Application of UV LEDs for Turbid Wastewater Disinfection
by
Chenghui Zeng
A thesis submitted in partial fulfillment of the requirements for the
degree of Master of Engineering in
Environmental Engineering and Management
Examination Committee: Prof. Chettiyappan Visvanathan (Chairperson)
Prof. Nguyen Thi Kim Oanh
Dr. Oleg Shipin
Prof. Kensuke Fukushi (External Expert)
Nationality: Chinese
Previous Degree: Bachelor of Engineering in Water Supply and
Sewerage Engineering
Harbin Engineering University
P.R. China
Scholarship Donor: China Scholarship Council (CSC)
Asian Institute of Technology
School of Environment, Resources and Development
Thailand
May 2014
ii
Acknowledgements
Completing my master thesis is a milestone in my academic career that would not have
been possible without the support of many people to whom I am indebted.
Prof. C. Visvanathan has been an ideal supervisor in every aspect. I learned so much from
him and I am very grateful for his professional guidance and strong support. Without his
consistent instructions, this thesis could not have reached its present form.
I would also like to extend my sincere thanks to Prof. Nguyen Thi Kim Oanh and Dr. Oleg
Shipin, my thesis committee members, for their invaluable comments and inputs all the
time.
My profound thankfulness goes to Prof. Kensuke Fukushi, my supervisor during the
participation of UEHAS program, for giving valuable research recommendations and
financial support throughout this research. Likewise, I would also like to thank Dr. Pu Jian,
for her kind help during my stay in the University of Tokyo.
The research group under the supervision of Prof. C. Visvanathan is a wonderful setting for
my graduate experience. Here, I want to say thank you to all the team members-Paul, Park,
Paru, Pik, Milk, Ter, Mov Chinmeng, Plat, Ben, Ellis, and Lina.
Part of the experiment was conducted in the nano-lab. My special thanks go to Ms.
Mayuree, the research associate in nano-lab, for her kind guidance on microbial
experiment and technical support.
I also wish to acknowledge the financial support from China Scholarship Council. With
this scholarship, I was able to pursue my master degree at AIT.
I wish also to express my love and gratitude to my parents, my older sisters and older
brother, whose love and support made it possible for me to pursue my interests and dreams.
I love you!
iii
Abstract
UV LEDs was thought to be a good alternative to conventional UV lamp. In this study, the
impact of turbidity on the disinfection performance of UV LEDs (282 nm) was examined.
Inactivation of Escherichia coli (E. coli) and total coliform was compared based the
exposure time. Actinometry methods have been applied to measure the UV fluence in the
reactor, including iodide-iodate and ferrioxalate actinometer. Both synthetic and real
wastewater have been used to conduct the disinfection test. Turbidity of synthetic
wastewater was 27, 70, 113, and 156 NTU and that for real wastewater was 57, 72, 86, and
130 NTU.
Irradiance of UV LEDs was found to be 0.4282 mW/cm2 (iodide-iodate) and 0.382
mW/cm2 (ferrioxalate). In synthetic wastewater of 27, 70, and 113 NTU, 5 log-reduction of
E. coli could be achieved and the inactivation kinetics was a first order reaction. However,
in real wastewater of 72, 86 and 130 NTU, only 3 log-reduction of both total coliform and
E. coli could be achieved. The reason is that part of the coliform bacteria or E. coli in real
wastewater was attached to the particles. These particles can protect the bacteria from
being exposed to UV irradiation, so bacteria can still be alive even though more UV
fluence is applied.
In conclusion, UV LEDs may not be able to disinfect the wastewater to meet the required
guidelines for wastewater reclamation unless a proper pretreatment is applied to reduce the
turbidity of wastewater to a certain level.
iv
Table of Contents
Chapter Title Page
Title Page i
Acknowledgements
Abstract
ii
iii
Table of Contents iv
List of Tables vi
List of Figures iv
List of Abbreviations ix
1 Introduction
1.1 Background
1
1
1.2 Objectives of Study 2
1.3 Scope of Study 2
2 Literature Review 4
2.1 Introduction 4
2.2 Ultraviolet Light-emitting Diodes (UV LEDs) 4
2.2.1 Fundamentals of UV LEDs
2.2.2 Advantages and disadvantages of UV LEDs
2.2.3 Current status and applications
4
5
6
2.3 Turbid Water Disinfection by Ultraviolet Light 6
2.3.1 Application of simplified wastewater treatment process
in developing countries
2.3.2 Another possible application of UV LEDs for turbid
water disinfection
6
9
2.4 UV Disinfection 10
2.4.1 Fundamentals
2.4.2 Sources of UV
2.4.3 UV disinfection mechanism and the followed
photoreactivation
2.4.4 Advantages and disadvantages of UV disinfection
10
10
12
14
2.5 Factors Affecting the Disinfection Efficiency of UV
Light
2.5.1 Subordinate factors
2.5.2 UV fluence
2.5.3 Wavelength
2.5.4 Absorbance and scattering by particles in water
2.5.5 The types of microorganisms
14
15
15
15
16
18
2.6 Methods for UV Fluence Determination
2.6.1 Biodosimetry method
2.6.2 Chemical actinometry
19
19
21
2.7 Summary and Research Needs 25
3 Methodology 26
3.1 Introduction 27
3.2 Experimental Set-up 27
3.3 UV Fluence Determination by Actinometry 28
v
3.3.1 Iodide-iodate actinometer
3.3.2 Ferrioxalate actinometer
29
32
3.4 Disinfection Test with Synthetic Wastewater
3.4.1 Preparation of synthetic wastewater
3.4.2 Preparation of E. coli for synthetic wastewater
3.4.3 Enumeration of E. coli
3.4.4 Procedure for disinfection test with synthetic wastewater
3.5 Disinfection Test with Real Wastewater
3.5.1 Wastewater sampling
3.5.2 Selection of challenge organism and its enumeration
3.5.3 Procedure for disinfection test
3.6 Summary
35
35
36
37
39
40
40
41
43
43
4 Results and Discussions 44
4.1 Characterization of UV LEDs 44
4.2 UV Fluence Determination by Actinometry Method 44
4.3 Disinfection Test with Synthetic Wastewater 48
4.4 Disinfection Test with Real Wastewater 51
5 Conclusion and Recommendations 55
5.1 Conclusions 55
5.2 Recommendations for Further Study 56
References
57
Appendix A
Appendix B
62
64
vi
List of Tables
Table Title Page
2.1 Required Dilution Factors for Both E. coli and COD to Reach a
Safe Level
8
2.2 Comparison between UV LEDs and Conventional UV Lamp 8
3.1 Value for Each Parameter in Equation 3.2 31
3.2 Value Adopted for Each Parameter in Equation 3.8 and 3.9 35
3.3 Summary of Methods 43
4.1 Absorbance of Idide-iodate Actinometer Solution at Different
Exposure Time
44
4.2 Absorbance of Ferrioxalate Actinometer Solution at Different
Exposure Time
46
4.3 Inactivation Kinetics of UV Irradiation in Different Turbid
Wastewater
49
4.4 Properties of Wastewater Samples 52
vii
List of Figures
Figure Title Page
2.1 Structure of PN junction 4
2.2 UV LEDs (282 nm) with 9 chips 5
2.3 Distribution by continent of the 1.5 billion people using sewerage
facilities with no treatment in 2010
7
2.4a
2.4b
A full wastewater treatment process
A simplified wastewater treatment process
7
7
2.5 Application of UV LEDs in SWTP 9
2.6 Working conditions of WWTP in post disaster period in Japan 9
2.7 UV light in the electromagnetic spectrum 10
2.8
2.9
2.10
2.11
2.12
The spectra of low-pressure and medium pressure UV lamp
Spectra of UV LEDs
The disinfection mechanism of chemical disinfectant and UV
irradiation
Disinfection mechanism of UVB and UVC
UV disinfection and photoreactivation
11
12
12
12
13
2.13
2.14
The UV fluence (UV dose)-response curve of E. coli Absorbance of DNA to UV light with different wavelengths
15
16
2.15 The absorbance and scattering of UV light 16
2.16
2.17
The dose-response of different microorganisms
Three kinds of inactivation kinetics
18
18
2.18 The quasi-collimated beam apparatus 19
2.19
2.20
A typical UV inactivation-fluence (dose) response curve for B.
subtilis spores
The actinometry method based on ferrioxalate actinometer
21
24
3.1 Research framework 26
3.2 Experimental set-up 27
3.3 Quasi-collimated beam apparatus 28
3.4 Comparison between reactor and spectroradiometer 28
3.5 Procedures for UV fluence determination by iodide-iodate
actinometer
30
3.6
3.7
The procedure for UV fluence determination by ferrioxalate
actinometer
Procedure for making synthetic wastewater
34
36
3.8 Determination of montmorillonite added into the wastewater 36
3.9 The Process for making glycerol stock from E. coli strain 37
3.10 The Chromocult Colifrom agar and the colony 37
3.11 Series dilution 38
3.12 Plating of dilutions 39
3.13 Procedure for disinfection test with synthetic wastewater 39
3.14
3.15
3.16
Research framework for disinfection test with synthetic wastewater
AIT wastewater treatment process
MPN method
40
40
42
3.17
3.18
The procedure for conducting disinfection test with real wastewater
Research framework for disinfection test with real wastewater
43
43
4.1 The emission spectrum of UV LEDs 44
viii
4.2 The procedure for UV fluence determination 45
4.3 UV fluence determined by iodide-iodate vs. exposure time 45
4.4 UV fluence determined by ferrioxalate vs. exposure time 47
4.5 Turbidity of synthetic wastewater vs. concentration of
montmorillonite
48
4.6 Time-response curve of E. coli in different turbid wastewater 49
4.7 Sensitivity of E. coli to UV irradiation in different turbid synthetic
wastewater
50
4.8 Time-response curve of E. coli in synthetic wastewater of 27 and 70
NTU
50
4.9 Time-response curve of total coliform in different turbid wastewater 52
4.10 Time-response curve of total coliform within 100 s UV exposure 53
4.11 Time-response curve of E. coli in different turbid wastewater 54
ix
List of Abbreviations
ABS
APHA
ASFB
BOD
B. subtilis
c
CBD
CFU
COD
CT
DBP
DC
DNA
DW
E. coli
I
IEQ
J
LB agar
LED
LP mercury lamp
MDG
MP mercury lamp
MS2
NTU
PBS
PR
Q β phage
QCB
RNA
SEQ
SWTP
THM
Absorbance
American Public Health Association
Aerobic Spore-forming Bacteria
Biochemical Oxygen Demand
Bacillus Subtilis
Speed of Light
Collimated Beam Device
Colony Forming Unit
Chemical Oxygen Demand
Collection Tank
Disinfection Byproducts
Direct Current
Deoxyribonucleic Acid
Distilled Water
Escherichia coli
Intensity
Inlet of Equalization Tank
Joule
Luria-Bertani Agar
Light-emitting Diode
Low Pressure mercury lamp
Millenium Development Goal
Medium Pressure mercury lamp
Male-specific-2
Nephelometric Turbidity Unit
Phosphate Buffered Saline
Photoreactivation
F-specific RNA Bacteriaphage
Quasi-collimated Beam
Ribonucleic Acid
Surface of Equalization Tank
Simplified Wastewater Treatment Process
Trihalomethane
US EPA
USPHS
UV
UVDGM
United States Environment Protection Agency
United States Public Health Service
Ultraviolet
Ultraviolet Disinfection Guidance Manual
1
Chapter 1
Introduction
1.1 Background
Disinfection technology has been applied in water and wastewater treatment for a long
time. In the past decades, several kinds of disinfection methods have been developed, such
as chlorination, UV irradiation, ozonization, etc.
When compared to other methods, UV irradiation has many advantages, such as almost
zero formation of disinfection byproducts, more user-friendly and high disinfection
efficiency. Furthermore, the UV light can even inactivate some chlorine-resistant
microorganisms, such as Giardia and Cryptosporidium (Caron et al., 2007), which is also a
big concern in current water supply system. These advantages make UV irradiation
become a promising technology for water and wastewater disinfection.
Currently, most of the UV lamps are low- or medium-pressure mercury lamps (LP and MP
lamps) and they have some sustainability issues because they use toxic mercury to generate
the UV light and the corresponding lifetime ranges from 4,000 h to 10,000 h (Autin et al.,
2013). So the disposal of this kind of lamp may cause some environmental pollution after
using up. Another shortcoming of this kind of lamp is the low energy efficiency as most of
the electricity was converted to heat, not the desired UV light, which makes it less cost-
competitive than chlorination.
UV light-emitting diodes (LEDs) offer a possible solution to this problem. UV LEDs do
not contain toxic mercury, have a longer lifetime and their flexibility is also very high due
to the compacted structure. Last but not the least, UV LEDs has the potential to offer a
high energy efficiency (Bowker et al., 2011; Wurtele et al., 2011). It is essentially a PN
junction opto-semiconductor that can emit the light with a defined wavelength when the
electricity is applied on the semiconductor in a forward biased direction. The UV LEDs
lamp remains cool when it works, so only a small amount of energy is converted to heat,
which can improve its efficiency significantly as compared to a low- or medium-pressure
mercury lamp.
Due to these advantages, UV LEDs has attracted a lot of attentions from both industry and
academic institutions and becomes a hot spot in the UV disinfection field. Some
investigations have already been done on the application of UV LEDs for disinfection
purpose (Chevremont et al., 2012; Oguma et al., 2013; Wurtele et al., 2011). Most of them
just focus on the comparison of UV LEDs with different wavelengths or their
combinations and disinfection performances on different kinds of microorganisms. The
indicating microorganisms include Escherichia coli (E. coli), fecal coliform and Bacillus
Subtilis spore (B. subtilis spore), etc. Little information has been paid to the investigation
of disinfection performance of UV LED under turbid wastewater conditions.
Application of UV irradiation is still limited by its high cost and it is even less cost-
competitive than chlorination nowadays. So, it is just used to disinfect the secondary
effluent in wastewater treatment, whose turbidity is relative lower. But UV LEDs will
make UV irradiation become a cheap technology in the future because it has high energy
efficiency. When the wastewater has a high turbidity, the disinfection efficiency of UV
2
LEDs might be reduced. But it may be still more cost-competitive than chlorination. So, its
application to disinfect the turbid wastewater could be expected.
In this research, the disinfection performance of UV LED (280 nm) under turbid
wastewater condition has been investigated. Currently, there are several kinds of UV LEDs
available, and the wavelengths are 255 nm, 265 nm, 280 nm, and 310 nm, etc. Now UV
LED (280 nm) has a relative longer lifetime when compared with UV LEDs in deeper UV
zone (less than 280 nm) and it also has a highest disinfection efficiency for time-based
inactivation (Oguma et al., 2013).
This research consists of three stages and total coliform and E. coli were selected as the
indicating organism. At first stage, actinometry methods have been applied to determine
the UV fluence. Actinometry methods are based on the photochemical reaction and more
details are provided in literature review and methodology part.
At the second stage, UV LEDs have been applied to disinfect synthetic wastewater with
different turbidity. Montmorillonite has been added into the distilled water to synthetize
the turbid wastewater and it was thought to be a good representative of particles in the
wastewater due to its tendency to swell and its surface active properties (Passantino et al.,
2004). The turbidity of synthetic wastewater was dependent on the amount of
montmorillonite added into the distilled water.
At third stage, disinfection test has been conducted with the real wastewater from AIT
wastewater treatment plant. Wastewater samples were taken from different parts of the
wastewater treatment process at different time. The quality of each sample is also different.
1.2 Objectives of Study
The main objective of this study was to check the disinfection performance of UV LEDs
under high turbid wastewater conditions. The specific objectives of this study are:
1) To develop a reactor for conducting the disinfection test with UV LEDs;
2) To determine the UV fluence in the reactor through two different actinometers and
compare the results;
3) To check the disinfection performance of UV LEDs with both synthetic turbid
wastewater and real turbid wastewater.
1.3 Scope of Study
This research was limited to the experimental scale, rather than full scale, as the UV LEDs
has not been implemented in industry yet. The UV fluence has been determined by the
actinometry methods, and the secondary data in previous literatures have been applied for
calculation. As the turbidity of real wastewater may just have short variation, the
disinfection performance of UV LEDs under turbid water conditions may not be well
understood. So, synthetic turbid water will be applied to get turbidity with a larger
variation, and give an aid to better understand the impact of turbidity on the disinfection
performance of UV LEDs. Additionally, the quality of wastewater varies seasonally and
geographically. So the results of disinfection test with real wastewater are not reproducible.
But disinfection test with synthetic wastewater can overcome this shortcoming. So,
conducting disinfection test with both synthetic and real wastewater may be the best way to
3
conduct this research - disinfection test with synthetic wastewater can supply the
reproducible results while the disinfection test with real wastewater can supply some
results as reference in real case.
4
Chapter 2
Literature Review
2.1 Introduction
This chapter provides a background of relevant information that relates to the research
topic. The first section provides the fundamentals of ultraviolet light-emitting diodes (UV
LEDs), including fundamentals of UV LEDs, its development, potential applications and
the advantages of UV LEDs over the conventional UV lamp. The second section explains
why UV LEDs have the potential to be applied for turbid wastewater disinfection. This is
followed by the introduction of fundamentals of UV disinfection, including the
classification of UV, UV sources, UV disinfections mechanism, and the advantages and
disadvantages of UV disinfection. The next section presents some factors that could affect
the disinfection efficiency of UV, such as UV fluence, wavelength of UV light, the type of
microorganism, scattering and absorbance by particles. This is followed by a section
introducing the UV fluence determination. Some methods used to determine UV fluence
were introduced, including biodosimetry and actinometry.
2.2 Ultraviolet Light-emitting Diodes (UV LEDs)
2.2.1 Fundamentals of UV LEDs
Light-emitting diode (LED) is a kind of semiconductor. When the electricity is applied on
the semiconductor in a forward biased direction, it will emit light with a specific
wavelength. The principle behind this kind of light could be described by Figure 2.1. When
the voltage is applied on the pn conjunction, the hole from p-type and the electron from the
n-type will combine together and generating a photon. The wavelength of the photon is
dependent on energy difference the carriers (hole and electron) overcome in order to
combine.
Figure 2.1 Structure of PN junction
The history of LED could be dated back to 1907. H. J. Round of MarconiLabs discovered
that some inorganic substances can emit light if an electric voltage is impress on them.
Nowadays, LED technology has already been applied in many fields successfully, such as
television, trafic light, and optical measurement systems, etc.
5
The invention of UV LEDs was just several decades later after the invention of LED. But
its commercialization has taken a long time. At the early stage of its development, the
energy efficiency is very low and the output power also could not satisfy the needs of
applications, which is still a limitation for its implementation in industry nowadays.
2.2.2 Advantages and disadvantages of UV LEDs
UV LEDs has attracted a lot of attentions due to its advantages, including:
1. It is expected to have high energy efficiency. The wall plug efficiency of UV LEDs
were supposed to achieve 75% in 2020 (Autin et al., 2013).
2. Long lifetime. Its lifetime could be up to 100,000 hours.
3. Compacted structure. Figure 2.2 shows the picture of UV LEDs with 9 chips and it
is quite robust.
4. No warm-up time. This makes it become very suitable for the frequent on-off
cycling system.
5. Lower voltage.
6. It could emit light with any wavelength ranges from 200 nm to 400 nm.
7. More environmentally friendly. They are made from aluminum nitride (AlN) or
gallium and aluminum nitride (AlGaN) and do not contain mercury, so it won’t
cause any disposal problems.
8. Less space requirement.
Figure 2.2 UV LEDs (282 nm) with 9 chips
However, UV LEDs still cannot be implemented in the industry due to its disadvantages,
including:
1. Low output optical power, especially in deep UV zone. But the increase of its
output power could be clearly expected through its physical improvement in a few
years. Nowadays, the output power of one UV LED could be several hundred
milliwatt (SeoulOptodevice, 2013). However, in 2011, an output power of 66 mW
can be regarded as leading light output (Hayward, 2013).
2. The price of UV LEDs is very high. Because now the technology of making UV
LEDs is still not mature, so its price is very high. The UV LEDs shown in Figure
2.2 costs about 1,000 $.
6
2.2.3 Current status and applications
UV LEDs could be applied in many areas, such as water disinfection, wastewater
disinfection, and UV curving, etc. Now the technology of making UV LEDs is still not
mature so it has not been applied in industry yet. Almost all the published papers on UV
LEDs are about research results at lab-scale, not the full-scale (Autin et al., 2013;
Chevremont, Farnet, Coulomb, et al., 2012; Hayward, 2013; Oguma et al., 2013; Wurtele
et al., 2011).
Now the lifetime of UV LEDs also varies. The lifetime of UV LEDs from different
companies are quite different due to different technologies applied to make UV LEDs.
Before its implementation in industry, it still has a long way to go. Undoubtedly, some
investigations on the application of UV LEDs can promote its implementation in industry.
2.3 Turbid Wastewater Disinfection by Ultraviolet Light
Generally, UV disinfection is just applied to disinfect the clean water. When turbidity of
water is higher than 2 NTU, a filter should be applied before UV irradiation. But in the
future, UV may be applied to disinfect the turbid water when the cost of UV irradiation
comes down.
One good example of the application change due to development of technology is
membrane. In the past, membrane technology was just used to filter the clean water
because it is a costly technology. But when the price of membrane comes down, it was
widely applied to filter the wastewater directly.
Following information explain the reason why UV LEDs has the potential to disinfect the
turbid wastewater.
2.3.1 Application of simplified wastewater treatment process in developing countries
Supplying improved sanitation and access to safe drinking water for human being has been
a goal for many international organizations and governments for a long time. A lot of work
has also been done in order to achieve this goal. The information from the United Nations
Millenium Development Goal (MDG) show that the target to reduce the proportion of
people without access to improved source of water by half was achieved, five years ahead
of the schedule. However, another alarming fact is that 2.5 billion people still have no
access to the improved sanitation facilities. On the way to reach the MDG, human being
still has a long way to go.
Currently, the cost of wastewater treatment is still quite high, and many cities or
communities in developing countries still cannot afford wastewater treatment. One
research reveals that 1.5 billion people used sewage connections without treatment (Baum
et al., 2013). The distribution of this number in each continent is illustrated in Figure 2.3.
From this figure, it can be found that Asia accounts for most of them, and the percentage is
as high as 68.5%.
7
Figure 2.3 Distribution by continent of the 1.5 billion people using sewerage facilities
with no treatment in 2010 (Baum et al., 2013)
Another research result also reveals that for towns in developing countries with a
population from 2,000 to 50,000 often falls into what has been termed as the management
gap: they are large and compact enough to have the centralized sanitation system, but they
are not large enough to have the resources to manage these highly mechanized
infrastructures (Pilgrim et al., 2008). In many developing countries, the sanitation facilities
often fail prematurely due to lack of maintenance and proper operations. This may be
caused several reasons, such as high initial investment, low initial revenues, and the
complex operation of WWTP.
Under such conditions, the full wastewater treatment process (Figure 2.4 a) may be not the
best choice for the developing countries as it is very costly and very complex to operate.
Some communities prefer to use a simplified wastewater treatment process (SWTP) to treat
the wastewater and then discharge it. This simplified wastewater treatment process just
consists of a screen and a tank for settling (Figure 2.4 b).
Figure 2.4 a A full wastewater treatment process
Figure 2.4 b A simplified wastewater treatment process
4%
68%
12%
5% 11%
0.10 Africa
Asia
Europe
North and Central America
South America
Oceania
Screen 1st clarifier Aeration 2
nd clarifier Disinfection
8
The simplified wastewater treatment process can remove most of suspended particles in the
wastewater. But the removal of dissolved pollutants may be very poor. So, the
concentration of pollutants (such as COD, BOD, N, and P) in the effluent may be still quite
high and they may cause pollution to the receiving water body.
Rather than these organic and inorganic pollutants, more attentions should be paid to the
pathogens remained in the effluent. For example, the COD of sewage ranges from 100 –
400 mg/L, and it could meet the standard for discharge after around 20-folds dilution.
However, the required dilution factor for fecal coliform in sewage could be as high as 1000
times. Furthermore, most of the pollutants contained in the wastewater are biodegradable,
while the pathogens may be quite persistent in the environment. So, the risk would be
reduced greatly if this kind of wastewater could be disinfected before discharge.
Table 2.1 Required Dilution Factors for Both E. coli and COD to Reach a Safe Level
Parameters Concentration Standard for
discharge
Required dilution
factors
COD (mg/L) 100 - 400 20-60 5-20
E. coli (CFU/100 mL) 106 - 10
8 100-1000 10
5 - 10
6
UV irradiation may be a better choice than chlorination in this simplified process because
the disinfection efficiency of chlorination could be affected by the pH, temperature, and
many components in the wastewater. It can also lead to the formation of disinfection
byproducts, which is carcinogenic. But UV irradiation is a physical method and it won’t
cause the formation of disinfection byproduct. Furthermore, it is relative easier to operate
when compared with chlorination.
But the cost may be a limitation factor for the application of UV irradiation. Now low
pressure UV lamp and medium pressure UV lamp are the most common UV lamps applied
in wastewater treatment field. They consume a large amount of energy and needs frequent
replacement. Now the UV irradiation is even less cost-competitive than chlorination.
Table 2.2 Comparison between UV LEDs and Conventional UV Lamp
UV LEDs Conventional UV lamp
Energy saving
Long lifetime
No mercury
High energy consumption and high cost
Frequent replacement
Contains mercury
But UV LEDs could overcome this shortcoming because it is expected to have a higher
energy efficiency and lower cost in the future. The comparison between UV LEDs and
conventional UV lamp is shown in Table 2.2. It can be concluded from this table that
conventional UV lamp will be hopefully replaced by UV LEDs. So, the application of UV
LEDs in SWTP is possible in the future (Figure 2.5).
9
Figure 2.5 Application of UV LEDs in SWTP
However, one thing must be kept in mind is that the effluent from the sedimentation tank
may still have a high turbidity. So, the impact of high turbidity in the primary effluent on
the disinfection performation of UV LEDs should be well known.
2.3.2 Another possible application of UV LEDs for turbid water disinfection
This simplified process may not just be used to in these cases. During post-disaster period,
it may be adopted as the disinfection method in temporary wastewater treatment. It has
been reported that some WWTPs in Japan were stopped by the earthquake and tsunami in
2011 because electricity was shut down after the earthquake and some infrastructures and
equipments were also damaged by the tsunami (Masaru et al., 2013). However, the
wastewater continues coming to the wastewater treatment plant as the every-day-life was
going on. So, the wastewater cannot be treated by the full treatment process. Under such
conditions, a large amount of wastewater was discharged to the adjacent water body after
the simple sedimentation in the primary sedimentation tank (Figure 2.6). Disinfection
before discharge can reduce the concentration of pathogens in the effluent, thus protect
public health.
Figure 2.6 Working conditions of WWTP in post disaster period in Japan
During post-disaster period, chlorination is not a suitable disinfection method because the
road is damaged by the earthquake and chemical reagent (such as Cl2, NaClO) cannot be
transported to the WWTPs. Ozonation is also not practical because its operation is quite
complex and consumes a large amount of electricity. UV irradiation may be the best choice
because it is relative easier to operate and the electricity can be supplied by temporary
electricity generator.
Based on above information, it can be expected that UV LEDs will be applied to disinfect
the turbid wastewater in the future due to its merits, such as less energy consumption, long
lifetime, and environmental friendly.
Stop working after the earthquake and tsunami Discharge
Disinfection by
UV LEDs
10
2.4 UV Disinfection
2.4.1 Fundamentals
According to the physical definition, the electromagnetic radiation with a wavelength
between 100 nm and 400 nm is called UV light (Figure 2.7). It is a nonvisible light. Due to
the different biological effects associated with different wavelengths of the UV light, it is
further classfied into 4 types: vacuum UV (100-200 nm), UV-C (200-280 nm), UV-B
(280-315 nm) and UV-A (315-400 nm).
Figure 2.7 UV light in the electromagnetic spectrum (courtesy of US EPA, 2006)
As it is known, the energy associated with a photon is inversely proportional to the
wavelegnth of light. This could be described by the following expression:
Equation 2.1
Where u is the energy (J) of a photon, h is the Planck constant (the value is 6.626×10-34
J∙s),
c is the speed of light (the value is 2.998 × 108 m/s) in the vacuum, and λ is the wavelength
(m) of light. Based on this, it can be easily concluded that the UV-A light is less energetic
than the UV-B and UV-C.
All kinds of UV lights are harmful to human being’s skin. UV-A can cause the tanning of
skin while the UV-B could cause the skin to burn and is known to eventually cause the
skin cancer. UV-C is more powerful than UV-A and UV-B and it can be absorbed by
proteins, ribonucleic acid (RNA) and deoxyribobucleic acid (DNA), then cause the cell
mutations and/or cell death. It has the highest germicidal ability. Vacuum UV is even more
powerful than UV-C. However, as it can be easily absorbed by the air and water, so it is
seldom used for disinfection purpose.
2.4.2 Sources of UV
UV light source can be classified into two sources: artificial and natural. The natural UV is
mainly from the sunshine. UV light accounts for 3% of the total sunshine on the surface of
the earth. Only a part of the UV light emitted by the sun could reach the surface of the
earth and most of them are UV-A and UV-B because the UV-C is completely absorbed by
the ozone layer and atmosphere. The artificial UV source includes many kinds of UV
lamps, such as low-pressure (LP) mercury vapor lamp, medium-pressure (MP) mercury
11
vapor lamp, UV lasers, and light emitting diodes (LED), etc. They use different materials
to emit the UV light, such as Argon, Xenon, and mercury vapor, etc. The mercury type UV
lamp is the one which is most commonly used in the water and wastewater treatment. The
mechanism of this kind of lamp is that when the electric current passes through the
mercury vapor, the mercury atoms are excited by the collisions with the electrons flowing
between the electrodes. The excited electrons returns to the particular electronic states in
the mercury atom and in doing so the electrons will release the energy they have absorbed
in the form of UV light.
Figure 2.8 The spectra of low-pressure and medium pressure UV lamp (Emperor
Aquatics, 2013)
Based on the vapor pressure of mercury, the mercury UV lamp can be further classified
into 3 types: low pressure mercury lamp, medium pressure mercury lamp, and high
pressure mercury lamp. High pressure mercury lamp is mainly used for emitting the visible
light after a modification rather than generating the UV light. The spectras of LP and MP
UV lamp were shown in Figure 2.8. For low pressure mercury lamp, the mercury vapor
pressure ranges from 0.14 to 14 Pa and the temperature of the mercury is 40 ℃ and it just
produces the monochromatic UV light with wavelength of 254 nm. For medium pressure
mercury lamp, the mercury vapor pressure is much higher (from 14 kPa to 1400 kPa) and
the operating temperature is also much higher (600-900 ℃) (USEPA, 2006).
UV LEDs is a new kind of UV source, whose spectra is shown in Figure 2.9. The
mechanism behind this kind of UV light is totally differently from the conventional
mercury lamp. It is a kind of semiconductor that when electricity is applied on it in a
forward direction, it can emit UV lights with a specific wavelength. However, at this stage
of development, the output power of UV LEDs is still very low and the cost is also quite
high, so it has not been fully implemented in industry yet.
12
Figure 2.9 Spectra of UV LEDs (DOWA, 2013)
2.4.3 UV disinfection mechanism and the followed photoreactivation
UV disinfection mechanism
Unlike the chemical disinfectants (such as chlorine and ozone), who kill microorganisms
mainly through damaging cell wall and some intracellular moleculesm (Figure 2.10), UV
light inactivates microorganism in a totally different way.
Figure 2.10 The disinfection mechanism of chemical disinfectant and UV irradiation
The disinfection mechanism varies based on the wavelength of UV light. For UV light
with shorter wavelength (UVB and UVC), the disinfection mechanism is mainly about the
formation of cis-syn cyclobutane pyrimidine dimers in the genome DNA or RNA of
organism. This result in the genetic disorder and then the replication, transcription and
reproduction process are stopped, and eventually lead to inactivation of microorganisms
(Figure 2.11). In this process, the photoproducts are also produced, but it is much less
important than the damage caused to nucleic acid (Oguma et al., 2002).
Figure 2.11 Disinfection mechanism of UVB and UVC
13
However, the principle behind the UVA disinfection is totally different. UVA is known to
inactivate the microorganism mainly through exciting the photosensitive molecules, such
as , H2O2, and , which can damage the genome and other intracellular modecules
and cause lethal or sublethal effects, such as mutation and growth delay. UVA LED has
already been tested as water disinfection technology by some researchers (Hamamoto et al.,
2007).
Photoreactivation
The past research results show that some microorganisms can repair the UV-damaged
DNA or RNA through many ways (Oguma et al., 2002, 2004, 2005; Zimmer et al., 2002).
But only the photoreactivation is discussed here because it can impair the UV disinfection
efficiency significantly within several hours after UV disinfection test. Other repair
mechanism, which is referred as dark repair, is less important than photoreactivation. But
in real wastewater or water treatment, photoreactivation was thought to be not as serious as
in lab because the water or wastewater will be discharged to the river or into the pipe, so
the visible light is isolated or reduced greatly. Without the energy from the light,
photolyase was unable to repair the dimmer in DNA. But attentions should be paid to the
photoreactivation if the disinfected water will be exposed to the room light.
Photoreactivation process is a totally inverse process of disinfection. The relationship
between the disinfection and photoreactivation (PR) could be described by Figure 2.12.
When DNA is exposed to the UV light, it can lead to the formation of dimmer, while under
white light condition this damage will be repaired by a kind of enzyme called photolyase.
The specific wavelength of light that can cause photoreactivation ranges from 310 – 480
nm. This means that UVA can disinfect the microorganism and cause photoreactivation
simultaneously.
Figure 2.12 UV disinfection and photoreactivation
The UV light with different wavelength can suppress the photoreactivation of
microorganism differently. Medium-pressure UV lamp was thought to be able to suppress
the photoreactivation of E. coli more than the monochromatic low-pressure UV lamp and
the UV light with broad spectra was also thought to be more effective on repressing
photoreactivation of E. coli. This suggests that medium-pressure UV lamp can offer an
advantage over low-pressure UV lamp in drinking water or wastewater treatment. This
T T
UV disinfection PR
DNA
DNA with dimer
T T
photolyase
White light
14
kind of effect might be attributable to the suppressing effect on photolyase caused by UV
light with relative longer wavelength. Someone even thought that the UV light at around
280 nm can repress the photolyase most effectively (Hu et al., 2008).
Additionally, high salinity was also found to be able to suppress the photoreactivation of
E. coli after UV irradiation. The NaCl solution at 2.4% or above (in weight/volume) can
suppress the photoreactivation of E. coli after UV irradiation significantly, but the NaCl
solution at 1.9% or lower did not shown such effect (Oguma. et al., 2013). This suggests
that the photoreactivation of E. coli potentially may occur in brackish and costal area
where the salinity is rather low.
For different microorganisms, the photoreactivation effects are also different. The
photoreactivation of Legionella pneumophila (L. pneumophila) after both low-pressure and
medium-pressure UV lamp are almost the same, which is quite different from that of E.
coli. This suggests that the E. coli cannot correctly indicate the fate of L. pneumophila in
UV disinfection system (Oguma et al., 2004).
2.4.4 Advantages and disadvantages of UV disinfection
UV disinfection has many advantages over other chemical disinfections methods. One
major advantage is that it can inactivate some chlorine-resistant pathogens, such as Giardia
cysts and Cryptosporidium oocysts (Craik et al., 2000; Craik et al., 2001). At the same time,
the chemical disinfection methods may change the water quality. For instance, chlorination
may lead to the formation of disinfection byproducts, such as trihalomethane (THM) while
UV disinfection won’t has such a problem. Low-and medium-pressure mercury UV lamps
did not have a significant impact on the formation of DBP when the dose is less than 500
mJ/cm2 (Liu et al., 2002). The recommended UV dose for disinfection in drinking water
treatment plant is 40 mJ/cm2, which is well below 500 mJ/cm
2 (Mosher et al., 2012).So
UV disinfection won’t cause the formation of DBP. Furthermore, the contact time for UV
disinfection is very short (generally a few seconds) so the space requirement of the
disinfection device is lower.
However, UV disinfection may also have some shortcomings. Unlike the chemical
disinfectants, UV won’t leave any residuals after disinfection in drinking water treatment,
so it cannot prevent the reproduction of microorganisms in the water after disinfection. In
practice, the UV disinfection is often combined with the chemical disinfection, such as
chlorination and ozonization. Additionally, UV disinfection is very sensitive to the
turbidity of water. High turbidity could reduce the disinfection efficiency significantly and
cause the scaling problem of the lamp envelope at the same time. The scaling problem
could be solved through the regular cleaning of lamp envelope. Last but not the least, some
microorganisms could be reactivated after the UV disinfection. The mechanisms include
photorepair and darkrepair. The reason why the microorganism could be reactivated is that
some enzyme system of the microorganism could repair the DNA or RNA damaged by the
UV light (USEPA, 2006).
2.5 Factors Affecting the Disinfection Efficiency of UV Light
The disinfection efficiency of UV light could be affected by many factors, such as UV
fluence, color, wavelength, particles in water, type of microorganism, fluence rate,
15
temperature and pH of the water. These factors can impact the disinfection performance of
UV light in different ways.
2.5.1 Subordinate factors
Generally, the impacts of UV intensity, temperature and pH of the water on the
disinfection performance of UV were thought to be negligible. When the UV intensity
ranges from 1mW/cm2 to 200mW/cm
2, the UV dose-response of microorganism follows
the Law of Reciprocity. The UV light with a low intensity can have the same disinfection
performance with a highed intense UV light when the UV dose is the same. The UV dose
required for a given log reduction of E. coli, Candida parapsilosis, and f2 bacteriaphage
increased slightly as the temperature decreased (Severin et al., 1983). pH was also thought
to be a negligible factor for UV disinfection because it can impact neither the transmission
of UV light in the water nor the intensity of UV light. Generally the pH of wastewater or
drinking water is around 7, and it does not have a significant impact on UV disinfection.
2.5.2 UV fluence
Both UV fluence and UV dose are used in the UV disinfection literature. But UV fluence
is the appropriate term for UV disinfection. UV dose means the energy absorbed by the
microorganism. In the case of microorganisms, almost all incident UV light passes through
the organism with only a few percent being absorbed. So, UV fluence is a more
appropriate term (Bolton et al., 2003).
It is the main factor that affecting the disinfection efficiency of UV light. UV fluence is
defined as the product of fluence rate (mW/cm2) and the exposure time (s). In North
America, the unit for UV fluence is mJ/cm2 while in Europe people prefer to use J/m
2.
Generally, the higher log-reduction could be achieved with the higher UV fluence. A UV
fluence-inactivation response curve of E. coli was shown in Figure 2.13. It can be found
from this figure that the log-reduction of E. coli was different at different UV fluence.
Figure 2.13 The UV fluence (UV dose)-response curve of E. coli (USEPA, 2006)
2.5.3 Wavelength
Wavelength is another important factor that could impact the disinfection efficiency of UV
irradiation. The UV light inactivates the microorganisms mainly through destroying the
DNA or RNA of the microorganism. The DNA absorbs the UV with a wavelength from
200 nm to 300 nm and tend to have a peak at around 260 nm, which is shown in
Figure 2.13 (USEPA, 2006). Microorganisms are most sensitive to the UV light with a
16
wavelength of 260 nm. So, it means that the UV light with a wavelength of 260 has the
highest germicidal ability. However, some kinds of virus may be more sensitive to the UV
light with a wavelength below 230 nm (Linden et al., 2001).
In the past, the most common UV lamps applied in water and wastewater engineering are
monochromatic low-pressure UV lamp and polychromatic medium pressure UV lamp, so
only 254 nm UV light and a polyspectra UV light are available in practical engineering.
But now, due to the development of light-emitting diodes, more and more monochromatic
UV lamps with different wavelength become available. Some investigations on the
disinfection performance of UV LEDs with different wavelength have already been done
(Chevremont, Farnet, Sergent, et al., 2012; Hamamoto et al., 2007; Oguma et al., 2013;
Wurtele et al., 2011). It can be expected that UV LED with a desired wavelength can be
used for the disinfection of a specific kind of wastewater, especially when the
microorganism requries the UV light with a specific wavelength.
Figure 2.14 Absorbance of DNA to UV light with different wavelengths
2.5.4 Absorbance and scattering by particles in water
UV disinfection can be impacted by the particles in the wastewater significantly. This is
done mainly through two mechanisms: scattering and absorbance by particles. The light
scattered by the particles can still inactivate the pathogens in the water, while the light
absorbed by particles was not able to. Additionally, the particles can also protect the
bacterial cells through shielding effects. When the particles’ size is big enough, the
bacterial cells can harbor inside the particles so that UV cannot inactivate them.
Another fact that should also be noticed is that the water can also absorb the UV energy.
The distilled water can absorb 8% of the UV light energy at a depth of 3 cm. When some
solids dissolved in the water, it may also contribute to this kind of blocking effect.
Figure 2.15 The absorbance and scattering of UV light
17
Particles in the wastewater can be measured in two ways: total solids (TS) and turbidity.
TS could be further classified into two types: total dissolved solids (TDS) and total
suspended solids (TSS). Both TS and turbidity can be used to monitor the wastewater
quality in wastewater treatment. TS is the measure of weight of particles in the wastewater,
so it can reflects the amount of particles in the wastewater exactly, regardless of its
properties and size. But it takes a long time to measure as the procedure is more complex
than that of turbidity. Turbidity is relative easier to monitor, but it cannot reflect the
amount of particles in the wastewater exactly, because it can be greatly impacted by the
surface properties of particle, particle size, etc. For example, 50 mg/L kaolin clay give a
turbidity reading of about 80 NTU, while 50 mg/L humic acid give a turbidity reading
slightly greater than 3 NTU. But turbidity can be monitored automatically, so sometimes
the technician prefers to use turbidity to describe the wastewater quality.
Particles with different size can impact the disinfection performance of UV light through
different mechanisms. Goethite particles (0.2 μm × 2 μm) can attach to the surface of
E. coli (0.5 μm × (1-3) μm), then protect the bacterial cell from UV disinfection even at a
low turbidity level (1 – 5 NTU). Some research results suggest that the coliform shielding
effect is mainly attributable to the particles with a diameter of 7-10 μm (Jolis et al., 2001;
Qualls et al., 1983). But for virus, such as MS2 coliphage (diameter = 23 nm) and
bacteriophage T4 (90 nm × (25- 200) nm), the particles with a diameter < 2 μm is large
enough to provide the shielding effect (Templeton et al., 2005).
The properties of particles can also affect the disinfection performance of UV significantly.
Particles with different composition can protect the bacteria cell differently. For example,
the humic acid can be linked with the E. coli through an affinity effect (Cantwell et al.,
2008). The humic acid can attach on the surface of E. coli and then protect them from UV
disinfection. The same effect was observed on virus. Humic acid and activated sludge floc
particles were found to be very effective on shielding the virus from UV disinfection,
while the kaolin clay provide no significant protection (Templeton et al., 2005). This can
be explained through the property of humic acid. The organic compounds contained in
humic acid can absorb the UV light, so UV light was absorbed before reaching the bacteria.
But kaolin clay does not absorb the UV heavily relative to humic substances(Bitton et al.,
1972).
To sum it up, two factors should be taken into consideration when investigating the impact
of particles on the disinfection performance: (1) the size of particles and target
microorganisms; (2) the interaction between the particles and target microorganisms. In
practice, the size and composition of particles varies seasonally and geographically, so it is
impossible to synthetize a turbid wastewater that can represent all kinds of wastewater. No
matter what kind of particles (such as kaolin clay, montmorillonite, humic acid) have been
used to synthetize the turbid wastewater, it still has some limitations to reflect the practical
conditions. But the benefit of using synthetic wastewater is that the results are reproducible,
which is an important factor to get a convincible research result. But if natural water and
wastewater sample are used, it would be impossible to carry out controlled and
reproducible experiments (Kollu et al., 2012). So, conducting disinfection test with
synthetic wastewater is still a reliable and popular way to investigate the impact of
turbidity on UV disinfection performance.
18
2.5.5 The types of microorganisms
Different type of microorganism could show a different fluence-inactivation response. This
means that the inactivation efficiency of UV may be also impacted by the type of
microorganism. Figure 2.15 shows the fluence-inactivation response of different
microorganisms (USEPA, 2006). This figure is quite old, so the fluence is till called dose
in this figure. It can be found that the E. coli is more sensitive than other species of
microorganisms and B. subtilis is most UV-resistant.
Figure 2.16 The dose-response of different microorganisms
The fluence-inactivation response of microorganisms is generally described through the
inactivation kinetics, which is a first-order model. It is the same with that of chemical
disinfection. This model could be described by the following equation:
(
) Equation 2.2
Where N and N0 are the concentrations of microorganism after and before disinfection,
respectively. Some researcher found that this model cannot be used to describe the
inactivation kinetics of microorganisms under all conditions. Later, two other models were
also developed, which are called shoulder model and tailing model (Figure 2.16). The
causes of tailing are still a matter of debate, but attachment of bacteria to particles was
supposed to be one of them. Shouldering was hypothesized to be due to the formation of
microorganism aggregates, photoreactivation or dark repair.
Figure 2.17 Three kinds of inactivation kinetics
19
2.6 Methods for UV Fluence Determination
Currently, three methods are available for the determination of UV fluence. They are: 1)
biodosimetry; 2) chemical actinometry; 3) mathematical model; 4) new validation method.
Among them, biodosimetry method and chemical actinometry method are more suitable
for bench scale, while the mathematical model method and the new validation method are
more widely used in practical engineering. Here, the chemical actinometry method has
been applied to determine the UV fluence in the reactor for exposure experiment.
2.6.1 Biodosimetry method
Biodosimetry method is based on the fluence-inactivation response of some
microorganisms. A quasi-collimated beam (QCB) bench scale apparatus must be used in
this kind of method to get the standard fluence-inactivation response curve. Then this curve
could be used as a reference for the UV fluence in other reactors.
This kind of method has three steps:
(1) Calculating the UV fluence in a QCB apparatus.
The QCB apparatus are schematically described in Figure 2.16. The UV lamp is put in
an enclosure. A long tube is connected with the enclosure, which is used to collimate
the UV light. So, on the surface of solution in the petri dish, the intensity of UV light is
quite uniform and could be measured by the spectroradiometer.
Figure 2.18 The quasi-collimated beam apparatus
For low-pressure UV lamp, it has a monochromatic emission and the average germicidal
fluence rate Eavg could be calculated by following equation (Bolton et al., 2003).
Equation 2.3
UV lamp
Lamp enclosure
Collimating tube UV light
Petri dish
Magnetic stirrer
20
Where:
1) E0 is the UV fluence rate measured by the spectrophotometer at the center of the
petri dish.
2) Petri factor is the ratio of the average of the incident irradiane over the area of the
petri dish to the irradiance at the center of the dish and is used to correct the
irradiance reading at the center of the petri dish to more accurately reflect the
average incident fluence rate over the surface area.
3) When UV light pass from one medium to another medium, a small part of the
light will be reflected off the interface between the two media. For UV light
between 200-300 nm, the reflection factor could be adopted as 0.975.
4) Water factor is used to correct the errors caused by the adsorption of water to UV
light. It is proportional to the depth of water. This factor could be calculated by
following equation.
Equation 2.4
a is the adsorbance for a 1 cm length path. l is the vertical path length (cm) in the
petri dish. For the MP UV lamp, the calculation for water factor is more complex
because the water absorbance to the UV light with different wavelength is
different. So, the correction must be made over a narrow band (1-5 nm) of
wavelength.
5) The divergence factor is used to correct the error caused by the unperfect
collimation by the beam. For finite distance between the petri dish and the UV
lamp, the irradiance falls off as the inverse square of the distance L from the UV
lamp to the surface of the cell suspension. Assume the irradiance of UV light at
distance L is I0. So the irradiance at L+X could be expressed as:
Equation 2.5
Divergence factor is the average of this function over the path length l of the cell
suspension and it could be expressed by following equation:
Equation 2.6
Theoretically, the divergence and the water absorbance should be considered
together to correct the error happened during the propagation of UV light. However,
for path length less than 5 cm, the errors involved in treating them separately are
negligible.
The UV fluence on the surface of petri dish could be calculated by the following
equation.
( ) Equation 2.7
For the medium pressure UV lamp, the determination of the UV fluence will be
much more complex because it has a polychromoatic emission and the germicidal
ability of UV light at different wavelength is different. So, only low pressure UV
lamp is used in biodosimetry.
21
(2) Plotting standard inactivation-fluence response curve of challenge microorganism
After getting the fluence rate above the surface of the solution in the petri dish, some
kinds of solution contains challenging microorganism will be used to do the
exposure test with the QCB. The challenge microorganism include F-specific RNA
bacteriophage Q β (Qβ phage), Bacillus Subtilis spore (BS spore), and MS2
coliphage, etc. Calculate the log-reduction of the challenge microorganism and the
corresponding UV fluence. Then plot the standard inactivation-fluence response
curve. A typical standard curve is shown in Figure 2.17. The corresponding equation
could also be obtained by doing the linear regression with the standard curve. This
equation could be expressed as following:
(
) Equation 2.8
(
) is the log reduction of microorganism. This parameter could be calculated
through the concentration of microorganism before (N0) and after (N) the exposure
test, which is relatively easy to obtain. Both k and b are the constant, which is
determined by the nature of microorganisms selected.
Figure 2.19 A typical UV inactivation-fluence (dose) response curve for B.
subtilis spores (Qualls et al., 1983)
(3) The last step of this method is conducting the UV exposure test with other reactors.
Repeat the same procedure in last step to calculate the log-reduction of challenge
microorganism. Then the corresponding UV fluence can be infered from the standard
curve or the corresponding equation. The UV fluence obtained here is called reduction
equivalent fluence.
However, the bioassay method is quite time-consuming and relatively expensive. To some
extent, this kind of method is not suitable for on-site measurement of UV. Furthermore,
this kind of method only provides the mean fluence in bench scale. In practice, the best
way for UV fluence determination is to combine the mathematical model with the
actinometry method.
2.6.2 Chemical actinometry
Chemical actinometry is another well-developed method for UV fluence measurement.
This method is based on the photochemistry. In this kind of method, some kinds of light-
22
sensitive chemicals are used to absorb the energy from UV light, including KI/KIO3
solution, uridine and potassium ferrioxalate, which are called actinometers. In this study,
two kinds of actinometers have been applied to determine the UV fluence. One is KI/KIO3
actinometer, and another one is potassium ferrioxalate actinometer.
(1) KI/KIO3 actinometer
The composition of KI/KIO3 actinometer is 0.6 M KI, 0.1 M KIO3 and 0.01 M Na2B4O7.
The role of Na2B4O7 in this actinometer solution is to maintain the solution has a constant
pH of 9.2, so the solution won’t turn to acid condition, which can lead to the oxidation of I-.
The principle behind this method is a photochemical reaction, which is shown as below:
8 I + IO3
+ 3 H2O + h 3 I3
+ 6 OH
Equation 2.9
In this chemical reaction, the number of I3- formed (N) has a linear relationship with the
number of photons absorbed by actinometer solution (P). This relationship could be
expressed by following formula:
N = P × Φ Equation 2.10
Where Φ is the quantum yield, moles I3-/mole photon.
The number of I3- formed in this reaction could be inferred from the increase of its
concentration (∆C, mole/L).
N = ∆C × V Equation 2.11
Where V is the volume of actinometer solution, L.
The concentration of I3 can be determined spectrophotometrically, and it is proportional to
increase of absorbance of actinometer solution at 352 nm, so:
∆C = ∆ ABS/ε Equation 2.12
Where ε is the adsorption coefficient, L ∙ mole-1
;
∆ ABS= ABS(before UV exposure) – ABS (after UV exposure).
Energy contained in UV light (E) is the product of moles of photon and the photon energy
(U), so:
E = P (moles) × U (J/mole) Equation 2.13
Assume that the exposure area is A and the exposure time is t, so the formula for the UV
fluence calculation could be expressed as following:
Equation
2.14
And the irradiance (I) could be expressed as:
23
Equaton 2.15
When using this to calculate the irradiance of UV lamp, the reflection of UV light by water
should also be taken into consideration. Assume the reflection factor is α, so the irradiance
of UV lamp (IL) is:
IL
Equaton 2.16
For the reflection between air and water, the correction factor is 0.975.
The procedure for conducting this experiment could be described as following:
1. Measure the absorbance of actinometer solution at 352 nm (ABS(blank));
2. Calculate the exposure area (A);
3. Add actinometer solution into the reactor, note the volume of actinometer as V;
4. Turn on the UV lamp and let UV irradiate to the actinometer solution;
5. After a period of time, turn off the UV lamp, note the exposure time as t;
6. Measure the absorbance of actinometer solution again at 352 nm (ABS(sample));
7. Calculate the UV fluence by Equation 2.14.
The procedure for preparation of actinometer solution and the selection of parameters for
calculation is detailedly introduced in methodology part.
This actinometer solution has a numerous advantages. Firstly, it is optically opaque to the
light with a wavelength shorter than 290 nm. So, all the UV light within the germicidal
range could be absorbed. It can be used as the photon counter to measure the UV fluence.
Secondly, KI/KIO3 solution is optically blind to the light with a wavelength longer than
330 nm, which means that it won’t absorb the room light. So, the fluence determination
experiment can be conducted in the presence of room light. Thirdly, all the chemicals used
in this kind of method are commercially available.
However, this kind of method also has some shortcomings. For example, KI is not easy to
store in the lab because it can absorb the moisture and then decompose to KOH and I2.
Furthermore, due to the slowly thermal reaction happened in this solution, it can just be
stored for 4 hours. So, the fresh solution needs to be made frequently. Last but not the least,
the quantum yield of KI/KIO3 at 282 nm still has not been accurately defined. Different
results will be obtained by adopting different quantum yield.
(2) Potassium ferrioxalate actinometer
Potassium ferrioxalate is another kind of actinometer that has been widely accepted as the
standard actinometer for UV fluence determination. Involved photochemical reaction can
be expressed as follows:
Equation 2.17
After exposure of a ferrioxalate solution to UV light, the will be converted to
(Figure 2.18). The generated could be determined through colorimetric method in
which complexed with o-phenanthroline.
24
Figure 2.20 The actinometry method based on ferrioxalate actinometer
The principle behind this kind of method is almost the same with that of KI/KIO3
actinometer. The UV fluence is proportional to the increase of concentration of Fe2+
formed during the UV exposure. And the concentration of Fe2+
could be determined
spectrophotochemically at 510 nm. So, the UV fluence could be inferred from the change
of absorbance of actinometer solution. Following is an example for this calculation.
Assume that the absorbance of actinometer solution before UV exposure is ABS510(blank)
and the absorbance of actinometer solution after UV exposure is ABS510nm(Sample). So the
absorbance of actinometer solution has increased by:
∆ABS= ABS510nm(Sample) - ABS510(blank) Equation 2.18
So the concentration of Fe2+
has increased by:
∆C(Fe2+
) = ∆ABS ∙ ε Equation 2.19
Where ε is the adsorption coefficient, mole-1
∙ cm-1
;
Then the total moles of Fe2+
formed during this chemical reaction are:
N = ∆C(Fe2+
) ∙ V(sample) Equation 2.20
Where V(sample) is the volume of sample exposed to the UV LEDs.
Assume that the quantum yield of this actinometer is Φ (moles I3-/mole photon.), so the
moles of photon (P) is:
Equation 2.21
Assume the photon energy is U (J/mole) and the exposure area is A, so the UV fluence (E)
is:
25
Equation 2.22
In conclusion, the formula for UV fluence calculation could be expressed as follows:
( )
Equation 2.23
When using this kind of actinometer to measure the UV fluence, there are several
precautions:
1. The phenanthroline solution cannot be exposed to the room light because they may
decompose and then cause interference to the final result;
2. The conversion of Fe3+
should be less than 5%;
3. Every operation should be done in red illumination condition.
2.7 Summary and Research Needs
In summary, UV LEDs are promising technologies in UV disinfection field because they
have a lot of advantages over conventional UV lamps, such as higher energy efficiency,
more environmental friendly, and potential longer lifetime, etc. Due to these advantages,
UV irradiation will become a cheap method for wastewater disinfection. It can be expected
that its implementation will not just be limited to the disinfection of clean water, but also
turbid wastewater. But the impact of high turbidity on the disinfection performance of UV
LEDs is still not well known.
Past research results show that the particles in the wastewater can impact the UV
disinfection performance significantly. The mechanism is dependent on the properties of
particles, such as composition, size distribution, and surface charge. But in research,
synthetic wastewater is always used in ordered to carry out the controlled experiments and
obtain reproducible research results.
One of the challenges to conduct the disinfection test with the reactor of UV LEDs is UV
fluence determination in the reactor. As the size of reactor is quite small and the UV
intensity distribution is very inhomogeneous, so spectroradiometer cannot be used to
measure the UV intensity. One possible way to measure the UV fluence is the actinometry
method. This method is less time-consuming, cheap and simple relative to biodosimetry.
In conclusion, as a promising technology for wastewater disinfection, some investigations
on the application of UV LEDs for turbid wastewater disinfection can provide some
valuable information on their application in the future.
26
Chapter 3
Methodology
3.1 Introduction
The objective of this study was to investigate the impact of turbidity on the disinfection
performance of UV LEDs (280 nm). The research framework is shown in Figure 3.1. At
first, the reactor has been built, based on the characteristics of UV LEDs, such as size, life
time and emission power, etc.
Figure 3.1 Research framework
After building the reactor, the actinometry method has been applied to measure the UV
fluence in the reactor. The actinometer used is iodide-iodate actinometer and ferrioxalate
actinometer. For iodide-iodate actinometer, the fluence determination experiment was
conducted under room light condition, while the ferrioxalate actinometer is quite sensitive
to the room light, so it has been conducted under the subdued red light condition.
At last, both synthetic wastewater and real wastewater have been used to conduct the
disinfection test. Synthetic wastewater was synthetized through adding montmorillonite
and E. coli into the distilled water. The wastewater with different turbidity was obtained
through controlling the amount of montmorillonite added into the distilled water.
Additional E. coli was added as the challenge microorganism.
27
In practice, the quality of wastewater may vary seasonally or geographically, so one kind
of synthetic wastewater cannot represent all kinds of wastewater. In order to check the
disinfection performance of UV LEDs with real wastewater, real wastewater with different
turbidities has also been used to conduct the disinfection test.
3.2 Experimental Set-up
When designing the reactor with UV LEDs, several factors must be taken into
consideration, including the emission power of UV LEDs, and the lifetime of UV LEDs.
Generally, the emission power of UV LEDs is still quite low and the lifetime of UV LEDs
is quite short at current stage of development.
In order to overcome such shortcomings of UV LEDs, the distance between UV LEDs and
surface of solution must be short enough, so the intensity of UV light won't attenuate too
much. Additionally, the volume of solution should also be small enough. A big volume of
solution may lead to a long exposure time, which may cause a big challenge to the lifetime
of UV LEDs.
In this study, the emission power of UV LEDs is 10.8 mW and lifetime is around 100
hours. Based on such facts, the volume of solution was defined as 5 mL and the distance
between the UV LEDs and the surface of solution is around 4 - 6 cm. The reactor was
shown in Figure 3.2. UV LEDs is placed above the vessel used to contain the actinometer
solution and water sample. A magnetic stirrer has been used to promote the dispersion of
solution in the vessel. The power was supplied by the DC power and the voltage and
current are 30 V and 0.06 A, respectively. The position of both vessel and UV LEDs are
fixed so the UV intensity in the vessel could keep constant all the time.
Figure 3.2 Experimental set-up
3.3 UV Fluence Determination by Actinometry
In quasi-collimated beam apparatus that composed of conventional UV lamp (Figure 3.3),
the UV intensity on the surface of petri dish is quite uniform and could be measured by the
spectroradiometer. The UV fluence could also be calculated through the mathematical
model, which was already introduced in biodosimetry method part.
2.2 cm
Volume
=5 mL Stirring bar
UV LEDs
Power source
28
Figure 3.3 Quasi-collimated beam apparatus
But UV intensity distribution in this reactor (Figure 3.2) is very inhomogeneous and the
diameter of reactor (φ = 2.2 cm) is even smaller than the diameter of probe of
spectroradiometer (φ = 2.5 cm) (Figure 3.4), so the UV intensity cannot be measured by
the spectroradiometer. Under such conditions, there are two possible ways to measure the
UV fluence inside the reactor. One way is biodosimetry method, and another way is
actinometry method. Biodosimetry method is more widely accepted, but it is also more
time-consuming and requires the cultivation of virus or other microorganisms. So, the
adopted one here is actinometry method. This kind of method is based on some
photochemical reaction and it could be adapted to vessel with any kinds of geometrical
shape. When compared with the biodosimetry method, it is also cheaper and less time-
consuming.
Figure 3.4 Comparison between reactor and spectroradiometer
Now many kinds of chemical actinometers are available for UV fluence determination,
such as uridine, KI/KIO3, ferrioxalate. In this research, two kinds of actinometers have
been applied to determine the UV fluence in UV LEDs reactor: iodide-iodate and
ferrioxalate actinometer. The reason why two kinds of actinometers are used to determine
the UV fluence is that actinometry is not a standard method for UV fluence determination,
so two-times UV fluence determination can make the result become more acceptable.
Φ=2.2 cm
Volume
=5 mL Stirring bar
UV LEDs
Power source
29
3.3.1 Iodide-iodate actinometer
The iodide-iodate actinometer solution consists of 0.6 M KI and 0.1 M KIO3 in 0.01 M
Na2B4O7 buffered solution. It has been developed for many years and becomes a popular
actinometer (Rahn, 1997, 2003, 2006, 2013). When this kind of actinometer solution is
exposed to the UV light, following chemical reaction happens:
8 I + IO3
+ 3 H2O + h 3 I3
+ 6 OH
Equation 3.1
The UV fluence is proportional to the amounts of chemical reactions happened. In the
products of this reaction, the concentration of I3 could be determined
spectrophotometrically. Its concentration is proportional to the change of absorbance of
actinometer solution at 352 nm.
(1) Preparation of iodide-iodate actinometer solution
The preparation of actinometer solution is the critical procedure in actinometry method.
Because I- could be easily oxidized by IO3
- under low pH condition, then lead to the
formation of I2. This is the reason why the 0.01 M Na2B4O7 buffered solution has been
used to dissolve the KIO3. Based on the experience obtained from this experiment,
following procedure was used to prepare the actinometer solution.
1. Dissolve 0.381 g Na2B4O7 in 60 mL ultrapure water in a beaker. Use a stirring bar
to promote the dissolution until it is totally dissolved. This solution should have a
pH of around 9.2.
2. Dissolve 2.14 g KIO3 in the solution obtained from the first step. The stirring bar
also could be used to promote the dissolution.
3. Dissolve 9.96 g KI in the solution of KIO3 and Na2B4O7.
4. Transfer the actinometer solution into a 100 mL volumetric flask, and add ultrapure
water into the solution until the total volume is 100 mL.
The order of adding chemicals into the water is very critical in this experiment. Chemicals
should not be added into the ultrapure water together as they may not dissolve in the water
quickly. If the KIO3 or Na2B4O7 cannot dissolve quickly, the pH of solution will be not 9.2
and the IO3- may oxidize I
- , eventually lead to the formation of I2. This phenomenon has
been observed for several times in this experiment. I2 could cause interference to final
result as they can absorb the UV light (352 nm) strongly.
Another important thing that must be kept in mind is that there is a slow thermal oxidation
happen in the actinometer solution, so it cannot be kept for a long time. The recommended
storage time for actinometer solution is 4 hours. The actinometer solution should be made
up freshly each time before experiment.
(2) Procedure for UV fluence determination by iodide-iodate actinometer
Actinometer solution (0.6 M KI/ 0.01 M KIO3) is quite sensitive to the UV light under 330
nm. Although there is no well-defined wavelength above which absolutely no absorption
takes place, it is assumed that light above 330 nm will not contribute the formation of I3-.
Hence, this experiment could be conducted with the presence of room light. However, it is
also suggested that one must avoid doing this experiment under sunlight or fluorescent
30
lamps without a plastic cover over the fixture to avoid exposure to light with wavelength
less than 330 nm (Rahn, 2013).
In order to reduce the impact of light with wavelength longer than 330 nm, all glassware
used in this experiment were covered by the aluminium foil. Moreover, the UV exposure
experiment was conducted under the yellow light condition. Yellow light has wavelength
ranging from 577 nm to 597 nm. Its possibility to cause photochemical reaction in the
actinometer solution is quite small. So, the impacts from the room light have been reduced
as much as possible.
Figure 3.5 Procedures for UV fluence determination by iodide-iodate actinometer
The procedure for UV fluence determination by iodide-iodate actinometer was
schematically described in Figure 3.5 and it could be described as follows:
1. Measure the absorbance of fresh actinometer solution in a 1.0 cm pathlength quartz
cell at 300 nm and 352 nm. These values should be around 0.58 and 0.02,
respectively. Call the later value A352 (blank).
2. Add 5 mL actinometer solution and a small stirring bar into the tube and turn on the
magnetic stirrer.
3. Turn on the UV LEDs and let the actinometer solution be irradiated by the UV light.
Note the exposure time as t.
4. Measure the absorbance of irradiated actinometer solution at 352 nm. Note this
value as A352(sample).
5. Calculate the UV fluence by following formula:
Equation 3.2
Where V is the volume of actinometer solution (L), U is the photon energy (J/mole), ε is
the adsorption coefficient (M-1∙ cm
-1), is the quantum yield of
(mole/einstein), S is the
area exposed to the UV light (cm-2
). The value of each parameter in this equation is shown
in Table 3.1.
The photon energy is energy contained in 1 mole of photon, and it can be calculated by
following formula:
Where h is the Planck constant (6.62606896×10-34 J∙s), c is the speed of light (2.9972458 ×
108 m/s ), Na is the Avogadro number (6.02214179 × 10
23 mole
-1), and λ is the
wavelength of UV light (282 × 10-9
m). So, the photon energy is:
31
⁄
= 424111.628 J/einstein
The quantum yield of has been updated for many times in last several decades due to the
development of technology (Bolton et al., 2011; Goldstein et al., 2008; Rahn et al., 2003).
The value of this parameter at 282 nm still has not been reported by any literatures yet. The
reason is that the actinometry has just been applied to measure the UV fluence from LP
UV lamp or calibrate the radiometer. As it is known, LP UV lamp is a kind of
monochromatic lamp, and its wavelength is 253.7 nm. So, most of the data published in
literatures are about the quantum yield at 253.7 nm. But the quantum yields at 280 nm and
284 nm have already been determined. The value at 280 nm and 284 nm are 0.37±0.01 and
0.30, respectively (Goldstein et al., 2008; Rahn et al., 2003). So, an average value of them
can be employed as the quantum yield at 282 nm, which is shown in table 3.1.
Table 3.1 Value for Each Parameter in equation 3.2
Parameter Value Unit References
Volume of sample (V) 0.005 L -
Area (S) 3.14 cm2 -
Adsorption coefficient (ε) 27,636 M-1∙cm
-1 Bolton et al., 2011
Photon energy (U) 424,111.628 Joul/einstein -
quantum yield (Φ) 0.335*
mole /einstein
Goldstein et al., 2008;
Rahn et al., 2003
Reflection factor 0.025 - Bolton et al., 2011 *Note: the average value of 0.30 and 0.37.
Using these values to replace the symbols in equation 3.2 and do calculations to simplify
the equation 3.2, then it becomes:
UV fluence = [A352(sample) – A352(blank) ] × 74.831 mJ/cm2 Equation 3.3
Following is an example of calculation for application of this formula (Bolton et al., 2011):
5.0 mL actinometer solution in a 10 mL beaker (cross-sectional area 3.80 cm2), the
absorbance at 352 nm (in 1 cm × 1 cm quartz cuvette) before irradiation is found to be
0.021- call this A352(blank). After irradiation for 3.0 min, the absorbance at 352 nm is
0.526-call this value A352(sample). The following calculations illustrate how the photon
irradiance and the irradiance are calculated:
Concentration of triiodide ion [I3-]=[ A352(sample)- A352(blank)]/adsorption coefficient
=[0.526-0.021]/27,636
=1.827×10-5
M
Moles of I3- =[I3
-]×V(L)
= 1.827×10-5
M×0.005 L
=9.137 × 10-8
moles
The quantum yield this reaction at 254 nm was 0.6 mole∙einstein-1
.
Einsteins (moles of photons) = moles of I3- /
32
= 9.137× 10-8
mole/(0.6 mole∙einstein-1
)
= 1.523× 10-7
einsteins
photon irradiance (Ep) = einsteins/(area time)
= 1,523 10-7
/ (3.80 cm2 180 s)
= 2,226 10-10
einstein s-1
cm-2
Irradiance (E) = Ep photon energy at 253.7 nm (U253.7)
The irradiance must be corrected for the 2.5% that is reflected from the water surface, so
the incident irradiance on the water surface is:
E(corrected) = E(uncorrected)/0.975
= (2.226 10-10
471,576)/0.975 W∙cm-2
= 1.077 10-4
W∙cm-2
= 0.1077 mW ∙cm-2
3.3.2 Ferrioxalate actinometer
Ferrioxalate actinometer is another actinometer that has already been widely accepted by
researchers. When compared with iodide-iodate actinometer, this actinometer has
advantages and disadvantages. The advantage is that the quantum yield of ferrioxalate
actinometer is well defined. Unlike the iodide-iodate actinometer whose quantum yield is
different at different wavelength, the quantum yield of ferrioxalate is a constant when the
wavelength ranges from 200 – 250 nm and 270 – 365 nm (Goldstein et al., 2008). So, it is
unnecessary to exactly determine the quantum yields at each wavelength, while this is
really a must when using the iodide-iodate actinometer. The disadvantage is that this kind
of actinometer is very sensitive to the room light, so all operations need to be done under
red light conditions.
(1) Preparation of actinometer solution
Before preparing the actinometer solution, following solution should be prepared (Bolton
et al., 2011).
a. Ferric sulfate solution (0.2 mole/L) in 1 mole/L H2SO4. The ferric sulfate
(Fe2(SO4)3) should be added into the H2SO4 solution, not the water, because the
Fe3+
hydrolysis at pH>2.3 and the slat does not dissolve. As most of the ferric salt
are a bit impure, so the concentration of Fe3+
should be determined by
phenanthroline again after making this kind of solution.
b. Potassium oxalate solution with a concentration of 1.2 mole/L. Use an electronic
balance to weigh out 55.26 g K2C2O4 (analytical grade) and dissolve it into a 250
mL volumetric flask.
c. Sodium acetate buffer solution at pH 4.5. 20.5 g of CH3COONa∙3H2O was
weighted out and transferred into a 250 mL volumetric flask. Use around 100 mL
distilled water to dissolve it and then add 2.5 mL concentrated sulfuric acid (96-
98%) into it.
d. 1,10-phenanthroline solution with a concentration of 0.2%. Dissolve 0.5 g 1,10-
phenanthroline into 250 mL water. Because the 1,10-phenanthroline is quite hard to
dissolve, so a magnetic stirrer was used to promote its dissolution. This kind of
solution must be kept in the dark place to prevent the photodecomposition of
phenanthroline.
e. Hydroxylamine hydrochloride (NH2OH) solution with a concentration of 1 mole/L.
Dissolve 6.95 g hydroxylamine hydrochloride in 100 mL distilled water. This
33
solution cannot be kept for a long time and should be prepared freshly before
experiment each time.
f. Sulfuric acid solution (1 mole/L). Add 14 ml concentrated sulfuric acid into a 250
mL volumetric flask, in which 125 mL distilled water was added in advance.
As mentioned before, the ferric salt is always a bit impure, so the concentration of Fe3+
in
the ferric sulfate solution should be determined again. The phenanthroline method has been
employed to determine the concentration of ferric ion. The procedure could be described as
follows:
1. Exactly 0.3 mL of Fe2(SO4)3 solution was transferred to a 100 mL volumetric flask.
Distilled water was added until the mark. Mix them thoroughly.
2. Take 0.8 mL of this solution into a 10 mL volumetric flask. And 2 mL distilled
water and 1 mL NH2OH solution was added.
4Fe3+
+ 2NH2OH 4Fe2+
+ N2O + H2O + 4 H+ Equation 3.4
3. After 2 min, 2 mL sodium acetate solution and 2 mL phenanthroline solution were
added. Then the solution was kept in the dark for 40 min. The ferrous ion can
complex with the phenanthroline.
Fe2+
+ 3C12H8N2 [Fe∙3 C12H8N2]2+
(red brown) Equation 3.5
4. Measure the absorbance of this solution at 510 nm. Note as ABS(sample).
5. Repeat above procedure without adding 1 mL NH2OH solution. The absorbance
was noted as ABS(blank).
6. Calculate the concentration of ferric ion by following formula:
[Fe3+
]=
⁄
= [ ] Equation 3.6
Among which, 11110 L/mole ∙ cm-1
is the adsorption coefficient of complex of ferric-
phenanthroline, other parts in this formula means the dilution in this procedure.
After the concentration of ferric ion was accurately determined, the ferric sulfate solution
was ready for being used to make ferrioxalate actinometer. The actinometer solution is
quite sensitive to visible and UV light, so all the operations were conducted under the red
light condition. A red lamp for developing the film was applied for illumination during
experiment. Following is the procedure for making actinometer solution.
1. 15.2 mL of potassium oxalate solution (1.2 mole/L) and 35 mL of sulfuric acid
solution (1 mole/L) were added into a 1 L volumetric flask.
2. 6/[Fe3+
] mL ferric sulfate solution (0.2 mole/L) was added into this volumetric flask
and mixed them thoroughly. [Fe3+
] is the concentration of ferric ion determined
before. Then the ferrioxalate actinometer solution was ready for use.
The final compositions of the actinometer solution could be described as 6 mM potassium
ferrioxalate in 0.1 N sulfuric acid (call this FeOx solution). The sulfuric acid is used here in
order to maintain the pH of solution at below 2. So the ferric sulfate would not hydrolyze.
34
When this actinometer solution was exposed to the UV light, the Fe3+
will be reduced to
Fe2+
, which could be described as following:
Equation 3.7
The number of Fe2+
formed during this reaction is proportional to the UV fluence exerted
to the actinometer solution. At 270-340 nm, the quantum yield Φ ) is 1.39 ± 0.02
mole∙einstein-1
(Goldstein et al., 2008).
(2) Procedure for UV exposure test
The procedure for UV fluence determination by ferrioxalate actinometry is described as
below:
a. Adding 5 mL FeOx Solution into the reactor and turn on the magnetic stirrer, so
the actinometer solution could be mixed thoroughly;
b. Turn on the UV LEDs and start the UV irradiation;
c. After exposure for a period of time (t), turn off the UV LEDs.
d. Take 1 mL sample for measuring the concentration of Fe2+
.
e. Use relative formula to calculate the UV fluence
This procedure is shown in Figure 3.6. The procedure is almost the same with that of
KI/KIO3 actinometer and just use the ferrioxalate actinometer to instead the KI/KIO3
solution.
Figure 3.6 The procedure for UV fluence determination by ferrioxalate actinometer
The method used to measure the concentration of Fe2+
is phenanthroline assay. This
method could be described as follows:
1. Prepare a set of labeled 10 mL volumetric flask before exposure test because the
UV fluence determination generally consists of several runs, and in each volumetric
flask, 2 mL sodium acetate buffer solution (pH=4.5) and 2 mL 1,10-phenanthroline
solution were added;
2. 1 mL FeOx solution is taken from the reactor after UV exposure test and added into
the prepared volumetric flask;
3. Add distilled water to the 10 mL mark and put the flask in a dark place for 40 min
to ensure that the Fe2+
can complex with 1,10-phenanthroline completely.
4. Transfer the solution to a quartz cuvette and measure the absorbance at 510 nm.
Note this value as ABS(510)(sample).
5. Take 1 mL FeOx solution without UV exposure, and repeat above procedure.
Measure the absorbance at 510 nm and note this value as ABS(510)(blank).
1. Addition of Fe x
Solution ( mL) 2. UV exposure . Take sample
(1 mL) at time t.
. Determine the
concentration of Fe2
through colorimetry.
35
For one single run of this experiment, the Fe2+
generated can be calculated from following
formula:
Equation 3.8
Where V is the total volume of FeOx solution irradiated, mL, V1 is the volume withdrawn
from the irradiated solution, mL, 11,110 M-1
cm-1
is the molar absorption coefficient of the
Fe-1,10-phenanthroine complex, M-1∙ cm
-1 , 10 is the volume of the volumetric flask, mL,
and 1000 is the transformation between liter and milliliter.
The UV fluence could is proportional to the numbers of Fe2+
formed in this reaction.
Following formula could be used to calculate the UV fluence:
Equation 3.9
Where is the quantum yield, mole∙einstein-1
; U is the photon energy, J/einstein; Area is
the area of solution surface irradiated by the UV light, cm2; Reflection factor is 2.5%,
which is caused by the reflection of water surface.
The values of each parameter used in equation 3.8 and 3.9 are shown in table 3.2.
Table 3.2 Value Adopted for Each Parameter in Equation 3.8 and 3.9
Parameter value unit References
V 0.005 L -
V1 0.001 L -
U 424,111.628 Joul/einstein -
Area 3.14 cm2 -
1.39 mole/einstein Bolton et al., 2011
Reflection factor 0.025 - -
Using these values to replace the symbols in equation 3.8 and 3.9 to simplify the
calculation, then the formula for UV fluence calculation becomes:
UV fluence = [A510(sample) – A510(blank)] × 448.5254 mJ/cm2
Equation 3.10
3.4 Disinfection Test with Synthetic Wastewater
3.4.1 Preparation of synthetic wastewater
The method to prepare the synthetic turbid wastewater is described in Figure 3.7. Synthetic
wastewater was prepared through adding montmorillonite and E. coli into the RO water. E.
coli was added as the challenge organism. Before UV exposure, synthetic wastewater was
mixed thoroughly by vortex mixture. Additionally, all operations for synthetizing
wastewater were done aseptically in order to avoid the contamination.
36
Figure 3.7 Procedure for making synthetic wastewater
Different kinds of turbid wastewater were obtained by controlling the amount of
montmorillonite added into the RO water. Figure 3.8 illustrates the determination of
amounts of montmorillonite added into the RO water for making synthetic wastewater.
Figure 3.8 Determination of montmorillonite added into the wastewater
3.4.2 Preparation of E. coli for synthetic wastewater
The E. coli (ATCC 29214) was obtained from Thailand Institute of Scientific and
Technological Research (TISTR). After getting pure culture of E. coli, it was inoculated in
LB broth to regain the activity. After cultivation for 24 hours, 1 mL LB solution of this LB
broth solution was inoculated in the new LB broth. After cultivation for 5 hours, it was
harvested to make the glycerol stock of E. coli.
The E. coli was stored in the glycerol solution ( 0%) so that it won’t be killed by low
temperature. Here, the glycerol stock of E. coli was obtained through mixing 50% E. coli
solution and 50% glycerol solution. A vortex mixer has been applied for mixing step so
that the E. coli solution and LB broth can be mixed thoroughly. After that, the glycerol
stock was distributed into the 1.5mL cyto tubes, and then stored in the freezer. The
temperature of freezer was kept at -20℃ and the storage time is up to 1 month. This
procedure was schematically described in Figure 3.9.
37
Figure 3.9 The process for making glycerol stock from E. coli strain
Before using E. coli for disinfection test, 1 mL E. coli stock solution was inoculated in the
LB broth. After cultivation in the new LB broth for a certain period, the E. coli at
approximately mid-exponential growth phrase was harvested for the disinfection test. 10
mL of the E. coli suspension was centrifuged at 5000 rpm for 20 minutes and the
supernatant was aseptically drawn off. Then, the remaining E. coli was re-suspended with
the phosphate buffered saline (PBS, 0.01M) and centrifuged again. As described above, the
washing procedure was repeated twice to remove any nutrient medium. After that, the
E. coli solution in PBS solution was read for use in the disinfection test.
3.4.3 Enumeration of E. coli
Pour plate technique has been applied to enumerate E. coli in synthetic wastewater before
and after UV exposure. One kind of selective medium-Chromocult Coliform agar (Merck,
Germeny)-has been used to cultivate the E. coli. Chromocult Coliform agar can promote
the growth of total coliform and E. coli, while suppress the growth of other
microorganisms. On this kind of agar, E. coli appear as dark violet colony (Figure 3.10),
and coliform bacteria appear as pink to red colonies. Other bacteria appear as colorless or
green colony.
The benefits of using this kind of agar to count the density of E. coli is that it can help to
avoid the contamination in experiment as most of the gram positive bacteria cannot grow
on this kind of agar. When combined with pour plate technique, the density determination
of E. coli in synthetic wastewater can be quite accurate. As only E. coli was added into the
synthetic wastewater, and Chromocult coliform agar is a selective medium for E. coli, so
this kind of modified method is quite good for synthetic wastewater.
Figure 3.10 The Chromocult Colifrom agar and the colony
38
The pour plate technique consists of two parts: series dilution of sample (Figure 3.11) and
the plating of dilutions (Figure 3.12). All the operations should be done aseptically.
Serial dilution is the stepwise dilution of E. coli solution and the dilution factor for each
step is 10-1
. The procedure could be described as following steps:
1. Sterilizing the glassware and other equipments, and put them into the laminar hood.
2. Clean hands and laminar hood with 70% alcohol and add 9 mL of sterile PBS
solution into each tube using 10 ml transfer pipette.
3. Labeling all tubes with 10-1
to 10-5
indicating the dilution factor.
4. Transfer 1 mL from the culture sample to the first tube (10-1
) by 1 mL auto-
micropipette and mix them gently by vortex.
5. Take 1 mL of the diluted E. coli solution from the first tube and add them to the
next tube (10-2
). Thereafter mix them gently.
6. Repeat the same procedure for the left four tubes (10-3
to 10-6
). Then the E. coli
solutions with dilution factors 10-1
, 10-2
, 10-3
, 10-4
, 10-5
, 10-6
are formed.
Figure 3.11 Series dilution
After finishing the serial dilution, each solution was inoculated in the Chromocult
Coliform agar. The plating of solution could be described as following steps:
1. Label all necessary information on the plate, such as microbes name, date, user’s
name and dilution factor.
2. Mix the last solution (10-6
) by using vortex mixer. Take 1 mL solution from it by
using a 1 mL micropipette and drop them slowly into the plate. Take other two
samples and inoculate them in other two plates, so each dilution was cultivated in
triplicates.
3. Repeat the same procedure for all other dilution samples and inoculate them into
the corresponding plates.
4. Pouring Chromocult Coliform agar into the plates and shake them gently to mix
the E. coli and agar.
5. Incubate them at 37 ℃ for 24 h. Then count the number of colonies in each plate
and calculate the numbers of E. coli per 1 mL. For example, the numbers of
colonies of one dilution are a1, a2, a3.
So the average number of colonies is:
Equation 3.11
39
If the dilution factor is α, so the concentration (CFU/mL) of E. coli in this dilution
is:
Equation 3.12
It must be noted that this method is just suitable for 30 - 300 colonies per petri dish. If the
number of colonies exceeds this range, the solution should be diluted (more than 300
CFU/dish) or concentrated (less than 30 CFU/dish).
Figure 3.12 Plating of dilutions
3.4.4 Procedure for disinfection test with synthetic wastewater
The procedure for conducting disinfection test with synthetic wastewater is illustrated in
Figure 3.13. Firstly, the synthetic wastewater was added into the reactor. Then UV LEDs
were turned on. After exposure to the UV LEDs for a period of time, the UV LEDs were
turned off and 1 mL water sample will be taken for determining the concentration of
E. coli.
Figure 3.13 Procedure for disinfection test with synthetic wastewater
Usually, the turbidity of secondary effluent is quite low, sometimes it is even could be less
than 20 NTU. Turbidity of primary effluent may depend on the quality of raw wastewater
and the performance of primary sedimentation tank. 50-100 NTU could be a rational range
of primary effluent’s turbidity. For raw wastewater, the turbidity may vary significantly.
Sometimes, it can be as high as 1000 NTU, while it can be also lower than 100 NTU.
Based on such information, the turbidity chose for synthetic wastewater are 27, 70, 113,
and 156 NTU, which have also been adjusted based on the disinfection performance of UV
LEDs in this study. For each kind of turbid wastewater, the applied UV fluence was
different in order to get the UV fluence-inactivation response curve in a suitable range. For
synthetic wastewater of 27 NTU, the required UV fluence for getting a UV fluence-
response curve should be much lower than that of wastewater of 156 NTU. The structure
of this experiment was shown as below:
40
Figure 3.14 Research framework for disinfection test with synthetic wastewater
3.5 Disinfection Test with Real Wastewater
3.5.1 Wastewater sampling
As the quality of real wastewater varies seasonally and geographically, one kind of
synthetic wastewater would not be able to represent all kinds of wastewater. In real case,
the quality of wastewater might be very different from synthetic wastewater used here, so
the disinfection performance of UV LEDs may be also quite different. In order to give out
a reference for real case, the disinfection test has also been conducted with several kinds of
real wastewater samples.
The wastewater samples were taken from the AIT wastewater treatment plant, whose
treatment process was shown in Figure 3.14. A sequencing batch reactor (SBR) has been
employed to treat the wastewater generated by the residents in AIT campus. Raw
wastewater flows into the collection tank first. Then it was lifted to the equalization tank
by a pump. The equalization tank can homogenize the quality of wastewater flowing into
the aeration tank so that the shocking load won’t happen in the aeration tank. The aeration
tank was operated in aeration-settling-discharge-filling-aeration mode. During aeration,
most of organic matters in the wastewater were removed by the activated sludge. When
aeration is stopped, the activated sludge would settle down and the anaerobic environment
in the tank is created at the same time, so the denitrification may happen, thus nitrogen in
the wastewater can also be removed. At last, the wastewater was discharged.
Figure 3.15 AIT wastewater treatment process
41
The aeration system in equalization tank and SBR tank works at the same time. So during
the settling stage of SBR tank, the aeration in equalization tank also stopped and the
particles in the wastewater can settle down. The wastewater sample taken from the surface
of equalization tank is much less turbid than the wastewater sample taken from the inlet of
equalization tank.
In real case, the wastewater disinfected by UV irradiation should be as less turbid as
possible. As it is known, sedimentation is a cheap process to remove part of the particles
and pathogens from the wastewater. Properly designed and operated sedimentation tank
should remove 50-70% of the suspended particles and 25-40% of the BOD (Metcalf &
Eddy, 2003). So, it is better to disinfect the turbid wastewater after sedimentation because
such pretreatment can improve the disinfection performance of UV LEDs.
However, primary sedimentation tank was eliminated in AIT wastewater treatment plant in
order to simplify the treatment process. But, as mentioned above, settling process also
happened in the equalization tank and the wastewater at the surface level is less turbid than
the raw wastewater. To some extent, the wastewater at the surface of equalization tank has
the same property of wastewater after settling. So, it is reasonable to use this kind
wastewater to instead the wastewater after sedimentation. Granted, the turbidity might be
different from the wastewater sample after real sedimentation as the time for settling is
different. But it will not matter because the turbidity of primary effluent from the real
wastewater treatment plants that using conventional activated sludge process may still vary.
In real case, the turbidity of primary effluent is dependent on both detention time and
initial concentration of total solids in raw wastewater.
Additionally, different turbid wastewater was obtained through mixing raw wastewater and
wastewater from the surface of equalization tank in different ratio. And all of them have
been used to conduct the disinfection test. The result obtained through disinfection test
with different turbid wastewater can shows the impact of turbidity of wastewater on
disinfection performance of UV LEDs better.
The method of taking water sample from the primary effluent is based on the standard
method of APHA. Water sample was stored in ice box and take to the lab within half hour.
The disinfection test and cultivation of E. coli should be done as soon as possible because
the long duration of storage may change the microbial properties of wastewater.
3.5.2 Selection of challenge organism and its enumeration
Both total coliform and E. coli in the wastewater has been selected as the challenge
organism. Both of them have been adopted as the microbial parameter for wastewater
discharge standard or reuse standard.
Unlike the synthetic wastewater, in which only E. coli exists, so that Chromocult Coliform
agar and pour plate technique can be used to count the density of E. coli and the obtained
results are also quite reliable. In real wastewater, millions of microorganism exists.
Although Chromocult Coliform agar with additional antibiotics (cufsulodin) has been
reported to be able to detect the fecal pollution very well (Byamukama et al., 2000), the
addition of antibiotics has not been approved by US EPA (Olstadt et al., 2007).
Furthermore, no articles related to using Chromocult Coliform agar for wastewater
42
examination have been published until now. So, it is better to use the standard method to
count the density of total coliform and E. coli in wastewater.
In this study, the method for counting density of total coliform is the most probable
number method (MPN method), which is adapted from the standard methods (APHA,
2005). In standard MPN method, the number of tubes for each set can be 3, 5, and 10. In
this study, 5 tubes were selected in order to obtain accurate results and control the
workload. This method is schematically described in Figure 16.
Figure 3.16 MPN method
In MPN method, 1 mL water sample was inoculated into the tube with 10 mL lactose broth.
After incubation at 37℃ for 24 hours, the tube with both turbidity and gas production was
reported as positive result. Otherwise, the result was recorded as negative result. All tubes
with negative results have been put into the incubator for another 24 hours to ensure that
they are negative results.
The detection of E. coli was done after getting the result for total coliform detection. The
enriched solution of total coliform was inoculated onto the Eosin Methylene Blue (EMB)
agar. In this step, only the solution showing positive result was inoculated onto the EMB
agar. EMB agar is a kind of selective medium and the color of E. coli colony appears as
metallic green sheen.
The most probable number of total coliform and E. coli was calculated by following
formula:
⁄ Equation 3.13
The value of table MPN can be read from the table in standard method.
43
3.5.3 Procedure for disinfection test
The procedure for conducting the disinfection test with primary effluent is shown in Figure
3.17. The procedure is almost the same with the disinfection test with synthetic wastewater
and the only difference is to use the real wastewater to instead the synthetic wastewater.
Figure 3.17 The procedure for conducting disinfection test with real wastewater
Different UV fluence has been applied to disinfect each kind of wastewater sample. As
illustrated in the last section, UV fluence required to obtain the UV fluence-inactivation
response curve is different. So, the time of UV exposure for each wastewater sample is
also very different. At the same time, the turbidity of each kind of wastewater was noted.
The research framework was shown as below:
Figure 3.18 Research framework for disinfection test with real wastewater
3.6 Summary
All the methods for conducting this research were summarized in Table 3.3.
Table 3.3 Summary of Methods
Parameters Units Methods Reference
E. coli CFU/mL Pour plate technique (Merck, Germeny)
MPN/100 mL MPN method (APHA, 2005)
Total Coliform MPN/100 mL MPN method (APHA, 2005)
Turbidity NTU Nephelometric (APHA, 2005)
Absorbance - Spectrophotometer (APHA, 2005)
44
Chapter 4
Results and Discussions
This chapter presents all the results for this study and the corresponding discussions to the
results. Firstly, the characterization of UV LEDs was introduced. This was followed by a
section to illustrate the results from two kinds of actinometers, which were used to
determine the UV fluence in the reactor. A comparison has been done between the results
of two kinds of actinometers and the discrepancy between them was also discussed. At last,
the results for UV disinfection test were shown, including disinfection test results for both
real wastewater and synthetic wastewater.
4.1 Characterization of UV LEDs
Before the fluence determination experiment and disinfection test, spectroradiometer
(Ocean Optics USB2000) has been used to measure the wavelength of UV LEDs. The
result is shown in Figure 4.1. From this figure, it can be found that the peak emission was
not at 280 nm exactly, the measured wavelength is more close to 282 nm. As the properties
of 280 nm and 282 nm UV light are almost the same, so it won’t cause a big impact to the
final results.
Figure 4.1 The emission spectrum of UV LEDs
4.2 UV Fluence Determination by Actinometry Method
1. Iodide-iodate actinometer
The procedure for UV fluence determination was shown in Figure 4.2. Generally, a full
UV fluence determination process consists of several runs of this procedure. In each run, 5
mL iodide-iodate actinometer solution was added into the reactor. The absorbance of
actinometer solution before and after exposure was measured by spectrophotometer at
352 nm. As illustrated in last chapter, the equation for calculating UV fluence can be
expressed as follows:
UV fluence = [A352(sample) – A352(blank) ] × 74.831 mJ/cm2 Equation 4.1
0
0.2
0.4
0.6
0.8
1
1.2
200 240 280 320 360
Re
lati
ve e
mis
sio
n p
ow
er
wavelength (nm)
45
Figure 4.2 The procedure for UV fluence determination
In this study, the UV fluence determination has been run for 18 times. The results were
shown in table 4.1. The absorbance of actinometer solution before UV exposure was 0.003.
Table 4.1 Absorbance of Iodide-iodate Actinometer Solution at Different Exposure
Time
Exposure time (s) 6.33 9.91 15.51 20.37 25.35 29.94 40.46 50.12 60.84
ABS 0.003 0.003 0.016 0.045 0.074 0.103 0.176 0.224 0.288
UV fluence (mJ/cm2) 0 0 0.97 3.14 5.31 7.48 12.95 16.54 21.33
Exposure time (s) 61.88 63.34 70.34 79.94 80.4 81.04 89.44 100.34 100.65
ABS 0.29 0.301 0.351 0.399 0.389 0.39 0.445 0.494 0.51
UV fluence (mJ/cm2) 21.48 22.30 26.04 29.63 28.88 28.96 33.08 36.74 37.94
After doing linear regression between the UV fluence and exposure time with the data
obtained after 15.51 s, following result was obtained (Figure 4.3). The UV fluence was
proportional to the exposure time, and the R2 was 0.9978. The relationship between UV
fluence and exposure time was:
UV fluence = 0.4282 × exposure time – 5.1583 Equation 4.2
The reason why the first two data were not used here is that they are abnormal result
because the ABS value did not change after UV exposure. This may be caused by
unexpected chemical reactions. The triiodide formed during UV exposure reacts with other
chemicals in the actinometer solution and then no absorbance change was observed after
UV exposure.
Figure 4.3 UV fluence determined by iodide-iodate vs. exposure time
46
In the reactor, the emission power of UV LEDs is a constant and the position of UV LEDs
in the reactor is fixed, so UV fluence in the reactor should be proportional to exposure time.
When the exposure time is 0 s, the UV fluence is 0 mJ/cm2. So the line in Figure 4.3
should be set to have an intercept of 0. Then equation 4.2 becomes:
UV fluence (mJ/cm2) = 0.4282 mW/cm
2 × exposure time (s) Equation 4.3
The equation shows that the irradiance the actinometer received from UV LEDs is
0.4282 mW/cm2. The area exposed to the UV LEDs is 3.14 cm
2, so the total irradiance is
1.34 mW. When compared with 10.8 mW, which is the emission power shown in the
instruction of UV LEDs, this is much lower. This attenuation is attributable to several
reasons. Firstly, there is attenuation during the transportation of UV light in the air.
Secondly, not all UV lights from the UV LEDs irradiated into the actinometer solution,
only a part of them were absorbed by the actinometer solution. Thirdly, a small part of the
UV lights were reflected by the actinometer solution.
2. Ferrioxalate actinometer
Just as illustrated in the methodology part, the procedure for this actinometry is the same
with that of iodide-iodate actinometer. The equation for the calculation of UV fluence was
shown as below:
UV fluence = [A510(sample) – A510(blank)] × 448.5254 mJ/cm2
Equation 4.4
In this experiment, ferrioxalate actinometer solution has been used to replace the iodide-
iodate actinometer solution, and it has been repeated for 12 times. In each run, the ferrous
ion concentration was analyzed by phenanthroline assay before and after UV exposure,
which was done through the spectrophotometric method. The absorbance measured in this
experiment was shown in table 4.2.
Table 4.2 Absorbance of Ferrioxalate Actinometer Solution at Different Exposure
Time
Exposure Time (s) 0 5.25 5.6 10.22 10.66 15.43
Absorbance 0.011 0.018 0.018 0.024 0.022 0.024
UV fluence (mJ/cm2) 0 3.14 3.14 5.83 4.93 5.83
Exposure Time (s) 19.84 20.25 25.03 30.31 40.25 50.12
Absorbance 0.032 0.03 0.035 0.039 0.048 0.056
UV fluence (mJ/cm2) 9.42 8.52 10.76 12.56 16.59 20.18
The results shown in above table were used to do linear regression. The obtained result was
shown in Figure 4.4.
From this figure, it can be found that the relationship between UV fluence and exposure
time could be expressed as following:
UV fluence (mJ/cm2) = 0.382 × exposure time + 1.0828 Equation 4.5
47
Before using this equation, the intercept of the trendline should be set to zero. So the
equation turns to:
UV fluence (mJ/cm2) = 0.382 × exposure time Equation 4.6
The irradiance of UV LEDs was found to be 0.382 mW/cm2. When it is multiplied by the
exposure area (3.14 cm2), the irradiance becomes 1.2 mW. This value is even lower than
the value measured by iodide-iodate actinometry. The reasons that should be responsible
for such kind of attenuation were already discussed in the iodide-iodate actinometry part.
Figure 4.4 UV fluence determined by ferrioxalate vs. exposure time
3. Comparison between the results from two actinometers and discussion
The irradiance of UV LEDs measured by iodide-iodate actinometer was 0.4282 mW/cm2,
while this value was 0.382 mW/cm2 in ferrioxalate actinometry. Based on the experience
from this study, two factors should be accountable for this discrepancy.
Firstly, the quantum yield used in iodide-iodate was not very accurate. Currently, quantum
yield of iodide-iodate is not available at 282 nm. The value applied here is the average
value of quantum yields at 280 nm and 284 nm, which is already illustrated in the chapter 3.
So, this may be also a resource of the discrepancy between them.
Secondly, iodide-iodate actinometry and ferrioxalate actinometry are two different
actinometry system. Actinometry methods are quite sensitive to procedural variation, so it
is normal to get two results that are slightly different.
Based on above discussion, following conclusion can be made:
1. The irradiance of the UV LEDs might be 0.382 mW/cm2 or 0.4282 mW/cm
2.
2. UV fluence measured by actinometry method was not constant. The result is very
sensitive to the procedural variation and the results obtained through different
actinometry are also different.
48
4.3 Disinfection Test with Synthetic Wastewater
Synthetic wastewater was synthetized through adding montmorillonite into the water that
purified through reverse osmosis. Different concentration of montmorillonite can result in
different kinds of turbid wastewater. The relationship between the turbidity of synthetic
wastewater and the concentration of montmorillonite was shown in Figure 4.5.
Figure 4.5 Turbidity of synthetic wastewater vs. concentration of montmorillonite
From this figure, it can be concluded that the turbidity of synthetic wastewater was
proportional to the concentration of montmorillonite (Cm) in the water, and the relationship
between them can be expressed as following:
Turbidity (NTU) = 0.1804 × Cm – 15.78 Equation 4.7
In this study, the turbidity chose for synthetic wastewater was 0, 27, 70, 113, and 156 NTU
and the concentration of montmorillonite was 0, 237, 475.5, 713.9, and 952 mg/L,
respectively. Disinfection test has been conducted with each kind of synthetic wastewater
and the result for disinfection test was shown in Figure 4.6.
As explained in chapter 2, the UV inactivation kinetics can be expressed by following
equation:
log (
) = k1 × Ft Equation 4.8
Where log (
) is the log-reduction of microorganism, C0 is the concentration of challenge
organism before UV exposure (CFU/mL), Ct is the concentration of challenge organism
after UV exposure (CFU/mL), Ft is the UV fluence at time t, and k is a constant (cm2/mJ).
In this reactor, UV irradiance from UV LEDs was a constant, and the UV fluence is the
product of irradiance and exposure time. So, equation 4.8 can be converted to:
log (
) = k2 × t Equation 4.9
y = 0.1804x - 15.778 R² = 0.9934
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Turb
idit
y (N
TU)
concentration of montmorillonite (mg/L)
49
Where k2 is a constant (s-1
), and t is the exposure time.
Through doing linear regression with result of each kind of synthetic wastewater, the
equation of inactivation kinetics under different turbid water condition was obtained and
they are shown in table 4.3. In order to make it simple, log (
) was expressed as L.
Figure 4.6 Time-response curve of E. coli in different turbid wastewater
Table 4.3 Inactivation Kinetics of UV Irradiation in Different Turbid Wastewater
Turbidity (NTU) Equation k2 (s-1
) R2
n
0 L = -0.13 × t + 1.53 -0.13 0.92 13
27 L = -0.06 × t + 1.09 -0.06 0.95 10
70 L = -0.05 × t + 0.08 -0.05 0.55 9
113 L = -0.03 × t + 1.33 -0.03 0.91 10
156 L = -0.02 × t + 0.02 -0.02 0.76 10
* Note: R2 is the coefficient in each linear regression;
n is the number of data used for linear regression.
A UV fluence of 40 mJ/cm2 is typically the minimum requirement for drinking water
disinfection plants, whereas recommended design UV fluence can go up to 100 mJ/cm2 for
reclaimed water systems, depending on the upstream treatment application (Kollu et al.,
2012). In this reactor, the maximum UV irradiance measured by actinometry was 0.4282
mW/cm2. The exposure time would be 233 s if a UV fluence of 100 mJ/cm
2 was exerted to
the synthetic wastewater solution. When E. coli in wastewater is exposed to the UV LEDs
for such a long time, the expected log-reduction would be at least 4 in all kinds of turbid
wastewater.
The absolute value of k2 in the equation shown in table 4.3 represent the sensitivity of
E. coli to UV irradiation. When increased the turbidity, the sensitivity of E. coli to UV
irradiation decreased significantly, which was shown in Figure 4.7. The absolute value of
-5.0
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160
Log-
red
uct
ion
Exposure time (s)
0 NTU
27 NTU
70 NTU
113 NTU
156 NTU
50
k2 was 0.13 when the turbidity was 0 NTU, while it was just 0.02 when the turbidity was
increased to 156 NTU, with a decrease of 85%. This means that E. coli became much less
sensitive to UV irradiation when the turbidity of wastewater was increased. This can be
due to several reasons: absorbance of particles to UV light, scattering effect of particles on
UV light and habitation of E. coli inside the particles.
Figure 4.7 Sensitivity of E. coli to UV irradiation in different turbid synthetic
wastewater
In Figure 4.8, it can be found that, when the turbidity of wastewater was 70 NTU, the
disinfection performance of UV LED was better that of 27 NTU. This is caused by error
in this experiment. The R2 was just 0.55 when the turbidity is 70 NTU, which can be found
in table 4.3. It means that the log-reduction did not show a linear relationship with UV
exposure time very well. However, it clearly shows that the disinfection performances of
UV LED on E. coli in both two kinds of turbid wastewater are very close to each other.
Figure 4.8 Time-response curve of E. coli in synthetic wastewater of 27 and 70 NTU
Based on the information above, it can be concluded that E. coli in turbid wastewater can
still be disinfected by UV LED effectively. But the sensitivity of E. coli to UV irradiation
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0 50 100 150 200
Sen
siti
vity
of
E. c
oli
to U
V (
s-1)
Turbidity (NTU)
-5.0
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100
Log-
red
uct
ion
Exposure time (s)
0 NTU
27 NTU
70 NTU
51
may decrease greatly when increasing the turbidity of wastewater. However, a slight
increase of turbidity of wastewater may not reduce the disinfection efficiency too much.
Wastewater reclamation became more and more popular around the world due to the water
scarcity. In developing countries, domestic wastewater is even used for agriculture
irrigation directly, without any pretreatment, which increases the risk to human being
significantly. Pretreatment of wastewater before reuse in strongly recommended.
The typical abundance of E. coli in domestic wastewater ranges from 106 to 10
8
CFU/100 mL (Kadam et al., 2008; Molleda et al., 2008; Vrhovšek et al., 1996; Zhang et al.,
2007). Based on the treatment performance, treated wastewater can be reused for many
purposes, such as agricultural irrigation, recreation and even discharged to the environment
directly. The requirements for wastewater that used in different purpose are quite different.
In Canada, the concentration of fecal coliform of treated wastewater for unrestricted use
must be less than 2.2 CFU/100 mL(Zhang et al., 2007). To meet such strict standard,
disinfecting wastewater directly may be not a good option.
However, the standard for reusing wastewater for agriculture irrigation is not so strict. In
Spanish, the microbial standard for reusing wastewater in agriculture irrigation is that
E. coli concentration should not exceed 100 CFU/100 mL, which requires around 5 log-
reduction of E. coli in wastewater. In Figure 4.6, it can be found that, in all turbid
wastewater samples (27, 70, 113, 156 NTU) except the turbid wastewater of 156 NTU, 5
log-reduction of E. coli is possible when UV exposure time is as long as 233 s. So,
wastewater disinfected by UV irradiation might be used for agriculture irrigation. It should
also be noted that composition of real wastewater is much more complex than the synthetic
wastewater. This may lead to a reduction of disinfection efficiency of UV LEDs.
Meanwhile, pathogens in real wastewater may be also more resistant to UV irradiation than
E. coli used in synthetic wastewater. Thus, whether it is possible to use UV LEDs to
disinfect the wastewater for reuse may depend on two aspects: one aspect is the quality of
wastewater and another one is the standard requirement.
To some extent, the result shown above proved that UV irradiation might be a possible
way to disinfect the turbid wastewater. However, as other pathogens may be more resistant
to UV irradiation than E. coli, so the disinfection efficiency of UV LEDs on UV-resistant
microorganism should also be checked, especially the virus and protozoa.
4.4 Disinfection Test with Real Wastewater
As mentioned in last section, the composition of real wastewater is much more complex
than synthetic wastewater. The disinfection efficiency of UV LEDs on real wastewater
should also be different from that of synthetic wastewater. So it is necessary to conduct the
disinfection test with real wastewater.
In this study, the wastewater sample from AIT wastewater treatment plant was used to
conduct the disinfection test. Wastewater was sampled at different parts of treatment
process at different time. The sampling site includes: surface of equalization tank (SEQ),
inlet of equalization tank (IEQ), and collection tank (CT). More detailed information could
be found from chapter 3. Properties of wastewater samples were described in table 4.4,
including the turbidity, TS, TSS.
52
Table 4.4 Properties of Wastewater Samples
No. of
sample Date
Sampling
site
Turbidity
(NTU)
TS
(mg/L)
TSS
(mg/L)
TDS
(mg/L)
1 April 9 IEQ 86 255 65 190
2 April 11 SEQ 57 396 75 321
3 April 14 CT 130 435 85 350
4 April 16 CT 72 398 65 333
*Note: SEQ: surface of equalization tank
CT: collection tank
From this table, it can be found that turbidity of wastewater did not show a proportional
relationship with total solids or total suspended solids in wastewater. This might be caused
by different size distribution or composition of particles in the wastewater. Turbidity is a
measure of the scatter of visible light (400-700 nm). This means the particles with size
ranges from 400 to 700 nm can contribute to the turbidity most, while particles with other
size distribution may contribute less. So, different size distribution might be a reason for
this phenomenon. Another reason might be the different composition of particles in
wastewater. Generally, inorganic particles can contribute more to turbidity than organic
particles. For example, 50 mg/L kaolinite gave a turbidity reading of about 80 NTU, while
50mg/L humic acid gave a turbidity reading only slightly greater than 3 NTU (Edzwald,
1987).
The results of disinfection test with each kind of wastewater were shown in Figure 4.9 and
4.11. Unlike the synthetic wastewater, the disinfection test can be repeated for many times
and enough amount of data can be obtained to do very accurate analysis. Disinfection test
with real wastewater needs to be done within 6 hours after sampling. Here, the disinfection
test for each wastewater sample has just been conducted for 4 to 5 times in order to finish
the experiments within 6 hours and control the workload. So, only 4 to 5 data for each trial
were used in Figure 4.9 and 4.11 to show the trend of result.
Figure 4.9 Time-response curve of total coliform in different turbid wastewater
53
From Figure 4.9, it can be found that the disinfection performances of UV LEDs on total
coliform in wastewater were quite similar to each other even though the turbidity of each
wastewater sample is different. When the exposure time was less than 100 s, the
log-reduction of total coliform shows a good linear relationship with exposure time,
regardless of the turbidity of wastewater. This could be further confirmed by Figure 4.10.
In Figure 4.10, all log-reduction of total coliform and the corresponding exposure time
were used to do linear regression. The obtained equation was:
Log-reduction = -0.03×exposure time + 0.39 Equation 4.10
And the trendline also has a R2 of 0.95. It means that the log-reduction of E. coli shows a
good linear relationship with exposure time. Actually, the size distribution of particles in
the four kinds of wastewater should be different, because one wastewater sample was taken
from the surface of EQ tank after sedimentation, while another sample was taken from the
inlet of EQ tank, which was without sedimentation. But coliform bacteria in all samples
showed the similar UV fluence-inactivation response curve. This result suggests that
slightly change of particle size distribution may not impact the disinfection performance of
UV LEDs significantly.
Figure 4.10 Time-response curve of total coliform within 100 s UV exposure
The maximum log-reduction of total coliform is nearly 3 and it could not be increased
further by prolonging the UV exposure time. This is caused by the attachment of coliform
on particles, which is also called “tailing” in UV inactivation kinetics. The coliform
bacteria in wastewater can be free or attached to particles. Free coliform bacteria could be
inactivated by UV easily. The inactivation kinetics is also a first order reaction. However,
coliform bacteria attached to the particles can be very resistant to UV irradiation as they
are sheltered inside the particles. Hence, even the UV exposure time has been increased,
the coliform bacteria can still be alive in the wastewater. This is the reason why log-
reduction did not increase further after reaching its maximum value at around 110 s.
Special attentions should be paid to the total coliform in the wastewater of 57 NTU as it
has reached the maximum log-reduction at 111.87 s, with a value of 4.12. This indicates
that the tailing effect may not appear or will be retarded if the turbidity of wastewater is
low enough. However, this may be also caused by the error of results because only one
y = -0.03x + 0.39
R² = 0.95 -2.7
-2.4
-2.1
-1.8
-1.5
-1.2
-0.9
-0.6
-0.3
0.0
0 10 20 30 40 50 60 70 80 90 100
Lo
g-r
ed
uct
ion
Exposure time (s)
54
data was obtained here. More investigations should be done to confirm whether it is caused
by error or the “tailing” effect would disappear or be retarded in UV disinfection of less
turbid wastewater.
Figure 4.11 Time-response curve of E. coli in different turbid wastewater
The time-response curve of E. coli in wastewater was shown in Figure 4.11. Due to the
statistically unreliable data, only two sets of data were presented. The result was a bit the
same as what has been observed in the disinfection test with total coliform, E. coli in
wastewater can be inactivated effectively at first, and then it became very resistant to UV
light due to the “tailing” effect. The maximum log-reduction of E. coli was nearly 3 and it
could not be increased further by prolonging the UV exposure time.
As mentioned in last section, a 5 log-reduction of E. coli is generally required for
wastewater reclamation. But the maximum log-reduction of both total coliform and E. coli
achieved in this experiment was about 3. Hence, after disinfection with UV directly, the
wastewater may still be unable to meet the required guidelines. However, if the bacteria
attached to the particulates could be removed, the maximum log-reduction could be
increased further and it is possible to meet the required guidelines. In practice, the particles
in wastewater can be removed through sedimentation or other ways. Properly designed
primary sedimentation tank can remove 50-70% suspended particles (Metcalf & Eddy,
2003). The turbidity of wastewater after primary sedimentation tank (primary effluent) can
be as low as 54.62 NTU (Ravazzini et al., 2005). Sometimes, even the turbidity of raw
sewage can be as low as 51 NTU (Bukhari, 2008). Figure 4.9 shows that the log-reduction
of total coliform can be as high as 4.12 when the turbidity of wastewater is 57 NTU. Thus,
if a proper pretreatment is applied (e.g. sedimentation), wastewater may reduce the
turbidity significantly and UV irradiation might be a possible way to disinfect the
wastewater.
-5.0
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160
Log-
red
uct
ion
Exposure time (s)
72 NTU 130 NTU Linear (clean water)
55
Chapter 5
Conclusions and Recommendations
5.1 Conclusions
This study focused on the disinfection performance of UV LEDs on turbid wastewater.
Both synthetic wastewater and real wastewater were used to conduct the disinfection test.
In synthetic wastewater, only E. coli was selected as the target organism, while both total
coliform and E. coli in real wastewater have been selected as the challenge organisms. The
conclusions for this study are as follows:
1. The reactor built for this experiment has met all requirements for conducting
disinfection test with both synthetic and real wastewater. 5 mL was a suitable
wastewater sample size for conducting disinfection test with UV LEDs.
2. The irradiance of UV LEDs measured by iodide-iodate actinometer was 0.4282
mW/cm2, while that value of ferrioxalate actinometer was 0.382 mW/cm
2. It
suggests that different actinometry system may lead to different results.
3. UV inactivation of E. coli in synthetic wastewater was a first order reaction and
there is no tailing effect. The log-reduction of E. coli could be increased by
prolonging the UV exposure time, regardless of the turbidity of synthetic wastewater.
4. With the increase of turbidity, E. coli in synthetic wastewater became less sensitive
to UV irradiation.
5. Slightly changes for a given turbid wastewater did not change the UV disinfection
performance in real wastewater.
6. Measurement of turbidity and suspended solids of real wastewater did not show any
direct correlation. This could be mainly due to the variation of particle size
distribution and its composition. As disinfection efficiency is more related to
particle size distribution than turbidity, there is need to measure this parameter.
7. The UV inactivation kinetics of total coliform in real wastewater was not completely
first order reaction. The tailing effect was observed due to the interaction of
coliform bacteria with particles. A maximum of ~ 3 log-reductions for total coliform
in real wastewater was observed, which could not be further increased by
prolonging the UV exposure time.
8. The inactivation kinetics of E. coli in real wastewater was similar with that of total
coliform. A maximum ~ 3 log-reduction was achieved and could not be increased
further with an increase in the exposure time.
9. The amount of coliform bacteria attached to the particles is the limiting factor for
further improvement of disinfection efficiency of UV LEDs. Higher disinfection
efficiency could be achieved if the coliform bacteria attached to the particles can be
removed by a proper way.
56
10. UV LEDs were not able to disinfect the real wastewater effectively at higher
turbidities. However, if coupled with proper treatment technologies to reduce
wastewater turbidity, UV LEDs should be able to disinfect real wastewater
effectively and meet the required WHO guidelines.
11. In some emergent cases (e.g. after earthquake), UV LEDs can be still be used for
disinfecting the wastewater before discharge so that the risks exposed to the
environment by wastewater can be reduced significantly.
5.2 Recommendations for Further Study
Based on the experience from this study, following recommendations are for the further
study:
1. UV light have a range of 200 - 400 nm wavelength. Thus assessing the effective
UV wavelength is vital as their disinfection performance on turbid wastewater
might vary greatly. Thus, disinfection performances of various UV lights should be
checked, especially the UV LED with wavelength of 265 nm as its wavelength is
closer to the germicidal peak (264 nm).
2. Scaling up the reactor from lab scale to bench scale and later full scale is necessary
to establish both technological and economical feasibility of the this treatment
system.
3. In practice, UV irradiation was applied to disinfect the water or wastewater in flow-
through condition, which is different from the batch reactor used in this study.
Previous study has already proved that the disinfection performance of UV LEDs in
flow-through conditions is quite different from that of batch mode. Moreover, the
behavior of particles in flow condition is also different from the behavior in batch
mode. Hence, a study on the disinfection performance of UV LEDs in flow-through
reactor is deemed necessary for real world application.
4. Previous studies on conventional UV lamps shown that their disinfection
efficiencies on different microorganisms were different. UV LEDs may have the
same properties. Hence, other kinds of microorganisms (e.g. protozoa and virus)
should also be used to conduct the disinfection test with UV LEDs.
5. Coupling low cost technologies like woven fiber membranes to remove turbidity
and UV LEDs to reduce pathogen inactivation should be conducted. As UV LEDs
cannot be used as a standalone technology for wastewater treatment. To increase
the overall treatment efficiency coupling technologies becomes predominant for
economical use.
57
References
Autin, O., Romelot, C., Rust, L., Hart, J., Jarvis, P., MacAdam, J., Parsons, S. A., and
Jefferson, B. (2013). Evaluation of a UV-light emitting diodes unit for the removal
of micropollutants in water for low energy advanced oxidation processes.
Chemosphere, 92(6), 745-751.
Baum, R., Luh, J., and Bartram, J. (2013). Sanitation: A Global Estimate of Sewerage
Connections without Treatment and the Resulting Impact on MDG Progress.
Environmental Science & Technology, 47(4), 1994-2000.
Bitton, G., Henis, Y., and Lahav, N. (1972). Effect of several clay minerals and humic acid
on the survival of Klebsiella aerogenes exposed to ultraviolet irradiation. Applied
microbiology, 23(5), 870-874.
Bolton, J. R., and Linden, K. G. (2003). Standardization of methods for fluence (UV dose)
determination in bench-scale UV experiments. Journal of Environmental
Engineering, 129(3), 209-215.
Bolton, J. R., Stefan, M. I., Shaw, P.-S., and Lykke, K. R. (2011). Determination of the
quantum yields of the potassium ferrioxalate and potassium iodide–iodate
actinometers and a method for the calibration of radiometer detectors. Journal of
Photochemistry and Photobiology A: Chemistry, 222(1), 166-169.
Bowker, C., Sain, A., Shatalov, M., and Ducoste, J. (2011). Microbial UV fluence-
response assessment using a novel UV-LED collimated beam system. Water
Research, 45(5), 2011-2019.
Bukhari, A. A. (2008). Investigation of the electro-coagulation treatment process for the
removal of total suspended solids and turbidity from municipal wastewater.
Bioresource technology, 99(5), 914-921.
Byamukama, D., Kansiime, F., Mach, R. L., and Farnleitner, A. H. (2000). Determination
of Escherichia coli contamination with chromocult coliform agar showed a high
level of discrimination efficiency for differing fecal pollution levels in tropical
waters of Kampala, Uganda. Applied and Environmental Microbiology, 66(2), 864-
868.
Cantwell, R. E., Hofmann, R., and Templeton, M. R. (2008). Interactions between humic
matter and bacteria when disinfecting water with UV light. Journal of applied
microbiology, 105(1), 25-35.
Caron, E., Cheurefils, G., Barbeau, B., Payment, P., and Prevost, M. (2007). Impact of
microparticles on UV disinfection of indigenous aerobic spores. Water Research,
41(19), 4546-4556.
Chevremont, A. C., Farnet, A. M., Coulomb, B., and Boudenne, J. L. (2012). Effect of
coupled UV-A and UV-C LEDs on both microbiological and chemical pollution of
urban wastewaters. Science of the Total Environment, 426, 304-310.
58
Chevremont, A. C., Farnet, A. M., Sergent, M., Coulomb, B., and Boudenne, J. L. (2012).
Multivariate optimization of fecal bioindicator inactivation by coupling UV-A and
UV-C LEDs. Desalination, 285(0), 219-225.
Craik, S. A., Finch, G. R., Bolton, J. R., and Belosevic, M. (2000). Inactivation of Giardia
muris cysts using medium-pressure ultraviolet radiation in filtered drinking water.
Water Research, 34(18), 4325-4332.
Craik, S. A., Weldon, D., Finch, G. R., Bolton, J. R., and Belosevic, M. (2001).
Inactivation of cryptosporidium parvum oocysts using medium- and low-pressure
ultraviolet radiation. Water Research, 35(6), 1387-1398.
DOWA (2013). Products instruction. Retrieved May 1, 2014, from
http://www.dowa.co.jp/index_e.html
Eddy, M., Tchobanoglous, G., Burton, F. L., and Stensel, H. D. (2003). Wastewater
engineering: Treatment and reuse (4th ed.). Boston: McGraw-Hill. ISBN:007-
124140-X.
Edzwald, J. K. (1987). Coagulation-sedimentation-filtration processes for removing
organic substances from drinking water. Control of Organic Substances in Water
and Wastewater,( pp 26-64). Noyes Data Corporation, Park Ridge New Jersey.
Emperor Aquatics, I. (2013). Effective Disinfection and Chlramines removal. Retrieved
January 18, 2014, from http://www.emperoraquatics-pool.com/sterilization.php.
Goldstein, S., and Rabani, J. (2008). The ferrioxalate and iodide–iodate actinometers in the
UV region. Journal of Photochemistry and Photobiology A: Chemistry, 193(1), 50-
55.
Hamamoto, A., Mori, M., Takahashi, A., Nakano, M., Wakikawa, N., Akutagawa, M.,
Ikehara, T., Nakaya, Y., and Kinouchi, Y. (2007). New water disinfection system
using UVA light‐emitting diodes. Journal of applied microbiology, 103(6), 2291-
2298.
Hayward, K. (2013). UV LEDs light the way for disruptive technologies. Water 21, 41-42.
Hijnen, W. A. M., Beerendonk, E. F., and Medema, G. J. (2006). Inactivation credit of UV
radiation for viruses, bacteria and protozoan (oo)cysts in water: A review. Water
Research, 40(1), 3-22.
Hu, J. Y., and Quek, P. H. (2008). Effects of UV radiation on photolyase and implications
with regards to photoreactivation following low-and medium-pressure UV
disinfection. Applied and Environmental Microbiology, 74(1), 327-328.
Kadam, A. M., Oza, G. H., Nemade, P. D., and Shankar, H. S. (2008). Pathogen removal
from municipal wastewater in constructed soil filter. Ecological engineering, 33(1),
37-44.
59
Kollu, K., and Örmeci, B. (2012). Effect of particles and bioflocculation on ultraviolet
disinfection of Escherichia coli. Water Research, 46(3), 750-760.
Linden, K. G., Shin, G., and Sobsey, M. D. (2001). Comparative effectiveness of UV
wavelengths for the inactivation of Cryptosporidium parvum oocysts in water.
Water Science and Technology, 43(12), 171-174.
Liu, W., Andrews, S. A., Bolton, J. R., Linden, K. G., Sharpless, C., and Stefan, M. (2002).
Comparison of disinfection byproduct (DBP) formation from different UV
technologies at bench scale. Water Supply, 2(5-6), 515-521.
Mamane, H., Ducoste, J. J., and Linden, K. G. (2006). Effect of particles on ultraviolet
light penetration in natural and engineered systems. Applied Optics, 45(8), 1844-
1856.
Masaru, N., and Miho, T. (2013). It will take a few years for wastewater infrastructure
rehabilitation, serious damages especially in coastal areas; Japanese: 下水道復旧
は数年かかる見通し 沿岸部に集中、被害深刻. Retrieved December 10,
2013, from http://www.asahi.com/special/10005/TKY201104020353.html
Molleda, P., Blanco, I., Ansola, G., and de Luis, E. (2008). Removal of wastewater
pathogen indicators in a constructed wetland in Leon, Spain. Ecological
engineering, 33(3), 252-257.
Mosher, J. J., and Gina, M. V. (2012). Ultraviolet disinfection guidelines for drinking
water. Fountain Valley, California, USA: National Water Research Institute.
ISBN:NWRI-2012-04.
Oguma, K., Izaki, K., and Katayama, H. (2013). Effects of salinity on photoreactivation of
Escherichia coli after UV disinfection. Journal of Water & Health, 11(3).
Oguma, K., Katayama, H., and Ohgaki, S. (2002). Photoreactivation of Escherichia coli
after low-or medium-pressure UV disinfection determined by an endonuclease
sensitive site assay. Applied and Environmental Microbiology, 68(12), 6029-6035.
Oguma, K., Katayama, H., and Ohgaki, S. (2004). Photoreactivation of Legionella
pneumophila after inactivation by low- or medium-pressure ultraviolet lamp. Water
Research, 38(11), 2757-2763.
Oguma, K., Katayama, H., and Ohgaki, S. (2005). Spectral impact of inactivating light on
photoreactivation of Escherichia coli. Journal of Environmental Engineering and
Science, 4(S1), S1-S6.
Oguma, K., Kita, R., Sakai, H., Murakami, M., and Takizawa, S. (2013). Application of
UV light emitting diodes to batch and flow-through water disinfection systems.
Desalination, 328, 24-30.
Olstadt, J., Schauer, J., Standridge, J., and Kluender, S. (2007). A comparison of ten
USEPA approved total coliform/E. coli tests. Journal of water and health, 5(2),
267-282.
60
Passantino, L., Malley, J. R., Knudson, M., Ward, R., and Kim, J. (2004). Effect of low
turbidity and algae on UV disinfection performance. Journal of American Water
Works Association, 96(6), 128-137.
Pilgrim, N., Roche, B., Kalbermatteni, J., Revels, C., and Kariuki, M. (2008). Town water
supply and sanitation: challenges, solutions, and guidelines. Retrieved December
3, 2013, from http://documents.worldbank.org/curated/en/2008/06/ 9677417/town-
water-supply-sanitation-challenges-solutions-guidelines
Qualls, R. G., Flynn, M. P., and Johnson, J. D. (1983). The role of suspended particles in
ultraviolet disinfection. Journal (Water Pollution Control Federation), 1280-1285.
Rahn, R. O. (1997). Potassium Iodide as a Chemical Actinometer for 254 nm Radiation:
Use of lodate as an Electron Scavenger. Photochemistry and Photobiology, 66(4),
450-455.
Rahn, R. O. (2013). Fluence Measurements Employing Iodide/Iodate Chemical
Actinometry as Applied to Upper‐Room Germicidal Radiation. Photochemistry
and Photobiology, 89(4), 816-818.
Rahn, R. O., Bolton, J. R., and Stefan, M. I. (2006). The lodide/lodate actinometer in UV
disinfection: Determination of the fluence rate distribution in UV reactors.
Photochemistry and Photobiology, 82(2), 611-615.
Rahn, R. O., Stefan, M. I., Bolton, J. R., Goren, E., Shaw, P. S., and Lykke, K. R. (2003).
Quantum Yield of the Iodide–Iodate Chemical Actinometer: Dependence on
Wavelength and Concentration. Photochemistry and Photobiology, 78(2), 146-152.
Ravazzini, A., Van Nieuwenhuijzen, A., and Van Der Graaf, J. (2005). Direct
ultrafiltration of municipal wastewater: comparison between filtration of raw
sewage and primary clarifier effluent. Desalination, 178(1), 51-62.
Seoul Optodevice, I. (2013). Product instruction. Retrieved from http://socled.com/en/
product/categorys.asp?catecode= 1002003
Severin, B. F., Suidan, M. T., and Engelbrecht, R. S. (1983). Effect of temperature on
ultraviolet light disinfection. Environmental Science & Technology, 17(12), 717-
721.
Templeton, M. R., Andrews, R. C., and Hofmann, R. (2005). Inactivation of particle-
associated viral surrogates by ultraviolet light. Water Research, 39(15), 3487-3500.
US EPA. (2006). Ultraviolet disinfection guidance manual for the final long term 2
enhanced surface water treatment rule. Washington D.C., USA: US EPA.
Vrhovšek, D., Kukanja, V., and Bulc, T. (1996). Constructed wetland (CW) for industrial
waste water treatment. Water Research, 30(10), 2287-2292.
61
Wu, Y., Clevenger, T., and Deng, B. (2005). Impacts of goethite particles on UV
disinfection of drinking water. Applied and Environmental Microbiology, 71(7),
4140-4143.
Wurtele, M. A., Kolbe, T., Lipsz, M., Kulberg, A., Weyers, M., Kneissl, M., and Jekel, M.
(2011). Application of GaN-based ultraviolet-C light emitting diodes - UV LEDs -
for water disinfection. Water Research, 45(3), 1481-1489.
Zhang, K., and Farahbakhsh, K. (2007). Removal of native coliphages and coliform
bacteria from municipal wastewater by various wastewater treatment processes:
implications to water reuse. Water Research, 41(12), 2816-2824.
Zimmer, J. L., and Slawson, R. M. (2002). Potential repair of Escherichia coli DNA
following exposure to UV radiation from both medium-and low-pressure UV
sources used in drinking water treatment. Applied and Environmental Microbiology,
68(7), 3293-3299.
62
Appendix A
Experimental data
63
The log-reduction of E. coli in clean water at different exposure time was presented in this
part. Pour plate technique has been applied to count the density of E. coli, and the agar for
pour plate was Chromocult Coliform agar.
Table B.1 Log-reduction of E. coli in Clean Water at 15.37 & 36.62 s
Time for experiment Mar. 4th
, 2014
UV exposure time (s) 0 15.37 36.62
Density of E. coli ( × 107 CFU/ mL) 5.3 1.04 0.0073
log-reduction 0 -0.71 -2.86
Table B.2 Log-reduction of E. coli in Clean Water at 20.28 & 40.28 s
Time for experiment Mar. 6th
, 2014
UV exposure time (s) 0 20.28 40.28
Density of E. coli (× 107 CFU/ mL) 1.48 0.15 0.00013
log-reduction 0 -0.99 -4.07
Table B.3 Log-reduction of E. coli in Clean Water at 20.28 & 40.28 s
Time for experiment Mar. 8th
, 2014
UV exposure time (s) 0 20.28 40.28
Density of E. coli (× 107 CFU/ mL) 1.48 0.23 error
log-reduction 0 -1.16 error
Table B.4 Log-reduction of E. coli in Clean Water at 17.36 & 37.28 s
Time for experiment Mar. 9th
, 2014
UV exposure time (s) 0 17.36 37.28
Density of E. coli (× 107 CFU/ mL) 6 1.07 0.00069
log-reduction 0 -0.75 -3.94
Table B.5 Log-reduction of E. coli in Clean Water at 24.31, 28.28 & 32.31 s
Time for experiment Mar. 11th
, 2014
UV exposure time (s) 0 24.31 28.28 32.31
Density of E. coli (× 107 CFU/ mL) 4.2 0.053 0.0096 0.019
log-reduction 0 -1.9 -2.64 -2.35
Table B.6 Log-reduction of E. coli in Clean Water at 24.28, 28.34 & 33.16 s
Time for experiment Mar. 12th
, 2014
UV exposure time (s) 0 24.28 28.34 33.16
Density of E. coli (× 107 CFU/ mL) 6.1 0.088 0.082 0.0046
log-reduction 0 -1.84 -1.87 -3.12
64
The log-reduction of E. coli in synthetic wastewater of 70 NTU were presented here and
montmorillonite was used to synthetize the turbid wastewater.
Table B.7 Log-reduction of E. coli in Synthetic Wastewater of 70 NTU (1st trial)
Time for experiment Mar. 14th
, 2014
UV exposure time (s) 0 75.16 125.16
Density of E. coli (× 107 CFU/ mL) 5.5 0.00018 0.000002
log-reduction 0 -4.8 >7
Table B.8 Log-reduction of E. coli in Synthetic Wastewater of 70 NTU (2nd
trial)
Time for experiment Mar. 15th
, 2014
UV exposure time (s) 0 32.03 42.56 52.09 59.31 77.31
Density of E. coli (× 107 CFU/ mL) 3.9 0.047 0.067 0.018 0.0075 0.0035
log-reduction 0 -2.22 -2.07 -2.63 -3.01 -3.35
Table B.9 Log-reduction of E. coli in Synthetic Wastewater of 70 NTU (3rd
trial)
Time for experiment Mar. 17th
, 2014
UV exposure time (s) 0 15.50 25.25 60.03 70.91
Density of E. coli (× 107 CFU/ mL) 4.6 >3 2.82 0.2 0.045
log-reduction 0 <0.48 -0.51 -1.67 -2.31
In this section, the data of disinfection test on synthetic wastewater of 156 NTU were
presented.
Table B.10 Log-reduction of E. coli in Synthetic Wastewater of 156 NTU (1st trial)
Time for experiment Mar. 18th
, 2014
UV exposure time (s) 0 21.25 30.41 48 62.19 71.38
Density of E. coli (× 107 CFU/ mL) 5.6 2.19 0.9 0.28 0.25 0.21
log-reduction 0 -0.41 -0.79 -1.30 -1.34 -1.42
Table B.11 Log-reduction of E. coli in Synthetic Wastewater of 156 NTU (2nd
trial)
Time for experiment Mar. 20th
, 2014
UV exposure time (s) 0 40.28 52.18 80.91 92.25 100.28
Density of E. coli (× 107 CFU/ mL) 0.74 0.26 0.21 0.049 0.015 0.0069
log-reduction 0 -0.45 -0.54 -1.18 -1.70 -2.03
65
The result of disinfection test with synthetic wastewater of 113 NTU were presented here.
Table B.12 Log-reduction of E. coli in Synthetic Wastewater of 113 NTU (1st trial)
Time for experiment Mar. 23rd
, 2014
UV exposure time (s) 0 30 50.03 70.19 93.84 109.88
Density of E. coli (× 107 CFU/ mL) 1 0.96 0.65 0.53 0.015 0.0073
log-reduction 0 -0.018 -0.19 -0.28 -1.84 -2.14
Table B.13 Log-reduction of E. coli in Synthetic Wastewater of 113 NTU (2nd
trial)
Time for experiment Mar. 25th
, 2014
UV exposure time (s) 0 69.97 89.88 110.34 135.78 150.47
Density of E. coli (× 107 CFU/ mL) 0.92 0.028 0.0129 0.0078 0.00017 0.0002
log-reduction 0 -1.52 -1.85 -2.07 -3.75 -3.66
The result of disinfection test with synthetic wastewater of 113 NTU were presented here.
Table B.14 Log-reduction of E. coli in Synthetic Wastewater of 27 NTU (1st trial)
Time for experiment Mar. 27th
, 2014
UV exposure time (s) 0 21.28 30.19 40.78 59.47 70.59
Density of E. coli (× 107 CFU/ mL) 2.9 1.16 0.81 0.19 0.016 0.0086
log-reduction 0 -0.4 -0.55 -1.18 -2.25 -2.53
Table B.15 Log-reduction of E. coli in Synthetic Wastewater of 27 NTU (2nd
trial)
Time for experiment Mar. 29th
, 2014
UV exposure time (s) 0 49.78 60.28 73.84 80.65 90.34
Density of E. coli (× 107 CFU/ mL) 4.4 0.073 0.036 0.01 0.00123 0.000125
log-reduction 0 -1.78 -2.08 -2.64 -3.55 -4.55
66
In this part, the result of disinfection test with real wastewater was presented, including the
result for total coliform and E. coli. The densities of both total coliform and E. coli were
counted by MPN method.
Table B.16 Log-reduction of total coliform in Real Wastewater of 86 NTU
Time for experiment April 10th, 2014
Turbidity (NTU) 86
exposure time (s) 0 31.65 55.4 72.41 90.81
total coliform (× 106 MPN/100 mL) 13 2.3 0.8 0.3 0.07
log-reduction 0 -0.75 -1.21 -1.64 -2.27
E. coli detection failed.
Table B.17 Log-reduction of total coliform in Real Wastewater of 57 NTU
Time for experiment April, 11th, 2014
Turbidity (NTU) 57
exposure time (s) 0 32.78 53.31 76.29 91.34 111.87
total coliform (× 106 MPN/100 mL) 17 5 0.8 0.23 0.07 0.0011
log-reduction 0 -0.53 -1.33 -1.87 -2.39 -4.19
E. coli detection failed.
Table B.18 Log-reduction of total coliform and E. coli in Real Wastewater of 130
NTU
Time for experiment April 14th, 2014
Turbidity (NTU) 130 NTU
exposure time (s) 0 40.25 70.34 90.22 119.97 149.72
total coliform (× 106 MPN/100 mL) 13 3 0.17 0.035 0.022 0.013
log-reduction 0 -0.64 -1.88 -2.57 -2.77 -3
E. coli (× 106 MPN/100 mL) 3 1.7 0.11 0.028 0.005 0.003
log-reduction 0 -0.27 -1.44 -2.03 -2.78 -3
Table B.19 Log-reduction of total coliform and E. coli in Real Wastewater of 72 NTU
Time for experiment April 17th
, 2014
Turbidity (NTU) 72
exposure time (s) 0 31.81 60.91 120.53 140.35
total coliform (× 106 MPN/100 mL) 5 1.3 0.09 0.009 0.0034
log-reduction 0 -0.59 -1.74 -2.74 -3.17
E. coli (× 106 MPN/100 mL) 2.2 0.35 0.005 0.0021 0.0034
log-reduction 0 -0.80 -2.64 -3.02 -2.81
67
Appendix B
Experimental set-up
68
Figure A.1 Details of experimental set-up