, a firrther reduction of at least 50% was measured when
compared to the combined ozonation and uncoated AZTC membrane filtration system
alone. A 5 kilodalton (kD) nominal molecular weight cut off (MW C0) AZTC membrane,
coated with 40 layers of Fe203 nanoparticles and sintered in air at 900 °C, combined with
ozonation (at a gaseous ozone concentration of 2.5 g/m3) produced a permeate water that
met the US EPA regulatory requirements for TTHMs of 80 ug/L and HAAs of 60 ug/L
set under Stage 2 D/DBPs Rule for drinking water (Karnik et al., 20050).

The effects of sintering temperature and coating layer thickness on the
microstructure of the commercially available AZTC membranes coated with sol
suspension processed Fe203 nanoparticles have previously been characterized in our
laboratory using atomic force microscopy (AFM), scanning electron microscopy (SEM)

and energy dispersive x-ray spectroscopy (EDS) (Karnik et al., 2006). Our results showed

139

decreasing surface roughness after Fe203 coating, while an increase in the F e203 coating
thickness caused a change in the microstructure from a fine-grained morphology at 20
coating layers (average grain size 27 i 10 nm) to a coarser grained morphology at 40
coating layers (average grain size 66 i 23 nm) with a corresponding increase in the
average pore size fi'om 57 d: 15 nm to 120 d: 40 nm. Optimum water quality was achieved
at a coating of 40 layers of R203, corresponding to a surface having a uniform, coarse-
grained (average grain size 66 :l: 23 nm) structure with open, nano-sized (66 :l: 23 nm)
interconnected pores (Karnik et al., 2006).

In this work, transmission electron microscopy (TEM) and x-ray diffraction
(XRD) were used to further investigate the characteristics of the F0203 nanoparticle
coated AZTC membranes. A suspension of F e203 nanoparticles was layered onto the
AZT C membrane surface 20, 30, 40, 45 or 60 times followed by sintering in air at 900
°C. The surface area measurements and average pore size of the AZTC membrane were
measured using Brunauer-Emmett-Teller (BET) and the Barett, Joyner and Halenda
(BJH) methods, respectively.

7.3 MATERIALS AND METHODS
7. 3. 1 Membrane Preparation

Tubular AZTC (a mixture of alumina, zirconia and titania) membranes (CéRAM
Inside, TAMI North America, St. Laurent, Quebec, Canada, shown in Figure 7.1) with a
clover-leaf design (containing three channels), with nominal molecular weight cut-offs of
5 kilodalton (kD) were used as a support for the Fe203 catalytic coatings. The external
diameter of each membrane was 10 mm and the active membrane length was 8 cm. The

total filtering area of each membrane was approximately 11 cm2. The initial permeability

140

of the membranes was determined using distilled deionized (DDI) water (Karnik et al.,
2005a).

A detailed description of the membrane preparation is available in our earlier
published work (Karnik et al., 20050). Briefly, the colloidal Fe203 nanoparticles used for
coating the AZT ceramic membranes were prepared using Sorum’s method (McKenzie et
al., 2002). Transmission electron microscopy (TEM) characterization of the F e203
nanoparticles showed that the average particle diameter was 4 to 6 nm (Figure 7.2). The
layer-by-layer technique used to coat the AZTC membranes is based on a protocol
described by McKenzie et al. (McKenzie et al., 2002) for coating doped tin oxide
electrodes. The AZTC membranes were immersed in the colloidal suspensions for one
minute and then rinsed with DDI water. Then, the membranes were immersed in aqueous
phytic acid (40 mM) for one minute and again rinsed with DDI water. This sequence was
repeated for the desired number of layers (20, 30, 40, 45 or 60). After coating, the Fe203
coated AZTC membranes were sintered in air at 900 °C for 30 minutes. This temperature
was chosen to produce membranes on which the R203 particles were sintered together as
well as to the underlying AZTC membrane surface. These sintered membranes were then

prepared for examination using TEM and XRD.

141

AZT Ceramic Iron oxide (Fe203)
membrane coating

I 5 b

 

 

Membranes from supplier Slice of membrane with
Coated by layer by layer method three channels
STEP: 2 AZT Ceramic membrane
Cut a X-section
Iron oxide (Fe203)
coafing
Slice of membrane X-section Sample
(1mm thickness) (3mm length x 1mm width thickness)
STEP:3
Copper grid Direction of viewing for

Microscope

. l

e --

Sampb '
TEM TEM
(Top View) (Front View)

Figure 7.1 Schematic Representation of sample preparation for TEM imaging

142

 

 

 

 

Figure 7.2 TEM image of F0203 nanoparticles processed from sol

7. 3.2 Membrane Characterization

The schematic representation of the procedure used to obtain images of the coated
surface of the tubular F0203 coated AZTC membranes is given in Figure 7.1. The
membrane was first sliced into circular discs of 1 mm thickness using a diarnond-
wafering saw. Subsequently, these 1 mm sections were sliced with a razor blade to form
small arcs, 3 mm in length and 1 mm in width. These arcs were then mounted onto
slotted copper grids, (3 mm inner diameter) perpendicular to the slot. These grids were

then subsequently mounted on stubs for further preparation.

143

The grids were next prethinned with hand polishing using 15, 6, 3 and then 1 pm
diamond paste to obtain a slice with a final thickness of approximately 70-100 pm. The
sample was then dimpled (GATAN Precision Dimple Grinder, Model 656) to thin the
center of the disk while minimizing the damage to the sample surface. The dimpler load
was controlled and approximately 40 pm of sample removed, making the final thickness

of the sample approximately 40 to 60 pm. The final thinning of the F0203 coated ceramic
membranes was done by ion milling, at an accelerating voltage of 4.5 keV with ion beam
inclination of 4°, to avoid preferential thinning, using a commercial ion mill (GATAN
Precision Polishing System Model 691). The thinned specimens were examined using
TEM on a JEOL 100CX at accelerating voltage of 100 kV and photomicrographs
collected using a Megaview 111 digital camera. Electron diffiaction patterns were
collected at a camera length of 100 cm in order to distinguish between amorphous and
crystalline particles.

XRD analysis using Cu Ker radiation was carried out on a Rigaku Rotaflex 200B
diffractometer with an accelerating voltage of 45 kV and a current of 100 mA. Samples
were scanned at angles ranging fi'om 20° to 80°, with a scanning angle speed of 2°/min
and a step size of 002° and the results analyzed using MDI Jade 6.5 XRD software.

Nitrogen adsorption-desorption isotherms were recorded on a NOVA 2000. The
samples of size 1 cm X 1 cm weighing 1.2 a: 0.05 g were dried at 150 °C under vacuum
overnight, prior to N2 sorption measurement. The specific surface area was calculated
using the multipoint Brunauer—Emmett-Teller (BET) method (Brunauer 0t al., 1938). Pore
size distributions were calculated by the Barrett, Joyner and Halenda (BJH) method

(Barrett et al., 1951).

144

7.4 RESULTS AND DISCUSSION
Figure 7.3a is a TEM photomicrograph of a cross section of an AZTC membrane
supplied by the manufacturer. This micrograph reveals a multi-layered structure showing
the underlying alumina-zirconia-titania ceramic membrane with a non-unifonn TiO2
filtration coating of thickness ~ 30 nm - 100 am. An electron diffraction pattern of this
surface shows the amorphous nature of the titania coating (shown in the insert of Figure
7.3a). After coating the AZTC membranes with 40 layers of F0203 nanoparticles, TEM
photomicrographs clearly show a second distinct surface layer having an average
thickness of 50 um i 5 nm (Figure 7.3b). The electron diffraction pattern of this coating
(shown in the insert of Figure 7.3b) demonstrates the crystalline nature of the hexagonal
closed packed (hcp) or-F0203 coating.

While not clear hour the TEM micrographs, evidence of F0203 diffusion into the
porous AZTC membrane was collected in the SEM, using EDS mapping which clearly
showed that iron had diffused into the membrane to a depth of at least 500 um (Karnik 0t

al., 2006).

145

The scale on the bar represents 100 nm.

AZTC Membrane

r‘“

, /=->r"

lozflltratlon coating

 

Figure 7.3 a) TEM cross section of the micro porous AZTC
membrane supplied from the manufacturer. The insert is electron
diffraction pattern of the Ti02 coating on the AZTC membrane

AZTC Membrane

 

Figure 7.3 b) TEM cross section of micro porous AZTC membrane
with 40 coatings of iron oxide sintered at 900 °C for 30 minutes

146

The scale on the bar represents 100 nm.

  

 

 

>5 . .. . - .3... a... :1 A ‘—

Figure 7.4 a) TEM cross section and Electron diffraction patterns
of a 5kD AZTC membrane with 20 coatings of iron oxide,
sintered at 900 °C for 30 minutes

 

 

Figure 7.4 b) TEM cross section and Electron diffraction patterns
of a 5kD AZTC membrane with 30 coatings of iron oxide,
sintered at 900 °C for 30 minutes

147

The scale on the bar represents 100 nm.

 

 

 

Figure 7.4 0) TEM cross section and Electron diffraction patterns of
a 5kD AZTC membrane with 45 coatings of iron oxide, sintered at
900 °C for 30 minutes

.5 a» a k: “_
.mesri .
. II ' T t

. , t 5
' I, ~' W}. ..“fl‘lvh :5.

-‘*'gw§:§%*0
s. a ”in“: *3“ f
“is fi‘fi’i‘k fi‘fifir

 

 

Figure 7.4 d) TEM cross section and Electron diffraction patterns
of a 5kD AZTC membrane with 60 coatings of iron oxide, sintered
at 900 °C for 30 minutes

148

Figures 7.3b, 7.4a, 7.4b, 7.4c and 7.4d illustrate the relationship between the
number of F e203 layers applied and the resulting thickness of the sintered Fe203 coating.
While the TEM photomicrograph for 40 layers of Fe203 (Figure 7.3b) shows a layer
thickness of 50 um i: 5 nm, the 20, 30, 45 and 60 layers of Fe203 coating (Figure 7.4a-d)
yielded a coating layer thickness of 20 um i: 3 nm, 30 :1: 4 nm, 55 nm 3: 5 nm, and 57 um
i: 5 nm respectively. In general, the coating thickness was found to have increased with
increasing number of layers. On average, the thickness of the coating increased by
approximately 10 nm of sintered F e203 for every 10 layers applied. This rate of increase
in the Fe203 layer thickness is not what would be expected if each coating resulted in the
deposition of a F e203 monolayer, given the 4-6 nm particle size of the F e203 that could
result in a Fe203 layer thickness of 28 to 42 nm for every 10 layers applied. However,
because of the aforementioned diffusion of iron into the AZTC membrane itself, we
know that some of the F6203 penetrated into the membrane surface (Karnik et al., 2006).
It is thought that capillary action through the abundant surface connected porosity during
the coating process, as well as diffusion during sintering, results in the uniform
distribution of iron oxide particles into the membrane to a depth of at least 500 um.
Further, as is typical for ceramics, upon sintering, shrinkages of 40-50% are common
(Barsoum, 2003). Therefore, it is reasonable to achieve only a 50 nm thick coating layer,
for 40 applied layers of Fe203, after taking into account the combined effects of diffusion
into the interior of the membrane and sintering.

The increase in the R20; coating layer thickness improved the AZTC membrane
performance in terms of water quality. This may occur because as the Fe203 layer

thickness increased, a greater surface area is available to facilitate surface catalytic

149

reactions that result in the degradation of contaminants in the treated water. The
corresponding water quality data for AZTC membranes with O, 20, 30, 40, 45 and 60
layers of F e203 are shown in Figure 7 .5. More detailed presentations of these water
quality results have been published elsewhere (Karnik et al., 2005a; 2005b; 2005c).
Comparing water quality data for the AZTC membranes receiving 20 and 30 R203 layers
shows no statistically significant differences in terms of water quality, which is linked to
a similar thickness of the Fe203 surface coating. Once the R203 layers were increased to
40, the resulting sintered Fe203 surface layer thickness begins to increase again, and we
find a corresponding improvement in water quality. However, 40 Fe203 layers was found
to be the optimum in terms of water quality improvements, as continuing to deposit
Fe203 at 45 layers, found no further gains in water quality as once again no significant
increases in Fe203 surface layer thickness were found. In fact, water quality from the
AZTC membranes coated with 60 Fe203 layers (Figure 7.5) also showed no gain in water
quality as the increase in the Fe203 surface thickness was small compared to that for the
membrane coated 45 times (Figure 7.4d).

Diffusion of Fe203 into the AZTC membrane also has long term implications for
the successful commercialization of this technology because over time, as the ceramic
membrane gradually erodes, the benefits of the catalytic action of the F e203 would be
expected to continue beyond the water quality improvements that result from just the

initial Fe203 coating layer.

150

 

 

200

 

 

 

 

 

 

 

 

 

 

 

 

6000 ‘1‘ Cl 0 nm F6203
a 5000 q E] 20 nm F8203
3 4000 i I 30 nm Fe203
150 - c
g 3000 y I 50 nm F6203
a 8 2000 I 55 nm F6203
3 5 T E 57 nm F8203
g e 1000 4
g 100 ‘ o t"
g Dissolved Organic Carbon
0
u

 

 

 

,
,\2.;_.
1 .04
in.
,t. .‘
f“... ‘
.. ... .
=12.“ .
.... . ‘
§ o.
., - H 1
.‘.
.:.
.0 _ .
..
t
a.
‘.'.‘
...
-.' r
4
....“
g... _«
e ' ‘
_\.
,, ... .‘
b: . r.
4 .
v‘L ‘
'5 “‘

 

 

 

 

 

Figure 7.5 Water Quality data for the permeate after combined ozonation
membrane treatment process.

The data for water quality parameters for a 5kD AZ T membrane uncoated, unsintered (0 nm coating
thickness), with 20, 30, 40, 45 and 60 coatings of iron-oxide, sintered at 900°C for 30 minutes with
respectively, 20, 30, 50, 55, 57 nm coating thickness. The insert in the graph is the plot of dissolved
organic carbon concentrations for the same membranes. Experiment details (Karnik et al., 2005c)

151

XRD characterization (Figure 7.6) of the samples showed that the uncoated,
unsintered AZTC membrane samples were a mixture of the anatase and rutile phase of
Ti02 as well as Zr02 and corundum, A1203. Following coating, XRD scans showed the
presence of a-Fe203 with no changes observed after sintering in air at 900 °C with a
preferred (111) orientation along the [111] direction aligned perpendicular to the
membrane surface. All XRD peaks were indexed as hcp a-Fe203 with a least squares best

fit lattice parameter of 0.50361 :t 0.00015 nm, 0.503631 1: 0.00020 nm, 1.37524 :t

 

 

 

 

 

 

 

 

 

 

0.00018nm.
<> TIC, 4} F920,

g <> 5* a a g

' .<>_. 3*. A B E A t
M ~ L 3 L~———————-—’:-

d O n g n :1
W A Aw 2%

c o 0 g a c

b ,I no. A x L J. i
g 4 A A : AL _I
20 30 40 50 60 70

20

 

 

 

Figure 7.6 X-ray diffraction patterns of AZTC membrane

a) 5kD M WCO AZTC membrane uncoated, b) 5kD AZTC membrane uncoated, sintered at 900°C
for 30 minutes c) 5kD AZT C membrane with 40 coatings of iron-oxide unsintered and d-g) 5kD
AZTC membrane with 20, 30, 40 or 45 coatings of iron oxide, sintered at 900 °C for 30 minutes.

152

The surface area and pore size measurements for the as received and the F6203
coated AZTC membranes are tabulated in Table 7 . 1. The BET surface area measurements
did not show any statistically significant difference in surface area measured for the
samples. The pore size distributions showing the average pore size of the as-received
AZTC membranes ranged from 4.6 nm 1: 0.02 for the uncoated and unsintered
membrane, to 4.8 nm :t 0.03 for both the 20 and 30 F6203 layer coated AZTC
membranes. Beyond 30 F6203 layers, the average pore size increased to 5.2 um :i: 0.03 for
40 F6203 layers, remaining constant at 45 and 60 F6203 layers. It should be further noted
that the additional heat treatment required to sinter the F6203 layers also serves to coarsen
the pores in the underlying AZTC membrane. These results suggest that increasing the
number of F6203 layers yields a corresponding increase in coarsening of the AZTC
membrane which also serves to facilitate the catalytic performance that was optimized at
40 F6203 layers, as well as to transform the surface from a relatively flat surface, to one
having surface undulations on the micron scale. This confirms our earlier AFM results
showing a decrease in submicron scale surface roughness with sintering and F6203
coating (Karnik et al., 2005c; Karnik 6t al., 2006). Further increasing the number of coats
F6203 does not improve the water quality which corresponds to no statistical increase in
the coating thickness at 45 or 60 layers of F6203 (Figure 7.46 and 7.4d), nor did we find
any statistical significant decrease in AFM roughness beyond 40 layers, as shown in our
previous AFM analysis (Karnik 6t al., 2006). Further these TEM photomicrographs
confirm, as do our other findings that the sintered F6203 coated AZTC membranes used
in our hybrid nanofiltration-ozonation study are indeed still operating in the nanofiltration

domain.

153

Table 7.1
Summary of surface area and average pore size

 

 

 

 

 

 

 

 

 

 

 

 

Sample Surface area (m2/ g) Averag(:rplc))re $126
5 kD-uncoated-unsintered 4.985 4.6 (i 0.02)
5 kD - 20 coats-900 °C 5.117 4.8 (i 0.03)
5 kD - 3O coats-900 °C 5.171 4.8 (:h 0.02)
5 kD - 4O coats-900 °C 4.452 5.2 (2!: 0.03)
5 kD - 45 coats-900 °C 4.523 5.2 (:t 0.01)
5 kD - 6O coats-900 °C 4.589 5.2 (i 0.02)
7.5 CONCLUSIONS

TEM observations revealed that coating the AZTC membrane with iron oxide
nanoparticles followed by sintering at 900 °C in air led to a transformation of the outer
most amorphous Ti02 filtration layer on the as received AZTC membrane to a crystalline
Ti02 structure coated by a crystalline a-F6203 surface layer. Increasing the number of
F6203 layers did not produce a one to one corresponding increase in the thickness of the
F6203 surface. The fact that the thickness of the coating is less than might be expected is
a result of both diffirsion of the F6203 nanoparticles into the membrane by capillary
action and subsequent densification of the porous F6203 layer during sintering.

In general the porosity of the AZTC membranes having a nanoscale F6203
coating, led to changes in the surface morphology and average pore size of the F6203
surface layer as a result of sintering and coarsening of the F6203 nanoparticles, while
maintaining nanoscale pore filtration capabilities. Capillary action during the coating
process, and/or diffusion during sintering, resulted in a uniform distribution of iron oxide

particles into the membrane to a depth of at least 500 um (Karnik 6t al., 2006). The

154

synergy of the catalytic effect of the F6203 nanoparticle coating on the membrane surface
and the diffused F6203 layer into the membrane enhances water filtration because of the
increased exposure to the catalytic iron oxide, not only at the membrane surface, but into
the membrane itself. The AZTC F6203 coated membrane exceeds the current EPA
regulatory requirements (U S-EPA), while still in the nanofiltration regime. Future
research will examine the mechanisms for the degradation of NOM and the removal of
harmful DBPs by the iron oxide coated AZTC ceramic membranes.

7.6 LIMITATIONS-PROBLEMS

The diameter of the smallest diffraction aperture size on the JEOL 100CX is 100nm. The
edge surface regions being imaged for electron diffiaction patterns are obtained by
carefully positioning the aperture at the extreme edge. However there are limitations to
the technique. A diffraction patterns could have emerged from the regions other the edge
surface of the sample. So the electron diffraction patterns obtained cannot be positively
said to come fiom only the outer surface. However the TEM images illustrating the
morphology and coating thiclmess of the samples are valid. However it should be noted
that while the surface region cannot definitely be said to be only F6203, an F6203 surface
layer is not ruled out. Research is ongoing to obtain nano beam diffraction patterns using
I EM-22OOF S which is a high resolution TEM with a resolution limit of 1 nm. This study
will help us confirm the electron diffraction patterns obtained for the ceramic membranes
and the catalyst coating on the lower resolution JEOL 100CX.

7.7 ACKNOWLEDGEMENTS

The authors would like to acknowledge the US Environmental Protection Agency (U 8

EPA) Science To Achieve Results (STAR) Program (Grant No. RD830090801) and

155

National Science Foundation Nanoscale Interdisciplinary Research Teams (N SF—NIRT)
Program (Grant No. BESOSO6828) for financial support of this work. Our thanks also go
to Ms. Alicia Pastor, from the Center for Advanced Microscopy and Mr. Robert Pcionek
for there assistance during TEM sample analysis. Mr. Rui Huang is also acknowledged
for XRD analysis of the samples. Thanks are also due to Yang Chen for performing the
BET measurements. We would also like to thank Mr. David Jackson for his assistance

with the membrane coating process.

156

7.8 REFERENCES

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ 03-H202 and 03/Ti02) for the oxidation of natural organic matter
in water” Ozone: Science and Engineering, 15(5), 419—432.

Barrett, E., Joyner, L., and Halenda, P. (1951). "The Determination of Pore Volume and
Area Distributions in Porous Substances. I. Computations from Nitrogen Isotherrns"
Journal of American Chemical Society, 73(1), 373-380.

Barsoum, M. (2003). In "Fundamentals of ceramics” (Institute of Physics Publishing,
2003), 13.336.

Brunauer, S., emmett, P., and Teller, E. (1938). "Adsorption of Gases in Multimolecular
Layers" Journal of American Chemical Society, 60(2), 309-319.

Castro, K. and Zander, A.K. (1995). "Membrane air-stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

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

Karnik, B.S., Davies, S.H., Chen, K.C., Jaglowski, D.R., Baumann, M.J. and Masten, SJ.
(2005). "Effects of ozonation and pH on the permeate flux of nanocrystalline ceramic
membranes" Water Research, 39(4), 728-734.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005b). "The effects of
combined ozonation and filtration on disinfection by-product formation" Water
Research, 39(13), 2839-2850.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, 8.]. (20056). "Fabrication of
catalytic membranes for the treatment of drinking water using combined ozonation
and ultrafiltration" Environmental Science and Technology, 39(19), 7656-7661.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, SJ. (2006). "AFM and SEM
Characterization of Iron Oxide Coated Ceramic Membranes" Accepted for
publication in Journal of Materials Science.

Keane, M. (2003). "Ceramics for catalysis" Journal of Materials Science, 38(23), 4661-
4675.

Kim, J. 0., Sorrriya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in

organic acid fermenter by intermittent ozonation", In the Proceedings of the 14’"
Ozone World Congress, Dearborn, MI, 131-143.

157

Kim, J .0. and Somiya, I. (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

List of drinking water contaminants and MCLS, EPA website
http://www.epa.gov/safewater/mcl.html#mcls

McKenzie, K.J., Marken, F., Hyde, M., and Compton, R.G. (2002). "Nanoporous
ironoxide membranes: Layer-by-layer deposition and electrochemical

characterization of processes within nanopores" New Journal of Chemistry, 26, 625-
629.

Neri, G., Rizzo, G., Galvagno, S., Loiacono, G., Donato, A., Musolino, M., Pietropaolo,
R. and Rombi, E. (2004). "Sol—gel synthesis, characterization and catalytic properties
of Fe—Ti mixed oxides" Applied Catalysis: A General, 274 (2004), 243-251.

PuhlfitrB, P., Voigt, A., Weber, R. and Morbé, M. (2000). "Microporous Ti02 membranes
with a cut off <500 Da " Journal of Membrane Science, 174 (1-2), 123-133.

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

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

Shen Z.S., Semmens M.J., and Collins A.G. (1990). "A novel-approach to ozone water
mass-transfer using hollow fiber reactors" Environmental Technology, 1990, 11, 597-
608.

Trudeau, M., and Ying, J. (1996). "Nanocrystalline materials in catalysis and
electrocatalysis: Structure tailoring and surface reactivity" Nanostructuted Materials,
7 (1-2) 245-258.

US EPA (2000). "Low-pressure membrane filtration for pathogen removal: Application,
implementation and regulatory issues " Office of Water 815-C-01-001.

158

CHAPTER EIGHT

Karnik, Bhavana 8., Davies, Simon H., Chen, Kuang C., Jaglowski, Dan R., Baumann,
Melissa J. and Masten, Susan J ., (2005). Removal and survival of Escherichia coli after
treatment using ozonation-ulnafiltration with iron oxide coated membranes. Manuscript
submitted to Ozone: Science and Engineering.

159

CHAPTER EIGHT

REMOVAL AND SURVIVAL OF ESCHERICHIA COLI AFTER
TREATMENT USING OZONATION-ULTRAFILTRATION WITH
IRON OXIDE COATED MEMBRANES

8.1 ABSTRACT

The effect of membrane filtration, ozonation, and combined ozonation-membrane
filtration on the removal of bacteria was studied. Commercially available ceramic
membranes with a molecular weight cutoff (MWCO) of 5kDa were used as is, and also
coated with iron oxide nanoparticles and sintered at 900°C. With membrane filtration and
ozonation-membrane filtration using the uncoated membrane, 7 log removal of bacteria
was achieved, as compared to 7.5 log removal with ozonation-membrane filtration with
the coated membrane. Only 4 log removal was achieved with ozonation alone. A Live-
Dead assay indicated that the mortality of bacteria in the product water was ~50%, ~85%,
and >99% with ozonation, combined ozonation-membrane filtration with the uncoated
membrane, and ozonation-membrane filtration with the coated membrane, respectively.
With the coated membrane, the concentration of assimilated organic carbon (AOC) was
reduced by up to 50% more than with the uncoated membrane filtration (both, with
systems operated with ozonation). It appears that the catalyst-coated membranes enhance
surface catalytic reactions that degrade microbial substrates, such as aldehydes and
ketoacids, thereby reducing the potential for microbial regrowth in the water distribution

system. Scanning electron micrographs (SEM) of the membrane surface show that the

160

surface morphology is changed as a result of the coafing the membrane. The SEM results
also show that the adhesion of particles found in the water to the membrane surface is
affected by the coating on the membrane.
Keywords: ceramic membranes, ozonation, nanofiltration, iron oxide, bacteria, catalytic
ozonation, nanofilrn, membrane filtration, scanning electron microscopy (SEM),
assimilated organic carbon (AOC)
8.2 INTRODUCTION

The implementation of sand filtration and disinfection for water treatment in the
early twentieth century significantly reduced the number of outbreaks of waterborne
diseases transmitted through drinking water. The effect was so great that by the early
1980’s the drinking water community assumed that problems related to microbial
contamination were essentially solved. This perspective changed when the Milwaukee
Cryptosporidium outbreak in 1993 resulted in an estimated 400,000 cases of
cryptosporidiosis and at least 50 deaths. More recently, in Walkerton, Ontario, at least
2,300 cases of gastroenteritis and seven deaths were attributed to drinking water
contaminated with E. coli 0157 (Finstein, 2004, Hrudey 6t al., 2003). The severe
consequences of accidental, as in these cases, or (potentially) deliberate, contamination of
drinking waters by pathogenic organisms has made it clear that better water treatment
technologies must be developed and implemented to reliably eliminate pathogens fi'om
drinking water supplies.

Ozone, a strong oxidant and a powerful disinfectant, is very effective for the
inactivation of bacteria, protozoa and viruses (6.g., see, Lee et al., 1999, Owens 6t al.,

1999, Liltved et al., 1995, Hehner et al., 1993). Membrane technology is considered to be

161

an effective alternative to conventional water treatment for the removal of particles,
microorganisms and organic matter (Cleveland, 1999, Nakatsuka et al., 1996). Several
researchers have reported greater than 4 log reduction for bacteria, viruses, and protozoa
(i.e., Cryptosporidium and Giardia) using membrane filtration (Hsu 6t al., 2003, Hagen,
1998, Otaki et al., 1998, 06 et al., 1996, Dumoutier 6t al., 1996, Madaeni 6t al., 1995).
However, the decrease in permeate flux resulting from membrane fouling remains a
major problem (e.g., see Lee 6t al., 2001; Cho 6t al., 2000a, 2000b; Field, 1996).

Ceramic membranes are ozone resistant and when combined with ozonation,
generate very high and stable permeate fluxes without causing membrane damage
(Karnik 6t al., 2005a, Schlichter et al., 2004, Kim et al., 2001, Kim et al., 1999). These
membranes have proved effective for the control of disinfection by-products (DBPs)
(Karnik 6t al., 2005a, Lee and Cho 2004, Benfer et al. 2004). A decrease in the
concentration of Simulated Distribution System total trihalomethanes (SDS TTHMs) and
Simulated Distribution System halo acetic acids (SDS HAAs) of up to 80% and 65% (as
compared to generated by ozonation alone), respectively, was reported with the
application of the ozonation-membrane filtration process (Karnik 6t al., 2005b). The
ceramic membranes when used for drinking water treatment also showed complete
removal of microorganisms, when measured for fecal coliforrns, total coliforrns and
E. coli, thus having tremendous potential for the application in drinking water treatment
process to improve the water quality (Bottino et al., 2001).

Ozone has been used in the presence of different metal oxide catalysts, including
manganese oxide, titania, alumina, and zirconia, to degrade refractory compounds,

including saturated carboxylic acids, phenols, aromatic hydrocarbons, dyes, humic

162

substances and herbicides (Beltran 6t al., 2003a, 2003b, Radhakrishnan and Oyama,
2001, Ni and Chen, 2001, Gracia 6t al., 1996, 2000a, 2000b, Legube and Karpel Vel
Leitner, 1999). Catalytic ozonation is a promising technology as the OH and oxygen
radicals generated at the surface can potentially react with both living and non-living
organic substrates, including bacteria and viruses, which would be killed or inactivated
on contact by the oxidants (Legube and Karpel Vel Leitner, 1999, Heinig, 1993). In our
earlier work with ozonation-membrane filtration, we showed that the catalyst (iron oxide)
coated membrane enhanced the degradation of the sorbed or trapped ozone byproducts,
such as aldehydes and ketoacids, which facilitate the regrowth of bacteria in water
distribution systems. When the coated membrane was used, the concentrations of
aldehydes, ketones, and ketoacids in the membrane permeate were reduced by >50% as
compared to those obtained with the uncoated membranes (Karnik 6t al., 20056).

Metal oxides, such as iron oxide, have been shown to retard the transport of bacteria
(Penn et al., 2002, Silliman et al., 2001, Johnson et al., 1996, Scholl et al., 1990, DeFlaun
6t al., 1997, Knapp et al., 1997). Bacteria adhere to metal oxides, reducing the levels of
the bacteria in the aqueous phase (Silliman 6t al., 2001). As such, it is expected that
adhesion to the iron oxide surface will increase the retention of bacteria at the membrane
surface, prolonging the contact of the bacteria with ozone, and, as a result, will improve
disinfection.

In this work we hypothesize that catalyzed ozonation and membrane filtration will
act synergistically, resulting in improved inactivation of and/or removal of bacteria from
the filtrate. We have studied the removal and survival of bacteria in these treatment

systems. The assimilated organic carbon concentrations (AOCs), a measure of bacterial

163

regrowth potential of heterotrophic bacteria, were also monitored. Scanning electron
microscopy was used to study the membrane surface before and after treatment.
8.3 MATERIALS AND METHODS

Experiments were carried out using samples taken from Lake Lansing (Haslett,
MI), which is a borderline eutrophic lake (Lake Lansing Watershed Advisory Committee
Progress Report, 1998). The samples were collected at the boat ramp at the Lake Lansing
Park-South, Haslett, MI in five-gallon polyethylene carboys and stored at 4°C for a
maximum storage period of seven days. The typical characteristics of Lake Lansing water
are given in Table 8.1.

Table 8.1
Typical Characteristics of Lake Lansing Water (Haslett, MI) '

Parameters Lake
TOC 8.6 to 11.6

7.7 to 8.6

145 to 157
UV-254 abs. 0.160 to 0.180

SDS 240

SDS HAA 75

BDOC 1.0 to 4.1

Nitrate 0.44

Total 0.06

Hardness as 190 to 198
“All data reported is obtained from the Lake Lansing Watershed Advisory Committee Report (1998) except
for SDS THM and SDS HAA, which were measured as part of this study
”SDS THM and SDS HAA were measured using Standard Method 5 710 and USEPA Method 552. 2,

respectively.

 

A protocol was developed to challenge the treatment system using E. coli strains (E.
Coli - DHI obtained from Department of Cell and Molecular Biology, Michigan State
University). Cultures of E. Coli (100 mL) were grown to late log phase in deionized
water that contained reagent grade NaCl (10 g/L), Bactotryptone (10 g/L), and Difco
yeast extract (5 g/L). The pH of the media was adjusted to 7.2 using 1 M NaOH before

being spiked with E. Coli.

164

Once the bacteria reached late log phase (1X108 bacteria/mL for E. coli corresponding
to an OD570 of ~0.03), 30 mL of this culture were concentrated by centrifugation at
10,000X g for 10-15 minutes. The supernatant was discarded and the pellet was
suspended in 10 mL of the deionized water that had previously been filtered through a 0.2
pm pore size filter to remove bacteria and particulate matter.

The untreated feed raw water sample was spiked with 100 mL of this bacterial suspension
per 3.5 liters of the feed raw water sample.

8. 3.2 Experimental Methods

The experimental setup used for the combined ozone membrane filtration is shown in
Figure 8.1. A tubular ceramic membrane in a stainless steel filter holder was used in these
experiments. Teflon® tubing and stainless steel or Teflon® joints and valves were used
throughout the system. Other components included: 3.5-liter and 1.5-liter water-jacked
glass reservoirs made of Pyrex® glass, and a simple Y inline mixer (Ozone Service,
Burton, BC, Canada). Ozone gas was added into the water stream through the simple Y
inline mixer just before the feed water entered the membrane module. The water level in
the 3.5-liter reservoir was maintained at a constant level during the experiments using a
peristaltic pump (Masterflex Model 7520-35, Cole-Parmer Co., Chicago, Illinois) to
transfer water from the 1.5-liter reservoir into the 3.5-liter reservoir. A constant water
temperature of 20°C was maintained using a recirculating water bath. The operating
conditions used are given in Table 8.2. The gaseous ozone concentration was 2.5 g/m3.
The experiments were performed with membrane cross flow velocity of 0.6 m/s; the flow

was turbulent with Reynolds number of approximately 6000.

165

Table 8.2
Operating Conditions for the Ozone —Membrane Filtration System

 

 

 

 

Water recirculation rate 2.75 LPM
Water temperature 20°C

Ozone gas flow rate 100 mL/min
TMP 0.5 bar

 

 

 

 

Permeate samples were collected in bottles covered with Parafilm® and stored in an
ice-bath for the duration of the experiment. The experiments were performed included
ozonation alone, membrane filtration alone, combined ozonation-nanofiltration using the
uncoated/unsintered membranes.

The membrane used in these experiments was a Clover-leaf design (containing three
channels) CéRAM membrane (TAMI North America, St. Laurent, Quebec, Canada) with
a molecular weight cut-off of 5 kDa. The external diameter of the membrane was 10 mm
and the active membrane length was 25 cm. The membrane had a total filtering area of
41.2 cmz. The membrane was used uncoated (as supplied by the manufacturer) and it

was also used coated with iron oxide.

166

Waste gas

Moisture

  

 

  

 

 

 

 

 

 

 

 

 

 

 

 

Adsorbent Spectrophotometer
Ozone
generator F—
Oxygen % |
Cyllnder w Kl I

 

 

D<l

  

7s— @4
fl

Membrane

 

  

 

 

—®

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Water 1 \
Tank ® Flowmeter
C 6’) Pressure gauge

 

 

 

B ‘I

O X Valve

Figure 8.1 Schematic representation of the ozone-membrane filtration system
(Karnik et al., 2005a)

 

 

 

 

167

The layer-by-layer technique used to coat the membranes is based on a protocol
described by McKenzie 6t a1. (McKenzie et al., 2002) for coating doped tin oxide
electrodes. The membrane was coated with 40 layers of iron oxide nanoparticles and
sintered at 900°C. A detailed description of the method used to prepare the iron oxide
coated membrane is available in our earlier work (Karnik et al., 20056). The colloidal
particles used for coating the membranes were prepared by Sorum’s method.
Transmission electron microscopy (TEM) characterization showed that the average
particle diameter was 4 to 6 nm. The choice of the conditions for coating and sintering
were based on results obtained in previous work on the removal of disinfection by-
products and on evidence obtained by microscopy regarding the morphology of the
membrane surface (Karnik 6t al., 2005b, 6).

To generate ozone, pure oxygen gas (99.999%) from a pressurized cylinder was dried
using a molecular sieve trap, and then fed to an ozone generator (Model OZZPCS,
Ozotech Inc., Yreka, Calif). The gaseous ozone concentration was controlled by varying
the voltage applied to the ozone generator. The excess gas was vented after passing the
gas through a 2% potassium iodide (KI) solution to destroy any residual ozone gas.

The experiments were performed for ozonation alone, membrane filtration alone,
ozonation membrane filtration and ozonation using catalyst coated membrane filtration.
All the experiments were canied out for eight hours. A gaseous ozone concentration of
2.5 g/m3 was used for all ozonation experiments, so the ozone dosage was the same for all
experiments.

Samples of the feed water, permeate and reject streams (for the filtration experiments)

and ozonated water (for the ozonation experiment) were analyzed using fluorescence

168

microscopy to determine the ratio of livezdead bacteria. Samples were also taken to
determine the concentration of assimilable organic carbon (AOC) in the water.

8. 3.3 Analytical Methods

8. 3. 3. 1 Bacterial analysis by fluorescence microscopy

A commercially available molecular probe kit (L-7012 LIV E/DEAD® BacLight,
Bacterial Viability Kit, Molecular Probes, Eugene, OR.) was used for the characterization
of bacteria as live or dead by fluorescence spectroscopy. The kit is used for monitoring
the viability of bacterial populations as a function of the membrane integrity of the 6611.
Under the experimental conditions and the range of excitation wavelengths used, cells
with a compromised membrane, those are considered to be dead or dying, will stain
fluorescent red, whereas cells with an intact membrane will stain fluorescent green.

An aliquot (10 mL) of the water sample from the challenge experiments was prepared
for fluorescence microscopy by concentrating the bacteria using centrifirgation at
10,000X g for 10—15 minutes. The supernatant was then removed and the pellet was
resuspended in 2 mL deionized water. One (1) mL of this suspension was added to each
of the 1.5 mL microcentrifuge tubes and centrifuged for 10 minutes. The supernatant was
decanted and resuspended in either 1 mL of filtered water for LIVE or 1 ml 70% ethyl
alcohol for DEAD cell counts. The samples were incubated at room temperature for 1
hour, with mixing accomplished every 15 minutes. The samples were then centrifuged at
10,000X g for 10-15 minutes. Each pellet was resuspended in 1 mL of filter sterilized
water.

Bacteria were stained with the dye available in the L-7012 kit in proportions given by

the manufacturer. One volume of SYTO 9 dye was combined with one volume of

169

propidium iodide stain in a 1.5 mL microcentrifuge tube and mixed thoroughly. For each
1 mL of the bacterial suspension, 3 uL of the dye mixture were added. When used at the
recommended dilutions, the reagent mixture contributed 0.3% dimethyl sulfoxide
(DMSO) to the staining solution. Higher DMSO concentrations were not used as these
may adversely affect staining. The suspension was mixed thoroughly and incubated at
room temperature in the dark for 15 minutes. The stained bacterial suspension was
pipetted in volumes of 5 [IL onto an ethanol cleaned slide and covered with an 18 mm
square coverslip. Live and dead bacteria were then observed simultaneously using a
fluorescence microscope (Leica DM4000 digital microscope) equipped with a standard
fluorescein longpass filter set (Filter 13). For each sample, three slides were prepared and
data from 10 fields were recorded. Digital images of the bacteria were also recorded.

To quantify the bacteria counts, fluorescence was measured using an excitation
wavelength of 470 nm, and emission wavelengths within range of 490-700 nm and
compared to a calibration curve prepared fiom lmown densities of a stained cell
suspension. The calibration curve was prepared by measuring the ratio of the integrated
intensity of the portion of each spectrum between 510—540 nm (emission for green) to the
intensity of the portion between 620—650 nm (emission for red) for each bacterial
suspension. This ratio of integrated green fluorescence to integrated red fluorescence (R
(Gm) versus percentage of live cells (% Bacteria counts) in the E. coli suspension was
plotted to give the calibration curve. The calibration equation obtained was
R (G/R) = 0.043 X percentage of live bacteria counts + 0.6466
where R(G/R) is the ratio of green fluorescence to red fluorescence and bacteria counts are

measured for live bacteria detected in fluorescence spectrophotometer (SpectraMax M2,

170

Molecular Devices). The results for the feed raw water sample showed that only 90% of
the bacteria in this water were alive. The results reported here do/do not account for the
number of bacteria that were dead in the raw water.

8. 3. 3.2 Regrowth potential of heterotrophic bacteria

The biodegradability of organics present in water was assessed using AOC (Van der
kooij et al., 1982). The concentration of AOC was measured using Method 9217
(Clesceri 6t al., 1998). Water samples were collected in organic-carbon-free glass bottles.
After the samples were rrrixed well, aliquots were decanted into test tubes. The test tubes
were capped immediately and pasteurized in a 70°C water bath for 30 minutes to
inactivate bacteria originally present in the sample. After pasteurization, samples were
cooled to room temperature and inoculated with 1 mL of Pseudomonas fluorescens strain
P17 and Spirillum strain NOX (each, at 500 colony-forming units (CFU)/mL). The
samples were then incubated at room temperature (20 °C) for seven days. Following this,
the Heterotrophic Plate Count (HPC) was determined by distributing aliquots of the
sample on predried R2A agar plates (Clesceri et al., 1998). The colonies were counted on
the plates after three days of incubation at room temperature and the concentration of
AOC was determined.

8. 3. 3.3 Microscopic observation of the membrane surface

To obtain the images of the membrane surface, the tubular ceramic membrane was first
sliced into circular disc sections of 1 mm thickness using a diamond wafering saw.
Subsequently, these sections were cut to form small arcs of length 3 mm and width of 1
mm. These arcs were then mounted, using carbon adhesive tape, on aluminum mounting

stubs for SEM examination.

171

To image the membrane surfaces after treatment, the bacteria and biomass present
were immobilized by immersing the membrane surface in 4% glutaraldehye, a standard
fixative, and 0.1M phosphate buffer for 45 minutes. These samples were next rinsed with
0.1M sodium phosphate buffer (pH 4) for 30 minutes and subsequently dehydrated in
25% ethanol (in water) by soaking for 15 minutes. Next, the samples were immersed in
50% ethanol for 15 minutes, sliced in rings of 1 cm thickness and placed in critical point
dryer baskets, standard practice when fixing biological samples for SEM analysis. The
dehydration was continued using 75, 95 and 100% ethanol, where samples were then
allowed to stand for 15 minutes after each step. The samples were finally dried using a
critical point dryer (Blazers Inc.). The dried samples were cut in arcs having a length of 3
mm and a width of 1 mm. These arcs were mounted, using carbon adhesive tape, on
aluminum mounting stubs for SEM examination.

SEM images of the mounted samples were obtained using a JEOL 6400V scanning
electron microscope with a LaB5 emitter at an accelerating potential of 15 kV and
magnifications ranging from 5,000X to 60,000X. The mounted samples were coated with
gold using Emscope SC 500 sputter coater at a rate of 7 nm/min with a 20 mA current.
The resulting digital micrographs were recorded using the AnalySIS® imaging system.

8.4 RESULTS AND DISCUSSION
A statistical analysis of the data presented was carried out using one way ANOVA
at the 95% confidence level (p<0.05).
8. 4. 1 Bacterial analysis by fluorescence microscopy
Comparison of the images obtained from fluorescence microscopy (see Figure 8.2)

reveal that ozonation resulted in a decrease in the number of live bacteria (those that

172

\

\

- Ozonation + Membrane filtration

-) Catalytic coated Membrane filtration
+ ozonation

 

Figure 8.2 Fluorescence images indicating bacteria presence in the permeate
after different treatments.

Experimental setup: see Fig. 8.] Operating Conditions: see Table 8.2. Membrane Size: 5 kDa,
Gaseous ozone concentration 2.5 g/m’. catalyst coated membrane: coated with 40 times with iron
oxide nanoparticles. and then sintered in air at 900"C.

173

stained green) as compared to that observed in the feed raw water from Lake Lansing
prior to ozonation treatment. In the combined ozonation-membrane filtration experiment,
the number of live bacteria in the permeate was firrther decreased, suggesting a synergy
between ozonation and membrane filtration. Further improvements were obtained when
using the iron oxide catalyst-coated membrane (see Figure 8.26).

The mortality of the bacteria in the product water (permeate or ozonated water) after
treatment using ozonation, combined ozonation-membrane filtration, and ozonation-
membrane filtration process with an iron oxide coated membrane was 50%, 85%, and
>99%, respectively (as shown in Table 8.3, Figure 8.3).

Table 8.3

Live-Dead bioassays using fluorescence spectroscopy
The precision of the analytical procedure is within 5 %.

 

 

 

 

 

 

 

 

Live-Dead assays using Fluorescence Spectroscopy Method
Permeate Stream Removal
Treatment “00°58 R(G/R) LIVE. R(G/R) LIVE.
(% Bacteria counts) (% Bacteria counts)

Membrane filtration 4'29 85 4.60 92
02°"a"°" 2.78 49.5 3.87 75
Ozonation-Membrane

filtration 1.23 13.5 2.15 35
Catalytic Ozonation-

Membrane filtration 0'69 1 1'16 12

 

 

 

 

 

 

 

174

 

 

[Ix mm

 

Clam:
Oauuflomllouhnm
mitten

 

 

Oauntbmllorlhnm
fllntlon

 

 

Ozonation

 

 

 

mum. Mutton

 

 

 

 

I I I I I I I I I I

0102030405060708090100
itBactorlaeounts

 

 

 

Figure 8.3 Percent of live — dead bacteria in the permeate after different

treatments.

Experimental setup: see Figure 8.1 Operating Conditions: see Table 8. 2. Membrane Size: 5 kDa,
Gaseous ozone concentration 2.5 g/m’, catalyst coated membrane: coated with 40 times with
iron oxide nanoparticles, and then sintered in air at 900°C. The results were determined using
fluorescence spectroscopy after staining with molecular probes. All values are the average of
triplicates within experiments.

The mortality of the bacteria in the reject stream was 25%, 65%, and 88% for
ozonation, combined ozonation-membrane filtration, and catalytic ozonation-membrane
filtration, respectively. It is evident that mortality of the bacteria was significantly higher

in combined ozonation-filtration experiments than that obtained in the ozonation

175

experiments and that the mortality of the bacteria is highest when the coated membrane
was used.

As the same ozone dosage was used in all experiments, it is apparent that disinfection
is more effective using ozonation-membrane filtration. The improved disinfection
observed using the combined process may be explained by the catalytic decomposition of
ozone at the iron oxide surface, which results in the formation of OH or other radical
species that inactivate the bacteria near the surface. Bacteria have been shown to adhere
strongly to iron oxide surfaces (Knapp 6t al., 1997). The adhesion of bacteria to the iron
oxide surface may also improve performance. The lower mortalities found in the reject
stream are also consistent with this explanation, since the bacteria in permeate are more
likely to have been at the membrane surface (and passed through imperfections in the
membrane) than the bacteria in the reject stream (which are in the recirculating water
stream).

The bacteria counts measured as HPC in the permeate stream are tabulated in Table
8.4. The higher bacteria counts were found in the permeate for the membrane filtration
alone (2 CFU/mL) when compared to that observed in the permeates from the hybrid
process using ozonation and membrane filtration with the uncoated membrane (1.67
CFU/mL) and also the hybrid process using ozonation and iron oxide coated membranes
(0.67 CFU/mL). These results show that removal of bacteria by filtration is at least 3 logs
better than that obtained with ozonation alone. For the ozonation-membrane filtration
experiment using the uncoated membrane the removal of bacteria is only slightly higher
than that observed in the filtration experiment. However, for the ozonation-membrane

filtration experiment using the coated membrane the removal of bacteria is significantly

176

higher. The higher removal found using the coated membrane is consistent with higher

mortality found using the Live-Dead assay.

 

 

 

 

 

 

 

Table 8.4
Reduction in E. coli Counts Using Heterotrophic plate count method
Heterotrophic Plate Count Method
Treatment Process 05,3353?" 66322332611 Removal Log Removal
water (CD). (Rmc’
(CFU/mL) (CFU/mL) R=1-c.=./cR RLw-logU-R)
Membrane "mam" 1.86E+07 2.00 0.99999989 6.97
02°"a"°" 1.83E+07 1670.00 0.99990891 4.04
3:33;"; Mafia" 1.83E+07 1.67 0.99999991 7.04
aifi'grnfé‘t’rggggn 1905+07 0.67 0.99999996 7.45

 

 

 

 

 

 

 

'All values are with 5% std. deviation

8. 4.2 Regrowth Potential of Heterotrophic Bacteria

As shown in Figure 8.4, the measured AOC concentrations decreased significantly in
the permeate using membrane filtration-ozonation compared to that obtained in the
product water when using ozone alone. AOC concentrations of 0.089 (:l: 0.0005) mg/L,
5.05 (i 0.04), and 3.75 (:l: 0.03) mg/L were observed in the permeate samples from
membrane filtration alone, ozonation/ membrane filtration using uncoated membranes,

and ozonation/filtration using coated membranes.

177

 

I Permeate Stream
0 Reject Steam

 

AOC concentration (mg/L)

 

 

 

 

 

 

Membrane filtration Ozonation Ozonation+Membrane Catalytic
filtration Ozonation+Membrane
filtration

 

Figure 8.4 Assimilated Organic Carbon (AOC) concentration after different
treatments for the permeate and reject streams.

Experimental setup: Fig. 8.1 Operating Conditions: Table 8.2. Membrane Size: 5 kDa, Gaseous
ozone concentration 2.5 g/m’, catalyst coated membrane: coated with 40 times with iron oxide
nanoparticles, and then sintered in air at 900”C. 'All values are the average of triplicates within
experiments.

This compares to an AOC concentration in the ozonated water of 9.44 (:t 0.05) mg/L. Th6
AOC concentrations measured for the reject stream were 6.19 (:t 0.008), 8.14 (:1: 0.004),
and 3.67 (t 0.03) mg/L, respectively, for membrane filtration alone, combined ozonation-
(uncoated) membrane filtration, and catalytic ozonation-membrane filtration.

For the iron oxide coated membrane experiments, the AOC concentrations in the
permeate were <50% of that that obtained in the experiments with the uncoated
membrane. This confirms our earlier work where it was found that the concentrations of
ozonation byproducts in the permeate were lower when ozonation-membrane filtration

was used as compared to the concentrations observed in the heated water with ozonation

178

alone (Karnik et al., 2005b). Work in our laboratory has also shown that a significant
reduction in the concentration of ozonation byproducts in the permeate occurs when the
membrane is coated with iron oxide (when compared to uncoated membranes) (Karnik 6t
al., 20056). The membrane surfaces appear to serve as a catalyst for the degradation of
ozone, thereby reducing the formation of the ozonation byproducts. The membrane
surface may also catalyze the degradation of some ozonation by-products (Karnik 6t al.,
2005b, 20056). The data indicates that there is a reduced potential for regrowth in water
treatment using the coated membranes. As is discussed above, the survival of bacteria is
also lower when these membranes are used. The lower survivals and AOC concentrations
observed using combined catalytic ozonation-membrane filtration suggest it is likely to
be a very effective process to both disinfect the water and control bacterial regrowth in
the distribution system.
8. 4.3 Microscopic observation of the membrane surface

Figure 8.5 shows SEM micrographs for an uncoated membrane with 5 kDa MWCO
and a 5 kDa membrane coated 40 times with iron oxide particles and sintered at 900°C.
Images are shown for the membrane both before and after the membranes were used in
the ozonation-membrane filtration treatment process. The SEM rrricrographs clearly show
that the morphology of the surface is altered as a result of the coating procedure. The
surface morphology changed from a fine-grained microstructure for the uncoated
membrane (Figure 8.5a) to a coarse-grained morphology with uniformly interconnected
pores on the iron oxide coated and sintered membrane (Figure 8.5d). As can be seen in
Figures 8.5b and 8.56, there are mineral deposits and possibly organic detritus on the

uncoated-unsintered membrane surface that had been used to filter water.

179

. ~--.

.‘ , _ r 1', biomass '

     
    

post treatment-
poet treatment croes sectlonal View

poet mamm-
crou sectlonal View

Figure 8.5 SEM images of membrane surface before and after treatment
SEM JEOL 6400V, accelerating voltage 15 kV, a—c) SEM images of an uncoated-unsintered
membrane from the manufacturer, d-fl SEM images of a membrane coated with 40 times with
iron oxide nanoparticles, then sintered in air at 900°C. Pretreatment and post-treatment refers to
samples before and after ozonation-membranefiltration hybrid process treatment

180

Mineral grains in the micron size range are found on this surface. However, there is
no evidence of such mineral deposits on the iron oxide coated-sintered membrane surface
(see Figures 8.56 and 8.50. The lack of mineral deposits may be a result of the alteration
in the surface charge or morphology of the membrane surface, which does not allow
mineral deposition onto the membrane surface. It appears that there is more organic
detritus on the coated membranes than on the uncoated membrane (compare Figure 8.5b
with Figure 8.56). Further research will be conducted to study the deposition of this
material on the membrane surface.

8.5 ACKNOWLEDGEMENTS

The authors would like to acknowledge the US Environmental Protection Agency (US
EPA) Science To Achieve Results (STAR) Program (Grant No. RD830090801) and
Michigan section of American Water Works Association (MI-AWWA) for financial
support of this work. Thanks are also due to Ms. Ewa Danielewicz, from the Center for
Advanced Microscopy for her assistance during SEM sample preparation. We would also
like to thank Mr. David Jackson for his assistance with the membrane coating process
and Mr. Michael Satoh and Yolanda Brooks for assistance in the laboratory analysis. Ms.
Deborah Rodrigues fi'om Microbiology and Molecular Genetics is also acknowledged for

her co-operation and assistance in using the equipments.

181

8.6 REFERENCES

Beltran, F. J., Rivas, F. J., and Montero-de-Espinosa, R. (2003a). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts.l. Homogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14), 3210—3217.

Beltran, F. J ., Rivas, F. J ., and Montero-de-Espinosa, R. (2003b). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts. 2. Heterogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14), 3218-3224.

Bottino, A., Capannelli, 0, Del Borghi, A., Colombino, M., and Conio, O. (2001).
"Water treatment for drinking water: ceramic microfiltration application"
Desalination, 141, 75-79 (2001).

Cho, J., Amy, G., Pellegrino, J. (20008). "Membrane filtration of natural organic matter:
Comparison of flux decline, NOM rejection, and foulants during filtration with three
UF membranes" Desalination, 127 (3): 283-298.

Cho, J ., Amy, G., Pellegrino, J. (2000b). "Membrane filtration of natural organic matter:
Factors and mechanisms affecting rejection and flux decline with charged
ultrafiltration (UF) membrane" Journal of Membrane Science, 164 (1), 89-110.

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

DeFlaun, M.F., Murray, C.J., Holben, W., Scheibe, T., Mills, A., Ginn, T., Griffin, T.,
Majer, E., and Wilson J .L. (1997). "Preliminary observations on bacterial transport in
a coastal plain aquifer" Federation of European Microbiological Societies
Microbiology Reviews, 20, 473-487.

Dumoutier, N., and Mandra, V. (1996). "Giardia and Cryptosporidium removal by water
treatment plants", Water Supply, 14 (34), 91-396.

Field, R.W. (1996). "Mass transport and the design of membrane systems. Industrial
Membrane Separation Technology", Scott, K. and Hughes,R., Eds., Blackie
Academic and Professional, Glasgow, UK.

Finstein, M. S. (2004). "Protecting watersheds from Cryptosporidium in manure: A
literature review ", Journal of American Water Works Association, 96(2), 114-10.

Gracia, R., Aragiies, J .L., and Ovelleiro, J. L. (1996). "Study of the catalytic ozonation of

humic substances in water and their ozonation byproducts" Ozone: Science and
Engineering, 18(3), 195-208.

182

Gracia, R., Cortes, S., Sarasa, J., Onnad, P., and Ovelleiro, J. L. (2000a). "Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Ormad, P., and Ovelleiro, J. L. (2000b). "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

Hagen, K., (1998). "Removal of particles, bacteria and parasites with ultrafiltration for
drinking water treatment", Desalination, 119 (1-3), 85-91.

Heinig, C., (1993). "03 or 02 and Ag: A new catalyst technology for aqueous phase
sanitation", Ozone: Science and Engineering, 15 (6), 533-546.

Helmer, DR, and Finch, GR, (1993). "Use of MS2 coliphages as a surrogate for enteric
viruses in surface waters disinfected with ozone", Ozone: Science and Engineering,
15 (4), 279-293.

Hrudey, S.E., Payment, P., Huck, P.M., Gillham, R.W., and Hrudey, E]. (2003). "A fatal
waterborne disease epidemic in Walkerton, Ontario: comparison with other
waterborne outbreaks in the developed world", Water Science and Technology, 47
(3), 7-14.

Hsu, B.M., and Yeh, H.H. (2003). "Removal of Giardia and Cryptosporidium in drinking
water treatment: a pilot-scale study", Water Research, 37 (5), 11 1 1-1 1 17.

Johnson, WP. and Logan, BE. (1996). "Enhanced transport of bacteria in porous media
by sediment phase and aqueous-phase natural organic matter", Water Research, 30
(4), 923-931.

Karnik, B.S.; Davies, S.H., Chen, K.C.; Jaglowski, D.R.; Baumann, M.J.; Masten, S.J.
(2005a). "Effects of ozonation and pH on the permeate flux of nanocrystalline
ceramic membranes" Water Research, 39(4), 728-734.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005b). "The effects of
combined ozonation and filtration on disinfection by-product formation" Water
Research, 39(13), 2839-2850.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S.J. (2005c). "Fabrication of
catalytic membranes for the treatment of drinking water using combined ozonation
and ultrafiltration" Environmental Science Technology, 39, 7656-7661.

Kim, J. 0., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14'"
Ozone World Congress, Dearborn, MI, 131-143.

183

Kim, J .O. and Somiya, 1., (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery fi'om coagulated
raw sludge" Environment and Technology, 22(1): 7-15.

Knapp, E.P., Herman, J.S., Homberger, G.M., and Mills, AL. (1997). "The effect of
distribution iron-oxyhydroxide grain coatings on the transport of bacterial cells in
porous media" Environmental Geology, 33 (4), 243-248.

Lake Lansing Watershed Advisory Committee. Progress Report, Ingham County Drain
Commissioner’s Office, Mason, Michigan, (1998).

Lee, J .Y., and Deininger, R.A. (1999). "Survival of bacteria after ozonation", Ozone:
Science and Engineering, 22 (1), 65-75.

Lee, K.H. (2001). "The effects of ozonation pathways on the formation of ketoacids and
assimilable organic carbon (AOC) in drinking water", PhD Dissertation, Michigan
State University, East Lansing, MI 48823.

Legube, B. and Karpel Vel Leitner, N. (1999). "Catalytic Ozonation: A promising
advanced oxidation technology for water treatmen " Catalysis Today, 53 (1), 61-72
(1999).

Liltved, H., Hektoen, H., and Efraimsen, H. (1995). "Inactivation of bacterial and viral
fish pathogens by ozonation and UV irradiation in water of different salini ",
Aquacultural Engineering, 14, 107-122.

Madaeni, S.S., Fane, A.G. and Grohmann GS. (1995). "Virus removal from water and
waste water using membranes" Journal Membrane Science, 102, 65-75.

McKenzie, K.J.; Marken, F .; Hyde, M.; Compton, R.G. (2002). "Nanoporous ironoxide
membranes: Layer-by-layer deposition and electrochemical characterization of
processes within nanopores" New Journal of Chemistry, 26, 625-629.

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

Ni, CH; Chen, J.N. (2001). "Heterogeneous catalytic ozonation of 2-chloroph6nol
aqueous solution with alumina as a catalyst" Water Science and Technology, 43 (2),
213-220.

06, T., Koide, H., Hirokawa, H., Okukawa, K. (1996). "Performance of membrane
filtration system used for water treatmen " Desalination, 106 (1), 107-113.

Otaki, M., Yano, K., and Ohgaki, S. (1998). "Virus removal in a membrane separation
process" Water Science and Technology, 37 (10), 107-116.

184

Owens, J .H., Miltner, R.J., Rice, E.W., Johnson, C.H., Dahling, D.R., Schaefer F.W. and
Shukairy, HM. (1999). "Pilot scale ozone inactivation of Cryptosporidium and other
microorganisms in Natural Water" Ozone: Science and Engineering, 22 (5): 501-517
(1999)

Penn, R.L., Zhu, Q, Xu, H., and Vebeblen, DR. (2002). "Iron oxide coatings on sand
grains from the Atlantic coastal plain: High resolution transmission electron
microscopy characterization" Geology, 29 (9), 843-846.

Radhakrishnan, R., Oyama, ST. (2001). "Ozone Decomposition over manganese oxide
supported on ZrO2 and TiO2: A kinetic study using in situ Laser Raman
Spectroscopy’ ’, Journal of Catalysis, 199(2), 282-290.

Schlichter, B., Mavrov, V. and Chmiel H. (2004) "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltmtion for drinking water production from
surface water" Desalination, 168, 307-317.

Scholl M.A., Mills, A.L., Herman, J .S., and Homberger, GM. (1990). "The influence of
mineralogy and solution chemistry on the attachment of bacteria to representative
aquifer materials" Journal of Contaminant Hydrology, 6 (4), 321-336.

Silliman, S.E., Dunlap. R., Fletcher, M., and Schneegurt, MA. (2001). "Bacterial
transport in heterogeneous porous media: Observations from laboratory experiments",
Water Resources Research, 37 (11), 2699-2707.

Standard Methods for Examination of Water & Wastewater (1998). 20th ed. Clesceri,
L.S., Greenberg, A. E., Eaton, A.D. (Eds.), (American Public Health Association,
1998)

Van der kooij, D., Visser, A., and Hijnen, W.A.M. (1982). "Determining the

Concentration of Easily Assimilable Orgarric- Carbon in Drinking-Water" Journal
American Water Works Association, 74 (10), 540-545.

185

CHAPTER NINE

USE OF SALICYLIC ACID AS A MODEL COMPOUND TO
INVESTIGATE HYDROXYL RADICAL REACTION IN
OZONATION-MEMBRANE FILTRATION HYBRID PROCESS

9.1 ABSTRACT

In this paper, we report on an investigation of the reaction mechanism which occurs in
the ozone-membrane filtration hybrid process. In this study experiments were conducted
using salicylic acid (SA) as a hydroxyl radical probe. Four processes were investigated:
ozonation alone, membrane filtration alone, the ozonation-ceramic membrane filtration
hybrid process and ozonation-iron oxide coated ceramic membrane filtration.
Experiments were conducted at two different pH values: pH ca. 2.5 and pH ca. 7.0. There
was no change in concentration of SA at pH below 3.0. The results show SA was not
removed by either molecular ozone reactions, filtration (including sorption), or the
combined processes at pH values less than 3.0 and that the reaction rates increase
significantly at pH values >7.0, demonstrating that hydroxyl radical reactions control the
reaction mechanism. The results also showed that the salicylate ion sorbed to the
membrane whereas salicylic acid did not. At pH above 7.0 in the hybrid process, there is
continuous decline in the concentration of SA suggesting a strong surface catalytic
reaction dominating the process. The iron oxide coated membranes combined with
ozonation result over 95% removal of SA after 240 minutes at an ozone dosage of 2.5

g/m3, compared with 80% percentage with the uncoated membrane. The presence of 2,

186

3-dihydroxy benzoic acid (2, 3-DHBA) and 2, 3-dihydroxy benzoic acid (2, 5-DHBA) in
the permeate was confirmed using GC/MS. 2,3-DHBA was found to be the predominate
byproduct and is known to form as a result of hydroxyl radical reaction with SA. A
reaction mechanism is suggested to explain the enhanced SA removal in the hybrid
process.
Keywords: ceramic membrane, iron oxide, catalytic coating, ozonation, catalytic
ozonation, salicylic acid, membrane filtration, high perform liquid chromatography
(I-IPLC), GC-MS
9.2 INTRODUCTION

Over the last decade there has been increasing interest in combining ozonation
with membrane technology for drinking water treatment. Few studies have investigated
effect of ozone on organic membranes as these membranes are not resistant to ozone and
are damaged during operation (Hashino et al., 2000; Shanbag et al., 1998; Castro and
Zander 1995; Shen et al., 1995). An alternate solution is to combine ozone with ceramic
membranes, as these membranes are ozone resistant and have demonstrated stable
permeate fluxes without membrane damage (Karnik et al., 2005a; Schlichter et al., 2004;
Kim et al, 2001, 1999; Allemane, 1993). Recent studies have shown that permeability
rates obtained with ceramic membranes are superior to those from polymeric membranes.
Ceramic membranes have also been found to be to be more effective than polymeric
membranes for the treatment of textile waste water and in alkaline and acidic solutions
(Weber 6t al., 2003). Lee and Cho (2004) found a ceramic tight-ultrafiltration membrane
had the same potential as a nanofiltration polymeric membrane, in terms of reducing the

formation of haloacetic acid formation, and were comparable to polymeric membranes

187

with similar molecular weight cut-off (MWCO) for the removal of natural organic matter
(NOM). Several researchers (Karnik et al., 2005b, 20056; Schlichter, 2004; Shioyama et
al., 2001) found that the combined process involving ceramic membrane filtration with
ozonation was more efficient for the elimination of disinfection byproducts (DBPS) and
other organic contaminants in raw water without membrane damage. Ceramic
membranes have robust performance when compared to polymeric membranes in terms
of increasing severity of operating parameters, including higher temperatures and
pressures, as well as a higher resistance to chemicals and overall durability, thus making
ceramic membranes a natural choice in spite of their higher initial costs (Benfer et a1.
2004; Zuzek et al., 2001). The metal oxide that makes up the base material for the
ceramic membrane matrix also acts as a catalyst and assists in the degradation of ozone
on the membrane surface. The ozone gas is adsorbed on the catalyst surface, resulting in
its decomposition and subsequent formation of -OH radicals (Acero et al., 1999). In the
presence of different metal oxide catalysts, such as iron oxide, manganese oxide, titania,
alumina and zirconia, ozone degrades NOM, recalcitrant organic compounds, including
saturated carboxylic acids, phenols, aromatic hydrocarbons, dyes, humic substances and
herbicides (Beltran 6t al., 2005, 2003a, 2003b; Trapido et al., 2005; Ni and Chen 2001;
Radhakrishnan and Oyama, 2001; Gracia et al., 2000a, 2000b, 1996). The enhanced
degradation of ozone on the metal oxide surface is thought to result from the adsorption
of organic compounds on the metal oxide and the subsequent decomposition of ozone,
resulting in the formation of hydroxyl radicals which further assist in the degradation of
these refractory organic compounds (Ernst 6t al., 2004; Pines and Reckhow, 2003; Ma et

al., 2000, 1999; Andreozzi 6t al., 2000, 1998a, 1998b, 1992).

188

Several researchers have attempted to study hydroxyl radical reactions in water
with the use of probe compounds, electron paramagnetic resonance (EPR) spectroscopy,
and ultrasound irradiation process (Han et al., 2002). The most commonly used probe
chemicals include para chlorobenzoic acid, salicylic acid, dimethyl sulfoxide (Park et al.,
2004; Xi et al., 2004; Stiener and Babbas, 1990). Since the ideal probe will react only
with OH radicals and not to an appreciable extent with ozone, the disappearance of the
probe is a measure of the OH radical concentration. While a number of probes may be
used, salicylic acid (SA) is preferred over other probe compounds because of i) its high
reaction rate with the OH radical (5><109 M'ls'I), ii) its use at concentrations sufficient to
compete with other scavengers present in the solution and iii) the stability of the resulting
products, which allows for their analysis and quantification (Punchard and Kelly, 1996).
The major products formed as a result of hydroxyl reactions with SA are 2, 3-dihydroxy
benzoic acid (2,3-DHBA), 2, 5-dihydroxy benzoic acid (2, 5-DHBA) and catechol
(Albarran and Schuler, 2003). This paper reports on our investigations into the catalytic
reaction in the ozone-membrane filtration hybrid process using salicylic acid as a probe
compound to quantify hydroxyl radicals and to determine the reaction mechanism in the
hybrid process.

9.3 MATERIALS AND METHODS
9. 3. 1 Ozone-Membrane Filtration Experiments

The experimental setup used for the combined ozone membrane filtration is shown in
Figure 9.1. Ozone gas was added into the water stream through the simple Y inline mixer
(Ozone Service, Burton, BC, Canada) just before the feed water entered the membrane

module. The water level in the 3.5-liter reservoir was maintained at a constant level

189

during the experiments using a peristaltic pump (Masterflex Model 7520-35, Cole-
Parrner Co., Chicago, Illinois) to transfer water from the 1.5-liter reservoir into the 3.5-
liter reservoir. The operating conditions used are given in Table 9.1. The experiments
were performed with membrane cross flow velocity of 0.6 m/s; the flow was turbulent

with Reynolds number of approximately 6000.

 

 

 

 

Table 9.1
Operating Conditions for the Ozone -Membrane Filtration System
Water recirculation rate 2.75 LPM
Water temperature 20 °C
Ozone gas flow rate 100 mL/min
TMP 0.5 bar

 

 

 

 

To generate ozone, pure oxygen gas (99.999%) fiom a pressurized cylinder was dried
using a molecular sieve trap, and then fed to an ozone generator (Model OZ2PCS,
Ozotech Inc., Yreka, Calif). The gaseous ozone concentration was controlled by varying
the voltage applied to the ozone generator. The excess gas was vented after passing the
gas through a 2% potassium iodide (KI) solution to destroy any residual ozone gas.
Teflon® tubing and stainless steel or Teflon®‘joints and valves were used throughout the
system.

A tubular ceramic membrane in a stainless steel filter holder was used in these
experiments. The membrane used in these experiments was a Clover-leaf design
(containing three channels) CéRAM membrane (TAMI North America, St. Laurent,
Quebec, Canada) with a molecular weight cut-off of 5 kDa. The external diameter of the
membrane was 10 rmn and the active membrane length was 25 cm. The membrane had a

total filtering area of 41 .2 cm2.

190

Waste gas

     

 

 

 

 

 

 

 

 

 

 

 

 

 

Moisture
Adsorbent Spectrophotometer
Ozone
generator ‘r—
Oxygen ‘ l |
Cylinder , __ ‘
Kl .

    
    

 

 

 

 

 

 

 

 

 

 

 

 

D<l
f\ Membrane
v 9..
Water \
Tank ® Flowmeter
L :5 ® Pressure gauge

 

 

 

 

 

 

g Pump
49' XValve

Figure 9.1 Schematic Representation of the Ozone-Membrane Filtration System
(Karnik et al., 2005a)

 

 

 

 

 

191

The membrane was used uncoated (as supplied by the manufacturer) and it was also
used after being coated with iron oxide. Colloidal iron oxide particles, having an average
particle diameter of 4 to 6 nm, were prepared by Sorum’s method and used for coating
the membranes following the layer-by-layer technique developed by McKenzie et al.
(2002) for coating doped tin oxide electrodes (Karnik et al., 20056). The membranes
were coated with 40 layers of iron oxide nanoparticles and sintered at 900 °C. The
conditions for coating and sintering were determined based on results obtained for the
removal of disinfection byproducts (Karnik et al., 2005b,c) and on the changes in the
morphology of the membrane surface as detected by electron microscopy (Karnik 6t al.,
2006)

In this study, the treatment processes evaluated included membrane filtration,
ozonation, and combined ozonation-membrane filtration using both uncoated/unsintered
membranes and iron oxide coated/sintered membranes. Deionized water was spiked with
salicylic acid (17 mg/L, 65 M) at a concentration equivalent to the total organic carbon
(TOC) concentration typically present in Lake Lansing, MI a border line eutrophic lake
which served as the water source for all our earlier work (Kamik et al., 2005a, b, c). All
experiments were carried out for four hours with a constant water temperature of 20 °C
and constant gaseous ozone concentration of 2.5 g/m3. Permeate samples were collected
in bottles covered with Parafilrn® and stored on ice for the duration of the experiment.
The permeate was collected and analyzed to determine the concentrations of salicylic

acid, 2,3-DHBA, 2,5-DHBA and catechol. The experiments were performed at pH 2.5-

192

3.0 and pH 7.9-8.1. Where pertinent t-butyl alcohol (t-BuOH) (ACS grade, 15.5 mg/L)
was added as an -OH radical scavenger to study the reaction mechanism.
9. 3.2 Quantification of salicylic acid and its byproducts

Salicylic acid (SA), 2,3-dihydroxy benzoic acid (2,3-DHBA), and 2,5-dihydroxy
benzoic acid (2,5-DHBA) were acquired from Sigrna-Aldrich Corp. (St. Louis, MO,
USA). A stock solution of SA (20 mg/L) was prepared weekly and stored in the dark at 4
°C. The concentration of SA, 2,3-DI-IBA , 2,5-DHBA and catechol was measured using
high performance liquid chromatography (HPLC).

HPLC analysis was carried out using a Perkin Elmer (Wellesley, MA, USA)
Binary LC 250 pump, Waters Millipore Lambda-Max Model 480 LC (Bedford,
MA,USA) spectrophotometer detector and PE series 200 autosampler. The wavelength
was set to 234 nm.

Separation of SA and its byproducts was achieved using a LiChrospher 100 C18
RP, 250><4.6 mm ID, 5 pm (Merck — 5 pm). Elution of the compounds was performed at
1 mL/min. The mobile phase was a solution of HPLC-grade water (supplier) containing
0.1 % phosphoric acid and HPLC-grade acetonitrile (supplier) prepared at a ratio 60:40.
Under these conditions, the retention times of SA, 2,3-DHBA, 2,5-DHBA and catechol
were 8.7, 3.4, 4.5 and 5.2 minutes, respectively. All samples were adjusted to pH 2.9 with
ACS reagent grade phosphoric acid before analysis. Calibration curves were obtained
following the analysis of five standards prepared from stock solution.

9. 3.3 Confirmation of the identity of byproducts using GC—MS
The permeate samples were evaporated under helium and the residue was dried

over P205 in a vacuum desiccator for one hour. The completely dried sample was

193

derivatized by silylation using bis(trim6thylsilyl)trifluoroacetamide (BSTFA) + 1%
trimethylchlorosilane (TMCS) (Regis, Morton Grove, IL) at 100 °C for 20 min to convert
all free -OH and -COOH groups into their volatile TMS-ether (-OSiM63) and TMS-ester
(-CO2SiMe3) derivatives, respectively. GC-MS was performed using a JEOL AX-505H
double-focusing mass spectrometer coupled with a Hewlett-Packard 5890J GC (Norwalk,
CT). The mass spectrometer was operated in electron impact mode and a 30 m X 0.32
mm ID, 0.25 pm DB-lms column (J&W Scientific, Folsom, CA) was used for the
analysis. The oven temperature was ramped from 150 °C to 299 °C at a rate of 10
°C/min. The injector temperature and separator temperature were 299-300 °C and 280
°C, respectively. The flow rate of the carrier gas (N2) was 20.0 mL/min.
9.4 RESULTS AND DISCUSSION

As shown in Figure 9.2, neither ozonation, neither membrane filtration-ozonation,
nor membrane filtration alone resulted in a significant decrease in the SA concentration
(within 5% std. deviation) after up to 240 minutes of treatment. As membrane filtration
(alone) did not result in a significant reduction in the SA concentration, it appears that at
low pH, SA is not sorbed within the system to any appreciable over the period of
experiment. It is also apparent that SA does not react with molecular ozone under these
experimental conditions. Removal of SA did not occur to any appreciable extent with the
ozone-filtration process, suggesting that significant decomposition of ozone and
formation of OH radicals did not occur with the uncoated membrane at a pH of 2.5-3.0.
This further substantiates our choice of SA as a model compound as it neither

substantially sorbs to the membrane surface nor undergoes direct reactions with

194

molecular ozone, thus the decrease in concentration of SA can be correlated to the

hydroxyl radical concentration.

 

.0-02 --e-MF+OZ -0-Mlethrecycle

 

Concentration 0.1M)
‘6’ 8

20*

1o-

 

 

 

o r r r r

O 50 100 150 200 250
Time (mine)

 

 

 

Figure 9.2 Disappearance of salicylic acid at pH 2.5-3 with different treatment
processes
Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1

Membrane filtration (MF), Ozonation (OZ), Ozone-uncoated unsintered Membrane filtration
(MF+0Z)

As shown in Figure 9.3, at pH values in the range of 7.9 to 8.1, membrane
filtration with recycle resulted in the sorption of SA on the membrane surface over the
first 50 minutes of treatment. The concentration of SA gradually reached steady state and
concentration of SA remained constant for the remainder of the experiment. When the

same treatment was repeated but without recycle of the reject stream, the capacity for SA

195

appeared to increase, although steady-state conditions were still reached after
approximately 30 minutes and further reductions in the SA concentration beyond the

initial loss did not occur.

 

+ MF + OZ
+ MF+OZ + Catalyst coated MF+OZ
+ Membrane filtration vdo recycle —><— OZ-MF-t-BOi-i

0 02 with t-BuOH

 

 

 

Concentration (31M)

 

 

 

 

o r r T r

0 50 100 150 200 250
Time (mine)

 

 

 

Figure 9.3 Disappearance of salicylic acid at natural pH with different

treatment processes

Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1, pH 7. 0-8. I

Membrane filtration (MF), Ozonation (OZ), Ozone-uncoated unsintered Membrane filtration
(MF+ OZ), Iron oxide catalyst coated membrane filtration combined with ozonation (Catalyst coated
MF+OZ)

Comparing these results to those obtained with membrane filtration at pH 2.5-3.0,
it appears that the salicylate ion more effectively sorbs to the membrane surface than does
the protonated form (the pKa for salicylic acid is 2. 9). At the higher pH, SA was

degraded by ozone to a greater extent than that observed at pH 2.5-3.0 (40% at pH 8 vs.

196

<5% at pH 2.8 after 240 minutes). The combination of ozonation and membrane
filtration resulted in further removals (>80% after 240 minutes), with combined
ozonation and membrane filtration using the iron oxide coated membrane being the most
effective of the processes studied (up to 95% after 240 minutes). In order to elucidate the
mechanism of the reaction, t-BuOH was added as an OH radical scavenger. The addition
of t-BuOH to the ozonation process alone and the ozonation—uncoated membrane
filtration system, resulted a significant decrease in the reactivity of SA even at higher pH
(<15% removal after 240 minutes), suggesting the importance of indirect oxidation
reactions rather than the effect of pH being the result of a greater reactivity of the
salicylate ion with ozone than salicyclic acid. The indirect reaction pathways result in the
formation of hydroxyl, super oxide and other radicals, which appear to accelerate the
decomposition of SA. The synergy between sorption and reactions involving the
hydroxyl radical appears to lead to a greater decrease in SA concentration in the hybrid
process than in the ozonation process (at the same ozone dose), This enhanced removal of
SA with the iron oxide coated membranes is believed to be a result of changes in the
surface morphology (Karnik et al., 2006), which affect the performance of the system by
improving surface catalytic properties of the membrane surface. This improved
performance with coated membranes when compared to uncoated membranes suggest a
strong correlation of surface chemistry and hydroxyl reaction in the hybrid process.

To substantiate our claims of the importance of hydroxyl radical reactions in the
process, we analyzed the permeate from different experiments for the products formed

from the reaction of SA with hydroxyl radicals. These products include 2, 3 DHBA, 2, 5-

197

DH'BA and catechol. In these experiments we detected only two byproducts 2, 3 DI-LBA

and 2, 5-DHBA, there was no evidence of catechol in any of the experiments conducted.

 

 

503 -o-oz ~4-m=+oz -o-m=

Concentration 03M)
8 8

N
O
1

 

 

 

10‘
....... 3""------------.3...........,____‘
...-a ------
OG—i-Iza. , a I 9 1 4 r
0 50 100 150 200 250
Tine(mhutee)

 

 

 

Figure 9.4 Formation of oxidation byproducts at pH 2.5-3 with difl'erent
treatment processes
Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1

Membrane filtration (MF), Ozonation (OZ), Ozone-uncoated unsintered Membrane filtration
(MF+ OZ)

Further the by products concentration are reported as a total by products concentration by
adding the molar concentration of the individual compounds. The individual
concentration of these compounds is reported in Table 9.2. At pH 2.5-3.0, there was no
evidence of SA byproducts in the treated stream from either ozonation or membrane
filtration alone (Figure 9.4). However, with the hybrid process SA byproducts were

formed at concentrations <7 .8 11M, supporting the suggested mechanism that adsorption

198

followed by the decomposition of ozone on the membrane surface leads to the formation

of hydroxyl radical, which react with SA to form the observed byproducts.

Table 9.2

Total molar concentration (11M) of the by—product in individual treatment process

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oz MF w/o MF+OZ Catalyst

Time pH ”5 ”52C? recycle /MF 1 w/ oz pH MF+OZ coated
(mins) 25- 2 g 3 0 p 3 0' pH 8.1 w "f” e t-BuOH 3.1 pH 8.1 MF+OZ

3.0 . - . . pH 8.1 pH 8.1

0 ND ND ND ND ND ND ND ND ND

20 ND ND ND ND ND 4.54 13.60 12.33 24.66
25 ND ND ND ND ND 4.54 14.30 22.06 24.33
30 ND ND 1.95 ND ND 4.54 14.20 22.06 30.17
60 ND ND 3.89 ND ND 4.54 20.71 35.66 33.74
120 ND ND 7.79 ND ND 4.54 20.30 36.98 35.04
180 ND ND 7.79 2.6 ND 4.54 21.60 I40.36 45.09
240 ND ND 7.79 2.6 ND I7.85 21.62 37.63 55.15

 

’ A ll concentrations are molar concentrations of 2, 3-DHBA expect in these cases where it is the sum of 2, 3-
DI-IBA and 2,5-Dl-IBA. There was no detection of catechol in any of the experiments.

For MF+OZ pH pH 8.1 at time 180 minutes, [2, 5-DHBA ] =2. 78 W and

For MF+OZ w/ t-BuOH at time 240 minutes, [2, 5-DHBA] =3. 32 W

Again, comparison of Figures 9.4 and 9.5 for hybrid process at lower pH < 3 and the
hybrid process with t-BuOH at pH > 7.0 respectively show similar concentration (~ 7.8
uM) of SA byproducts, supporting the theory that the reaction was the result of hydroxyl
radical initiated decomposition rather than sorption alone. Also we found in experiment
where t-BuOH was added to ozonation alone, there was no significant concentration of
the by-products concentration reported firrther supporting our hypothesis of hydroxyl

radical reaction mechanism.

199

 

 

 

 

 

 

 

 

 

 

 

+ MF vale recycle pH 8.1 -><— OZ+MF N t-BuOH
+ 02 pH 8.1 -~X~ MP wtrecycie pH pH 8.1
+ OZ+MF pH pH 8.1 + Catalyst coated MF+OZ pH 8.1
60.00 0 02 11W t-BuOH
50.00 ~
g 40.00 -
C
O
E 30.00 -
c
i a“-.. _
8 20.001 “I” '“ ‘ ”M”
10.00 i /
0.00 a [I , ;_
1 200 250
Time (mine)

 

 

Figure 9.5 Formation of byproducts at pH 8 with different treatment process
Experimental Setup: Fig 9.1, Operating Conditions: Table 9.1, pH 7. 0-8.1

Membrane filtration (MF), Ozonation (OZ), Ozone-Membrane filtration (MF +02), Iron oxide catalyst
coated membrane filtration combined with ozonation (Catalyst coated MF+OZ)

From Figure 9.5, we can see there is a significant decrease in SA byproducts formed from
membrane filtration alone with or without recycle than with ozonation where we have
reported. With the hybrid process with uncoated/unsintered or coated/sintered
membranes the concentration of SA byproducts increased to ~ 40 pM and ~ 55 11M,
respectively compared to ~21. M with ozonation.

GC-MS analysis verified the identity of the SA byproducts as 2,3—DHBA and
2,5—DHBA. There was no evidence of any catechol formed in the process. The
concentrations of these compounds were verified with HPLC and indeed coincided with

the results of GC-MS. There was no evidence of 2,5—DHBA in any of the experiments

200

except for the hybrid process and the concentration was detected only after two hours of
experiment, making 2,3—DHBA the most dominant byproduct obtained in this study.
Figure 9.6a shows the total ion chromatograrn obtained from the feed water, containing
SA acid. Figures 9.6b and 9.66 show the results of the permeate analysis after treatment
with the ozonation-membrane filtration and ozonation-iron oxide coated membrane
filtration. There are clear peaks of SA marked as 457, 2, 3-DHBA marked as 625 and 2,
5-DHBA marked as 649. The mass spectra (Figure 9.7 a-c) confirm the identity of these
compounds. Quantification of these compounds using HPLC analysis confirmed the

dominance of 2, 3-DHBA isomer as the major reaction product in the permeate.

201

a)

 

 

 

 

 

SA
285 457
4|llllll
122
1‘111111
11111111 182
IIIIII m
m “..1.-.” . . 1. -,. 24778244
5“" ‘ " 5°” " " "' 55!]. 1000 "001300 13001400
Mn 5 8 10 1'2 14 1'6 1'8 2'0 2'2 2'4 2's
1) 2, 3-DHBA
) SA 625
288 457
'4IIIIII
m n, 2, 5-DHBA
12111111 1 I 649
:111111 as $3 1154
1"ch l L‘sa . [1-9. L . .-., film
$933111 11 11 11 11 11 11 11 '11 1am 11m 1m 13m 14m
1431 i 8 1'0 1'2 14 1'6 1'8 2'0 2'2 24 2'6
c) 2, 3-DHBA
s A 625
457 ,5-DHBA
649
$2 810
..71.‘ "mutt???" “11"”“‘¢‘l ...... ‘ ..................... ' .................. -3“3"6°
' ' 550' ' 71101....7950' -1019... .1139.-. 1.599..-. 13001400
1'0 1'2 1'4 1'6 1'8 2'0 2'2 2'4 2'6

 

Figure 9.6 Total ion chromatograms (TIC) for Salicylic acid (SA), 2, 3- dihydroxy
benzoic acid (2, 3-DHBA), and 2, 5- dihydroxy benzoic acid (2, 5-DHBA)
derivatized by BSTFA+ TMCS.
a) TIC obtained from the sample before treatment, b) TIC obtained fiom the permeate from ozone-
membrane filtration process and c) TIC obtained from the permeate from ozone-catalyst coated
membrane filtration process

202

a) Salicylic acid derivative

 

 

 

 

 

 

 

A 1 73.1 267.0
?
a 1490
. 91.0 ' 193.0 209.0
g 11 .1 111,0 ‘21-'21?” 11158.9 130. 1 122822330 2491-0 112809
M/Z
b) 2,3 DHBA derivative
... 10 731 355.1
E
i a
x
a 193.0
'0 2 "7'" 2420 2550
a 59.0 91011191] ‘ Iii-0L 1550 mf-Ozzgu .11 1" 281.0 J37?”
M/Z

6) 2,5 DHBA derivative

 

 

 

"" 10 731 355.1
a
g a
:g 6

7.
a
a 2 3700
. 117.0 .
3 48859.0 - 105019.11?" uI6?.0179.1133.183§0¥%‘231f25102328102910 32103320 h

M/Z

Figure 9.7 Mass spectra (MS) for compounds derivatized by BSTFA+

TMCS.
a) Mass Spectra for Salicylic acid, b) Mass spectra for 2.3-DHBA and c) Mass spectra for 2,5-
DHBA

203

From our results we can suggest that the probable mechanism occurring in the
ozone-membrane filtration hybrid process is a heterogeneous phase reaction involving the
enhanced removal of SA by the adsorption of SA to the membrane followed by the
decomposition of ozone at the membrane surface resulting in the formation of OH
radicals which then react with the sorbed SA. It is thought that the active sites on the
ceramic membrane surface or the modified surface morphology (pore size, grain size,
surface area) and properties (surface charge, hydrophobicity) resulting from coating and
sintering (Karnik et al., 2006) accelerates the decomposition of ozone and the reaction of
these radicals, thereby promoting the reaction of SA. This could possibly explain the
decrease in the dissolved ozone concentration in the treated Lake Lansing water for
ozonation alone and the hybrid process with uncoated/unsintered membranes and coated
sintered membranes combined with ozonation. The results are tabulated in Table 9.3; we
found a decrease in dissolved ozone concentration in hybrid process when compared to
ozonation alone. In addition with coated membranes the dissolved ozone concentration
decreased more than 40 (:l: 1.0) % when compared to the uncoated membranes.

Table 9.3

Dissolved ozone concentration in the permeates of Lake Lansing water for different
treatment processes.

 

 

 

 

 

 

 

Treatment Process Dissolved ozone
concentration (rig/L)
Ozonation alone 0.15 (i 0.01)
Uncoated-unsintered membrane filtration + ozonation 0.12 (:t 0.03)
Coated-sintered membrane filtration + ozonation 0.07 gt 0.01)

 

The suggested mechanism is depicted in Figure 9.8. When the system is operated
in the continuous recirculation operation mode, the dissolved ozone adsorbs on the

ceramic membrane surface and rapidly decomposes due to the presence of reactive

204

surface groups. It is known that ozone in gaseous phase rapidly decomposes on metal
oxide surface (e.g., Ernst 6t al., 2004, Acero et al., 1999). The decomposition of ozone is
well documented (Bablon et al., 1991). Based on the published mechanisms for the
decomposition of ozone, we hypothesize the decomposition of ozone generates active
atomic oxygen and reacts with the membrane surface (TiO2 in case of uncoated
membranes and TiO2-F62O3 in case of coated ceramic membranes) to produce 02H
anions which subsequently react with the dissolved O3 molecule to generate the ozonide
anion (O3) radical. This radical further decomposes to form the hydroxyl radical which
can oxidize the SA. However, further investigations are necessary to confirm the

oxidation mechanism.

205

Step: I Step: II

03 0’
OH OH '02 6- “0211 OH
A A A A
Step: 111 Step: IV
....y 02

................ ”a, r
o, ,, / \\
'OH on SA SA“
A A

Figure 9.8 Schematic of the suggested mechanism during ozonation membrane
filtration hybrid process.

A: is T iO2 in case of uncoated/unsintered membranes and is T iO2-F 6203 in case of iron
oxide coated sintered membrane.

206

9.5 ACKNOWLEDGMENTS

The authors would like to acknowledge the US Environmental Protection Agency (U 8
EPA) Science To Achieve Results (STAR) Program (Grant No. RD830090801) for
financial support of this work. We acknowledge Ms. Beverly Chamberlain from the Mass
Spectrometry Core of the Research Technology Support Facility, Michigan State
University for her assistance in the GC-MS analysis and Mr. Cun Liu from the
Department of Crops and Soil Sciences for generously allowing us to use the HPLC
facility in his lab. We would also like to thank Mr. David Jackson for his assistance with

the membrane coating process.

207

9.6 REFERENCES

Acero, J .L. and Gunten, U.V. (1999). "Influence of carbonate on the ozone/hydrogen
peroxide basedadvanced oxidation process for drinking water treatment" Ozone:
Science and Engineering, 22, 305-328.

Albarran, G., and Schuler, R. (2003). "Concerted effects in the reaction of °OH radicals
with aromatics: radiolytic oxidation of salicylic acid" Radiation Physics and
Chemistry, 67(3), 279-285.

Allemane H., Deloune B., Paillard H., and Legube B. (1993). "Comparative efficiency of
three systems (03/ O3H2O2 and O3/TiO2) for the oxidation of natural organic matter in
water" Ozone: Science and Engineering, 15(5), 419-432.

Andreozzi, R.; Insola, A.; Caprio, V.; D’Amore, M.G. (1992). "The kinetics of Mn(II)
catalyzed ozonation of oxalic acid in aqueous solutions" Water Research, 26, 917-
921.

Andreozzi, R.; Caprio, V.; Insola, A.; Marotta, R. Tufano, V. (1998a). "The Ozonation of
Pyruvic acid in aqueous solutions catalyzed by suspended and dissolved manganese"
Water Research, 32, 1492-1496.

Andreozzi, R.; Caprio, V.; Insola, A.; Marotta, R.; Tufano, V. (1998b). "The use of
manganese dioxide as a heterogeneous catalyst for oxalic acid ozonation in aqueous
solutions" Applied Catalysis: A, 131, 75-81.

Andreozzi, R.; Marotta, R.; Sanchirico, R. (2000). "Manganese catalyzed ozonation of
glyoxlic acid in aqueous solutions" Journal of Chemical Technology &
Biotechnology, 75, 59-65.

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

Beltran, F. J ., Rivas, F. J ., and Montero-de-Espinosa, R. (20038). "Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts]. Homogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research, 42(14), 3210—3217.

Beltran, F. J., Rivas, F. J ., and Montero-de-Espinosa, R. (2003b)."Ozone enhanced
oxidation of oxalic acid in water with cobalt catalysts. 2. Heterogeneous catalytic
ozonation" Industrial and Engineering Chemistry Research. 42(14), 3218-3224.

Beltran, F. J., Rivas, F. J., and Montero-de-Espinosa, R. (2005) "Iron type catalysts for
the ozonation of oxalic acid in water" Water Research 39, 3553-3564.

208

Benfer, S., Arki and Tomandl, G. (2004). "Ceramic membranes for filtration

applications-Preapration and Characterization" Advanced Engineering Materials,
6(7), 495- 500.

Castro K. and Zander A.K. (1990). "Membrane air-stripping - effects of pretreatment"
Journal of American Water Works Association, 87(3), 50-61.

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

Ernst M., Lurot F ., and Schrotter J .C. (2004). "Catalytic ozonation of refractory organic

model compounds in aqueous solution by aluminum oxide" Applied Catalysis B:
Environmental, 47(1), 15-25.

Gracia, R., Cortes, S., Sarasa, J., Ormad, P., and Ovelleiro, J. L. (20008). "Catalytic
ozonation with supported titanium dioxide: The stability of catalyst in water" Ozone:
Science and Engineering, 22 (2), 185-193.

Gracia, R., Cortes, S., Sarasa, J ., Ormad, P., and Ovelleiro, J. L. (2000b). "Heterogeneous
catalytic ozonation with supported titanium dioxide in model and natural waters"
Ozone: Science and Engineering, 22(5), 461-471.

Han, S.K., Nam, SN, and Kang, J .W. (2002) "OH radical monitoring technologies for
AOP advanced oxidation process" Water Science and Technology, 6(11-12), 7-12.

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

microfiltration system for drinking water treatment" Water Science and Technology,
41(10-11),17-23.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, SJ. (2005b). "The effects of
combined ozonation and filtration on disinfection byproduct formation" Water
Research, 39(13), 2839-2850.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, S]. (20056). "Fabrication of
catalytic membranes for the treatment of drinking water using combined ozonation
and ultrafiltration", Environmental Science Technology 39(19), 7656-7661.

Karnik, B.S.; Davies, S.H.; Baumann, M.J.; Masten, SJ. (2006). "AFM and SEM
Characterization of Iron Oxide Coated Ceramic Membranes", In Press, Journal of
Materials Science.

Kim, J. 0., Somiya, I. and Fujii, S. (1999). "Fouling control of ceramic membrane in
organic acid fermenter by intermittent ozonation", In the Proceedings of the 14'”
Ozone World Congress, Dearborn, MI, 131-143.

209

Kim, .10. and Somiya, I., (2001). "Effective combination of microfiltration and
intermittent ozonation for high permeation flux and VFAs recovery from coagulated
raw sludge" Environment and Technology, 22 (1): 7-15.

Lee, S. and Cho, J. (2004). "Comparison of ceramic and polymeric membranes for
natural organic matter (N OM) removal" Desalination, 160, 223-232.

Ma, J.; Graham, N.J.D. (1999). "Degradation of atrazine by manganese-catalyzed
ozonation-influence of hurrric substances" Water Research, 33(3), 785-793.

Ma, J .; Graham, N.J.D. (2000). "Degradation of atrazine by manganese-catalyzed
ozonation-influence of radical scavengers" Water Research, 34 (15), 3822-3828.

McKenzie, K.J.; Marken, F.; Hyde, M.; Compton, R.G. (2002). "Nanoporous ironoxide
membranes: Layer-by-layer deposition and electrochemical characterization of
processes within nanopores" New Journal of Chemistry, 26, 625-629.

Ni, CH; Chen, J .N. (2001). "Heterogeneous catalytic ozonation of 2-6hlorophenol
aqueous solution with alumina as a catalyst" Water Science and Technology, 43 (2),
213-220.

Park, J., Choi, H., Cho, J. (2004). "Kinetic decomposition of ozone and para-
chlorobenzoicacid (pCBA) during catalytic ozonation" Water Research, 38 (9), 2285-
2292.

Pines, D.S., and Reckhow, D.A., (2003). "Solid Phase Catalytic Ozonation Process for
the Destruction of a Model Pollutant" Ozone: Science and Engineering, 25(1), 25-39.

Punchard, N., and Kelly, F. (EDs.) (1996). "Free radicals: a practical approach" Oxford
University Press, Oxford, 1996.

Radhakrishnan, R., Oyama, ST. (2001). "Ozone Decomposition over manganese oxide
supported on ZrO2 and TiO2: A kinetic study using in situ Laser Raman
Spectroscopy”, Journal of Catalysis, 199(2), 282-290.

Shanbhag P.V., Guha A.K., and Sirkar K.K. (1998). "Membrane-based ozonation of
organic compounds Membrane-based ozonation of organic compounds" Industrial
and Engineering Chemistry Research, 37 (11), 4388 -4398.

Schlichter, B., Mavrov, V. and Chmiel H. (2004). "Study of a hybrid process combining
ozonation and microfiltration/ultrafiltration for drinking water production from
surface water" Desalination, 168, 307-317.

Shen Z.S., Semmens M.J., and Collins A.G. (1990). "A novel-approach to ozone water

mass-transfer using hollow fiber reactors" Environmental Technology, 1990, 11, 597-
608.

210

Shioyama, M., Kawanishi, T., Yokoyama, S., Nuno, M., and Yarnarnoto, T. (2001).
"Development of advanced ceramic membrane filtration system combined with

ozonation and powdered activated carbon treatment" Water Science and Technology,
1 (5-6), 91-96.

Steiner, M., and Babbs, C. (1990). "Quantification of the hydroxyl radical by reaction
with dirnythl sulfoxide" Archives of Biochemistry and Biophysics, 278(1), 478-481.

Trapido, M., Veressinina, R., Munter, R., and Kallas, J. (2005). "Catlytic ozonation of m-
Dinitrobenzene" Ozone: Science and Engineering, 27, 359-363.

Weber, R., Chmiel, H., and Mavrov, V. (2003). "Characteristics and application of new
ceramic nanofiltration membranes" Desalination, 157, 113-125.

Xi, Y., Zhan, M., Kong, L., and Wang, L. (2004). "Determination of hydroxyl radicals
with salicylic acid in aqueous nitrate and nitrite solutions" Journal of Environmental
Sciences, 16(4), 687-689.

Zuzek, E., Catén, S., Arciprete, C., Dimitrijewits, M., Almandoz, M., and Marchese, J.

(2001). "Asymmetric ceramic rrricroporous membranes" Granular Matter, 3(1-2),
145-148.

211

CHAPTER TEN

CONCLUSIONS AND RECOMMENDATIONS

10.1 CONCLUSIONS

1.

For the experimental setup described in this work increases in gaseous ozone
concentration beyond a particular ozone gas concentration (5 g/m3), had no effect on
the level of permeate flux recovery, thus a minimum threshold concentration could
achieve complete recovery.

The pH of the feed water affected the permeate flux. The improved permeate flux
recovery at lower pH is a result of higher dissolved ozone concentration under these
conditions, leading to increased ozone reaction with the NOM species that are
responsible for fouling.

Use of the combined ozonation/filtration treatment system resulted in significant
improvements in water quality compared to the filtered raw water and to that using
either ozonation or membrane filtration alone.

Using a membrane with a lkD MWCO, the minimum gaseous ozone concentration
required to bring about effective NOM degradation and meet regulatory requirements
for chlorinated DBPs was 2.5 g/m3.

With iron oxide coated ceramic membrane, the concentration of dissolved organic
carbon was reduced by >85% and the concentrations of simulated distribution system
total trihalomethanes and simulated distribution system halo acetic acids decreased by

up to 90% and 85%, respectively. In addition the concentrations of aldehydes,

212

ketones, and ketoacids in the permeate were reduced by >50% as compared to that
obtained with the uncoated membranes. Also on coating the membranes a 5kD
MWCO membrane can produce water quality to meet pertinent regulatory standards.

6. With coating and sintering the membrane surface there is an evidence of change in
surface morphology, which leads to changes in the surface properties, like changes in
pore size, roughness, grain size which subsequently led to improvement in surface
properties of the membrane which enhanced the catalytic reactions on the membrane
surface thereby leading to improvements in water quality.

7. With coating and sintering the membranes led to modification in surface morphology
which accelerated decomposition of ozone on the membrane surface and reaction of
radicals formed as a result of this decomposition, this was demonstrated by decrease
in dissolved ozone concentration for the permeates of coated-sintered membranes
when compared to the uncoated-unsintered membranes.

8. Surface characterization demonstrated beyond 40 coats of iron oxide on the
membrane there is no evidence of significant decrease in surface roughness, or
increase in the thickness of the catalyst coating, make the 40 coats the optimum.

9. Diffusion and capillary action during the coating process not only leads to uniform
coating on the membrane surface but also results in distribution of iron oxide
nanoparticles into the membrane surface.

10. The catalyst coated membrane filtration results in greater 7 log removal for bacteria,
and also reduces the risk of potential regrowth of bacteria in the distribution system

after treatment.

213

11. The changes in the surface properties of the membrane degrade the ozone on the

membrane surface and accelerate the hydroxyl radical reactions which further
degrade organic compounds on the membrane surface consequently reducing the
extent the membrane fouling, maintain stable permeate fluxes and giving improved
performance in terms of water quality to meeting US-EPA regulatory standards for

drinking water.

10.2 RECOMMENDATIONS FOR FUTURE RESEARCH

1.

The current system was operated at low transmembrane pressure, thus it is necessary
to upgrade the ozone nanocerarnic system to high pressure system and evaluate its
performance in terms of permeate flux, water quality and membrane fouling if
present.

The system should be scaled to pilot-scale to evaluate the performance of the
ozonation nano ceramic membrane filtration system. The study should be carried out
to evaluate the performance by varying operating parameters like higher flow rates,
higher pressures, temperatures and ozone dosage.

Ceramic membranes coatings tested in this work were commercially available and
Ti02 membranes coated with iron oxide nano particles. Different metal oxide
coatings like Mn02 and Zr02 should be investigated to determine their effect on the
treated water quality.

The economic analysis and feasibility studies need to be conducted to investigate
application of this system as an option to conventional treatment methods for drinking

water.

214

. Further studies should be done to investigate the mechanism involved in catalytic
oxidation.

The experiments should be carried out with different coating and sintering conditions

for different metal oxide coatings. Specifically effort could be directed to investigate
how porosity, average grain sizes and microstructure of the coating affect the
performance of the membranes. If there is a possibility of microcracking with
different coating materials that could possibly affect the performance of the system.
. This work found increased removal of ozonation by-products with increase in the
number of coatings; however there was no enhanced removal of disinfection by-
products with increase in number of coatings, a study is required to investigate this
different removal efficiency with different number of coats.

The surface charge of the membrane affects the performance of the membrane in
terms of catalytic reactions and membrane fouling. A study should be carried out to
modify the surface charge and surface properties and evaluate the performance of the
system to these conditions.

. The Lake Lansing water contains low bromide concentration. As bromide not only
reacts with ozone and forms bromate, but also affects the formation of other DBPs, it
is important to investigate the applicability of this system on waters containing high

bromide concentrations.

10. A study should be carried out to investigate the efficiency of the system for the

control and removal of existing or potential contaminants having an effect on human

health and environment.

215

APPENDIX A

The Supporting Information section consists of five pages, including eight figures
showing water quality data for permeate 1 showing effect on number of coatings on the
permeate of 15 kD and 5kD molecular weight cut off membranes, the effect of sintering
at two different temperatures on the permeate water quality.

216

 

 

 

LE] Uncoated I 20-coatings llll 40-coatings

 

 

 

 

 

% Remaining

 

 

 

 

 

 

 

 

-100

 

Fig.Al Effect of the number of catalyst coatings on the water quality of P1
Experimental setup: Fig. 5.2. Operating Conditions: Table I. Membrane Size: 15 kD, 20 or 40
layers, oven baked at 500°C. '21]! values are averages of triplicates within experiments.

 

 

 

D Uncoated I 20-coatinga I 40-coatlngs 1

 

 

Te“
8

 

T8’
8

 

Concentration.de

 

 

 

 

P1-Aidehydee&Ketonee P1-Ketoaclde

 

Fig.A2 Effect of the number of catalyst coatings on the concentrations of

ozonation by-products in the permeate P1
Experimental setup: Fig. 5.2. Operating Conditions: Table 1. Membrane Size: 15 kD, 20 or 40 layers,
oven baked at 500°C. 'All values are average of triplicates within experiments.

217

 

 

 

 

 

D Uncoated I 20-coatlngs I 40-coatings

 

100
75
50
25

-25
-50

% Remaining
O

   

-100

 

 

Fig.A3 Effect of the number of catalyst coatings on the water quality of P1
Experimental setup: Fig. 5.2. Operating Conditions: Table 1. Membrane Size: 5 kD, 20 or 40 layers,
oven baked at 5 00°C. .A II values are averages of triplicates within experiments.

 

 

 

El Uncoated I 20-coatlnge I 40-coatinge

 

125.00
100.00
75.00
50.00
25.00

Concentration, 139/L

 

P1-Aldehydee81Ketonee P1 -Ketoaclde

 

i

Fig.A4 Effect of the number of catalyst coatings on the concentrations of
ozonation by-products in the permeate P1

Experimental setup: Fig. 5.2. Operating Conditions: Table I. Membrane Size: 5 kD, 20 and 40
layers, oven baked at 500°C. .21” values are average of implicates within experiments.

218

 

 

 

I 15-20-500 CJ 15-20-900
I 15-40-500 I 15-40-900

 

 

 

 

100

 

 

 

 

 

 

 

% Remdning

 

 

 

 

 

 

 

-100

 

 

Fig.ASWater quality for PI for two different sintering temperatures Experimental
setup: Fig. 5.2. Operating Conditions: Table I. Membrane Size: 15 kD, 20 or 40 coatings, oven-baked at
500°C or sintered at 900°C. 'All values are average of triplicates within experiments. Explanation of the
legend is described in the caption of Fig. l

 

 

I 15-20-500 El 15-20-900
I 1540-500 I 1540-900

 

 

 

 

300

 

 

 

  

 

 

 

 

Pt-Ndehydeel-Ketonee P1-Keblclds

 

 

Fig.A6 Concentrations of ozonation by-products in PI for two different sintering

temperatures

Experimental setup: Fig. 5.2. Operating Conditions: Table 2. Membrane Size: 15 kD, 20 or 40 coatings,
oven-baked at 5 00°C or sintered at 900°C. ‘All values are average of triplicates within experiments.
Explanation of the legend is described in the caption of Fig. l

219

 

 

 

 

I 5-20-500 CI 5-20-900
I 5-40-500 I 5-40-900

if,” I
1“" . I; IN ii
11 1 "'

TTHM

 

 

 

 

100

 

 

 

 

     

°/e Remaining
o

 

 

 

-75
-1 00

 

 

 

 

 

Fig. A7 Water quality for PI for two different sintering temperatures

Experimental setup. Fig. 5. 2. Operating Conditions: Table I. Membrane Size: 5 kD, 20 or 40 layers,
oven-baked at 500°C or sintered at 900°C. All values are average of triplicates within experiments.
Explanation of the legend is described in the caption of Fig I

 

 

I 5-20-500 El 5-20-900
I 5-40-500 I 5-40-900

 

 

 

 

200

 

Concentration, ngL
A
c
o

 

 

 

"W' ..... f

Pt-AldehydeeaKetonee P1-Ketoacide

 

 

Fig.A8 Concentrations of ozonation by-products in P1 for two different

sintering temperatures

Experimental setup: Fig. 5. 2. Operating Conditions: Table I. Membrane Size. 5 kD, 20 or 40 layers,
oven-baked at 5 00°C or sintered at 900°C. All values are average of triplicates within experiments
Explanation of the legend ts described in the caption of F 1g I

220

         
  

  

111111111111111;11111

028 5