U UNIVERSITY OF CINCINNATI
Date: 05/14/2009
I, Rangesh Srinivasan , hereby submit this original work as part of the requirements for the degree of:
Doctorate of Philosophy (Ph.D) in Environmental Engineering
It is entitled: Treatment of Microcontaminants in Drinking Water
Rangesh Srinivasan Student Signature:
This work and its defense approved by:
Committee Chair: Dr. George A. Sorial Dr. Dionysios D. Dionysiou Dr. Mingming Lu Dr. E. Sahle Demessie
Approval of the electronic document:
I have reviewed the Thesis/Dissertation in its final electronic format and certify that it is an accurate copy of the document reviewed and approved by the committee.
Committee Chair signature: George A. Sorial Treatment of Microcontaminants in Drinking Water
A dissertation submitted to the
Division of Research and Advanced Studies of
the University of Cincinnati
in partial fulfillment of
the requirements for the degree of
DOCTORATE OF PHILOSOPHY (Ph.D.)
in the Department of Civil and Environmental Engineering of
the College of Engineering
2009
By
Rangesh Srinivasan
M.S. University of Cincinnati, Cincinnati, 2003
B.S. Civil Engineering, Anna University, Madras, India, 2000
Committee Chair: Dr. George A. Sorial
ABSTRACT
Most of the challenges in drinking water treatment arise out of contaminants that are
present in extremely low concentrations and have physical/chemical properties that make them
extremely difficult to be removed by conventional treatment methods. This study tries to look at
the treatment of two such microcontaminants – Perchlorate and MIB (Methyl Isoborneol)/
Geosmin.
- Perchlorate (ClO4 ) is a major inorganic contaminant in drinking water and has been
detected in a number of public drinking water systems across the country and has serious health
impacts associated with it. A complete and critical review on this intriguing contaminant is
presented including a detailed discussion on sources of contamination, the policy aspects
including regulation and the available treatment technologies and their feasibility. Although
some technologies have become more established than the others, it is clearly evident that a
single technology cannot be directly applied to a drinking water treatment system for compete
removal of perchlorate and it is highly likely that a combination of these technologies that would
have to be employed to overcome this challenge.
Although some of the removal technologies such as ion exchange are effective for
perchlorate, they invariable result in brines highly concentrated in perchlorate that pose further
challenges due to disposal. Hence a destruction technology that completely reduces perchlorate
to harmless chloride is preferred. This study investigates the potential removal of perchlorate ion in drinking water by combining electrochemical reduction with zero-valent iron reduction. The
I use of zero-valent iron in the process is expected to enhance electrochemical reduction of the perchlorate ion and the effectiveness of this technology will serve as a basis to further develop the proposed methodology to purify surface or ground water in target zones at a larger scale.
Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) and MIB (2-methyl isoborneol–
C11H20O) are organic semi-volatile chemicals that can seriously influence the finished quality of drinking water by imparting taste and odor to it even at extremely low concentrations. A critical review of these two taste and odor causing compounds in drinking water is presented with emphasis on their sources, health and regulatory implications and recent developments in their analysis. The relevant treatment alternatives are described in detail focusing on their respective advantages and problems associated with their implementation in a full-scale facility. Although some of these technologies are more effective and show more promise than the others, much work remains to be done in order to optimize these technologies so that they can be retrofitted or installed with minimal impact on the overall operation and effectiveness of the treatment system.
Studies have shown that activated carbon adsorption is the most effective technology currently available for treatment of these compounds. The impact of adsorbent pore size distribution on activated carbon adsorption of MIB and geosmin was evaluated through single solute and multicomponent adsorption of these compounds on three types of activated carbon fibers (ACFs) and one granular activated carbon (GAC). The effect of the presence of natural organic matter (NOM) on MIB and geosmin adsorption was also studied for both the single solute and binary systems. The single solute adsorption isotherms fit Myers and Freundlich equations reasonably well and pore size distribution significantly influenced adsorption on the
II ACFs and GAC. It was also seen that presence of NOM significantly reduced its adsorption capacity due to competitive adsorption.
The binary adsorption of MIB and geosmin on ACFs and GAC was well defined by the ideal adsorbed solute theory (IAST), which is a well established thermodynamic model for multicomponent adsorption. There were no significant differences in the binary isotherm between the oxic and anoxic conditions, indicating that adsorption was purely through physical adsorption and no oligomerization was taking place. Binary adsorptions in the presence of NOM did result in deviation from IAST behavior in case of two adsorbents.
III
IV Acknowledgement
I would like to express my deepest appreciation to my advisor, Dr. George A. Sorial, for his guidance, encouragement, and support throughout my years at the University of Cincinnati. I would also like to thank my committee members; Dr. Dionysios D. Dionysiou, Dr. Mingming Lu and Dr. E. Sahle Demessie for their feedback and for letting me use their laboratory facilities.
Special thanks to Choi, Venu and Yongjun for their valuable contribution to my research efforts.
I would like also to thank Charlie, Quili, Niranjan, Ashraf, Hafiz and Kavitha for their continuous support, help and cooperation. I would also thank all the people in the department for their help and friendship over the last few years. I acknowledge my deepest gratitude to my wife
Subha, my parents and my sister, for their unconditional love, faith and support.
V Table of Contents
Abstract...... i
Acknowledgement ...... v
Table of Contents ...... vi
List of Tables...... xi
List of Figures...... xiii
CHAPTER 1...... 1
INTRODUCTION...... 1
1.1 BACKGROUND...... 1 1.1.1 Perchlorate ...... 2 1.1.2 MIB/Geosmin ...... 5
1.2 SIGNIFICANCE OF THE STUDY...... 7 1.2.1 Sources and current regulation...... 7 1.2.1.1 Perchlorate ...... 7 1.2.1.2 MIB/Geosmin ...... 8 1.2.2 Treatment Technologies...... 9 1.2.2.1 Perchlorate ...... 9 1.2.2.2 Geosmin/MIB Treatment Technologies ...... 10
1.3 OBJECTIVES OF THE STUDY...... 11 1.3.1 Perchlorate ...... 11 1.3.2 MIB/Geosmin ...... 12
1.4 STRUCTURE OF DISSERTATION ...... 14
1.5 REFERENCES ...... 15
CHAPTER 2...... 23
MATERIALS AND METHODS ...... 23
VI 2.1 INTRODUCTION...... 23
2.2 MATERIALS...... 23 2.2.1 Perchlorate ...... 23 2.2.2 MIB/Geosmin ...... 24
2.3 EXPERIMENTAL METHOD...... 24 2.3.1 Perchlorate ...... 24 2.3.2 MIB/geosmin ...... 27
2.3 ANALYTICAL METHOD...... 28 2.3.1 Perchlorate ...... 28 2.3.2 MIB/Geosmin ...... 29
2.4 REFERENCES ...... 30
CHAPTER 3...... 31
TREATMENT OF PERCHLORATE IN DRINKING WATER: A CRITICAL REVIEW31
3.1 ABSTRACT...... 31
3.2 INTRODUCTION...... 31
3.3 SOURCES ...... 33
3.4 CONTAMINATION ...... 35
3.5 ANALYSIS ...... 37
3.6 REGULATIONS ...... 40
3.7 EXPOSURE/ HEALTH EFFECTS ...... 42
3.8 TREATMENT ...... 44 3.8.1 Activated Carbon ...... 45 3.8.2 Ion Exchange ...... 47 3.8.3 Membrane Technologies...... 51 3.8.4 Chemical Reduction...... 53 3.8.5 Electrochemical Reduction ...... 57 3.8.6 Microbial Remediation ...... 59 3.8.7 Integrated Technologies...... 67
3.9 CONCLUSION...... 73
3.10 REFERENCES ...... 76
VII CHAPTER 4...... 93
REMOVAL OF PERCHLORATE AND CHLORATE IN AQUATIC SYSTEMS USING INTEGRATED TECHNOLOGIES...... 93
4.1 ABSTRACT...... 93
4.2 INTRODUCTION...... 94
4.3 MATERIALS AND METHODS ...... 97 4.3.1 Materials ...... 97 4.3.2 Analytical Method ...... 97 4.3.2.1 Perchlorate/Chlorate concentration...... 97 4.3.2.2 Iron characterization ...... 98 4.3.3 Acid Washing of Iron Filings ...... 98
4.3.4 TiO2 coating of electrodes ...... 99 4.3.5 Coating of electrodes with iron oxide sol-gel...... 99 4.3.6 Preparation of heterogeneous Rhenium-Palladium-Carbon catalyst ...... 99 4.3.7 Experimental Setup...... 99 4.3.8 Experimental Method...... 100
4.4 RESULTS AND DISCUSSIONS...... 101 4.4.1 Perchlorate Removal...... 101 4.4.2 Chlorate...... 106 4.4.2.1 Chlorate Batch Experiments ...... 106 4.4.2.2 Electrochemical Reduction of Chlorate...... 107
4.5 CONCLUSIONS ...... 118
4.6 ACKNOWLEDGEMENTS...... 119
4.7 REFERENCES ...... 120
CHAPTER 5...... 126
TREATMENT OF TASTE AND ODOR CAUSING COMPOUNDS IN DRINKING WATER: ...... 126
5.1 ABSTRACT...... 126
5.2 INTRODUCTION...... 127
5.3 SOURCES AND CONTAMINATION OF MIB/GEOSMIN IN WATER...... 129
VIII 5.4 DEVELOPMENT IN ANALYSIS OF MIB/GEOSMIN IN WATER SAMPLES...... 131
5.5 HEALTH EFFECTS/REGULATION...... 133
5.6 TREATMENT TECHNOLOGIES...... 134 5.6.1 GAC Adsorption ...... 135 5.6.2 Advanced Oxidation Processes (AOP) ...... 140 5.6.3 Bioremediation...... 144 5.6.4 Integrated Technologies...... 146 5.6.5 Other Novel Treatment Methods ...... 147
5.7 CURRENT STATUS ...... 148
5.8 CONCLUSION...... 149
5.9 REFERENCES ...... 152
CHAPTER 6...... 161
SINGLE SOLUTE ADSORPTION OF MIB AND GEOSMIN ON ACTIVATED CARBON FIBERS ...... 161
6.1 ABSTRACT...... 161
6.2 INTRODUCTION...... 162
6.3 MATERIALS AND METHODS ...... 164 6.3.1 Adsorbates...... 164 6.3.2 Adsorbents ...... 164 6.3.3 Adsorption Isotherm Procedure ...... 164 6.3.4 Analytical procedure...... 166
6.4 RESULTS AND DISCUSSION...... 167
6.5 CONCLUSION...... 174
6.6 REFERENCES ...... 175
CHAPTER 7...... 177
ADSORPTION OF GEOSMIN AND MIB ON ACTIVATED CARBON – SINGLE AND BINARY SOLUTE SYSTEM ...... 177
7.1 ABSTRACT...... 177
7.2 INTRODUCTION...... 178
IX 7.3 EXPERIMENTAL ...... 181 7.3.1 Materials ...... 181 7.3.1.1 Adsorbates...... 181 7.3.1.2 Adsorbents ...... 181 7.3.2 Methods...... 181 7.3.2.1 Adsorption Isotherm Procedure ...... 181 7.3.3 Analytical procedure...... 182
7.4 RESULTS AND DISCUSSION...... 183 7.4.1 Single solute adsorption...... 183 7.4.2 Binary adsorption...... 188
7.5 CONCLUSION...... 197
7.6 REFERENCES ...... 198
CHAPTER 8...... 202
CONCLUSIONS AND RECOMMENDATIONS...... 202
8.1 CONCLUSIONS ...... 202 8.1.1 Perchlorate ...... 202 8.1.2 MIB/ Geosmin ...... 203
8.2 RECOMMENDATIONS FOR FUTURE RESEARCH...... 204 8.2.1 Perchlorate ...... 204 8.2.2 MIB/ Geosmin ...... 205
X List of Figures
Figure 2.1 Schematic of the Experimental Setup…………………………………………...……25
Figure 3.1 Perchlorate Contamination Sites………………………………………………….….35
Figure 3.2 Structure of the Perchlorate Ion………………………………………………………54
Figure 3.3 Factors Influencing Microbial Perchlorate Reduction…………………………….....61
Figure 4.1 Electrochemical reduction of perchlorate with iron filings as cathode……………..103
Figure 4.2 Perchlorate reduction in a zero-valent iron column (C0 : 10 mg/L)………………...103
Figure 4.3 Electrochemical reduction of perchlorate with various electrodes………………….104
Figure 4.4. a. Chlorate degradation (C0: 10 mg/L) with iron filings at two iron dosages with and
without DO removal; b. Zero order reaction kinetics for chlorate reduction………107
Figure 4.5 Electrochemical reduction of chlorate with various electrodes……………………..109
Figure 4.6 Electrochemical reduction of chlorate with ‘glassy carbon’ electrode……………..109
Figure 4.7 Electrochemical reduction of chlorate with Ni electrode spray-coated with TiO2….111
Figure 4.8 Electrochemical reduction of chlorate with Ni electrode dip-coated with TiO2……113
Figure 4.9 X-Ray Diffraction spectrum of zero-valent iron……………………………………116
Figure 5.1 Molecular structure of MIB and geosmin…………………………………………..128
Figure 5.2 Pathway of MIB/ Geosmin Formation……………………………………………...130
Figure 6.1 MIB adsorption isotherms in organic free water and humic acid solutions………...169
Figure 6.2 Comparison of MIB adsorption isotherms………………………………………….170
Figure 6.3 Geosmin adsorption isotherms in organic free water and humic acid solutions……171
Figure 6.4 Comparison of Geosmin adsorption isotherms……………………………………..173
Figure 7.1 Single solute adsorption isotherm of MIB…………………………………………..186
Figure 7.2 Single solute adsorption isotherm of geosmin………………………………………187
XI Figure 7.3 Binary adsorption isotherms of MIB and Geosmin on ACC-15……………………192
Figure 7.4 Binary adsorption isotherms of MIB and Geosmin on F-400………………………193
Figure 7.5 Binary adsorption isotherms of MIB and Geosmin on ACC-20……………………194
Figure 7.6 Binary adsorption isotherms of MIB and Geosmin on ACC-25……………………195
Figure 7.7 Binary adsorption isotherms of MIB/Geosmin in the presence of humic acid....…..196
XII List of Tables
Table 3.1 Standard Methods for Analysis of Perchlorate in Water……………………………..38
Table 3.2 Recent Standard Methods for Analysis of Perchlorate in Water……………………..41
Table 3.3 GAC Adsorption of Perchlorate………………………………………………………47
Table 3.4 Perchlorate Removal with Ion-Exchange……………………………………………..51
Table 3.5 Perchlorate Treatment with Membrane Technologies………………………………..52
Table 3.6 Chemical/ Catalytic reduction of perchlorate…………………………………………55
Table 3.7 Electrochemical reduction of perchlorate……………………………………………..59
Table 3.8 Microbial Perchlorate Reduction……………………………………………………...62
Table 3.9 Perchlorate removal with Integrated technologies…………………………………....73
Table 4.1 First order reaction rates (hr-1) for electrochemical reduction of chlorate …………..106
Table 5.1 Activated carbon adsorption of MIB/geosmin……………………………………….138
Table 5.2 Removal of MIB/geosmin by AOPs…………………………………………………142
Table 5.3 Biological removal of MIB/geosmin………………………………………………...145
Table 6.1. Physical properties of the adsorbents used in the study……………………….…….165
Table 6.2. Freundlich constants for MIB and geosmin adsorption …………………………….168
Table 7.1. Myers isotherm equation parameters………………………………………………..188
XIII CHAPTER 1
Introduction
1.1 Background
Most of the water treatment unit processes followed in treatment plants, at least in the
developed world have been well established over the years. Consequently, for most people,
treatment and supply of drinking water seems to be a rather uncomplicated and straightforward
process. But for people actively involved in this field, either in research or in designing and
operating water treatment plants, there are certain issues arising out of the presence of certain contaminants that can significantly influence the quality of the treated water. These types of contaminants are invariably present in extremely low concentrations and have physical/chemical
properties that make them extremely difficult to be removed by conventional treatment methods
that are already in place [1]. This study tries to look at the treatment of two such microcontaminants. The first one is perchlorate which is an inorganic contaminant that has serious health impacts associated with it and is very challenging to remove. The second type of contaminants, MIB and Geosmin are organic semi-volatile chemicals that can seriously influence the finished quality of drinking water by imparting taste and odor even at extremely low concentrations.
1.1.1 Perchlorate
In the last few years, perchlorate has become a major inorganic contaminant in drinking
water and has been detected in a number of public drinking water systems, especially in the
southwest. Perchlorate and its salts are mainly used in solid propellants and for manufacturing
matches, rockets, explosives and fireworks. It also occurs naturally. Perchlorate contamination in
drinking water and food supplies has recently become a concern for human health, as studies
have shown that it can interfere with the normal iodine uptake by the thyroid gland, resulting in decreased thyroid production. Thyroid is important for normal growth, development and metabolism in the human body and the effects can be significant in case of pregnant women and fetuses. Application of a sensitive new detection method has revealed widespread perchlorate contamination of groundwater, particularly in the southwestern United States. Perchlorate contamination has also become an issue in some surface waters, such as Lake Mead, which is the primary source of drinking water for Las Vegas and southern California. Although the U.S.EPA has not established maximum contamination level and maximum contamination level goal for perchlorate at this time, several states have already established their action levels against perchlorate contamination in their drinking water supply.
- The perchlorate anion (ClO4 ) in its salt form is extremely soluble in both water and polar
organic solvents. It is exceedingly mobile and persistent in groundwater under typical
environmental conditions. Most standard physical and chemical water and wastewater treatment
processes are not generally applicable to remove or destroy the perchlorate ion [2]. The standard
redox potential of perchlorate is +1.38 V [3], which renders it a strong oxidizing agent; however,
due to its high activation energy of 120 kJ mol-1 [4] its oxidizing power is retarded. The high
2 activation energy is mainly caused by the tetrahedron structure of perchlorate in which chlorine
is surrounded by four oxygen atoms. As a result, perchlorate is extremely slow to react and
cannot be reduced with common reducing agents or be precipitated using commonly available
treatment reagents.
Currently, perchlorate treatment technologies are generally classified into two categories, destruction technologies or removal technologies. The destructive processes can be divided further into two sub categories: biological and chemical treatments [5]. The non-destructive removal processes or physical processes mainly focus on ion exchange resins. Physical removal processes are suitable for perchlorate removal in relatively small scale water treatment plants.
However, such processes have some disadvantages, such as lack of selectivity as well as the generation of a concentrated residue which requires further treatment. As a result, destructive processes which reduce perchlorate to harmless chloride are preferred.
The pathway of perchlorate reduction has been studied by many researchers ([4, 6, 7] and
they all agree that perchlorate will lose an oxygen atom to form chlorate at the first or rate
limiting step in the reduction process; further reduction kinetics to the chloride ion is very fast.
Perchlorate reduction by biological processes under anaerobic conditions has been widely
investigated for many years ([8-11]. It is postulated that special enzymes in the microorganisms’
cells serve as catalysts during the degradation of perchlorate, which effectively decrease the
activation energy of perchlorate reduction. Due to the effectiveness of biological processes, the
U.S. Army and Air Force have already installed perchlorate treatment facilities at several
military manufacturing factories [8].
3 However, in the case of drinking water treatment and groundwater remediation,
biological treatment has its own limitations due to the presence of unknown pathogens and the
requirement of nutrient addition. An effective chemical reduction process which is favorable for
drinking water treatment and groundwater remediation is, therefore, necessary. Many common
reducing reagents like ferrous ion [12] and hydrogen gas [7] have been tested without observable
perchlorate reduction. The key is to find a way to overcome the energy barrier. The answer may
be found at the surfaces of certain minerals. Experimental evidence from the last two decades has
shown that aqueous mineral surfaces provide additional accelerated pathways for certain redox
processes, such as the oxygenation of metal ions and the oxidation of organic pollutants [13].
The reaction mechanism at the aquatic mineral interface is an active topic as the reaction is
complicated and the mechanism is not completely understood until now, although it has already
been applied in the field and in the laboratory, especially with respect to the use of zero valent
iron. For example, zero valent iron has been studied for its capability to reduce
chlorohydrocarbon ([14, 15]. The most common application of zero valent iron is in packing iron
filings as a permeable reactive barrier. Some studies have been conducted on the removal of
perchlorate on iron surface. Lang and Horanyi reviewed the conversion of perchlorate at
different metal surfaces and found iron to have the fastest reaction rate [7]. Gurol and Kim, and
Moore and coworkers discovered removal of perchlorate in water by cast iron filings respectively [12, 16]. Moore and coworkers tried to confirm the reduction effect of iron fillings by chlorine mass balance. Huang studied the effect of pH and water hardness on the kinetics of perchlorate reduction by zero-valent iron by conducting batch and column experiments [17].
Through batch and column experiments, he was able to determine reaction kinetics and understand perchlorate reduction mechanism at the iron surface and was able to demonstrate that
4 iron serves as a reducing agent and a catalyst simultaneously during the reaction at the iron
surface. Although iron filings were effective in reducing perchlorate, the reaction rates were
extremely slow and were significantly affected by the presence of other common anions. This
project investigates the possibility of removal of perchlorate ion in drinking water by enhancing
zero-valent iron reduction with the use of catalysts and by combining zero-valent iron reduction
with electrochemical reduction. The effectiveness of this technology will serve as a basis to
further develop the proposed methodology to purify surface or ground water in target zones at a larger scale.
1.1.2 MIB/Geosmin
It is not uncommon for water utilities to be flooded with complaints from consumers
about taste and odor in their drinking water, especially during warmer weather. For an average
consumer, off-flavors are the only way of determining the safety of tap water [18]. As a result,
drinking water utilities throughout the world are facing the necessity for more innovative and
cost-effective technologies for removal of taste and odor removal in water treatment and
purification. Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) and 2-methyl isoborneol–
C11H20O (MIB) are the major taste-and-odor-causing compounds in drinking water obtained from surface water and are associated with earthy and musty odors [19]. Geosmin and MIB in
surface water mainly result from the biodegradation of certain types of cyanobacteria that normally bloom in the presence of nutrients at warmer temperatures. There are currently no
regulations for these two compounds as they have not been associated with any health effects
[20]. The main problem of the presence of geosmin, 2-MIB and other odor-causing substances is
associated with their extremely low odor threshold concentrations (OTC) and their persistence to
5 elimination by conventional water treatment processes such as coagulation, sedimentation,
filtration and chlorination. The OTC for geosmin and MIB is 4 ng/L and 9 ng/L, respectively
[21]. Another factor that adds to this challenge faced by drinking water facilities, when it comes
to removing taste and odor compounds, is the presence of natural organic matter (NOM). NOM, a complex mixture of organic compounds derived primarily from the decay of plant and animal materials, is invariably present in all water sources and at much higher concentrations than geosmin or MIB.
Adsorption by granular activated carbon (GAC) is considered as one of the best available
technologies for removal of organic contaminants from water. Numerous studies have looked into GAC and PAC (powdered activated carbon) adsorption of MIB and Geosmin [9, 12, 21-23].
However, NOM levels of 3–10 mg/L competitively reduce activated carbon adsorption capacity for MIB or Geosmin [21, 24, 25]. Compared with GAC, activated carbon fibers (ACFs) have attracted increasing attention due to their excellent surface properties, high adsorption capacity and are an ideal adsorbent for targeting the impact of pore size [26]. The low ash and surface oxide content and controlled pore structure are key advantages of these types of adsorbents over
GAC. Pelekani and Snoeyink looked at activated carbon fiber adsorption of an organic micropollutant atrazine and the mechanism of competitive adsorption between atrazine and another compound much larger in size [27].
This study aims to study the adsorption of two taste and odor causing compounds
geosmin and MIB on a series of activated carbon fibers with different micropore size
distributions. In order to better understand the effect of pore size characteristics on adsorption,
6 activated carbon fiber adsorption was compared to adsorption with a commercially available
granular activated carbon. Experiments were also conducted to look at the effect of natural
organic matter on geosmin and MIB adsorption on activated carbon fibers.
1.2 Significance of the Study
1.2.1 Sources and current regulation
1.2.1.1 Perchlorate
A major source of perchlorate contamination is the manufacture of ammonium
perchlorate which is used as the oxidizer component and primary ingredient in solid propellant for rockets, missiles, and fireworks. Perchlorate salts are also used in pyrotechnics and fireworks, blasting agents, solid rocket fuel, matches, lubricating oils, nuclear reactors, air bags and certain types of fertilizers [28]. Prior to 1997, perchlorate was discharged into sewage systems or natural waters without any treatment. In fact, eleven U.S. states have been reported to have such discharging sites [5]. Due to the mobility of perchlorate, California public water suppliers detected 33 out of the 110 investigated wells had perchlorate concentration greater than 18 μg/L, with the highest concentration at 280 μg/L [29]. As of April 2003, the EPA listed over 150 separate locations in 25 different states with known perchlorate soil and groundwater contamination [28, 30].
Recognized as an endocrine disruptor chemical recently, perchlorate is a potential human
health hazard in drinking water which may inhibit normal iodide uptake by thyroid glands to
7 cause mental retardation and, hearing and speech degradation [31]. Perchlorate is listed by the
U.S.EPA in the drinking water contaminant candidate list [32]. Therefore, the maximum
contamination level (MCL) and maximum contamination level goal (MCLG) are still not
available at this time. EPA has established an official reference dose (RfD) of 0.0007 mg/kg/day
of perchlorate [28, 30]. Several states have also established their own action levels against
perchlorate contamination in their drinking water supply. California lowered its action level from
18 to 4 μg /L in 2004. Similarly, Arizona has adopted an action level of 14 μg/L, while Texas
adopted a level of 4 μg/L [33].
1.2.1.2 MIB/Geosmin
Geosmin and MIB are produced primarily by cyanobacteria (blue-green algae) and are
released into the water during the biological decomposition of these cyanobacteria.
Cyanobacteria blooms, which are common in surface water sources during the summer months when there is excessive sunlight and an overabundance of nutrients, result in significant
MIB/geosmin production. Water utilities are flooded with consumer complaints about taste when
and odor when the concentrations exceed the odor threshold.
Numerous studies have shown that the presence of these taste and odor causing
compounds is mainly an aesthetic concern and has not been associated with any health effects
[20]. MIB and geosmin have also not been correlated to presence of cyanobacteria toxins which
are extremely toxic at even low concentrations [35]. As a result, there is no MCL or MCLG for
either geosmin or MIB.
8 1.2.2 Treatment Technologies
1.2.2.1 Perchlorate
Perchlorate treatment technologies can be generally classified into categories of
destruction or removal technologies. Destruction processes include biological, chemical and
electrochemical reduction whereas removal technologies comprise of ion exchange, granular
activated carbon (GAC) adsorption and membrane filtration (including reverse osmosis and
nanofiltration). Among the physical removal technologies, only ion-exchange has been
successfully used to treat perchlorate contamination. Ion exchange has been implemented for a
long time and is well tested and there are also several anion exchange resins available
commercially for removal of perchlorate from water/wastewater [2, 36, 37]. However, the most important concern with the use of ion exchange for perchlorate removal is the resultant brine wastewater that is high in perchlorate concentration, and has to be somehow disposed or treated.
Another issue is the selectivity of these resins; they need to be modified in order to increase their selectivity towards perchlorate. Other removal technologies such as granulated activated carbon
(GAC)[38-40] and membrane filtration have had limited application [41, 42]. Because of the
concerns associated with the disposal of perchlorate-laden wastes, a process involving
destruction of the perchlorate ion is preferred over physical removal. As far as the destructive
processes are concerned, biological treatment appears to hold a lot of promise for perchlorate treatment. Several species of micro-organisms have been developed to effectively remove perchlorate from water [8-11, 43-49]. It is a fairly cost-effective technology and has been found to have good selectivity for perchlorate. However, the biggest issue with bioremediation is public acceptance, for the intentional introduction of micro-organisms into the drinking water system, considering that some of these could be pathogenic. Chemical reduction of perchlorate is limited
9 by its high activation energy (120 kjmol-1) [4]. Consequently, most of the common reducing
agents have been tested without any perchlorate reduction. Numerous studies have shown that perchlorate can be reduced electrochemically to Cl- [6, 7, 50-55] and electro-reduction also does
not involve any major waste products. It involves low maintenance and is easy to implement in
large systems. However the reaction rates associated with electrochemical reduction of
perchlorate are extremely slow and have to be improved for electrochemical reduction to be
practical. Experimental evidence has shown that aqueous mineral surfaces provide additional accelerated pathways for redox processes [13]. Lang and Horanyi reviewed the conversion of perchlorate at different metal surfaces and found iron to have the fastest reaction rate [7] Studies have shown that zero-valent iron successfully reduces perchlorate to chloride by providing the necessary catalytic effect [16, 17, 56]. This study investigates the removal of perchlorate by an integrated process that combines zero valent iron with electrochemical reduction.
1.2.2.2 Geosmin/MIB Treatment Technologies
The only treatment methods that have been successfully employed by water treatment plants to remove MIB and Geosmin are adsorption by activated carbon or oxidation by strong
oxidants such as ozone. Ferguson et al.[57], Glaze et al. [58] and Bruce et al. [19] studied and
demonstrated MIB and Geosmin removal using oxidants such as ozone, hydrogen peroxide and
UV. Addition of chemicals however is expensive and can result in formation of disinfection
byproducts (DBPs), which are unacceptable due to health and regulatory concerns. Studies have
shown that conventional water treatment processes such as coagulation, sedimentation and
filtration are unable to achieve any significant removal of MIB and Geosmin [19]. Adsorption by
activated carbon, either granular activated carbon (GAC) or powdered activated carbon (PAC) is
10 considered as the best available technology for removal of organic contaminants from water.
Numerous studies have looked into GAC and PAC adsorption of MIB and Geosmin [9, 12, 21-
23]. However, NOM levels of 3–10 mg/L competitively reduce activated carbon adsorption capacity for MIB or Geosmin [21, 24, 25]. In natural waters, the size and concentration of NOM particles is many folds higher than that of MIB or Geosmin, and as a result a large volume of the
GAC is not utilized for MIB/Geosmin adsorption, thereby significantly reducing the GAC adsorption capacity. Compared with GAC, activated carbon fibers (ACFs) have attracted increasing attention due to their excellent surface properties, high adsorption capacity and are an ideal adsorbent for targeting the impact of pore size [59]. This study aims to study the adsorption of two taste and odor causing compounds geosmin and MIB on a series of activated carbon fibers with different micropore size distributions.
1.3 Objectives of the Study
1.3.1 Perchlorate
This project investigates the potential removal of perchlorate ion in drinking water by combining electrochemical reduction with zero-valent iron reduction. The use of zero-valent iron in the process is expected to enhance electrochemical reduction of the perchlorate ion. The effectiveness of this technology will serve as a basis to further develop the proposed methodology to purify surface or ground water in target zones at a larger scale.
Specific objectives of the study are:
• Huang [60] studied perchlorate reduction with zero-valent iron filings through batch and
column experiments. However the reaction rates obtained were too slow for treatment
11 applications. One of the objectives of this study will be to enhance perchlorate
remediation by zero-valent iron by using cobalt based catalysts.
• The performance of the integrated process (zero-valent iron + electro-reduction) will
depend on various parameters (current density, ionic strength of the electrolyte, pH,
dissolved oxygen and concentration of the perchlorate). Experiments at various
conditions will determine the impact of the variation of one parameter at a time on the
performance of the electrochemical process in order to determine its role on the reduction
of perchlorate.
• To design and setup an electrochemical reactor to investigate reduction of perchlorate
with the integrated process. The effect of various process parameters such as pH, current
density, ionic strength of electrolyte, dissolved oxygen and initial perchlorate
concentration, on the overall efficiency of the process will be studied.
• Experiments will also be conducted to study electrochemical reduction without zero-
valent iron and with different types of electrodes such as Nickel, Copper, Titanium,
Glassy Carbon. Previous studies have shown that metal oxide films can provide a
catalytic effect during reduction of perchlorate. Use of metal oxide coatings such as
titanium dioxide and ferric oxide will be used on the electrodes to enhance
electrochemical reduction.
1.3.2 MIB/Geosmin
This study looks at the removal of two taste and odor causing compounds, namely MIB and Geosmin in drinking water by activated carbon fiber adsorption. The primary objective of this study is to investigate the impact of pore size of activated carbon on the adsorption of MIB
12 and geosmin. The controlled structure and uniform pore size distribution of activated carbon fibers (ACFs) is ideal to study the impact of pore size on adsorption [61]. In order to better understand the impact of the pore size distribution of ACFs, three types of ACFs with different pore sizes were chosen in this study.
Specific objectives are:
• To conduct single solute adsorption isotherms for MIB and geosmin on three types of
activated carbon fibers, ACC-15, ACC-20 and ACC-25.
• To conduct solute adsorption isotherms for MIB and geosmin on a type of GAC, a
bituminous base GAC namely F-400, in order to compare ACF adsorption to GAC
adsorption.
• In natural waters, activated carbon adsorption of MIB and geosmin is limited by the
presence of NOM which are present at a much higher concentration. In order to better
understand the impact of NOM, adsorption isotherms experiments will be conducted for
the three ACFs, ACC-15, ACC-20 and ACC-25 and the GAC, F-400 in the presence of
humic acid, which is typically used to represent NOM.
• Binary solute adsorption isotherms will be conducted for MIB and geosmin to simulate
practical usage of activated carbon with the three ACF adsorbents and one GAC
adsorbent. The single solute adsorption isotherms are typically used to predict
multicomponent adsorption behavior. The effectiveness of the models to predict binary
adsorption of MIB and geosmin will also be examined in this study.
13 1.4 Structure of dissertation
The dissertation is organized by chapters in the following manner.
Chapter 1 provides introduction, significance of the research, and objectives of the study.
Chapter 2 provides the materials and methods used in this study. For perchlorate, the various
chemicals, electrodes and the experimental setup of the electrochemical reactor are discussed.
For the MIB/geosmin adsorption study, the adsorbates, adsorbents, isotherm and extraction
procedure, and analytical procedure are stated in this chapter.
Chapter 3 offers a comprehensive literature review on perchlorate, focusing on contamination,
regulation and recent developments in treatment technologies.
Chapter 4 looks at an integrated approach to treat perchlorate in water by combining
electrochemical reduction with zero-valent iron remediation and looks at the various factors
affecting perchlorate removal.
Chapter 5 presents a detailed critical review on MIB and geosmin which are the major taste and odor causing compounds in water. The review covers the current state of these contaminants with a focus on the available treatment technologies.
Chapter 6 investigates single solute adsorption of MIB/ geosmin on four adsorbents (GAC F400
and three types of ACFs with different activation levels, namely ACC-15, ACC-20 and ACC-25)
14 with the intent of studying the effect of pore size distribution and presence of NOM on adsorption of these compounds.
To better understand competitive adsorption, chapter 7 covers binary adsorption of MIB and geosmin on the same set of adsorbents to determine if there were deviation from IAST predictions in the presence and absence of NOM.
1.5 References
1 Speth, T. F. and Schock, M. R. (2007) Removing Esoteric Contaminants from Drinking
Waters: Impacts of Treatment Implementation. Journal of Environmental Engineering,
665-669
2 Urbansky, E. T. (1998) Perchlorate Chemistry: Implications for Analysis and
Remediation. Bioremediation Journal 2, 81-95
3 Espenson, J. H. (2000) The Problem and Perversity of Perchlorate. In Perchlorate in the
Environment (Urbansky, E. T., ed.), pp. 1-7, Kluwer Academic/Plenum, New York
4 Gu, B., Dong, W., Brown, G. M. and Cole, D. R. (2003) Complete Degradation of
Perchlorate in Ferric Chloride and Hydrochloric Acid under Controlled Temperature and
Pressure. Environmental Science & Technology 37, 2291-2295
5 Urbansky, E. T. and Schock, M. R. (1999) Issues in managing the risks associated with
perchlorate in drinking water. Journal of Environmental Management 56, 79-95
6 Almeida, C. M. V. B., Gianetti, B. F. and Rabockai, T. (1997) Electrochemical study of
perchlorate reduction at tin electrodes. Journal of Electroanalytical Chemistry 422, 185-
189
15 7 Lang, G. G. and Horanyi, G. (2003) Some interesting aspects of the catalytic and
electrocatalytic reduction of perchlorate ions. Journal of Electroanalytical Chemistry 552,
197-211
8 Wallace, W., Beshear, S., Williams, D., Hospadar, S. and Owens, M. (1998) Perchlorate
reduction by a mixed culture in an up-flow anaerobic fixed bed reactor. Journal of
Industrial Microbiology and Biotechnology 20, 126-131
9 Kim, K. and Logan, B. E. (2001) Microbial Reduction of Perchlorate in Pure and Mixed
Culture Packed-Bed Bioreactors. Water Research 35, 3071-3076
10 Gingras, T. M. and Batista, J. R. (2002) Biological reduction of perchlorate in ion
exchange regenerant solutions containing high salinity and ammonium levels. Journal of
Environmental Monitoring 4, 96-101
11 Okeke, B. C., Giblin, T. and Frankenberger, W. T. (2002) Reduction of perchlorate and
nitrate by salt tolerant bacteria. Environmental Pollution 118, 357-363
12 Moore, A. M. and Young, T. M. (2005) Chloride Interactions with Iron Surfaces:
Implications for Perchlorate and Nitrate Remediation Using Permeable Reactive Barriers.
Journal of Environmental Engineering 131, 924-933
13 Wehrli, B. (1990) Redox Reactions of Metal Ions at Mineral Surfaces. In Aquatic
Chemical Kinetics (Stumm, W., ed.), pp. 311-337, Wiley Interscience, New York
14 Su, C. and Puls, R. W. (1999) Kinetics of trichloroethene reduction by zerovalent iron
and tin: Pretreatment effect, apparent activation energy, and intermediate products.
Environmental Science & Technology 33, 163
16 15 Szecsody, J. E., Fruchter, J. S., Williams, M. D., Vermeul, V. R. and Sklarew, D. (2004)
In Situ Chemical Reduction of Aquifer Sediments: Enhancement of Reactive Iron Phases
and TCE Dechlorination. Environmental Science & Technology 38, 4656-4663
16 Gurol, M. G. and Kim, K. (2000) Investigation of Perchlorate Removal in Drinking
Water Sources by Chemical methods. In Perchlorate in the Environment (Urbansky, E.
T., ed.), pp. 99-107, Kluwer Academic/Plenum, New York
17 Huang, H. and Sorial, G. A. (2007) Perchlorate Remediation in Aquatic Systems by Zero
Valent Iron. Environmental Engineering Science 24, 917-926
18 McGuire, M. J. (1995) Off-flavor as the consumer's measure of drinking water safety.
Water Science & Technology 31, 1-8
19 Bruce, D., Westerhoff, P. and Brawley-Chesworth, A. (2002) Removal of 2-
methylisoborneol and geosmin in surface water treatment plants in Arizona. Journal of
Water Supply: Research and Technology - AQUA 51, 183-197
20 Dionigi, C. P., Lawlor, T. E., McFarland, J. E. and Johnsen, P. B. (1993) Evaluation of
geosmin and 2-methylisoborneol on the histidine dependence of TA98 and TA100
Salmonella typhimurium tester strains Water Research 27, 1615-1618
21 Pirbazari, M., Ravindran, V., Badriyha, B. N., Craig, S. and McGuire, M. J. (1993) GAC
adsorber design protocol for the removal of off-flavors. Water Research 27, 1153-1166
22 Cook, D., Newcombe, G. and Sztajnbok, P. (2001) The application of powdered activated
carbon for MIB and geosmin removal: predicting PAC doses in four raw waters. Water
Research 35, 1325-1333
17 23 Ng, C., Losso, J. N., Marshall, W. E. and Rao, R. M. (2002) Freundlich adsorption
isotherms of agricultural by-product-based powdered activated carbons in a geosmin -
water system. Agricultural Wastes 85, 131-135
24 Newcombe, G., Morrison, J., Hepplewhite, C. and Knappe, D. R. U. (2002) Simultaneous
adsorption of MIB and NOM onto activated carbon-II. Competitive effects. Carbon 40,
2147-2156
25 Newcombe, G., Morrison, J. and Hepplewhite, C. (2002) Simultaneous adsorption of
MIB and NOM onto activated carbon. I. Characterisation of the system and NOM
adsorption. Carbon 40, 2135-2146
26 Lu, Q. and Sorial, G. A. (2004) Adsorption of phenolics on activated carbon--impact of
pore size and molecular oxygen. Chemosphere 55, 671-679
27 Pelekani, C. and Snoeyink, V. L. (2000) Competitive adsorption between atrazine and
methylene blue on activated carbon: the importance of pore size distribution. Carbon 38,
1423-1436
28 http://www.epa.gov/fedfac/documents/perchlorate.htm
29 Motzer, W. E. (2001) Perchlorate: Problems, Detection, and Solutions. Environmental
Forensics 2, 301-311
30 USEPA (2007) Perchlorate. (Reuse, F. F. R. a., ed.), Federal Facilities Restoration and
Reuse
31 Clark, J. J. J. (2000) Toxicology of Perchlorate. In Perchlorate in the Environment
(Urbansky, E. T., ed.), pp. 15-29, Kluwer Academic/Plenum, New York
32 Perciasepe, R. (1998) Announcement of the drinking water contaminant candidate list;
notice. Federal Register 63, 10273
18 33 http://www.swhydro.arizona.edu/archive/V2_N6/dept-r&d.pdf).
34 http://www.swhydro.arizona.edu/archive/V2_N6/dept-r&d.pdf
35 Zimmerman, L. R., Ziegler, A. C. and Thurman, E. M. (2002) Method of Analysis and
Quality-Assurance Practices by U.S. Geological Survey Organic Geochemistry Research
Group--Determination of Geosmin and Methylisoborneol in Water Using Solid-Phase
Microextraction and Gas Chromatography/Mass Spectrometry. 1-12
36 Gu, B., Ku, Y.-K. and Brown, G. M. (2002) Treatment of Perchlorate-Contaminated
Groundwater Using Highly Selective, Regenerable Ion-Exchange Technology: A Pilot-
Scale Demonstration. Remediation Journal 12, 51-68
37 Gu, B., Ku, Y. K. and Brown, G. M. (2005) Sorption and Desorption of Perchlorate and
U(VI) by Strong-Base Anion-Exchange Resins. Environmental Science & Technology
39, 901-907
38 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. I: Characterization of NH3 thermally tailored GACs.
Carbon 43, 573-580
39 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. II: Perchlorate adsorption. Carbon 43, 581 - 590
40 Parette, R. and Cannon, F. S. (2005) The removal of perchlorate from groundwater by
activated carbon tailored with cationic surfactants. Water Research 39, 4020-4028
41 Yoon, Y., Amy, G., Cho, J., Her, N. and Pellegrino, J. (2002) Transport of perchlorate
- (ClO4 ) through NF and UF membranes. Desalination 147, 11-17
42 Yoon, J., Yoon, Y., Amy, G., Cho, J., Foss, D. and Kim, T.-H. (2003) Use of surfactant
modified ultrafiltration for perchlorate (ClO4-) removal. Water Research 37, 2001-2012
19 43 Xu, J., Trimble, J. J., Steinberg, L. and Logan, B. E. (2004) Chlorate and nitrate reduction
pathways are separately induced in the perchlorate-respiring bacterium Dechlorosoma sp.
KJ and the chlorate-respiring bacterium Pseudomonas sp. PDA. Water Research 38, 673-
680
44 Xiaohua, L., Roberts, D. J., Hiremath, T., Clifford, D. A., Gillogly, T. and Lehman, S. G.
(2007) Divalent Cation Addition (Ca2+ or Mg2+) Stabilizes Biological Treatment of
Perchlorate and Nitrate In Ion-Exchange Spent Brine. Environmental Engineering
Science 24, 725-735
45 Steinberg, L. M., Trimble, J. J. and Logan, B. E. (2005) Enzymes responsible for chlorate
reduction by Pseudomonas sp. are different from those used for perchlorate reduction by
Azospira sp. FEMS Microbiology Letters 247, 153-159
46 Nerenberg, R., Kawagoshi, Y. and Rittmann, B. E. (2006) Kinetics of a hydrogen-
oxidizing, perchlorate-reducing bacterium. Water Research 40, 3290-3296
47 Herman, D. C. and Frankenberger, W. T., Jr. (1999) Bacterial reduction of perchlorate
and nitrate in water. Journal of Environmental Quality 28, 1018-1024
48 Giblin, T., Herman, D., Deshusses, M. A. and Frankenberger, W. T., Jr. (2000) Removal
of perchlorate in ground water with a flow-through bioreactor. Journal of Environmental
Quality 29, 578-583
49 Coates, J. D., Michaelidou, U., Bruce, R. A., O'Connor, S. M., Crespi, J. N. and
Achenbach, L. A. (1999) Ubiquity and Diversity of Dissimilatory (Per)chlorate-Reducing
Bacteria. Applied & Environmental Microbiology 65, 5234-5241
20 50 Horanyi, G. and Bakos, I. (1993) Study of the relationship between voltammetric
behavior and electroanalytic activity in the reduction of ClO4- ions at platinized platinum
electrodes. Journal of Electroanalytical Chemistry 347, 383 - 391
51 Horanyi, G., Bakos, I., Szabo, S. and Rizmayer, E., M., (1992) New observations in the
field of electrochemistry of technetium and rhenium: electrocatalytic reduction of
perchlorate ions at electrosorbed and electrodepositied Tc and Re layers in acid medium.
Journal of Electroanalytical Chemistry 337, 365-369
52 Colom, F. and Gonzalez - Tejera, M. J. (1985) Reduction of Perchlorate Ion on
Ruthenium Electrodes in Aqueous Solutions. Journal of Electroanalytical Chemistry 190,
243 - 255
53 Brown, G. M. (1986) The Reduction of Chlorate and Perchlorate at an Active Titanium
Electrode. Journal of Electroanalytical Chemistry 198, 319-330
54 Wasberg, M. and Horanyi, G. (1995) The reduction of ClO4- ions on Rh electrodes.
Journal of Electroanalytical Chemistry 385, 63-70
55 Theis, T. L., Zander, A. K., Xiang, L., Jeosadaque, S. and Anderson, M. A. (2002)
Electrochemical and Photocatalytic Reduction of Perchlorate Ion. Journal of Water
Supply: Research and Technology - AQUA 51, 367 - 374
56 Moore, A. M., DeLeon, C. H. and Young, T. M. (2003) Rate and Extent of Aqueous
Perchlorate Removal by Iron Surfaces. Environ. Sci. Technol. 37, 3189-3198
57 Ferguson, D. W., McGuire, M. J., Koch, B., Wolfe, R. L. and Aieta, E. (1990)
Comparing PEROXONE and Ozone for Controlling Taste and Odor Compounds,
Disinfection By-Products, and Microorganisms. Journal of the American Water Works
Association 82, 181-191
21 58 Glaze, W. H., Zarnoch, J. J., Ruth, E. C., Chauncey, W. and Schep, R. (1990) Evaluating
Oxidants for the Removal of Model Taste and Odor Compounds from a Municipal Water
Supply. Journal of the American Water Works Association 82, 79-84
59 Lu, Q. and Sorial, G. A. (2004) The role of adsorbent pore size distribution in
multicomponent adsorption on activated carbon. Carbon 42, 3133-3142
60 Huang, H. (2005) Reduction of Perchlorate by Zero Valent Iron. In Environmental
Engineering, pp. 87, University of Cincinnati, Cincinnati, OH
61 Pelekani, C. and Snoeyink, V. L. (2001) A kinetic and equilibrium study of competitive
adsorption between atrazine and Congo red dye on activated carbon: the importance of
pore size distribution. Carbon 39, 25-37
22 CHAPTER 2
Materials and Methods
2.1 Introduction
This chapter provides a brief description of the materials and experimental procedures that were a part of this study. For the perchlorate portion of the study, the various chemicals and electrode materials are listed along with the experimental procedure for running the zero-valent iron adsorption and electrochemical experiments. For MIB/geosmin, the adsorbates, adsorbents and the isotherm procedure are discussed. The analytical method for both perchlorate and
MIB/geosmin are also presented. Materials and methods unique to a chapter are described in detail in the methodology section of the corresponding chapter.
2.2 Materials
2.2.1 Perchlorate
1. Milli-Q water: Milli-Q water produced by a Millipore system (Billerica, MA, USA) with
18.2 MΩ-cm resistance was used for preparation of solutions.
2. Iron Filings: Degreased Iron filings (Peerless Metal Powders & Abrasives, Detroit, MI),
screened to retain particles between 16 mesh and 30 mesh was used.
3. Hydrofluoric acid: Hydrofluoric (HF) acid, obtained from Fisher Scientific was used for
acid washing of the iron filings.
4. Sodium Perchlorate: HPLC grade sodium perchlorate monohydrate (Fisher Scientific,
PA) was used for the preparation of perchlorate stock solutions.
5. Pretreatment Cartridges: 1.0 cc OnGuard II H cartridges (Dionex, Sunnyvale, CA) were
used to remove any dissolved iron species from the sample prior to analysis by IC.
23 6. Electrodes: Metallic nickel (Alfa Aesar, Ward Hill, MA), titanium rods (Alfa Aesar,
Ward Hill, MA), glossy carbon rods (zero porosity) (HTW, Germany) and platinum wire
(Alfa Aesar, Ward Hill, MA) were used as electrodes.
2.2.2 MIB/Geosmin
Adsorbates: Geosmin (Sigma Aldrich, St Louis, MO) was obtained as a solution with
concentration of 2 mg/mL in methanol. MIB (Sigma Aldrich, St Louis, MO) was also obtained
as a solution with concentration of 10 mg/mL in methanol. Humic acid (Sigma Aldrich, St
Louis, MO) was used as a representative of NOM in powder form with purity over 99%.
Adsorbents: Three microporous phenolic resin-based activated carbon fibers (ACFs) with
increasing degrees of activation designated as ACC-15, ACC-20, and ACC-25 (Nippon Kynol,
Japan) were used in the study. They were received as twilled-weave fabrics. F400 (Filtrasorb
400, Calgon, Pittsburgh, PA), which is a bituminous base activated carbon, was chosen as a typical GAC. All the ACFs and F-400 used in the study were dried in an oven at 105 oC
overnight, and then stored in a desiccator until use.
2.3 Experimental Method
2.3.1 Perchlorate
A schematic of the column experimental setup that was used for reduction of perchlorate
ions on iron fillings is shown in Figure 2.1. The feed container for perchlorate is a sealed 4L glass container which was kept under a head pressure of nitrogen to keep the system under
24 anoxic condition. A multichannel variable speed digital pump was used to pump the perchlorate
solution through the column. The column, made of plexiglass, consisted of two chambers. The
portion on the left is the cathode and was filled with the required amount of acid washed iron
fillings. The section on the right was primarily made up of a Platinum wire with the same length
as that of the cathode. The two chambers were separated by means of a nafion membrane
(Dupont, Fayetville, NC) that allowed only the passage of positive ions. A conductor wire was introduced into the center of the left chamber and was connected to the platinum wire in the right
chamber. The portion of the column that consisted of the iron fillings, acted as the cathode when current was applied and the platinum wire acts as the anode.
Figure 2.1 Schematic of the Experimental Setup
The iron filings were washed using 1N Hydrofluoric (HF) acid prior to use to remove
iron oxide and other oily contaminants on the surface. The acid washed iron filings were rinsed a
number of times by de-aerated Milli-Q water until the rinse water is free of brown color and its
25 pH is neutral. Finally, the washed iron filings were dried under nitrogen gas. The color of the
treated iron filings should be black.
Acid washed iron filings were added so as to completely pack the cathode section of the
electrochemical setup shown in Figure 2.1. A main stock solution with a concentration of 1g L-1
- ClO4 was prepared by adding 1.412g of sodium perchlorate monohydrate in Milli-Q water. This
stock solution was used for preparing the required dilutions for each of the experiments
discussed in section 2.1. The perchlorate feed solution was prepared by diluting the stock
- solution to get a volume of 4L and an inlet concentration of 500 or 1000 ppb ClO4 . Sodium sulfate (Na2SO4) was added to the feed solution as the electrolyte to a molar concentration of 9
mM corresponding to an ionic strength of 0.027 M. The feed solution was thoroughly purged
with nitrogen gas to achieve dissolved oxygen (DO) concentration ≤ 0.8 mg/L in the solution
before transferring it to the feed container. The feed container for perchlorate was immediately
brought under a head pressure of nitrogen to keep the system under anoxic condition. A 0.25 M
sodium sulfate solution (anodic solution) was used as the electrolyte solution in the anode
compartment of the reactor and was prepared by adding 35.525g of Na2SO4 per liter of DI water.
This salt solution was monitored on a daily basis for its conductivity and was changed when the
conductivity dropped below 30 mS/cm.
The multichannel variable speed digital pump set at the desired flow rate (11 ± 0.5
µL/min) was used to pump the perchlorate feed solution through the reactor into the holding
tank. A 500 mL holding tank made of stainless steel with a sealed lid was used. The holding tank
had two ports - one port was used for pH measurement and the other port was fitted with a
26 rubber septum for manual injection of acid for pH adjustment (this port was also be used as a sampling port). The pump for recirculating the anode electrolyte solution through the system was turned on at this stage. The anodic solution was thereby allowed to flow continuously through the anode compartment in a closed loop. Once there was flow in both the anodic and cathodic sections of the reactor, the current supply to the system was started by turning on the constant current supplier unit. In order to simulate batch conditions, the feed solution was pumped into the system (reactor + holding tank) until there was a constant effluent flow. The flow of the feed solution was turned off and the feed pump was used to pump the solution in the holding tank back into the reactor.
Samples were collected on frequent basis from the sampling port in the holding tank through the rubber septum and their pH and conductivity measured. The sample portion for perchlorate analysis was filtered through a 0.45μm membrane and then treated by an OnGuard II
H cartridge (Dionex, Sunnyvale, CA) to remove any dissolved iron species which may interfere in the perchlorate analysis and might significantly reduce the sensitivity of the IC column and accumulate in the suppressor. The samples were analyzed for perchlorate in an IC using standard
EPA method 314.0.
2.3.2 MIB/geosmin
Adsorption Isotherm Procedure: The bottle point method was used for conducting the adsorption isotherms at room temperature. Various initial concentrations of adsorbate, ranging from 50µg/l to 1000 µg/l were used. An initial concentration of 10 mg/L humic acid was used for adsorbate solutions containing NOM. Each set of bottles was accompanied by two blanks to
27 check for any volatilization, adsorption onto the walls of the bottles, and biodegradation of adsorbate. Accurately weighed (±0.1 mg) masses of activated carbon fibers or GAC were placed in 250-ml glass-amber bottles. The bottles were tightly sealed with Teflon lined caps after purging. Adsorbate solutions were prepared in autoclaved de-ionized water buffered with potassium phosphate (KH2PO4) and the pH adjusted to 7.0 by adding sodium hydroxide. The glass amber bottles containing the different masses of carbon were then completely filled with the adsorbate solution, sealed tightly with Teflon lined caps and covered by parafilm. The isotherm bottles were placed in a rotary shaker and allowed to equilibriate for eight days. After
equilibration, all samples were filtered through 0.45-µm nylon filters (Micron Separation, Inc.)
prior to analysis in order to minimize interferences of the carbon fines with analysis. The first 5
ml of the filtered samples was discarded before samples were taken for analysis in order to
minimize the impact of potential adsorption of the adsorbate on the filter membrane.
2.3 Analytical Method
2.3.1 Perchlorate
Perchlorate concentration was measured using EPA method 314.0 [1-3] with a manual
injection Dionex DX-600 (Sunnyvale, CA, USA) ion chromatography system with a GP-50
gradient pump, an anion self-regenerate suppressor ultra (ASRS-ultra), a CD-25 conductivity detector and a LC-20 chromatography enclosure. The injection volume was 1000 µL, controlled by a 1000 µL sample loop. Anion separation was obtained on a Dionex IonPac AS16 separation column (4.0 x 250 mm), with an IonPac AG16 guard column. The eluent was 50 mM NaOH.
The flow rate of eluent was 0.9 ml min-1 and a 100 mA current was applied on the ASRS-ultra to reduce the back ground conductivity to 4 μS. The ASRS-ultra was regenerated by external water
28 mode at water flow rate of 5 mL min-1 controlled by a pressure regulator. The pH measurement
was performed by an Accumet AR50 dual channel pH meter obtained from Fisher Scientific.
2.3.2 MIB/Geosmin
Analysis of MIB and geosmin were carried using solid-phase microextraction/gas
chromatography (SPME-GC) [4, 5]. For Geosmin analysis, 20 ml of sample was added to a 40
ml septum capped vial containing 6 g sodium chloride and a magnetic stir bar. In case of MIB,
25 ml of sample was used with 7.5 g of sodium chloride. The vials were then placed in a water
bath on a magnetic stir plate heated to 65 ± 1.5 °C. A SPME fiber (Supelco, Bellefonte, PA) was
introduced into the head space gas through the septum and the sample stirred for 30 min. The
fiber was removed from the vial and inserted into the GC (Agilent 6890 Series; Agilent
Technologies, Wilmington, DE) inlet port and left exposed for 10 minutes for complete
desorption of the target compounds. The GC was equipped with Agilent DB-5 column (30 m x
0.32 mm inside diameter x 0.25 µm film thickness) and flame ionization detector (FID). The
injection port and detector temperature were set at 250 °C. The GC oven temperature was
maintained at 60 °C for the first 4 min, then ramped to 270 °C at 10 °C/min and kept at 270 °C
for 1 min. The flow rate of the carrier gas (N2) was set at 1.5 ml/min. The detector makeup gas
(N2) flow rate was set at 43.5 ml/min. The flame gases hydrogen and air flow rates were set at 40
and 450 ml/min, respectively. The retention time for geosmin was 14.1 min and for MIB, 11.0
min. The fiber was then retracted into the holder, removed from the GC inlet and reused for the
next sample.
29 2.4 References
1 Hautman, D. P., Munch, D., Eaton, A. D. and Haghani, A. W. (1999) U.S. EPA Method
314.0. USEPA
2 Dionex (2000) Determination of low concentrations of perchlorate in drinking and
ground waters using ion chromatography. . In Application Note 134, pp. Dionex Corp.,
Sunnyvale.
3 Dionex (1998) Analysis of low concentrations of perchlorate in drinking water and
ground water by ion chromatography. Note 121. Dionex Corp, Sunnyvale
4 Lloyd, S. W., Lea, J. M., Zimba, P. V. and Grimm, C. C. (1998) Rapid analysis of
geosmin and 2-methylisoborneol in water using solid phase micro extraction procedures.
Water Research 32, 2140-2146
5 Watson (2000) Quantitative analysis of trace levels of geosmin and MIB in source and
drinking water using headspace SPME. Water Research 34, 2818-2828
30 CHAPTER 3
Treatment of Perchlorate in Drinking Water: A Critical Review
3.1 Abstract
Because of its extremely low concentrations and strong resistance to most treatment technologies, perchlorate has become one of the biggest challenges currently being faced by the drinking water industry. Although significant research has been performed to evaluate different treatment technologies for perchlorate removal from drinking water, there has not been a holistic review performed recently. A complete and critical review on the intriguing contaminant
‘perchlorate’ is presented. The sources of perchlorate along with the degree of contamination are discussed. The policy aspects including the regulation and toxicity in addition to the most recent developments in perchlorate analysis are also considered. The applicable treatment technologies including their feasibility are discussed in detail. Although some technologies such as microbial reduction and ion-exchange have become more established than the others, there is still not a single technology that can be directly applied to a drinking water treatment system for compete removal of perchlorate. Although significant research is still being conducted to come up with a novel technology for perchlorate remediation, it is highly likely that it would not be a single novice or conventional technology but a combination of these technologies that would have to be employed to overcome this challenge.
3.2 Introduction
In the last few years, perchlorate has become a major inorganic contaminant in drinking water and has been detected in a number of public drinking water systems throughout the Unites
31 States. Perchlorate and its salts are mainly used in solid propellants and for manufacturing matches, rockets, explosives and fireworks. It also occurs naturally. The main health concern regarding perchlorate is its ability to interfere with iodine uptake by the thyroid gland resulting in a decreased production of the thyroid hormones [1]. Thyroid is important for normal growth, development and metabolism in the human body and the effects can be significant in case of pregnant women and fetuses. Perchlorate is listed by the U.S.EPA in the drinking water contaminant candidate list [2]. EPA has established an official reference dose (RfD) of 0.0007 mg/kg-day of perchlorate [3] and is in the process of evaluating a maximum contamination level
(MCL). Several states have also established their own action levels against perchlorate contamination in their drinking water supply. However a recent ruling by the USEPA concluded that regulation for perchlorate would not be required at this time. This has serious implications for drinking water treatment plants across the country, that have been evaluating treatment alternatives. Ion exchange and biological treatment are currently the most widely used perchlorate treatment technologies. One concern with the use of ion exchange is the resultant brine wastewater that is concentrated in perchlorate, and has to be somehow disposed or treated.
Also, background ions present in the influent water can reduce significantly the driving force for removal of perchlorate by ion exchange due to the competition among various ions for the available adsorption sites. On the contrary, the biggest issue with bioremediation is public acceptance and the introduction of pathogenic micro-organisms into the drinking water system
[1]. For drinking water systems, a chemical reduction process needs to be developed that can completely reduce perchlorate to harmless chloride. Although perchlorate is a strong oxidizing agent by virtue of its redox potential (+1.38 V [4]), its oxidizing power is retarded because of its high activation energy (120 kjmol-1)[5]. It is the atomic structure of perchlorate that causes this
32 high activation energy. The chlorine atom is in the center of a tetrahedron structure surrounded by four oxygen atoms which essentially block common reductants from directly attacking the chlorine. As a result, most common reducing agents have been tested without any significant perchlorate reduction [6, 7]. The only solution is to find a technique that can overcome this energy barrier.
The issues surrounding perchlorate, including regulations, toxicity and treatment
technologies have been reviewed in a number of peer reviewed journal articles and books [1, 8-
11]. Perchlorate has received tremendous attention in the recent years, mainly due to the challenges faced by the drinking water industry regarding its treatment. There have been a number of interesting developments in the last few years, both from a policy and a technology
perspective. This review provides an elaborate discussion on these recent issues and attempts to
provide a holistic approach, by including different perspectives. Firstly, the sources and nature of
perchlorate contamination are discussed. This is followed by a discussion on the recent
advancements in perchlorate analysis, health affects and a summary on the current perchlorate
regulation, where there have been some rather interesting and possibly ‘game changing’
developments. Finally the various pertinent treatment technologies are discussed with regards to
the recent developments, current status and feasibility for application in drinking water treatment
systems.
3.3 Sources
- Perchlorate (ClO4 ) has been found to occur both naturally and as a man made
manufactured substance. Most of the naturally occurring perchlorate has been found to exist in
33 the Chilean nitrate deposits in the Atacama Desert [12]. This Chilean nitrate has been imported
into the Unites States for over a century mainly to be used in fertilizers. Naturally occurring
perchlorate has mainly been associated only with geographies of extremely arid climates. Some
studies have shown that perchlorate detected in certain parts of the Unites States may not be
entirely from anthropogenic sources like previously thought but might actually be from natural
sources. Research has shown that perchlorate can be produced naturally by certain atmospheric
processes. Jackson et al. [13, 14] studied natural or non-anthropogenic occurrence and sources of
perchlorate. They found perchlorate in parts of Texas, in areas which had no history of crop use
or other perchlorate discharge sources and the authors attributed this to atmospheric formation of perchlorate. In a related study Kang et al. demonstrated in a lab scale study that perchlorate can
be generated by ozone oxidation of chloride [15]. Although the authors did not quantify the
perchlorate generated by this process, they concluded that this does serve as significant evidence
for the hypothesis that perchlorate is formed naturally by atmospheric processes.
The main source of anthropogenic perchlorate in the United States is the manufacturing
of Ammonium perchlorate [1, 16]. Being a strong oxidizer, ammonium perchlorate is widely
used in solid propellants for rockets and missiles. Most of the perchlorate contaminated sites are
found in or close to NASA facilities or military establishments where wastewater containing
perchlorate was discharged to the ground without proper treatment. As of March 2005, the EPA
listed over 150 separate locations in 25 different states with known perchlorate soil and
groundwater contamination [17, 18]. The following map (Figure 3.1) shows perchlorate detection sites throughout the United States as of September 2004 [9]. Since a majority of these contaminates sites were operated by either NASA or the Department of Defense (DoD), both
34 organizations have started conducting research looking into perchlorate alternatives to be used
in their operations [9]. Perchlorate contamination is also associated with the manufacture, storage
and use of commercial explosives and fireworks. Perchlorate has also found limited application in certain pharmaceuticals, laboratories and other industries.
Figure 3.1 Perchlorate Contamination Sites (September 2004) [9]
3.4 Contamination
Prior to 1997, perchlorate was discharged into sewage systems or natural waters without any treatment. In fact, eleven U.S. states have been reported to have such discharging sites [1].
Shortly after the development of the low level analytical method, perchlorate began to be detected in ground and surface drinking water supplies across the country [19]. Perchlorate detection was the highest in the Southwest, especially the lower Colorado River region where perchlorate contamination is well documented. California public water suppliers detected 33 out of the 110 investigated wells had perchlorate concentration greater than 18 μg/L, with the highest concentration at 280 μg/L [18]. Due to its mobility, the perchlorate ion is readily absorbed by
35 agricultural produce and cattle exposed to perchlorate. As a result significant levels of perchlorate have been detected in dairy products, vegetables and even breast milk [11, 20-24].
Sanchez et al. looked at perchlorate exposure in the dairy industry in the southwestern
United States [22]. Because significant levels of perchlorate have been detected in the Colorado
River and because a large variety of dairy feed is grown in this area, there is potential for significant accumulation of perchlorate in cattle and in turn in the milk produced in this region.
The authors conducted sampling of water, forage, cattle and milk to estimate the extent of perchlorate contamination. The concentration of perchlorate found in crops in this region ranged from 43 to 2196 µg/kg whereas the concentration in the river water was between 2 and 5 µg/L.
The authors also found through statistical analysis that the perchlorate concentration in milk was directly related to the perchlorate content of the food. Kirk et al conducted perchlorate analysis on seven popular brands of milk in Texas and found concentrations as high as 6 µg/L [23]. The authors conclude that their results confirm that exposure to perchlorate occurs not only through drinking water but also through food and milk and hence total exposure should be considered while devising control and regulation strategies. In another study Kirk et al. looked at correlation between perchlorate and iodide concentrations in breast milk [24]. By collecting samples from
18 states, their results indicated that perchlorate concentrations in breast milk were on an average five times higher than in dairy milk. More importantly, the perchlorate concentration was inversely related to iodide concentration which let them to conclude that perchlorate resulted in lower iodide concentrations and these concentrations were low enough to be of serious concern.
Shi et al. [25] were the first to conduct extensive research on perchlorate contamination in China.
Because China is one of the largest manufacturers of fireworks, there is significant potential for
36 perchlorate to be present in the environment. The authors selected four matrices to study
perchlorate exposure - sewage, rice, bottled water and milk. Extensive sampling was conducted
all over China, in as many as 26 cities to determine the extent of perchlorate contamination. The
analysis of the sewage samples indicated that perchlorate contamination was widespread. The average concentration was found to be 22 µg/kg and concentrations as high as 380 µg/kg were being detected. Perchlorate was detected in all of the rice samples as well, although at much
lower concentrations, with the highest concentration being 5 µg/kg. Both dairy milk and breast
milk also showed presence of perchlorate with highest concentrations of 11 and 92 µg/L,
respectively.
Results from these studies confirm that perchlorate contamination is widespread and
because of the mobility of the ion, the problem is not limited to drinking water. This has to be taken into consideration when making policy decisions regarding the regulation and treatment of
perchlorate.
3.5 Analysis
Prior to 1997, perchlorate analysis was performed mainly by using gravimetric liquid-
liquid extraction and spectrophotometry and as a result only concentrations down to 100 µg/L
could be measured in water with reasonable degree of accuracy [18]. In 1997, a new method for
analysis of perchlorate in drinking water was developed by the California Department of Health
Services (CDHS) that utilized ion chromatography (IC) with suppressed conductivity. This
method used a specially developed anion exchange column to separate perchlorate was from
commonly occurring anions and used conductometric detection for measurement of the anions.
37 This method utilized a Dionex IonPac AS5 column, 120 mM sodium hydroxide and 2 mM p- cyanophenol as the eluent, and a 740 µL sample loop was able to obtain a method detection limit
(MDL) of 4 µg/L. This method found widespread application in detecting perchlorate in drinking water and groundwater samples throughout the country. Dionex through development of new anion-exchange columns played a crucial role in the development of new methods of perchlorate analysis in the subsequent years. In 1998, right after EPA added perchlorate to the contaminants candidate list, Dionex came up with a new improved method for perchlorate analysis that utilized a 1000 µL sample loop and had a MDL of 88 µg/L [26]. Although these methods worked with drinking water samples, the sensitivity of the methods was significantly altered by the presence of solvents and other chemicals, especially in industrial wastewater samples. In order to improve the application of the method for non-drinking water analysis, Dionex developed a new column
As-16, which had significantly higher capacity and better sensitivity than the previous two columns [27]. This led US EPA to establish method 314.0 for analysis of perchlorate in water samples in 1999. Since then, there have been a number of new standard methods developed with the capability of measuring perchlorate concentrations in the ng/L range.
Table 3.1 Standard Methods for Analysis of Perchlorate in Water
Method Key Reference Analytical Method Detection Limit (LCMRL1) Name 314.0 Hautman et al. [28] IC 4 µg/L 314.1 Wagner et al. [29] Inline Column Concentration/Matrix 0.140 µg/L (AS-16 column) Elimination IC With Suppressed 0.130 µg/L (AS-20 column) Conductivity 314.2 Wagner et al. [30] Two dimensional IC with suppressed 55 ng/L conductivity 331 Wendelken et al. Liquid Chromatography Electrospray 0.022 μg/L (MRM2 m/z3 83) [31] Ionization Mass Spectrometry (LC-EI- 0.056 μg/L (SIM4 m/z 101) MS) 332 Hedrick et al. [32] IC With Suppressed Conductivity And 0.1 μg/L Electrospray Ionization Mass Spectrometry (IC-EI-MS) 1: LCMRL: Lowest concentration minimum reporting level, 2: MRM: Multiple reaction monitoring, 3: m/z: Mass to charge ratio, 4: SIM: Selected ion monitoring
38
The main standard methods developed include 314.0 [28], 314.1 [29], 314.2 [30], 331.0 [31],
332.0 [32]. Table 3.1 tracks the development of these standard methods over the last few years.
There have been more developments in the area of low level analytical methods for
perchlorate. Kim et al developed a highly selective method for analysis of perchlorate in water
[33]. This method uses an ion-selective electrode that measures electrochemically extremely low
concentrations of ‘redox-inactive’ perchlorate i.e. without any electrolysis. By using a thin-layer
liquid membrane, this method significantly improves perchlorate detection in terms of selectivity
and response time and at the same time uses very small amount of sample. Although the study
did demonstrate analysis of tap and bottled water samples and was able to detect concentrations
as low as 20 ng/L, all the results were in terms of voltammograms and does not provide the
reader any insight on the detection limits through traditional means such as calibration curves
and MDLs. It would have been much more beneficial to run standards and samples to develop
calibration curve and statistical analysis to determine the detection limits.
Wagner et al. [30] developed an automated two-dimensional ion chromatography (2D-
IC) method for measuring perchlorate with suppressed conductivity detection that is currently being prepared for publication as EPA Method 314.2. According to the authors, two-dimensional
IC is essentially an automated “heart-cutting”, column concentration and matrix elimination technique. The first dimension involves injection of a large sample volume onto a first separation column and the subsequent diversion of separated matrix ions to waste while perchlorate and small amount of some other ions are diverted and trapped on a concentrator column. The second
39 dimension involves eluting and diverting the contents of the concentrator column onto a second higher resolution column for separation and quantitation of perchlorate. The use of two columns with different selectivities and formats results in high sensitivity analysis. The authors found that establishing the time intervals for the cut windows for both dimensions was the most crucial step in the analysis. Analyses were performed on different samples ranging from laboratory prepared reagent samples to actual surface water samples from different sources. With this method, they reported a Lowest Concentration Minimum Reporting Level (LCMRL) of 55 ng/L for perchlorate in drinking water samples. This study suggests that this new method has comparable sensitivity and selectivity and is simpler and more economical than IC-mass spectrometric (MS) or IC-MS–MS techniques.
3.6 Regulations
Since the development of a new analytical method in 1997, very low concentrations of perchlorate in water, up to 4 µg/L were detected. Perchlorate began to be detected in water supplies across the country. In March 1998, EPA added perchlorate to the drinking water
Contaminants Candidate List (CCL) [1, 9, 19] based on a Safe Drinking Water Act (SDWA) recommendation. The SDWA required the EPA to conduct further research on perchlorate occurrence and associated health effects before establishing an MCL. The USEPA released an interim guidance assessment on perchlorate and recommended a provisional Reference dose
(RfD) of 0.0001-0.0005 mg/kg-day. In 2005 EPA established a RfD for perchlorate of 0.0007 mg/kg/day [3].
40 Because of no decisive action taken by USEPA, some states started establishing advisory
levels and some even started considering setting MCLs on a state level. Tikkanen [34] looked at
perchlorate contamination in California and how it promulgated the California Department of
Health Services (CDHS) to establish its own action level against perchlorate. In addition to discussing number of contamination sites and the level of contamination, the study discussed perchlorate analytical methods and their development over the last few years. The author also tabulated the action levels that the various states have established although some of these numbers have changed since 2006 and have been updated in Table 3.2.
Table 3.2 Recent Standard Methods for Analysis of Perchlorate in Water
State Standard Regulation Year Source (µg/L) http://www.cdph.ca.gov/certlic/drinkingwater/Pages/Pe California 6 MCL 2007 rchlorate.aspx http://www.mass.gov/dep/service/regulations/310cmr2 Massachusetts 2 MCL 2006 2.pdf http://chppm- Texas 4 DWAL 2002 www.apgea.army.mil/documents/FACT/31-003- 0502.pdf Arizona 14 HBGL 2003 Tikkanen et al. [34] Nevada 18 AL 2005 http://ndep.nv.gov/BCA/perchlorate02_05.htm http://www.scag.ca.gov/wptf/pdfs/wptf092106_Stewart New Mexico 1 DWSL 2006 Perchlorate.pdf 5 DWPL https://www.denix.osd.mil/portal/page/portal/denix/env New York 2008 18 PNL ironment/MERIT/EC/ECAL/Perchlorate/StatesReg MCL: Maximum contaminant level; DWAL: Drinking water action level; HBGL: Health based guidance level; Al: Action level; DWSL: Drinking water screening level; DWPL: Drinking water planning level; PNL: Public notification level
On the federal level, there has been significant back and forth communications between
EPA and Congress over perchlorate regulation in the last few years. In November 2007, the
House Energy and Commerce Environment and Hazardous Materials Subcommittee did approve
a bill that would require the EPA to regulate perchlorate in drinking water [35]. However, in a
big blow to environmental policy makers and academicians supporting perchlorate regulation,
41 US EPA concluded in October of 2008 that perchlorate levels found in over 99% of drinking water systems are not of any concern to public health and recommended that perchlorate in drinking water not be regulated at a national level [36, 37]. Experts felt that the decision was not based on scientific data from numerous studies that prove the contrary but was due to the fact that cleanup of perchlorate contaminated sites arising out of regulation would cost millions of dollars [38].
3.7 Exposure/ Health Effects
Studies have shown that the primary route of exposure to perchlorate is through ingestion of water or food contaminated with perchlorate although dermal absorption and inhalation can also be considered as minor exposure pathways [39]. Perchlorate is known to interfere with the normal functioning of the thyroid by inhibiting the normal uptake of iodide. The thyroid plays an important role in the normal metabolism and growth of a human body. Thyroid is essential during fetal development of the central nervous system [9]. Studies have shown that exposure to perchlorate in pregnant women can result in serious and irreversible effects on the fetus.
Extensive animal toxicity testing has been performed to study the health effects of perchlorate in human beings. Paulus et al. studied the inhibition of iodide uptake in rats exposed to perchlorate
[40]. In order to better understand the impacts, they conducted their study on both normal and iodide deficient rats. As demonstrated by other studies, they found that perchlorate significantly impacted iodide uptake in the normal rats. Interestingly, however, they found that iodine- deficient rats demonstrated significant resistance to inhibition of iodide uptake and only the highest perchlorate concentrations had any adverse effect in this sample set. The authors went on to say that although these tests cannot be conducted directly on human subjects, it might be
42 possible that iodide deficient individual might have better resistance to perchlorate. Park et
al.[41] and Bradford et al. [42] were one of the first researchers to look at uptake of perchlorate
in different species of fish. They compared perchlorate uptake in fish exposed in laboratory
conditions with ones in actual field studies. They also looked at the various factors influencing
uptake kinetics and also looked at how the uptake varied between the different organs. They
concluded that perchlorate was taken up and eliminated rapidly from fish tissue without any
bioaccumulation. Although the authors contend that the results might be important when looking
at fate, effects and transport of the perchlorate ion in natural water systems, they did not delve
into the details of how this is important relative to dietary exposure to perchlorate through
consumption of fish, which in our opinion is more important. Liu et al. found that perchlorate
exposure in addition to influencing thyroid functioning, increased arsenate toxicity and also
resulted in significant growth retardation in zebra fish [43]. Fisher et al. through clinical
laboratory testing on adult rats, claim that exposure to perchlorate not only reduced the iodide
uptake by the thyroid gland but additionally altered the functioning of the thyroid itself [44].
At the same time, there are certain studies that contend the theory that exposure to
perchlorate results in serious health effects. Charlnley [11] argues that the relationship between
perchlorate exposure and inhibition of the thyroid uptake is inconclusive at best, based on the data from the studies that have been undertaken so far. The author feels strongly that the US EPA
RfD of 0.007 mg/kg/day might be ‘unnecessarily stringent’ and even questions whether perchlorate should be regulated. By evaluating the data from various recent studies, where exposure to high perchlorate concentrations has not resulted in adverse effects on public health or fetal development. The author even provides data from studies done in Chile and Israel where
43 even exposure to perchlorate concentrations of more that 100 ug/L has not affected the normal
functioning of the thyroid. The author presents the presence of other ubiquitous contaminants
and statistical misrepresentations as a reason for studies that show a positive relationship
between exposure and reduction in Thyroid uptake. Braverman studied the effect of perchlorate on thyroid function on ammonium perchlorate production workers [45]. The author looked at workers exposed intermittently to airborne perchlorate and also by exposing workers to perchlorate in drinking water over a short and long test period. The author contends that it was
difficult to find volunteers for the latter because of the adverse publicity that perchlorate has
received in the last few years. Although perchlorate was detected in serum of all the workers,
there were no changes in the thyroid function tests that are a measure of the normal functioning
of the thyroid. Although this study provides a new angle to perchlorate and its health impacts, it
should be noted that it was sponsored by a major aeronautical firm. Based on US EPA’s latest
verdict on the regulation, it appears that researchers on this side of the debate have proved their
point
3.8 Treatment
A significant amount of research has been performed to evaluate treatment alternatives
for perchlorate remediation in drinking water in the last decade. Current technologies available for perchlorate remediation in drinking water can be classified into removal or destruction technologies [1]. The major removal technologies would comprise of adsorption with activated carbon, ion-exchange and filtration. Destruction technologies for perchlorate would include biological treatment, chemical reduction, and electrochemical reduction. Each of these technologies is discussed in detail in the following sections.
44
3.8.1 Activated Carbon
On the biggest advantages with Granular Activated Carbon (GAC) adsorption is that it is
widely used in the drinking water treatment industry and it would be easy to retrofit to target perchlorate in the water. Also effective regeneration of the spent carbon makes this technology
economically feasible. Although virgin GAC has not been found to very effective for perchlorate
adsorption, tailoring of GAC had made it comparable to ion-exchange and microbial treatment.
Parette et al. [46] demonstrated perchlorate removal on GAC tailored with cationic surfactants.
Tailoring the carbon increased the breakthrough time by a factor of 30 and up to 34000 bed
volumes could be treated before breakthrough. The authors were able to achieve these results
with actual reservoir water in which other anions were present at a much higher concentration
than perchlorate. In a related study, Parette et al. [47] looked at the effectiveness of GAC
preloaded with CTAC (cetyltrimethylammonium chloride) for perchlorate adsorption. The
authors used Massachusetts groundwater for their Rapid Small Scale Coumn Test (RSSCT)
adsorption experiments. It was demonstrated that with the CTAC tailored carbon, the
breakthrough time for perchlorate increased by more than 8 times when compared to virgin
GAC. However the water also contained nitroorganics RDX and HMX and tailoring the carbon
significantly affected their adsorption. With virgin GAC, up to 300,000 bed volumes of water
containing RDX and HMX could be treated. However this number decreased sharply to less than
8000 with CTAC tailored carbon. The authors recommend using a polishing virgin GAC bed
following the CTAC tailored bed. With this setup, they were able to show that simultaneous
removal of both the perchlorate and the nitroorganics took place with similar breakthrough times.
45
Na et al. [48] evaluated GAC beds preloaded with iron-oxalic acid for perchlorate adsorption.
Through RSSCTs, the authors were able to demonstrate that preloaded GAC enhanced
perchlorate adsorption by an average of 42%. With a 0.34 mg/g adsorption capacity, the tailored
GAC could provide a service life of up to 70 days when compared to only 40 days with virgin
GAC. Also regeneration with sodium borohydride was able to replenish much of the adsorptive
capacity. The authors did not go into any details regarding the practical implications with the
disposal of a concentrated brine stream with perchlorate concentrations of 75-100 mg/L.
Chen et al [49] conducted research on ammonia tailoring of GAC to enhance its capacity
for perchlorate removal. According to the authors, perchlorate adsorption by GAC is directly
related to the surface chemical properties of the tailored carbon. With ammonia tailoring they
were able to increase the nitrogen content of the carbon thereby significantly increasing the
positive charge of the carbon surface, which is the most important surface property influencing
adsorption. They were able to demonstrate a four fold increase in perchlorate adsorption in one
commercially available GAC [50]. They also optimized the temperature for ammonia tailoring
that would result in the highest perchlorate adsorption capacity. In another study the authors also
looked at regeneration of the ammonia tailored GAC with ammonia and carbon dioxide [51].
When compared to steam regeneration, which resulted in loss of nitrogen and positive surface
charge, regeneration with ammonia and carbon dioxide was able to preserve most of the nitrogen
content and hence resulted in increased regeneration capacity for perchlorate adsorption. The key references and their findings for studies looking at perchlorate removal with GAC adsorption are summarized in Table 3.3.
46
Table 3.3 GAC Adsorption of Perchlorate
Key References Findings Parette et al. [46, 47] Virgin GAC adsorption of perchlorate is not cost effective Cationic surfactants increase GAC capacity by 8 folds Successfully demonstrated in a full-scale setup Adsorption capacity for organics diminishes due to tailoring Two beds in series required for complete removal of all contaminants Chen et al. [49-51] Ammonia tailoring increases GAC adsorption capacity by 4 folds Tailoring improves nitrogen content and surface positive charge Temperature during tailoring is critical Regeneration with ammonia and carbon dioxide resulted in increased adsorption capacity when compared to steam regeneration Na et al. [48] GAC tailoring with iron-oxalic acid Improved perchlorate adsorption by ~42% Regeneration with sodium borohydride extremely effective
In summary, virgin GAC has limited capacity for perchlorate adsorption and requires
tailoring to make it practically feasible. However this makes it more expensive and also
influences the adsorptive capacity of the carbon for some of the other contaminants. GAC
adsorption also requires disposal of the perchlorate laden spent carbon or regenerative brine.
3.8.2 Ion Exchange
Ion-exchange is currently the most effective and commonly used technology for perchlorate removal from drinking water. A wide variety of strong basic anion-exchange (SBA) resins have been developed that have very high selectivity for perchlorate. Ion-exchange is a physico-chemical process based on exchanging an anion (typically Cl-) with the perchlorate ion
in the water (equation 1). The ion-exchange resins used to target perchlorate are typically
47 polymers with a strong positively charged functional group, typically a quaternary amine
+ (R4N )[9].
−+ − + − − 4 4 ⇔+ 4 4 + ClClONRClOClNR (3.1)
Gu and Gilbert have done extensive research on perchlorate removal in water using ion
exchange [5, 52-55]. Based on previous research they had done on a radioactive contaminant
similar to perchlorate, they found that the key factor resulting in improved effectiveness of a
membrane is a balance between selectivity and reaction kinetics. They developed bi-functional
resins with two quaternary ammonium groups, the first one with a long chain for high selectivity and the second one with a short chain for better reaction kinetics [56]. The bi-functional resins showed high selectivity and reaction rates especially at low influent concentrations. They also developed a novel regeneration technique for resins loaded with perchlorate [57]. They used the
- tetrachloroferrate (FeCl4 ) ion, formed by the reaction of ferric chloride (FeCl3) with excess
chloride to displace perchlorate from resins. They observed that perchlorate was desorbed
rapidly, with close to 100% recovery and with a very small volume of water required for
regeneration. In another study they found that the same FeCl3 – HCl solution that was used for
regeneration, could be used to completely destroy perchlorate in ion exchange brine solutions
[5]. They demonstrated that at higher temperatures (~200 °C) and high pressure (~20 atm), the
regenerant would completely reduce perchlorate to chloride without altering the properties of the
regenerant, which could then be reused. The authors were able to demonstrate this technology in
a pilot scale reactor in an Air Force base where they not only treated 40000 BVs of 450 µg/L
perchlorate laden wastewater but also regenerated the resin with close to 100% perchlorate
recovery. They were able to achieve these results even when the groundwater contained
48 relatively high concentrations of sulfate and chloride [52]. However, significantly high amounts
of uranium were also concentrated in the resin and the authors had to find a way to separate the
two ions that had accumulated on the resin, before disposal [53]. Gu et al. also demonstrated
perchlorate removal in a field scale setup in a rocket manufacturing facility using these highly
selective ion exchange membranes [54]. They were able to destroy significant amount of
perchlorate in the waste stream by using the above mentioned technology. However, they did not
discuss the cost implications because of the high temperature and pressure requirements and the
feasibility on a large scale. They developed another lab scale alternative where they were also
able to recover perchlorate by precipitation as salts.
Xiong et al. [58] investigated in detail different types of ion exchangers, with different
matrices and functional group and their effect on perchlorate adsorption. They looked at strong
base anion (SBA) resins, weak base anion (WBA), bi-functional resin, polymeric ligand
exchanger (PLE) and an ion exchange fiber (IXF) with regards to their adsorption capacity,
regeneration efficiency, underlying mechanisms and chemical properties influencing perchlorate
removal. It was concluded that it is a combination of factors such as hydrophobicity, intraparticle
diffusion, surface area, pore accessibility that needs to be considered before determining the
suitability of a resin. The most significant finding was that the IXF resins displayed rapid
kinetics for perchlorate sorption and at the same had significantly higher regeneration efficiencies (~85%).
Kim et al. [59] synthesized ammonium functional groups on a mesoporous media and
evaluated perchlorate removal kinetics on these media. They found that reaction rates with this
49 mesoporous media was more than 30 times faster than with a conventional ion exchange resin.
Through detailed pore structure analysis, they attributed this to a significant higher surface area
in the mesoporous media. Also the bi-functionalized media, with two ammonium groups,
performed significantly better than a mono-functonalized media. Although the authors suggested
that this synthesized media could be used in an ion exchange column, they did not discuss the
applicability of this process outside the laboratory.
Dilip et al. found that chromatographic resins used to remove mercury from water also
had moderate capacity for perchlorate uptake and thereby had implications for contaminated sites
[60].
In 2004, when high levels of perchlorate, up to 43 µg/L were detected in the wells
supplying water to the town of Millbury, Massachusetts [61], the water supply had to be shut
down until authorities installed a perchlorate treatment system. A full scale ion exchange system
from Siemens was selected for this particular application as the most effective and economical
alternative, with the perchlorate in the spent resin being destroyed by incineration.
The key references and their findings for studies looking at perchlorate removal with Ion-
Exchange are summarized in Table 3.4.
50
Table 3.4. Perchlorate Removal with Ion-Exchange
Key References Findings Gu et al. [5, 46, 47, 52-57] Balance between selectivity and reaction kinetics is key Developed highly selective Bi-functional resins Demonstrated good selectivity and high reaction rates even at low concentrations Ferric chloride and excess chloride for complete sorption of perchlorate, recovery rates close to 100% Regenerant solution can also be used for completed destruction of perchlorate ion in the brine solution Successful full scale demonstration of this technology in the presence of other competing anions
Although ion-exchange has been well tested and has been found to be the most effective
technology for perchlorate removal in water, there are still some major obstacles to its
application in the field. The biggest issue is the disposal of the spent brine which is concentrated
in perchlorate. Also, presence of other commonly occurring anions at much higher
concentrations than perchlorate can result in competition for the available sites on the resin,
thereby reducing the capacity for perchlorate.
3.8.3 Membrane Technologies
Membrane technologies are based on the principle of employing a semi-permeable
membrane that prevents the passage of certain ions thereby treating the water. Technologies such
as ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO) have been somewhat
effective for perchlorate removal.
Huq et al. looked at polyelectrolyte enhanced ultrafiltration (PEUF) for perchlorate
removal [62]. PUEF is based on the principle of adding polymers to the water which form complexes with the target ion and are retained on a filtration membrane. With water containing
51 only perchlorate, up to 90% removal was achieved at low polymer doses. However in the presence of sulfate and nitrate, perchlorate removal decreased and a higher polymer dosage had to be used. Roach et al also looked at colloid enhanced ultrafiltration for perchlorate removal
(CEUF) [63]. They looked at various factors affecting CUEF and showed up to 95% perchlorate removal even in the presence of other competing anions. This technology has some serious drawbacks when it comes to feasibility; not only does the perchlorate laden brine have be disposed but also unknown chemicals are being introduced into the system.
Yoon et al studied hindered diffusion of perchlorate through ultrafiltration and
nanofiltration membranes [64, 65]. The study looked at the effects of factors such as pH,
conductivity, pore surface properties on diffusion. However, no effort was made to highlight or
make the reader understand the implications of this research for removal of perchlorate from
water. Lee et al. performed some preliminary studies using nanofiltration (NF) for perchlorate
removal [66]. They found that some ‘tight’ NF membranes with low molecular weight cutoffs
were able to remove up to 90% of the influent perchlorate. Some of the key references and their
findings for studies looking at perchlorate removal using membrane filtration are summarized in
Table 3.5.
Table 3.5 Perchlorate Treatment with Membrane Technologies
Key References Findings Huq et al. [62] Polyelectrolyte enhanced ultrafiltration (PEUF) High perchlorate removal at low polymer doses Very high dosage required in the presence of other anions Roach et al. [63] Colloid enhanced ultrafiltration (CEUF) Upto 95% perchlorate removal even with other anions present in the water Feasibility is questionable because of introduction of unknown particles Yoon et al. [65, 67] Diffusion of perchlorate ion through UF/NF membranes Factors affecting diffusion
52 These pressure driven filtration technologies suffer from certain drawbacks that makes their
application in large scale systems challenging. They suffer from fouling issues and generate large
volumes of reject streams that would require further treatment or disposal. Electrodialysis has
been found to be more effective than conventional membrane filtration for perchlorate although
it has some exorbitant operation costs associated with it.
3.8.4 Chemical Reduction
Although the physical removal technologies are capable of removing perchlorate from
water, they do not offer a complete solution, as they only transfer the perchlorate from one
stream to another, necessitating further treatment and disposal. A destruction technology that
completely reduces perchlorate in water to harmless chloride is definitely preferable. A chemical
reduction technology that is able to achieve this objective would be the most optimum.
Perchlorate is a strong oxidizing agent based on the central chlorine atom in its highest
oxidizes state (+VII). Both the thermodynamics (ΔE<0) and reduction potentials favor reduction
of perchlorate to chloride [68].
− −+ − 4 88 +⇔++ 4 2OHCleHClO E0 = 1.287 V (3.2)
However, perchlorate does not exhibit these oxidizing properties at conditions found in
contaminated water and as a result is not reduced by common reducing agents, primarily because
of the reaction kinetics. The reaction rate is controlled by the high activation energy resulting from the tetrahedral structure where the central chlorine atom is surrounded by the four oxygen
53 atoms (Figure 3.2). The only mechanism of reducing aqueous perchlorate is by overcoming this
energy barrier. Extensive research has been done to develop catalysts that could be used to
improve the reduction kinetics.
Cl
Figure 3.2 Structure of the Perchlorate Ion
Hurley et al developed a new heterogeneous catalysts for reduction of perchlorate [69].
This Pd/Rh-C bimetallic catalyst demonstrated rapid perchlorate reduction in the presence of
hydrogen and was able to achieve more than 99% reduction in 5 hours. Although the reaction
rate achieved with this catalyst was comparable to that achieved with other effective
technologies, there were two major limitations. The reaction occurred at a pH of less than 3 and
at a pressure of 5 bars, conditions not normally found in conventional water treatment systems.
Abu Omar et al. studied reduction of perchlorate using a methyloxerhenium oxides as
catalysts [70, 71]. They demonstrated pseudo first order reaction rates of ~5.5 s-1 that were
significantly faster than common reducing agents. However, the reduction occurred at a pH of 0 and the reactions governing the formation of the catalysts themselves were complex and difficult to control. Another study by Abu Omar looked at mechanisms of oxygen atom transfer reaction between perchlorate and Rhenium oxazoline catalyst [72-75]. Although the authors claim that the catalyst is easy to prepare by a series of chemical reactions, they acknowledge that practical application will be significantly challenging.
54 The key references and their findings for studies looking at perchlorate removal with chemical reduction are summarized in Table 3.6.
Table 3.6. Chemical/ Catalytic reduction of perchlorate
Key References Findings Hurley et al. [69] Heterogeneous Pd/Rh-C catalyst for perchlorate reduction High perchlorate reduction kinetics Drawbacks: Very low pH, high pressure and hydrogen gas Abu-Omar et al. [73, 74] Mthyloxyrhenium oxide catalyst Reasonable high raction rates Limited practical application Wang et al. [76-78] Monometallic catalysts and hydrogen gas High reaction rates on Ti-TiO2 surface Electrodialytically assisted catalytic reduction Extremely high rates of upto 5.5 day-1 achieved Moore et al. [6, 79, 80] Perchlorate reduction on zero valent rion Iron surface provides a catalytic effect for complete reduction of perchlorate Reaction kinetics extremely slow for practical implementation
Makris et al. [81] studied perchlorate removal by aluminum based drinking water treatment residuals. In an interesting lab scale study they found that, these residuals which are essentially waste products generated from water treatment plants are good adsorbents due their highly porous structure and high surface area. They observed an 80% reduction in perchlorate after 24 hours through sorption followed by reduction, although these high reduction rates were observed with perchlorate concentrations of 250 mg/L, which are too high for practical purposes.
Wang investigated Catalytic reduction of perchlorate with hydrogen gas [76]. They looked at aqueous perchlorate reduction by hydrogen in the presence of various catalysts (78 different catalytic systems) in a pressurized reactor. The reaction rates with most catalysts were extremely slow. Although on Ti-TiO2 surface, with hydrogen gas and a small current, they were able to demonstrate 90% perchlorate reduction in a matter of hours, even at a 10 ppm initial concentration at a first order rate of 0.8 day-1, which is comparable to biological reduction. Wang et al [77] looked at perchlorate reduction in a electrodialytically assisted catalytic membrane
55 (EDACM) reactor. Application of current on the catalytic membrane coated with nanoparticles
of a monometallic membrane resulted in generation of atomic hydrogen, which reduced
perchlorate. The authors looked at a number of similar monometallic catalyst assisted rapid
reactions even at low concentrations. Wang et al. also looked at electrodialytically assisted
catalytic reduction (EDACR) reduction of perchlorate, where electrodialysis facilitated
movement to perchlorate towards a cathode coated with nanometallic catalysts that eventually reduces it [78]. They were able to achieve first order reaction rates as high as 5.5 day-1.
Mahmudov et al. [82] studied perchlorate reduction on mono and bimetallic catalysts supported on activated carbon with hydrogen gas. Although they obtained fast reaction kinetics, perchlorate reduction in gas phase was much faster than in aqueous phase.
Studies have shown that aqueous mineral surfaces provided the necessary catalytic effect to improve the kinetics of certain redox processes such as oxygenation of metal ions and organic pollutants [83, 84]. Research has been performed on the reduction of perchlorate on iron surface
[6, 7, 85]. Moore et al. [6, 79, 80] has studied extensively perchlorate reduction on zero valent iron. Huang et al. [86] looked at perchlorate reduction with zero-valent iron filings. Through batch and column experiments, the authors were able to determine reaction kinetics and understand perchlorate reduction mechanism at the iron surface and were able to demonstrate that iron serves as a reducing agent and a catalyst simultaneously during the reaction at the iron surface. Although iron filings were effective in reducing perchlorate, the reaction rates were extremely slow and were significantly affected by the presence of other common anions. Studies have also demonstrated perchlorate reduction on nano-scale iron particles [87, 88]. Nano particles have shown high reactivity especially at high temperatures [89].
56 In summary, the kinetics associated with most of the chemical reduction processes is too slow to be practical and requires the use of catalysts. Catalysts require use of chemicals and conditions that are not normally present in drinking water systems and invariably contain heavy metals. Although use of hydrogen could be justified economically, it presents a significant explosion hazard. Although chemical reduction technology might be effective for a small system treating perchlorate laden wastewater, it is not ready to be used on large scale drinking water systems unless an effective catalyst that is completely harmless is discovered. However, this method has found some application in the form of PRBs for treating groundwater contaminated with perchlorate.
3.8.5 Electrochemical Reduction
Electrochemical reduction shows a lot of promise as a removal technology in the sense
that it completely destroys perchlorate without the use of catalysts. However, implementing it on
a large scale is a completely different story.
Numerous studies have shown that perchlorate can be reduced electrochemically to Cl-.
Horanyi and Bakos [90] studied the influence of the method of platinization of platinum
- electrodes on reduction of ClO4 ions. They concluded that the platinization technique determines
- the electrocatalytic properties of the electrode and hence its ClO4 reduction capacity. Horanyi et.
- al [91] used technetium (Tc) and Rhenium (Re) electrodes with a ClO4 based electrolyte and
- were able to conclude that a significant reduction of ClO4 took place at these electrodes with a
- strong dependence on the ClO4 ion concentration of the electrolyte. According to Lang and
Horanyi [7], the presence of sulfate or chloride ions can have an inhibiting effect on the
57 - reduction of the ClO4 ions at electrodes, thereby hinting at competitive adsorption that could
- exist during the electrochemical process as well. Clom et al.[92] studied ClO4 reduction on
- Ruthenium electrodes and found out that ClO4 reduction was a very slow process but the rate appeared to increase notably with an increase of the solution temperature. They also observed the
- reduction to be a function of hydrogen adsorption and ClO4 concentration. Almeida et. al.[93]
- looked at ClO4 reduction at tin electrodes and concluded that the quantity of chloride formed
- during the reduction was a function of the pH and the ClO4 reduction was faster in more acidic
- solutions. Wasberg and Horanyi [94] studied the ClO4 reduction on Rhodium (Rh) electrodes.
- They looked at reasons for the low electrochemical reactivity of ClO4 and how it is related to the
structure of the ion. They went on to conclude that the whole reduction process was a surface
phenomenon and the reaction itself was very complicated because as many as eight electrons are
involved. It should be noted that all the above mentioned studies were performed by
electrochemists and show the reduction in terms of voltammetric curves. These results are not
very clear for understanding reaction kinetics and for comparison with other treatment
technologies.
Theis et. al.[95] conducted a study on the electrochemical and photocatalytic reduction of
- ClO4 ion using titanium electrodes. They performed experiments at different initial perchlorate
concentrations and found out that as the initial perchlorate concentration decreased, the percent
reduced increased. They also observed that better reduction occurred at lower pH, as suggested
by other researchers. They concluded that the limiting factor in perchlorate reduction was
competition among anions for active sites on the electrode surface. The chloride ion formed from
- the reduction reaction also competed with ClO4 for adsorption sites, thereby slowing the reaction
58 with time. In order to study photochemical reduction, the electrodes were coated with TiO2 and
UV photocatalysis was employed. Doping of the electrodes with vanadium resulted in a substantial increase in the percent reduced. The authors went on to conclude that to make this technology practically feasible, the number of reactive sites on the electrode surface relative to the solution volume (A/V) will have to be increased.
The key references and their findings for studies looking at perchlorate removal with electrochemical reduction are summarized in Table 3.7.
Table 3.7 Electrochemical reduction of perchlorate
Key References Findings Theis et al. [95] Electrochemical reduction – extremely slow kinetics Electrodes coated with TiO2 showed improved reduction Reaction rates too slow to be practical Horanyi et. al [91] Significant perchlorate reduction at various electrodes Lang and Horanyi [7] Competitive adsorption on the electrodes between perchlorate and other common anions during electrochemical reduction
Like chemical reduction, reduction of perchlorate by electrical means also is too slow to
be practical. There has not been a single pilot scale demonstration of this technology for perchlorate remediation.
3.8.6 Microbial Remediation
Microbial perchlorate reduction has shown a lot of promise for large scale applications.
Several microorganisms have been identified to successfully reduce perchlorate to chloride.
Since this technology is well established now and has been successfully demonstrated in a
multitude of large full-scale studies, the technology has undergone significant optimization. It is
based on the principle that these bacteria contain special enzymes that lower the activation
59 energy required for perchlorate reduction, thereby using the perchlorate ion as an oxidant for
their metabolism.
Chaudhari et al. and Wu et al. studied factors that influence microbial perchlorate reduction [96, 97]. Chaudhari et al. investigated two strains of a perchlorate reducing bacteria and found that anaerobic conditions were essential along with presence of Molybdenium for perchlorate reduction by either strain. However, one strain was able to reduce perchlorate even in the presence of nitrate whereas the other one reduced nitrate before perchlorate. The authors concluded that although microbial perchlorate reduction is effective, a lot of attention needs to be paid to these controlling environmental factors.
Since in most waters perchlorate is present along with other anions such as nitrate, it is
important to understand how microbial perchlorate reduction occurs in the presence of other
anions. Xu et al looked at reduction of perchlorate and nitrate in the presence of bacteria
acclimated to only one of the two electron acceptors [98]. They found that the enzymes for
perchlorate and nitrate reduction were separately expressed. They found that bacteria that were
grown in the presence of both nitrate and perchlorate had better perchlorate reduction when
compared to bacteria grown only on perchlorate. Also, bacteria grown on only nitrate did not
show any perchlorate reduction. These results are important with regards to enzymatic reactions
in the presence of other anions and to develop more efficient strains for microbial perchlorate
reduction. Coates et al. also looked into the details of the mechanism microbial perchlorate
reduction along with the factors influencing this process [99] and developed a flowchart showing
the perchlorate reduction mechanism as shown in Figure 3.3.
60
Figure 3.3 Factors Influencing Microbial Perchlorate Reduction [99]
Coates et al. investigated the uniqueness of microorganisms responsible for perchlorate reduction [100]. According to the authors, previous studies looked at on only six types of bacteria for perchlorate reduction. Through their research, they demonstrated that bacteria from many diverse environments could reduce perchlorate under anaerobic conditions and as a result, these bacteria were not as unique as previously thought.
Giblin et al. investigated perchlorate reduction in a flow through bioreactor [101, 102].
They used Celite based columns inoculated with the strain Perclace, to study perchlorate removal in groundwater from wells in California. With a flow rate corresponding to a retention time of about 5 hours, they were able to reduce perchlorate to below detectable levels of 4 µg/L.
At a lower retention time of 2.5 hours, perchlorate removal of about 95% was achieved. Nitrate concentrations present at ~50 folds higher than perchlorate did not influence the result and both
61 contaminants were reduced to BDL. The authors found that the microbially active zone was in the first 25% of the column and the process needs to be optimized with active biomass throughout the column.
The key references and their findings for studies looking at microbial perchlorate removal are summarized in Table 3.8.
Table 3.8. Microbial Perchlorate Reduction
Key References Findings Giblin et al. [102, 103] Reduction in flow through bioreactor Very effective for complete redction of perchlorate Nitrate present at much higher concentrations did not affect reduction rates significantly Xu et al. [98] Nitrate and perchlorate enzymatic reduction pathways are separate Nerenberg et al. [104-106], Dudley Kinetics of hydrogen oxidizing perchlorate reducing et al. [107] bacteria Observed chlorate accumulation during reduction Demonstrated perchlorate reduction with this bacteria in a pilot scale hollow fiber membrane biofilm reactor Use of hydrogen is the major drawback Frankenberger et al., Okeke et al. Extensive genetic analysis of enzymes responsible for [108, 109] perchlorate reduction Hatzinger et al. [110-112] Full scale demonstration of a microbial FBR Consistent perchlorate removal over a long period of operation High perchlorate removal even in the presence of nitrate
Nerenberg et al. determined kinetic parameters for a hydrogen–oxidizing perchlorate reducing bacteria [106]. They found that based on kinetics, at low concentrations of perchlorate, other electron acceptors such as oxygen and nitrate are required to sustain high perchlorate reduction. Interestingly, they also observed chlorate accumulation during perchlorate reduction and found high chlorate accumulation inhibited perchlorate reduction until the chlorate was reduced. In a related study Dudley et al. [107] observed chlorate accumulation in perchlorate reduction with HCAP, which is a novel perchlorate reducing bacteria. Based on the kinetics, they
62 found that HCAP would not be very effective at low perchlorate concentrations. However at higher concentrations and in the presence of conventional perchlorate reducing bacteria, significantly higher overall reduction rates could be achieved due to synergistic effects. In another study, Nerenberg et al. also demonstrated perchlorate reduction in a hydrogen – oxidizing hollow fiber membrane biofilm reactor [104]. Perchlorate was reduced to below detectable concentrations in an actual groundwater with an influent concentration of 100 µg/L without the use of specialized perchlorate reducing bacteria. Presence of nitrate did reduce the perchlorate reduction rate although perchlorate did not influence the denitrification rate and supply of hydrogen gas was the only special requirement in this system. Although the authors claim that hydrogen is effective and at the same time safe and inexpensive to use, the explosiveness hazard cannot be ignored as far as practical application of this technology is concerned. In a recent study, Nerenberg studied in detail the ecology of the bacteria used in the hydrogen based membrane biofilm reactor [105]. They concluded that oxygen and nitrogen served as the primary electron acceptors although presence of perchlorate enriched the biofilm significantly through respiration. They also observed that oxygen was a better electron acceptor than nitrate and resulted in higher perchlorate reduction rates. Padhye et al. [113] also investigated perchlorate reduction in a bioreactor with autohydrogenotrophic bacteria. In this setup, microorganisms reduced the contamination in the water from non-oxidized substrate.
Although they demonstrated perchlorate reduction in this reactor, they could not get complete control over the biological process and growth of sulfate reducing bacteria affected the performance. However, there was enough hydrogen for simultaneous removal of nitrate to occur.
It was concluded that further research is required for any practical application of this technology, especially in the presence of other competing anions. Song et al found that presence of cyanide
63 had a significant inhibitory effect on growth of perchlorate reducing bacteria [114]. However
they conclude that cyanide did not result in loss of perchlorate reductase activity but rather
resulted in dissolved oxygen accumulation which prevented anaerobic cell growth.
Frankenberger has done extensive research on microbial perchlorate reduction [102, 103,
108, 109]. In one study, the authors were able to extract important biochemical and molecular data on the perchlorate reductase enzyme through detailed genetic analyses [108]. They were able to arrive at the DNA sequence of this enzyme which is critical in developing a highly efficient microbial perchlorate reduction system. It was found that although the perchlorate reductase enzyme is active on both perchlorate and nitrate, separate terminal reductases are present in the perclace bacteria to reduce the two anions [103, 109]. Bender et al. through genetic analysis of the perchlorate reductase enzyme attempted to target the gene that was directly responsible for perchlorate reduction functions [115]. It was concluded that there is a
common reductase ancestor which acquired mutations specific for utilization of a specific contaminant. Kengen et al. discovered that it was one enzyme that was responsible for reduction
of both chlorate and perchlorate [116]. They were the first to purify and characterize this
enzyme, perchlorate reductase. They also found key differences between perchlorate and nitrate
reductases.
Atikovic et al. studied perchlorate along with RDX reduction in an anaerobic fluid bed
reactor since both contaminants are commonly found together in ammunitions wastewater [117].
They observed competition for electron donors between the two contaminants with perchlorate
reduction occurring at lower dosages of the electron donor and RDX requiring much higher
64 dosages. They concluded that anaerobic treatment of the two contaminants in separate reactors
would be more efficient than treating them together. Wang et al. evaluated kinetic parameters for
microbial perchlorate reduction through batch experiments [118]. They also summarized kinetic
parameters comparing their results with those obtained from previous studies. They conducted
experiments to study the variation in the specific substrate removal rate with pH and fit the
results into a Gaussian distribution model. The results indicated a pH of 7 to be the most
optimum for microbial perchlorate reduction and an increase or decrease in pH directly
influenced the removal rate. Studies have shown that microbial communities can reduce
perchlorate to chloride even in deep unsaturated contaminated soils thereby providing insitu
remediation [119].
Wu et al [97] demonstrated perchlorate reduction by microbial cultures inoculated with
sludge from a domestic wastewater plant. Within 24 hours, perchlorate reduction to non detectable limits was achieved from an initial concentration of 50 mg/L. They also looked at various factors that influence perchlorate reduction and found that acetate was a very effective carbon source and maximum perchlorate reduction occurred at a near neutral pH at a temperature of 40 °C. Ju et al. [120] also demonstrated perchlorate reduction by chemolithotrophic bacteria present in wastewater sludges using two different inorganic electron donors. They found that the highest kinetic rates were observed with bacteria that were enriched using elemental sulfur as the electron donor.
Hatzinger et al. has done extensive research on demonstration of microbial perchlorate
reduction [110-112]. They were the first to demonstrate full-scale bioremediation of perchlorate
65 in a fluidized bed reactor (FBR) [110]. With GAC as the FBR support media, they were able to achieve perchlorate reduction to less than 4 µg/L, at an influent concentration ranging from 0.5 to 1 mg/L. These results were observed consistently over a period of two months even with high levels of nitrate in the influent groundwater. They also investigated in laboratory scale situ perchlorate reduction by indigenous bacteria present in the aquifer from where the groundwater sample was collected. By studying the reduction kinetics and by performing microcosm studies they were able to conclude that with the right substrate, these indigenous bacteria were very effective in reducing perchlorate. The results from a full scale demonstration of in situ perchlorate biodegradation in a naval facility were also published [112]. The microcosm studies at the site showed that maximum perchlorate reduction occurred at a pH of 7.3 with lactate as the electron donor. As a result, the acidic aquifer was buffered to get a near neutral pH and was also injected with lactate as the electron donor. Results showed more than 95% perchlorate reduction in the monitoring wells showing successful demonstration of the technology. Even nitrate levels in the test plot were reduced to below detection limits.
According to Holdren et al., a good example of the occurrence of natural microbial reduction of perchlorate is in Salton Sea in California [121]. Although up to 4 µg/L of perchlorate has been detected in the Colorado River, which feeds the Salton Sea through a few tributaries, no perchlorate has been detected in the Salton Sea itself. The study has found that low levels of dissolved oxygen and redox potential were found in the bottom waters of the sea, thereby providing conditions that are extremely favorable for bacterial reduction of perchlorate.
This theory is further confirmed by the fact that perchlorate was detected in the evaporation ponds where biological reduction could not have occurred because the pans were made of
66 stainless steel. Also the perchlorate concentration in the tributaries has also reduced over the last few years and the authors attribute this to the same reduction phenomenon taking place in the river sediments as well. One point that could be raised against this theory is the perchlorate reduction that could be occurring in the rivers due to control measures being implemented upstream.
In summary, there is no shortage of literature on microbial perchlorate reduction with plethora of studies conducted on this technology. There have also been full scale demonstrations showing the effectiveness of this technology. However microbial remediation has never been used directly in drinking water treatment systems because of the issues of public acceptance.
There is still limited information on the health effects associated with these microorganisms and moreover it would be extremely difficult to convince the public regarding the safety of the water with these organisms being present. Although microbial perchlorate reduction is almost perfect for treating perchlorate containing wastewater, its direct application in large scale drinking water systems still remains questionable at best.
3.8.7 Integrated Technologies
Based on the previous sections, it can be concluded that for various reasons, no single technology is ready to be applied in the field for perchlorate removal in a drinking water system.
However numerous studies have shown that a combination of technologies might be more feasible.
67 Choi et al. [122] demonstrated perchlorate removal on biologically active carbon (BAC), which is nothing but granular activated carbon (GAC) supporting biological growth. Not only
does the carbon act as the support media but it also provides for sorption of perchlorate resulting
in better microbial removal. In order to better understand the results the authors ran the same set
of experiments on support media with no sorptive capacity. Although complete perchlorate
removal was observed in both reactors under ideal dissolved oxygen (DO) conditions and with
the right substrate conditions, the differences were observed under non-ideal conditions. Even at
high influent DO concentrations and low substrate conditions, the BAC was able to achieve
complete perchlorate reduction. This high removal efficiency was attributed to chemisorption of
oxygen occurring on the BAC, which reduced the DO concentration thereby enhancing microbial
perchlorate reduction. Unlike plug flow BAC reactors, where significant desorption of
perchlorate has been observed after backwashing, in the completely mixed BAC used in this
study, there was an even distribution of biomass and backwashing resulted in improved
perchlorate removal due to increased adsorptive capacity for both perchlorate and for
chemisorption of oxygen. The authors as future research recommended that more work is
required to determine the regeneration of the sorptive capacity before practical application of this
BAC can be considered.
Brown et al. [123] studied perchlorate removal on GAC and through simple batch and
column experiments found that removal was mainly by ion-exchange and not by chemical
reduction. They also concluded that the sorptive capacity of GAC for perchlorate was very
limited. They also observed competitive adsorption from other contaminants and that perchlorate
removal was the highest at a low pH. They also looked at perchlorate reduction on BAC and
68 found that perchlorate was reduced to BDL and also that this reduction could be sustained over a
long period of time. They concluded that BAC had significant potential to become a feasible
treatment technology for perchlorate in drinking water. Brown et al. conducted a pilot study on
groundwater to understand perchlorate reduction in a fixed bed bioreactor [124]. They demonstrated perchlorate reduction to BDL by using a GAC supported microbial reactor that used indigenous organisms. By having both microbial reduction and GAC adsorption, the system ensured the prevention of any breakthrough. Even after a shutdown, backwash and in the presence of other anions, the performance was not affected.
In a recent study, Hristovski et al. [125] investigated simultaneous removal of perchlorate
and arsenate on a hybrid ion exchange (HIX) resin, which was prepared by impregnation iron oxide nanoparticles into a strong base ion-exchange resin. The resins were selected for removal
and perchlorate and arsenate was expected to be removed by binding to the iron oxide
nanoparticles. The authors conducted batch adsorption to determine the kinetic parameters for
both contaminants followed by a fixed-bed adsorption column study in simulated groundwater.
The authors also looked at different treatment methods of precipitating the iron oxide
nanoparticles on the ion exchange resin. They found that arsenate breakthrough took
significantly longer in the HIX media when compared to the virgin resin. It would have been
better had they compared breakthrough curves for HIX and virgin resin for perchlorate
separately as well. The column experiment with water containing both arsenate and perchlorate showed that the mass transfer kinetics were rapid enough for simultaneous removal of both contaminants. Although the study did address the uncertainty regarding the health effects of nanoparticles, it concluded that HIX resins will be more effective and economical than using two
69 separate column in series in cases where removal of both arsenate and perchlorate might be
necessary.
Lehman et al. [126] investigated the feasibility of using an ion exchange process followed
by biological treatment of the concentrated brine solution for simultaneous removal of
perchlorate and nitrate. The pilot study used a commercially available resin in a column followed
by a microbial sequencing batch reactor (SBR). The concentrated brine has perchlorate
concentrations as high as 3 mg/L and the SBR was able to reduce it to less than 500 µg/L and in
over 20 cycles of regeneration, perchlorate was reduced to below the detectable limit and there
was no accumulation of perchlorate in the system. The authors through a cost evaluation
concluded that this method would be 20% cheaper than conventional ion exchange followed by brine disposal. However the process is limited by the number of cycles and eventually perchlorate will start accumulating on the resin and will have to be disposed. Also, some other contaminant might accumulate in the process deeming the brine a hazardous waste. In a related study, Lin et al. found that presence of divalent cations such as Ca2+ and Mg2+ in the brine
resulted in notable increase in the microbial perchlorate reduction rates [127]. Their results
corroborated earlier studies that found metals present in the suitable concentrations are important
for cell growth in perchlorate reducing bacteria. Maximum reduction was observed at Mg to Na
ratio of 0.11. Chung et al. [128] investigated microbial perchlorate treatment of ion exchange
brine in a hydrogen based membrane biofilm ractor (MBfR). They found that perchlorate
reduction rates were significantly influenced by salt concentrations in the brine with higher salt
concentrations having an inhibitory effect on the microbial growth. Perchlorate reduction was
also directly related to the hydrogen pressure in the reactor. Similar results were observed by
70 Van Ginkel et al. [129] who also looked at perchlorate and nitrate reduction in a MBfR. They
also found that the highest removal rates for both nitrate and perchlorate were observed in cases
where the inoculum was well acclimatized over a long period of time.
Matos et al. [130] evaluated an ion exchange membrane bioreactor (IEMB) for
simultaneous perchlorate and nitrate removal. The IEMB is based on driving the target anions
through an anion exchange membrane into a bioreactor where perchlorate is reduced to chloride
by microbial activity. The process was able to reduce perchlorate in simulated polluted water
from 100 ppb to less than 4 ppb and the reactor was able to operate with the same membrane for
over 2 months without any fouling issues. The IEMB was able to perform to the same levels
even under substrate limiting conditions. The authors also discussed briefly how the process
compared with other perchlorate removal technologies although no effort was made to discuss
the feasibility of using this technology in the field.
Yu et al investigated microbial perchlorate reduction in the presence of zero valent iron
(ZVI) [131]. They used the autotrophic Dechloromonas strain that used H2 gas as the electron
source and CO2 as the carbon source. Through batch experiments with only ZVI and ZVI and
bacteria, they found that perchlorate degradation rate increased from 0.0033 hr-1 to 0.0758 hr-1.
In the presence of nitrate at much higher concentrations, they observed a notable decrease in perchlorate reduction rates. They authors simplified the process into three main steps and the corresponding factors: 1. Reactivity at the ZVI surface, 2. H2 mass transfer and 3. Microbial cell
density. By conduction experiments by varying these factors, they concluded that the process
was limited by bacterial kinetics and increasing cell density had a direct impact with an increase
71 in the perchlorate reduction rates. Son et al also demonstrated similar results for microbial
perchlorate reduction in the presence of ZVI in batch and column experiments [132]. They found
that H2 gas from anaerobic iron corrosion supported the microbial mixtures that completely
removed 65 mg/L of perchlorate in 8 days. The study claimed that reduction kinetics were
comparable with that achieved with acetate or hydrogen gas fed systems that can be expensive
to operate and suffer from explosion hazards respectively. The authors favor integrated ZVI-
microbial processes for perchlorate remediation over other systems. In a stark contrast to this
study, Shrout et al. found that addition of ZVI to a mixed bacterial culture significantly inhibited
perchlorate reduction [133]. Through characterization studies, they inferred that addition of ZVI
increased the pH and also resulted in formation of iron precipitates on the cell surface that
inhibited microbial perchlorate reduction by encapsulating the bacteria.
The key references and their findings for studies looking at perchlorate removal with
integrated technologies are summarized in Table 3.9.
It can be seen that integrated technologies are certainly more effective than any single
technology for perchlorate removal. It would not be a stretch of imagination by any means to say that if a technology is applied for perchlorate remediation in the near future, it would be a combination of two or more conventional treatment technologies.
72 Table 3.9. Perchlorate removal with Integrated technologies
Key References Technologies Findings Choi et al. [122], Brown Microbial reduction supported on Sorption followed by microbial et al. [123, 124] GAC media reduction Highly effective even under non- ideal conditions of high DO and low substrate concentrations GAC regenerations and competition from another anions are the limiting factors Full scale demonstration perchlorate removal with indigenous bacteria sustained over a long period of time Hristovski et al. [125] Iron oxide nanoparticles Hybrid resin showed better impregnated on ion exchange resin breakthrough behavior when compared to the virgin resin Lehman et al. [126], Lin Ion exchange followed by Microbial batch reactor reduced the et al.[127], Chung et al. microbial treatment of the brine perchlorate effectively [128], Van Gunkel et al No accumulation of perchlorate in [129]. the system 20% cheaper than a conventional iox exchange followed by brine disposal Hydrogen based membrane bioreactors effective for complete reduction of perchlorate Optimum concentrations of metals and salt, and well acclimatized inoculum maximize reduction rates Yu et al. [131, 134], Son Microbial reduction on Zero Demonstrated complete perchlorate et al. [132], Shrout et al Valent Iron (ZVI) reduction [133]. Reaction kinetics comparable with other technologies External H2 gas not required Contrasting results showing that iron inhibited bacterial growth by forming precipitates
3.9 Conclusion
Perchlorate has been the most challenging contaminant to treat and regulate for the policy makers and the environmental technology community in recent times. Perchlorate and its salts are mainly used in solid propellants and for manufacturing matches, rockets, explosives and fireworks and it also occurs naturally. Perchlorate contamination in drinking water and food supplies has recently become a concern for human health, as studies have shown that it can
73 interfere with the normal iodine uptake by the thyroid gland, resulting in decreased thyroid
production. Application of a sensitive new detection method has revealed widespread perchlorate
contamination of groundwater, particularly in the southwestern United States. Although the US
EPA has not established maximum contamination level for perchlorate at this time, several states
have already established their action levels against perchlorate contamination in their drinking
water supply.
- The perchlorate anion (ClO4 ) is exceedingly mobile and persistent in groundwater under
typical environmental conditions. Most standard physical and chemical water and wastewater treatment processes are not generally applicable to remove or destroy the perchlorate ion. The
standard redox potential and thermodynamics render perchlorate a strong oxidizing agent; however, due to its high activation energy of 120 kJ mol-1 [5] its oxidizing power is retarded.
The high activation energy is mainly caused by the tetrahedron structure of perchlorate in which
chlorine is surrounded by four oxygen atoms. As a result, perchlorate is extremely slow to react
and cannot be reduced with common reducing agents or be precipitated using commonly
available treatment reagents.
Currently, perchlorate treatment technologies are generally classified into two categories, destruction technologies or removal technologies. The destructive processes can be divided further into two sub categories: biological and chemical treatments [1]. The non-destructive removal processes or physical processes mainly focus on ion exchange resins. Physical removal processes are suitable for perchlorate removal in small scale water treatment plants. However, such processes have some disadvantages, such as lack of selectivity as well as disposal or
74 treatment of the concentrated residue. As a result, destructive processes which reduce perchlorate to harmless chloride are preferred.
Perchlorate reduction by biological processes under anaerobic conditions has been widely investigated for many years ([135-138]. It is postulated that special enzymes in the microorganisms’ cells serve as catalysts during the degradation of perchlorate, which effectively decrease the activation energy of perchlorate reduction. Due to the effectiveness of biological processes, the U.S. Army and Air Force have already installed perchlorate treatment facilities at several military manufacturing factories [135]. However, in the case of drinking water treatment and groundwater remediation, biological treatment has its own limitations due to the presence of unknown pathogens and the requirement of nutrient addition. An effective chemical reduction process which is favorable for drinking water treatment and groundwater remediation is, therefore, necessary. Many common reducing reagents like ferrous ion [79] and hydrogen gas [7] have been tested without observable perchlorate reduction. The key is to find a way that can overcome the energy barrier.
Although a plethora of new technologies have been identified and shown to remediate perchlorate in water, none of these technologies are currently at a stage were they can be implemented in the field. A lot of research still requires to be done to make them cost effective and practical for field application purposes. It is highly unlikely that a single technology would be feasible to completely remove perchlorate and an integrated technology that combines two or more different technologies might be necessary. Also, the recently passed decision by US EPA not to regulate perchlorate is a complete about turn by the agency after spending more than a
75 decade studying the related technological, health and regulatory aspects. This would also have
serious implications on the future of perchlorate treatment technologies and it would be interesting to see how the future pans out for this widely researched contaminant.
3.10 References
1 Urbansky, E. T. and Schock, M. R. (1999) Issues in managing the risks associated with
perchlorate in drinking water. Journal of Environmental Management 56, 79-95
2 Perciasepe, R. (1998) Announcement of the drinking water contaminant candidate list;
notice. Federal Register 63, 10273
3 USEPA (2005) EPA Sets Reference Dose for Perchlorate.
4 Espenson, J. H. (2000) The Problem and Perversity of Perchlorate. In Perchlorate in the
Environment (Urbansky, E. T., ed.), pp. 1-7, Kluwer Academic/Plenum, New York
5 Gu, B., Dong, W., Brown, G. M. and Cole, D. R. (2003) Complete Degradation of
Perchlorate in Ferric Chloride and Hydrochloric Acid under Controlled Temperature and
Pressure. Environmental Science & Technology 37, 2291-2295
6 Moore, A. M., DeLeon, C. H. and Young, T. M. (2003) Rate and Extent of Aqueous
Perchlorate Removal by Iron Surfaces. Environ. Sci. Technol. 37, 3189-3198
7 Lang, G. G. and Horanyi, G. (2003) Some interesting aspects of the catalytic and
electrocatalytic reduction of perchlorate ions. Journal of Electroanalytical Chemistry 552,
197-211
8 Urbansky, E. T. (2000) Perchlorate in the Environment. Kluwer Academic/ Plenum
Publishers, New York, NY
76 9 Perchlorate-Team (2005) Perchlorate: Overview of Issues, Status, and Remedial Options.
pp. 1-90, The Interstate Technology and Regulatory Council Washington, DC
10 Gu, B. and Coates, J. D. (2006) Perchlorate: Environmental Occurence, Interactions and
Treatment. Springer Science, New York, NY
11 Charnley, G. (2008) Perchlorate: Overview of risks and regulation. Food and Chemical
Toxicology 46, 2307-2315
12 Urbansky, E. T., Brown, S. K., Magnuson, M. L. and Kelty, C. A. (2001) Perchlorate
levels in samples of sodium nitrate fertilizer derived from Chilean caliche. Environmental
Pollution 112, 299-302
13 Jackson, A. W., Anderson, T. A., Harvey, G., Orris, G. J., Rajagopalan, S. and Kang, N.
(2006) Occurence and Formation of Non-Anthropogenic Perchlorate. In Perchlorate:
Environmental Occurence, Interactions and Treatment (Gu, B. and Coates, J. D., eds.),
pp. 49-66, Springer Science, New York, NY
14 Jackson, A. W., Anandam, S. K., Anderson, T., Lehman, T., Rainwater, K., Rajagopalan,
S., Ridley, M. and Tock, R. (2005) Perchlorate occurrence in the Texas Southern High
Plains Aquifer System. Ground Water Monitoring & Remediation 25, 137-149
15 Kang, N., Jackson, W. A., Dasgupta, P. K. and Anderson, T. A. (2008) Perchlorate
production by ozone oxidation of chloride in aqueous and dry systems. Science of the
Total Environment 405, 301-309
16 USEPA (2007) Perchlorate. (Reuse, F. F. R. a., ed.), Federal Facilities Restoration and
Reuse
17 USEPA (2005) Known Perchlorate Releases in the US.
77 18 Motzer, W. E. (2001) Perchlorate: Problems, Detection, and Solutions. Environmental
Forensics 2, 301-311
19 Pontius, F. W., Damian, P. and Eaton, A. D. (2000) Regulating Perchlorate in Drinking
Water. In Perchlorate in the Environment (Urbansky, E. T., ed.), pp. 31-36, Kluwer
Academic/ Plenum Publishers, New York, NY
20 Seyfferth, A. L. and Parker, D. R. (2006) Determination of Low Levels of Perchlorate in
Lettuce and Spinach Using Ion Chromatography−Electrospray Ionization Mass
Spectrometry (IC-ESI-MS). Journal of Agricultural and Food Chemistry 54, 2012-2017
21 Sanchez, C. A., Krieger, R. I., Khandaker, N., Moore, R. C., Holts, K. C. and Neidel, L.
L. (2005) Accumulation and Perchlorate Exposure Potential of Lettuce Produced in the
Lower Colorado River Region. Journal of Agricultural and Food Chemistry 53 5479-
5486
22 Sanchez, C. A., Blount, B. C., Valentin-Blasini, L., Lesch, S. M. and Krieger, R. I. (2008)
Perchlorate in the Feed - Dairy Continuum of the Southwestern United States. Journal of
Agricultural and Food Chemistry 56, 5443-5450
23 Kirk, A. B., Smith, E. E., Tian, K., Anderson, T. A. and Dasgupta, P. K. (2003)
Perchlorate in Milk. Environmental Science & Technology 37, 4979-4981
24 Kirk, A. B., Martinelango, P. K., Tian, K., Dutta, A., Smith, E. E. and Dasgupta, P. K.
(2005) Perchlorate and Iodide in Dairy and Breast Milk. Environmental Science &
Technology 39, 2011-2017
25 Shi, Y., Zhang, P., Wang, Y., Shi, J., Cai, Y., Mou, S. and Jiang, G. (2007) Perchlorate in
sewage sludge, rice, bottled water and milk collected from different areas in China.
Environment International 33, 955-962
78 26 Dionex (1998) Analysis of low concentrations of perchlorate in drinking water and
ground water by ion chromatography. Note 121. Dionex Corp, Sunnyvale
27 Dionex (2000) Determination of low concentrations of perchlorate in drinking and
ground waters using ion chromatography. . In Application Note 134, pp. Dionex Corp.,
Sunnyvale.
28 Hautman, D. P., Munch, D., Eaton, A. D. and Haghani, A. W. (1999) U.S. EPA Method
314.0. USEPA
29 Wagner, H. P., Pepich, B. V., Pohl, C., Later, D., Joyce, R., Srinivasan, K., Thomas, D.,
Woodruff, A., DeBorba, B. and Munch, D. J. (2006) US Environmental Protection
Agency Method 314.1, an automated sample preconcentration/matrix elimination
suppressed conductivity method for the analysis of trace levels (0.50μg/L) of
perchlorate in drinking water. Journal of Chromatography A 1118, 85-93
30 Wagner, H. P., Pepich, B. V., Pohl, C., Later, D., Srinivasan, K., Lin, R., DeBorba, B.
and Munch, D. J. (2007) Selective method for the analysis of perchlorate in drinking
waters at nanogram per liter levels, using two-dimensional ion chromatography with
suppressed conductivity detection. Journal of Chromatography A 1155, 15-21
31 Wendelken, S. C., Munch, D. J., Pepich, B. V., Later, D. W. and Pohl, C. A. (2005) U.S.
EPA Method 331.0. USEPA
32 Hedrick, E., Behymer, T., Slingsby, R. and Munch, D. (2005) U.S. EPA Method 332.0.
USEPA
33 Kim, Y. and Amemiya, S. (2008) Stripping Analysis of Nanomolar Perchlorate in
Drinking Water with a Voltammetric Ion-Selective Electrode Based on Thin-Layer
Liquid Membrane. Analytical Chemistry 80, 6056-6065
79 34 Tikkanen, M. W. (2006) Development of a drinking water regulation for perchlorate in
California. Analytica Chimica Acta 567, 20-25
35 Goto, S. (2007) Energy Subcommittee Votes Limits On Perchlorate Levels. In
CongressDaily - 19366132
36 Office of Water, U. (2008) Fact Sheet: Preliminary Regulatory Determination for
Perchlorate. In EPA 815-F-08-009
37 USEPA (2008) Drinking Water: Preliminary Regulatory Determination on Perchlorate.
FEDERAL REGISTER
38 Eilperin, J. (2008) EPA Unlikely to Limit Perchlorate in Tap Water. In Washington Post,
pp. A09
39 Mattie, D. R., Strawson, J. and Zhao, J. (2006) Perchlorate Toxicity and Risk
Assessment. In Perchlorate: Environmental Occurence, Interactions and Treatment (Gu,
B. and Coates, J. D., eds.), pp. 169-196, Springer Science, New York, NY
40 Paulus, B. F., Bazar, M. A. and Salice, C. J. (2007) Perchlorate Inhibition of Iodide
Uptake in Normal and Iodine-deficient Rats. Journal of Toxicology and Environmental
Health, Part A 70, 1142-1149
41 Park, J.-W., Bradford, C. M., Rinchard, J., Liu, F., Wages, M., Waters, A., Kendall, R. J.,
Anderson, T. A. and Theodorakis, C. W. (2007) Uptake, Elimination, and Relative
Distribution of Perchlorate in Various Tissues of Channel Catfish. Environmental
Science & Technology 41, 7581-7586
42 Bradford, C. M., Park, J.-W., Rinchard, J., Anderson, T. A., Liu, F. and Theodorakis, C.
W. (2006) Uptake and elimination of perchlorate in eastern mosquitofish. Chemosphere
63, 1591-1597
80 43 Liu, F., Gentles, A. and Theodorakis, C. W. (2008) Arsenate and perchlorate toxicity,
growth effects, and thyroid histopathology in hypothyroid zebrafish Danio rerio.
Chemosphere 71, 1369-1376
44 Fisher, J. and McLanahan, E. (2008) Evidence for perchlorate altering thyroid function.
Toxicology Letters 180, S182
45 Braverman, L. E. (2007) Clinical Studies of Exposure to Perchlorate in the United States.
Thyroid 17, 819-822
46 Parette, R. and Cannon, F. S. (2005) The removal of perchlorate from groundwater by
activated carbon tailored with cationic surfactants. Water Research 39, 4020-4028
47 Parette, R., Cannon, F. S. and Weeks, K. (2005) Removing low ppb level perchlorate,
RDX, and HMX from groundwater with cetyltrimethylammonium chloride (CTAC) pre-
loaded activated carbon. Water Research 39, 4683-4692
48 Na, C., Cannon, F. S. and Hagerup, B. (2002) Perchlorate removal via iron-preloaded
GAC and borohydride regeneration. J. AWWA 94
49 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. I: Characterization of NH3 thermally tailored GACs.
Carbon 43, 573-580
50 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. II: Perchlorate adsorption. Carbon 43, 581 - 590
51 Chen, W. and Cannon, F. S. (2005) Thermal reactivation of ammonia-tailored granular
activated carbon exhausted with perchlorate. Carbon 43, 2742-2749
81 52 Gu, B., Ku, Y.-K. and Brown, G. M. (2002) Treatment of Perchlorate-Contaminated
Groundwater Using Highly Selective, Regenerable Ion-Exchange Technology: A Pilot-
Scale Demonstration. Remediation Journal 12, 51-68
53 Gu, B., Ku, Y. K. and Brown, G. M. (2005) Sorption and Desorption of Perchlorate and
U(VI) by Strong-Base Anion-Exchange Resins. Environmental Science & Technology
39, 901-907
54 Gu, B., Brown, G. M. and Chiang, C. C. (2007) Treatment of Perchlorate-Contaminated
Groundwater Using Highly Selective, Regenerable Ion-Exchange Technologies.
Environmental Science & Technology 41, 6277-6282
55 Gu, B. and Brown, G. M. (2006) Field Demonstration using Highly Selective,
Regenerable Ion Exchange and Perchlorate Destruction Technologies for Watre
Treatment. In Perchlorate: Environmental Occurence, Interactions and Treatment (Gu, B.
and Coates, J. D., eds.), pp. 253-277, Springer Science, New York, NY
56 Gu, B. and Brown, G. M. (2006) Recent Advances in Ion Exchange for Perchlorate
Treatment, Recovery and Destruction In Perchlorate: Environmental Occurence,
Interactions and Treatment (Gu, B. and Coates, J. D., eds.), pp. 209-249, Springer
Science, New York, NY
57 Gu, B., Brown, G. M., Maya, L., Lance, M. J. and Moyer, B. A. (2001) Regeneration of
Perchlorate (ClO4-)-Loaded Anion Exchange Resins by a Novel Tetrachloroferrate
(FeCl4-) Displacement Technique. Environmental Science & Technology 35, 3363-3368
58 Xiong, Z., Zhao, D. and Harper, W. F. (2007) Sorption and Desorption of Perchlorate
with Various Classes of Ion Exchangers: A Comparative Study. Industrial &
Engineering Chemistry Research 46, 9213-9222
82 59 Kim, T.-H., Jang, M. and Park, J. K. (2008) Bifunctionalized mesoporous molecular
sieve for perchlorate removal. Microporous and Mesoporous Materials 108, 22-28
60 Dilip, M., Griffin, S. T., Spear, S. K., Rijksen, C., Rodri, guez, H., ctor and Rogers, R. D.
(2008) Dual Nature of Polyethylene Glycol-Based Aqueous Biphasic Extraction
Chromatographic (ABEC) Resins: Uptakes of Perchlorate versus Mercury(II). Industrial
& Engineering Chemistry Research 47, 7390-7396
61 Siemens, T. (2006) Ion-exchange system removes perchlorate. Membrane Technology
2006, 3-4
62 Huq, H. P., Yang, J.-S. and Yang, J.-W. (2007) Removal of perchlorate from
groundwater by the polyelectolyte-enhanced ultrafiltration process. Desalination 204,
335-343
63 Roach, J. D. and Tush, D. (2008) Equilibrium dialysis and ultrafiltration investigations of
perchlorate removal from aqueous solution using poly(diallyldimethylammonium)
chloride. Water Research 42, 1204-1210
64 Yoon, Y., Amy, G. and Yoon, J. (2005) Effect of pH and conductivity on hindered
diffusion of perchlorate ions during transport through negatively charged nanofiltration
and ultrafiltration membranes. Desalination 177, 217-227
65 Yoon, Y., Amy, G., Cho, J., Her, N. and Pellegrino, J. (2002) Transport of perchlorate
- (ClO4 ) through NF and UF membranes. Desalination 147, 11-17
66 Lee, S., Quyet, N., Lee, E., Kim, S., Lee, S., Jung, Y. D., Choi, S. H. and Cho, J. (2008)
Efficient removals of tris(2-chloroethyl) phosphate (TCEP) and perchlorate using NF
membrane filtrations. Desalination 221, 234-237
83 67 Yoon, J., Yoon, Y., Amy, G., Cho, J., Foss, D. and Kim, T.-H. (2003) Use of surfactant
modified ultrafiltration for perchlorate (ClO4-) removal. Water Research 37, 2001-2012
68 Urbansky, E. T. (1998) Perchlorate Chemistry: Implications for Analysis and
Remediation. Bioremediation Journal 2, 81-95
69 Hurley, K. D. and Shapley, J. R. (2007) Efficient heterogeneous catalytic reduction of
perchlorate in water. Environmental Science & Technology 41, 2044
70 Abu-Omar, M. M. and Espenson, J. H. (1995) Facile Abstraction of Successive Oxygen
Atoms from Perchlorate Ions by Methylrhenium Dioxide. Inorganic Chemistry 34, 6239-
6240
71 Abu-Omar, M. M., Appelman, E. H. and Espenson, J. H. (1996) Oxygen-Transfer
Reactions of Methylrhenium Oxides. Inorganic Chemistry 35, 7751-7757
72 Abu-Oma, M. M. (2003) Swift oxo transfer reactions of perchlorate and other substrates
catalyzed by rhenium oxazoline and thiazoline complexes. Chemical Communications
(Cambridge) 2003, 2102-2111
73 Abu-Omar, M. M. (2003) Effective and Catalytic Reduction of Perchlorate by Atom
Transfer-Reaction Kinetics and Mechanisms. Comments on Inorganic Chemistry 24, 15-
37
74 Abu-Omar, M. M. (2000) Clean and Efficient Catalytic Reduction of Perchlorate.
Angewandte Chemie 39, 4310
75 Arias, J., Newls, C. R. and Abu-Omar, M. M. (2001) Kinetics and Mechanisms of
Catalyti c Oxygen Atom Transfer with Oxorhenium(V) Oxazoline Complexes. Inorganic
Chemistry 40, 2185-2192
84 76 Wang, D. M., Shah, S. I., Chen, J. G. and Huang, C. P. (2008) Catalytic reduction of
perchlorate by H2 gas in dilute aqueous solutions. Separation and Purification
Technology 60, 14-21
77 Wang, D. M., Huang, C. P., Chen, J. G., Lin, H. Y. and Shah, S. I. (2007) Reduction of
perchlorate in dilute aqueous solutions over monometallic nano-catalysts: Exemplified by
tin. Separation and Purification Technology 58, 129-137
78 Wang, D. M. and Huang, C. P. (2008) Electrodialytically assisted catalytic reduction
(EDACR) of perchlorate in dilute aqueous solutions. Separation and Purification
Technology 59, 333-341
79 Moore, A. M. and Young, T. M. (2005) Chloride Interactions with Iron Surfaces:
Implications for Perchlorate and Nitrate Remediation Using Permeable Reactive Barriers.
Journal of Environmental Engineering 131, 924-933
80 Moore, A. M. (2003) Anion Reactions at Iron Surfaces: Implications for Perchlorate
Remediation Using Permeable Reactive Barriers. In Civil and Environmental
Engineering, pp. 142, University of California, Davis
81 Makris, K. C., Sarkar, D. and Datta, R. (2006) Aluminum-based drinking-water treatment
residuals: A novel sorbent for perchlorate removal. Environmental Pollution 140, 9-12
82 Mahmudov, R., Shu, Y., Rykov, S., Chen, J. and Huang, C. P. (2008) The reduction of
perchlorate by hydrogenation catalysts. Applied Catalysis B, Environmental 81, 78-87
83 Davies, S. H. R. and Morgan, J. J. (1989) Manganese(II) oxidation kinetics on metal
oxide surfaces. Journal of Colloid and Interface Science 129, 63-77
84 Wehrli, B. (1990) Redox Reactions of Metal Ions at Mineral Surfaces. In Aquatic
Chemical Kinetics (Stumm, W., ed.), pp. 311-337, Wiley Interscience, New York
85 85 Gurol, M. G. and Kim, K. (2000) Investigation of Perchlorate Removal in Drinking
Water Sources by Chemical methods. In Perchlorate in the Environment (Urbansky, E.
T., ed.), pp. 99-107, Kluwer Academic/Plenum, New York
86 Huang, H. (2005) Reduction of Perchlorate by Zero Valent Iron. In Environmental
Engineering, pp. 87, University of Cincinnati, Cincinnati, OH
87 Cao, J., Elliott, D. and Zhang, W.-x. (2005) Perchlorate Reduction by Nanoscale Iron
Particles. Journal of Nanoparticle Research 7, 499-506
88 Xiong, Z., Zhao, D. and Pan, G. (2007) Rapid and complete destruction of perchlorate in
water and ion-exchange brine using stabilized zero-valent iron nanoparticles. Water
Research 41, 3497-3505
89 Oh, S.-Y., Chiu, P. C., Kim, B. J. and Cha, D. K. (2006) Enhanced reduction of
perchlorate by elemental iron at elevated temperatures. Journal of Hazardous Materials
129, 304-307
90 Horanyi, G. and Bakos, I. (1993) Study of the relationship between voltammetric
behavior and electroanalytic activity in the reduction of ClO4- ions at platinized platinum
electrodes. Journal of Electroanalytical Chemistry 347, 383 - 391
91 Horanyi, G., Bakos, I., Szabo, S. and Rizmayer, E., M., (1992) New observations in the
field of electrochemistry of technetium and rhenium: electrocatalytic reduction of
perchlorate ions at electrosorbed and electrodepositied Tc and Re layers in acid medium.
Journal of Electroanalytical Chemistry 337, 365-369
92 Colom, F. and Gonzalez - Tejera, M. J. (1985) Reduction of Perchlorate Ion on
Ruthenium Electrodes in Aqueous Solutions. Journal of Electroanalytical Chemistry 190,
243 - 255
86 93 Almeida, C. M. V. B., Gianetti, B. F. and Rabockai, T. (1997) Electrochemical study of
perchlorate reduction at tin electrodes. Journal of Electroanalytical Chemistry 422, 185-
189
94 Wasberg, M. and Horanyi, G. (1995) The reduction of ClO4- ions on Rh electrodes.
Journal of Electroanalytical Chemistry 385, 63-70
95 Theis, T. L., Zander, A. K., Xiang, L., Jeosadaque, S. and Anderson, M. A. (2002)
Electrochemical and Photocatalytic Reduction of Perchlorate Ion. Journal of Water
Supply: Research and Technology - AQUA 51, 367 - 374
96 Chaudhuri, S. K., O’Connor, S. M., Gustavson, R. L., Achenbach, L. A. and Coates, J. D.
(2002) Environmental Factors That Control Microbial Perchlorate Reduction. Applied &
Environmental Microbiology 68, 4425-4430
97 Wu, D., He, P., Xu, X., Zhou, M., Zhang, Z. and Houda, Z. (2008) The effect of various
reaction parameters on bioremediation of perchlorate-contaminated water. Journal of
Hazardous Materials 150, 419-423
98 Xu, J., Trimble, J. J., Steinberg, L. and Logan, B. E. (2004) Chlorate and nitrate reduction
pathways are separately induced in the perchlorate-respiring bacterium Dechlorosoma sp.
KJ and the chlorate-respiring bacterium Pseudomonas sp. PDA. Water Research 38, 673-
680
99 Coates, J. D. and Achenbach, L. A. (2004) Microbial perchlorate reduction: rocket-fueled
metabolism. Nature Reviews. Microbiology 2, 569-580
100 Coates, J. D., Michaelidou, U., Bruce, R. A., O'Connor, S. M., Crespi, J. N. and
Achenbach, L. A. (1999) Ubiquity and Diversity of Dissimilatory (Per)chlorate-Reducing
Bacteria. Applied & Environmental Microbiology 65, 5234-5241
87 101 Herman, D. C. and Frankenberger, W. T., Jr. (1999) Bacterial reduction of perchlorate
and nitrate in water. Journal of Environmental Quality 28, 1018-1024
102 Giblin, T., Herman, D., Deshusses, M. A. and Frankenberger, W. T., Jr. (2000) Removal
of perchlorate in ground water with a flow-through bioreactor. Journal of Environmental
Quality 29, 578-583
103 Giblin, T. and Frankenberger, W. T. (2001) Perchlorate and nitrate reductase activity in
the perchlorate-respiring bacterium perclace. Microbiological Research 156, 311-315
104 Nerenberg, R., Rittmann, B. E. and Najm, I. (2002) Perchlorate reduction in a hydrogen-
based membrane-biofilm reactor. Journal AWWA 94, 103-114
105 Nerenberg, R., Kawagoshi, Y. and Rittmann, B. E. (2008) Microbial ecology of a
perchlorate-reducing, hydrogen-based membrane biofilm reactor. Water Research 42,
1151-1159
106 Nerenberg, R., Kawagoshi, Y. and Rittmann, B. E. (2006) Kinetics of a hydrogen-
oxidizing, perchlorate-reducing bacterium. Water Research 40, 3290-3296
107 Dudley, M., Salamone, A. and Nerenberg, R. (2008) Kinetics of a chlorate-accumulating,
perchlorate-reducing bacterium. Water Research 42, 2403-2410
108 Okeke, B. C. and Frankenberger, W. T. (2003) Molecular analysis of a perchlorate
reductase from a perchlorate-respiring bacterium Perc1ace. Microbiological Research
158, 337-344
109 Frankenberger, W. T. (2003) Perchlorate Removal in groundwater by perchlorate
reductases from the perchlorae respiring bacterium, perclace. In Technical Completion
Reports, pp. 1-18, University of California, Riverside, CA
88 110 Hatzinger, P. B., Whittier, M. C., Arkins, M. D., Bryan, C. W. and Guarini, W. J. (2002)
In-Situ and Ex-Situ Bioremediation Options for Treating Perchlorate in Groundwater.
Remediation Journal 12, 69 - 86
111 Hatzinger, P. B., Diebold, J., Yates, C. A. and Cramer, R. J. (2006) Field Demostration of
In Situ Perchlorate Bioremediation in Groundwater. In Perchlorate: Environmental
Occurence, Interactions and Treatment (Gu, B. and Coates, J. D., eds.), pp. 311-340,
Springer Science, New York, NY
112 Hatzinger, P. B. (2005) Perchlorate Biodegradation for Water Treatment. Environmental
Science & Technology 39, 239A-247A
113 Padhye, L., Rainwater , K., Jackson, W. A. and Morse, A. (2007) Kinetics for a
Membrane Reactor Reducing Perchlorate. Water Environment Research 79, 140-147
114 Song, Y. and Logan, B. E. (2004) Inhibition of aerobic respiration and dissimilatory
perchlorate reduction using cyanide. FEMS Microbiology Letters 239, 229-234
115 Bender, K. S., Shang, C., Chakraborty, R., Belchik, S. M., Coates, J. D. and Achenbach,
L. A. (2005) Identification, Characterization, and Classification of Genes Encoding
Perchlorate Reductase. Journal of Bacteriology 187, 5090-5096
116 Kengen, S. W. M., Rikken, G. B., Hagen, W. R., Van Ginkel, C. G. and Stams, A. J. M.
(1999) Purification and Characterization of (Per)Chlorate Reductase from the Chlorate-
Respiring Strain GR-1. Journal of Bacteriology 181, 6706-6711
117 Atikovic, E., Suidan, M. T. and Maloney, S. W. (2008) Anaerobic treatment of army
ammunition production wastewater containing perchlorate and RDX. Chemosphere 72,
1643-1648
89 118 Wang, C., Lippincott, L. and Meng, X. (2008) Kinetics of biological perchlorate
reduction and pH effect. Journal of Hazardous Materials 153, 663-669
119 Gal, H., Ronen, Z., Weisbrod, N., Dahan, O. and Nativ, R. (2008) Perchlorate
biodegradation in contaminated soils and the deep unsaturated zone. Soil Biology and
Biochemistry 40, 1751-1757
120 Ju, X., Sierra-Alvarez, R., Field, J. A., Byrnes, D. J., Bentley, H. and Bentley, R. (2008)
Microbial perchlorate reduction with elemental sulfur and other inorganic electron
donors. Chemosphere 71, 114-122
121 Holdren, G. C., Kelly, K. and Weghorst, P. (2008) Evaluation of potential impacts of
perchlorate in the Colorado River on the Salton Sea, California. Hydrobiologia 604, 173 -
179
122 Choi, Y. C., Li, X., Raskin, L. and Morgenroth, E. (2008) Chemisorption of oxygen onto
activated carbon can enhance the stability of biological perchlorate reduction in fixed bed
biofilm reactors. Water Research 42, 3425-3434
123 Brown, J. C., Snoeyink, V. L. and Kirisits, M. J. (2002) Abiotic and Biotic Perchlorate
Removal in an Activated Carbon Filter. Journal AWWA 94, 70-79
124 Brown, J. C., Anderson, R. D., Min, J. H., Boulos, L., Prasifka, D. and Juby, G. J. G.
(2005) Fixed-bed Biological Treatment of Perchlorate-Contaminated Drinking Water.
Journal AWWA 97, 70-81
125 Hristovski, K., Westerhoff, P., ouml, ller, T., Sylvester, P., Condit, W. and Mash, H.
(2008) Simultaneous removal of perchlorate and arsenate by ion-exchange media
modified with nanostructured iron (hydr)oxide. Journal of Hazardous Materials 152, 397-
406
90 126 Lehman, S. G., Badruzzaman, M., Adham, S., Roberts, D. J. and Clifford, D. A. (2008)
Perchlorate and nitrate treatment by ion exchange integrated with biological brine
treatment. Water Research 42, 969-976
127 Lin, X., Roberts, D. J., Hiremath, T., Clifford, D. A., Gillogly, T. and Lehman, S. G.
(2007) Divalent Cation Addition (Ca2+ or Mg2+) Stabilizes Biological Treatment of
Perchlorate and Nitrate In Ion-Exchange Spent Brine. Environmental Engineering
Science 24, 725-735
128 Chung, J., Nerenberg, R. and Rittmann, B. E. (2007) Evaluation for Biological Reduction
of Nitrate and Perchlorate in Brine Water Using the Hydrogen-Based Membrane Biofilm
Reactor. Journal of Environmental Engineering 157, 157-165
129 Van Ginkel, S. W., Ahn, C. H., Badruzzaman, M., Roberts, D. J., Lehman, S. G., Adham,
S. S. and Rittmann, B. E. (2008) Kinetics of nitrate and perchlorate reduction in ion-
exchange brine using the membrane biofilm reactor (MBfR). Water Research 42, 4197-
4205
130 Matos, C. T., Velizarov, S., Crespo, J., atilde, o, G. and Reis, M. A. M. (2006)
Simultaneous removal of perchlorate and nitrate from drinking water using the ion
exchange membrane bioreactor concept. Water Research 40, 231-240
131 Yu, X., Amrhein, C., Deshusses, M. A. and Matsumoto, M. R. (2006) Perchlorate
Reduction by Autotrophic Bacteria in the Presence of Zero-Valent Iron. Environmental
Science & Technology 40, 1328-1334
132 Son, A., Lee, J., Chiu, P. C., Kim, B. J. and Cha, D. K. (2006) Microbial reduction of
perchlorate with zero-valent iron. Water Research 40, 2027-2032
91 133 Shrout, J. D., Williams, A. G. B., Scherer, M. M. and Parkin, G. F. (2005) Inhibition of
bacterial perchlorate reduction by zero-valent iron. Biodegradation 16, 23 - 32
134 Yu, X., Amrhein, C., Deshusses, M. A. and Matsumoto, M. R. (2007) Perchlorate
Reduction by Autotrophic Bacteria Attached to Zerovalent Iron in a Flow-Through
Reactor. Environmental Science & Technology 41, 990-997
135 Wallace, W., Beshear, S., Williams, D., Hospadar, S. and Owens, M. (1998) Perchlorate
reduction by a mixed culture in an up-flow anaerobic fixed bed reactor. Journal of
Industrial Microbiology and Biotechnology 20, 126-131
136 Kim, K. and Logan, B. E. (2001) Microbial Reduction of Perchlorate in Pure and Mixed
Culture Packed-Bed Bioreactors. Water Research 35, 3071-3076
137 Gingras, T. M. and Batista, J. R. (2002) Biological reduction of perchlorate in ion
exchange regenerant solutions containing high salinity and ammonium levels. Journal of
Environmental Monitoring 4, 96-101
138 Okeke, B. C., Giblin, T. and Frankenberger, W. T. (2002) Reduction of perchlorate and
nitrate by salt tolerant bacteria. Environmental Pollution 118, 357-363
92 CHAPTER 4
Removal of Perchlorate and Chlorate in Aquatic Systems Using Integrated
Technologies
4.1 Abstract
Because of its extremely low concentrations and strong resistance to most treatment
technologies, perchlorate has become one of the biggest challenges currently being faced by the
drinking water industry. This study looked at the potential to develop an integrated method for
removal of perchlorate in water by combining zero-valent iron (Fe0) and electrochemical
reduction. Experiments with Fe0 have been conducted in a previous study and so this part of the
study looked at electrochemical reduction of perchlorate in a dual cell reactor. However, the use
of Fe0 in electrochemical process did not produce expected results as the high pH and the presence of hydrogen gas rendered the Fe0 ineffective for any perchlorate reduction. The surface
characteristics of the iron were studied to determine the reasons for this reduced capacity. Batch experiments and electrochemical experiments were conducted to better understand reduction of chlorate and the effect of various process parameters on chlorate reduction kinetics. As expected, the reduction rates for chlorate were significantly higher when compared with perchlorate and the electrolyte ionic strength and pH were found to be the most important factors influencing chlorate reduction.
93 4.2 Introduction
In the last few years, perchlorate has become a major inorganic contaminant in drinking
water and has been detected in a number of public drinking water systems, especially in the
southwestern states. A major anthropogenic source of perchlorate contamination is the
manufacture of ammonium perchlorate which is used as the oxidizer component and primary
ingredient in solid propellant for rockets, missiles, and fireworks. As of April 2003, the EPA
listed over 150 separate locations in 35 states throughout the United States with known
perchlorate soil and groundwater contamination [1]. Perchlorate has been identified as an
endocrine disruptor chemical and is a potential human health hazard in drinking water. Exposure
to perchlorate results in inhibition of normal iodide uptake by thyroid glands to cause mental
retardation and, hearing and speech degradation [2].
Perchlorate treatment technologies can be generally classified into two categories of destruction or removal technologies. Among the physical removal technologies, only ion- exchange has been successfully used to treat perchlorate contamination in water or wastewater
[3-5]. However, the most important concern with the use of ion exchange technology for perchlorate removal is the disposal and treatment of resultant brine wastewater that is high in perchlorate concentration, and has to be somehow disposed or treated. However these resins need to be modified in order to increase their selectivity towards perchlorate. Other removal technologies such as granulated activated carbon (GAC)[6-8] and membrane filtration have limited applications [9, 10]. Because of the concerns associated with the disposal of perchlorate- laden wastes, a process involving destruction of the perchlorate ion is preferred over physical removal. Several species of micro-organisms have been developed to effectively remove
94 perchlorate from water [11-21]. Bioremediation is a cost-effective technology and has been
found to have good selectivity for perchlorate. However, the biggest issue with bioremediation is
public acceptance, for the intentional introduction of micro-organisms into the drinking water
system, considering that some of these could be pathogenic.
An effective chemical reduction process which is favorable for drinking water treatment
and groundwater remediation is, therefore, necessary. Many studies have looked at the pathway
of perchlorate reduction ([22-24] and have concurred that perchlorate ion loses an oxygen atom
- to form chlorate (ClO3 ) at the first step, which is the rate limiting step in the reduction process.
This step onward, the reduction to chloride proceeds at a rapid rate. Although, perchlorate is a
strong oxidizing agent by virtue of its redox potential (+1.38 V) [25], its oxidizing power is
retarded because of its high activation energy (120 KJmol-1) [23]. It is the atomic structure of
perchlorate that causes this high activation energy. The chlorine atom is in the center of a
tetrahedron structure surrounded by four oxygen atoms which essentially block common
reductants from directly attacking the chlorine. As a result, most common reducing agents have
been tested without any significant perchlorate reduction [3].
Numerous studies have shown that perchlorate can be reduced electrochemically to Cl-
[22, 24, 26-31] and electro-reduction does not involve any major waste products. It involves low maintenance and is easy to implement in large systems. However, the reaction rates associated with electrochemical reduction of perchlorate are extremely slow and have to be improved for the process to be practical. Experimental evidence has shown that aqueous mineral surfaces provide added accelerated pathways for certain redox processes [32]. Lang and Horanyi
95 reviewed the conversion of perchlorate at different metal surfaces and found Iron to have the
fastest reaction rate [24]. Studies have shown that zero-valent iron (Fe0) successfully reduces
perchlorate to chloride by providing the necessary catalytic effect [33-35].
Previously we looked at perchlorate reduction with Fe0 filings through batch and column
experiments [35]. The reaction kinetics were used to shed some light on perchlorate reduction
mechanism at iron surface and to demonstrate that iron served as a reducing agent and a catalyst
simultaneously. Although iron filings were effective in reducing perchlorate, the reaction rates
were extremely slow and were significantly affected by the presence of other common anions. It
is well known that once perchlorate reduces to chlorate in the first reduction step, the reaction
proceeds rapidly to chloride. However, there is limited published data on abiotic removal of
chlorate in the literature. Chlorate is found in drinking water as a byproduct of the use of chlorine
dioxide as a disinfectant and during ozonation of water containing hypochrite ion [36]. Chlorate
has been found to have adverse health effects, especially through its reduction to chlorite [36].
Westerhoff [37] demonstrated chlorate removal in water by Fe0 through batch and column experiments.
This study investigates electrochemical reduction of perchlorate and chlorate under
different process conditions and at the same time evaluating the potential of achieving enhanced
perchlorate/chlorate reduction by combining electrochemical and Fe0 reduction in an integrated
process. For the electrochemical experiments, effects of various process parameters such as
electrode material, current density, ionic strength of the electrolyte, and pH on the
electrochemical reduction of perchlorate and chlorate were studied. The effectiveness of this
96 technology will serve as a basis to further develop the proposed methodology to purify surface or
ground water in target zones at a larger scale.
4.3 Materials and methods
4.3.1 Materials
Hydrofluoric (HF) acid (Fisher Scientific, PA) was used for acid washing of the iron
filings (40 mesh, Fisher Scientific, PA). HPLC grade sodium perchlorate monohydrate and sodium chlorate (Fisher Scientific, PA) were used for the preparation of stock solutions. To remove any dissolved iron species from the sample prior to analysis by Ion Chromatography
(Dionex, Sunnyvale, CA), 1.0 cc OnGuard II H cartridges (Dionex, Sunnyvale, CA) were used.
Milli-Q water produced by a Millipore system (Billerica, MA) with 18.2 MΩ-cm resistance was used for preparation of solutions. For electrodes, metallic nickel (Alfa Aesar, Ward Hill, MA) and titanium (Alfa Aesar, Ward Hill, MA) rods, glossy carbon rods (zero porosity) (HTW,
Germany) and platinum wire (Alfa Aesar, Ward Hill, MA) were used.
4.3.2 Analytical Method
4.3.2.1 Perchlorate/Chlorate concentration
Perchlorate concentration was measured with a manual injection of samples to Dionex
DX-600 (Sunnyvale, CA, USA) ion chromatography system equipped with a GP-50 gradient
pump, an anion self-regenerate suppressor ultra (ASRS-ultra), a CD-25 conductivity detector and a LC-20 chromatography enclosure. Dionex IonPac anion exchange column, AS-16 and guard
97 column AG-16 were used with 50 mM NaOH as eluent. The eluent flow rate was set at 0.9
mL/min. For analysis of chlorate, AS-14 and AG-14 columns were used with 3.5 mM Na2CO3/ 1
mM NaHCO3 as eluent, with an eluent flow rate of 1.2 mL/min. An Accumet AR50 dual channel
meter obtained from Fisher Scientific was used for pH and conductivity measurements. A
Corning Checkmate II dissolved oxygen sensor module (Cole Parmer, Vernon Hills, IL) was
used for measuring dissolved oxygen (DO) concentrations.
4.3.2.2 Iron characterization
X-ray diffraction was used to study the effect of the various processes on the iron
particles. XRD analysis was performed with a GBC Minimaterial analyzer (GBC scientific
equipment, Dandenong, Australia) operating at a scan speed of 1° 2θ/min from 25° to 90° and a
0.02 step size, with Cu K-α radiation (λ=1.54056 Å). All the iron samples were ground for 5 minutes with an agate mortar and pestle before analysis.
4.3.3 Acid Washing of Iron Filings
The iron filings were washed using 1N hydrofluoric (HF) acid prior to use to remove iron
oxide and other oily contaminants on the surface. The acid washed iron filings were rinsed
multiple times by de-aerated Milli-Q water until the rinse water was free of brown color and its
pH was neutral. Finally, the washed iron filings were dried under nitrogen gas. The color of the
treated iron filings was black.
98 4.3.4 TiO2 coating of electrodes
For the purpose of coating electrodes with thin titanium dioxide films, TiO2 sol-gel was prepared by following the procedure used by Chen [38]. A dip coating apparatus was used to dip the electrode in the sol-gel and pull it at a preset withdrawal velocity. The coated electrode was then allowed to dry at room temperature followed by calcination at 500 °C to achieve anatase phase of the TiO2 coating.
4.3.5 Coating of electrodes with iron oxide sol-gel
Electrodes were coated with iron oxide by using iron oxide sol-gel that was prepared
using the method developed by Gash et al. [39]. A coating method similar to the one used with
TiO2 films was followed.
4.3.6 Preparation of heterogeneous Rhenium-Palladium-Carbon catalyst
The Re-Pd-C catalyst was prepared using materials and methods described elsewhere
[40].
4.3.7 Experimental Setup
A schematic diagram of the column experimental setup that was used for electrochemical
reduction of perchlorate ions is shown in Figure 1.1. The divided cell setup consisted of two
chambers made of Plexiglas that were separated by means of a nafion membrane (Dupont,
Fayetville, NC) which allowed only the passage of positive ions. The left portion of the chamber
acted as the cathode and held a metallic rod (Nickel, Platinum or Glossy Carbon) which was
99 connected to a current source. The section on the right was primarily made up of a platinum wire
connected to a current source and had the same length as that of the cathode. A Honeywell UDC
3000 Universal Digital Controller (Honeywell, Philadelphia, PA) was used as a constant source
of current supply.
In order to simulate batch conditions, the perchlorate/chlorate feed solution was recycled
through the cathodic section with a positive displacement pump (Barnant Company, Barrington,
IL). The feed container for perchlorate/chlorate was a sealed 500 mL glass container with four
ports. The first two ports were used for inflow and outflow of the feed solution. The third port
held a pH probe (AP50 pH/ATC combination electrodes (Fisher Scientific, Pittsburgh, PA)) connected to a pH meter. The last port was fitted with a rubber septum for manual injection of acid for pH adjustment and was used as a sampling port. A multichannel variable speed digital pump was used to pump the anodic electrolyte (0.025 M Na2SO4) solution through the anode.
4.3.8 Experimental Method
The respective electrode was placed in the cathode section and the reactor was sealed
after the Nafion membrane was placed between the cathode and the anode. A stock solution with
- a concentration of 1g/L ClO4 was prepared by adding 1.412g of sodium perchlorate
monohydrate in Milli-Q water. The perchlorate feed solution was prepared by diluting the stock
- solution to get an inlet concentration 1000 ppb ClO4 . For chlorate experiments, a feed solution
was prepared with a chlorate concentration of 10 mg/L. Sodium sulfate (Na2SO4) was added to
the feed solution as the electrolyte to a molar concentration of 9 mM corresponding to an ionic
strength of 0.027 M or 18 mM for an ionic strength of 0.054 M. The feed solution was
100 thoroughly purged with nitrogen gas to achieve dissolved oxygen (DO) concentration ≤ 0.8 mg/L
in the solution before transferring it to the feed container. The feed container for
perchlorate/chlorate was immediately sealed to keep the system under anoxic condition. A 0.25
M sodium sulfate solution (anodic solution) was used as the electrolyte solution in the anode
compartment of the reactor and was prepared by adding 35.525g of Na2SO4 in a liter of DI water.
The two pumps for recirculating the feed solution through the cathode and for recirculating the
anode electrolyte solution through the anode system were turned on at this stage. Once there was
flow in both the anodic and cathodic sections of the reactor, the current supply (200 mA unless
specified otherwise) to the system was started by turning on the constant current supplier unit.
4.4 Results and discussions
4.4.1 Perchlorate Removal
One of the objectives of this study was to investigate the potential removal of perchlorate
ion in drinking water by combining electrochemical reduction with Fe0 reduction. Acid washed
iron filings weighing 450 g were packed into the cathode part of the reactor. Initially the reactor
was run without any current to determine perchlorate removal in the cathode section due to Fe0 reduction. Milli-Q water containing 1000 µg/L perchlorate was recirculated through the column simulating a batch reactor setup. The total volume of perchlorate laden solution recirculating through the column was around 600 mL thereby resulting in an iron dosage of 0.75 g/mL. This test was run for two weeks with samples collected at regular intervals. The performance of the column filled with Fe0 filings was close to the values predicted by the previous batch experiment
results [35]. The electrochemical reaction was started where the first experiment was run at 30
mA current with an initial perchlorate concentration of 1000 µg/L. Only 5% reduction in
101 perchlorate concentration was observed even after a 100 hour run. At this time, there was no
effort made to measure or control the pH of the recirculating solution. The experiment was repeated with the perchlorate solution buffered with acetate buffer to get a pH of 5. The perchlorate removal obtained at a pH of 5 was a little over 5%. Because the reactor was running continuously, there was no pH adjustment during the overnight hours and the pH increased to around 11. It was suspected that recirculating a high pH solution through the iron filings packed in the reactor over a sustained period of time significantly altered the surface properties of the iron filings thereby reducing its capacity for perchlorate reduction. Continuous interaction with hydrogen gas being generated at the cathode was also responsible for altering the properties of the Fe0 filings. This was confirmed when perchlorate solution was recirculated through same iron
filings without any current. The lack of pH control resulted in little or no removal of perchlorate,
as shown in Figure 4.1. and the performance was below that of a corresponding batch study [35].
XRD analysis of the iron filings before and after the electrochemical chemical experiment was
performed to address these issues and the results are discussed in the section 3.3.
Another factor potentially influencing electrochemical reduction of perchlorate would be
the presence of hydrogen gas which is generated at the cathode during the electrochemical
process. In order to better understand this phenomenon, a column experiment was conducted to
study the effect of perchlorate reduction with Fe0 in the presence of hydrogen gas. For the first
experiment, perchlorate solution was recirculated though a column packed with Fe0 over a two
week period.
102 1.00
0.95
0 0.90 C/C
0.85
No pH control pH 5 0.80 No current (Post electro-reduction)
0.00 0210 0430 050 100 Time (hr)
Figure 4.1 Electrochemical reduction of perchlorate with iron filings as cathode
(C0 : 1000 µg/L, 30 mA current)
1.00
0.95
0.90
0.85 0 C/C 0.80
0.75 Fe0 column 0.70 0 Fe column + H2 (5 mL/min) 0.00 0 100 200 300 400 Time (Hr)
Figure 4.2. Perchlorate reduction in a zero-valent iron column (C0 : 10 mg/L)
103 1.00
0.95
0 0.90 C/C
Pt cathode + CoCl2 catalyst 0.85 Ni cathode Ni cathode + CoCl2 catalyst C cathode
0.80 TiO2 coated Ni cathode + CoCl2 catalyst
Fe2O3 coated Ni cathode + CoCl2 catalyst 0.00 0246810121416
Time (Hr)
Figure 4.3 Electrochemical reduction of perchlorate with various electrodes
(C0 : 1000 µg/L, 75 mA current)
For the second experiment, the same setup was used with fresh Fe0 and hydrogen gas was
continuously bubbled into the feed solution. For both experiments, the glass column (155 mm id,
450 mm length) was packed with 250 g of acid washed iron and a perchlorate laden solution with
an initial concentration of 10 mg/L was recirculated through the column at a flow rate of 10
mL/min. For the second experiment hydrogen gas was continuously fed into the feed container at
a flow rate of 5 mL/min controlled with a mass flow controller. From Figure 4.2, it can be seen
that for the column with only Fe0 the reduction of perchlorate was about 21% after 300 hours.
This reaction rate is similar to our previous batch experiments [35]. However in the presence of
hydrogen gas, the perchlorate removal rate dropped to about 11% after a 300 hour run. Initial adsorption of perchlorate to iron was significantly higher when hydrogen was not used. Presence of hydrogen gas was adversely affecting perchlorate reduction with Fe0. Chen et al studied
104 trichloroethylene (TCE) reduction with Fe0 and found that TCE removal capacity decreased in
the presence of hydrogen gas [41].
In the case of electrochemical experiments with Fe0, cementation of the iron filings was
observed as there was an increased resistance being offered to the flow of perchlorate solution
resulting in a pressurized zone and ultimately causing the reactor to leak. Electrochemical reduction of perchlorate on other electrodes such as Platinum, Nickel and Glassy Carbon are shown in Figure 4.3. Cobalt chloride was used in some experiments to test if it promoted any catalytic effect to perchlorate reduction. With a platinum cathode, there was no measurable removal of perchlorate observed after an eight hour run (Figure 4.3). When nickel and carbon
electrodes were used, there was a marginal increase in perchlorate reduction to about 4% after an
eight hour run. Use of CoCl2 with nickel electrode did not result in any significant increase in
perchlorate reduction. Perchlorate reduction with the nickel electrode followed first order
kinetics with a rate constant reaction of 0.0047 hr-1 which is comparable to the rate obtained for
reduction of perchlorate with Fe0 at the higher iron dosages (0.0035-0.0037 hr-1) [35]. Iron oxide
film on Fe0 has been found to be extremely critical as it catalyzes the reduction of perchlorate
[34, 35, 42]. Electrochemical experiments conducted with nickel electrodes coated with thin
films of titanium dioxide and ferric oxide resulted in upto 10% removal of perchlorate after a 5
hour run (Figure 4.3). The oxide film resulted in an increased reduction rate with a first order
rate of 0.0117 hr-1 for the titanium dioxide coated nickel and 0.0207 hr-1 for the ferric oxide
coated electrode. Although these rates are 3 to 5 times higher than those obtained with Fe0, the
reaction rates (Table 4.1) are still extremely slow for any practical treatment consideration.
105 Since electrochemical reduction of perchlorate with different electrodes resulted in low removal efficiencies, and slow reaction rates, it was decided to investigate and better understand chlorate reduction.
Table 4.1 First order reaction rates (hr-1) for electrochemical reduction of chlorate
Coating Catalyst Electrolyte pH 3.75 5 7.3 11-12
Glassy ------0.0026 - - Na2SO4 .0209 0.0157 - 0.0107 Carbon - Co(NO3)2 Na2SO4 .0348 - - - - CoCl2 Na2SO4 .0455 ------0.0077 - - NaCl - 0.0264 - - # Ti02-spray Na2SO4 0.0331 - - - Na2SO4 0.0396 - - - Ti02-spray CoCl2 Na2SO4 0.0808 - - - Nickel CoCl2 Na2SO4 - - .045 - Na2SO4 0.0445 - - - Na2SO4* 0.0482 - - - Ti02-dip Pd/C-Re Na2SO4 0.0548 - - - Na2SO4* 0.0805 - - - CoCl2 Na2SO4 0.0607 - - - FeO3 CoCl2 Na2SO4* 0.0653 - - - * 0.054 M Electrolyte Ionic Strength # 2 Layers of TiO2 coating
4.4.2 Chlorate
4.4.2.1 Chlorate Batch Experiments
Batch experiments were conducted for reduction of chlorate with Fe0 filings following a procedure similar to that used for perchlorate [35]. Two levels of iron dosages (20 and 40 g/L) were selected and the batch experiments were run in the presence and absence of dissolved oxygen. The degradation of chlorate with time in the presence or absence of dissolved oxygen are shown in Figures 4.4a and 4.4b. Upto 70% of the chlorate was removed after eight hours at
106 the 40g/L iron dosage, whereas around 40% removal was observed at the lower dosage. Under similar conditions, it took almost 15 days to achieve a similar reduction in perchlorate concentration with an iron filings dosage 50 times higher. Chlorate reduction with iron filings is almost 2 orders of magnitude faster when compared to perchlorate reduction rate. Surprisingly,
dissolved oxygen did not result in any significant difference in chlorate reduction (Figure 4.4a).
Higher removal at a higher DO level was observed in another study that looked at Fe0 reduction of nitrates [43]. This was attributed to the formation of certain unstable surface iron precipitates in the presence of oxygen that might facilitate contaminant reduction. Figure 4.4b shows first
order reaction kinetics fit the data for chlorate reduction well at the two iron dosages.
1.0 0.0
-0.2 0.8 -0.4
0.6 -0.6
-0.8 C/Co
0.4 ln(C/Co) -1.0 20g/l - DO <1 mg/L 20g/l - DO < 1mg/L 0.2 40g/l - DO < 1mg/L -1.2 40g/l - DO < 1mg/L 20g/l - saturation DO 20g/l - saturation DO 40g/l - saturation DO -1.4 40g/l - saturation DO 0.0 02468100246810 Time (hr) Time (hr)
Figure 4.4. a. Chlorate degradation (C0: 10 mg/L) with iron filings at two iron dosages with
and without DO removal; b. Zero order reaction kinetics for chlorate reduction
4.4.2.2 Electrochemical Reduction of Chlorate
Experiments were conducted to study the electrochemical reduction of chlorate on nickel
and carbon electrodes. The reduction of chlorate with nickel and carbon electrodes at current of
107 75 mA and a six hour run were 4% and 11%, respectively (Figure 4.5). Higher current (200 mA)
and introducing electrolyte (0.027M NaCl) in the cathode increased chlorate reduction to around
10% after a 4 run.
Experiments were conducted with a glassy carbon rod (HTF, Germany) electrode in the
at 200 mA current, at different pHs, in the presence and absence of cathodic electrolyte and with
Cobalt based catalysts. The results of the tests are shown in Figure 4.6. In absence of pH control, the pH of the system quickly increased to 12. At this pH, no reduction of chlorate was observed after a 4 hour run. Sodium sulfate (0.027 M ionic strength) was introduced as an electrolyte and was used in all subsequent experiments in place of sodium chloride. Use of Na2SO4 electrolyte
increased the chlorate reduction by about 5%. Repeating the experiment at pH = 5 (acetate
buffer) and pH = 3.75 (formate buffer) resulted in a marginal increase in chlorate reduction,
therefore all subsequent experiments were conducted at this pH = 3.75. The use of 100 mg/L
CoCl2 or Co(NO3)2 as catalysts increased the chlorate reduction to about 15%. About 10 %
reduction in nitrate concentration was observed which suggested that nitrate was competing with
chlorate for electrochemical reduction. The first order reaction rate constants for the various
electrochemical experiments are shown in Table 4.1.
108 1.00
0.95 0 C/C 0.90
C electrode, 75 mA Ni electrode, 75 mA Ni electrode - NaCl electrolyte, 200 mA 0.85
0.00 0 1 2 3 4 5 6 7
Time (hr)
Figure 4.5 Electrochemical reduction of chlorate with various electrodes
1.00
0.95
0.90
0 0.85 C/C
0.80 pH 11-12 pH 11-12; Na2SO4 electrolyte pH 5; Na 0.75 2SO4 electrolyte pH 3.75; Na2SO4 electrolyte pH 3.75; Na2SO4 electrolyte -CoNO3 catalyst 0.70 pH 3.75-Na2SO4 electrolyte -CoCl2 catalyst 0.00 0 1 2 3 4 5 Time (hr)
Figure 4.6 Electrochemical reduction of chlorate with ‘glassy carbon’ electrode
109 The observed electrochemical reduction rates were still very low. One reason could be
the chlorate ions were unable to adsorb onto the electrode surface prior to electrochemical reduction. A thin oxide film would highly enhance the adsorption of chlorate ions on the electrode surface. This phenomenon was observed with zero valent iron where an iron film provided a catalytic effect during the reduction of both perchlorate and chlorate ions. To improve catalytic activity the nickel electrodes were coated with a thin titanium oxide (TiO2) film from a
TiO2 sol-gel and the thin film was calcined at 500 ºC before being used for the electrochemical
experiments. The results for electrochemical experiments conducted with these TiO2 coated nickel electrodes are shown in Figure 4.7. Most of the experiments were conducted at 200 mA current, at pH 3.75 and with 0.027 M ionic strength Na2SO4. Because sulfuric acid had to be
constantly added to adjust the pH, the experiments were conducted only for 4 to 5 hours to limit
interference during analysis of samples by ion chromatography.
Chlorate reduction of 13 % was observed after a four hour run as (Figure 4.7) and Table
4.1 shows the first order reaction rate constants. However at the end of the experiment, upon
closer inspection, it was observed the film had worn off in certain places and was not uniform
anymore. Hence, 8 additional layers were coated on the electrodes and the experiment was
repeated. With the additional layers, the chlorate reduction increased to about 15% with the
-1 -1 reduction rate increasing from 0.0331 hr to 0.0396 hr . Addition of CoCl2 to the initial chlorate
solution further increased the removal to around 26% and doubled the reduction rate to 0.0808
hr-1 after 4 hours. However, when the experiment was repeated at a pH of 7.3 (HEPES buffer),
the decrease in chlorate reduction was marginal (20%) after 5 hours with the first order reaction
rate of 0.045 hr-1.
110 One drawback of the spray coating of nickel electrodes was lack of controlling the film
thickness, which resulted in uneven coating thickness. It was also difficult to achieve uniform
coating with this method. In order to eliminate these difficulties, a dip coating method was used.
The dip coating method allowed better control over the speed of dipping, thereby resulting in uniform and consistent TiO2 films. Therefore dip coating method was used for all subsequent
coatings. Hurley et al. [40] developed a new heterogeneous catalyst that promoted efficient
reduction of perchlorate in the presence of hydrogen. Using this Palladium (Pd), carbon and
Rhenium (Re) catalyst reduced perchlorate at rates significantly faster when compared to
conventional reducing agents. Since the electrochemical process generated hydrogen gas in the
1.00
0.95
0.90
0.85
0 0.80
C/C 0.75
0.70 Ni electrode - pH 5 Ni electrode -NaCl - pH 5 TiO (2 layers) - Na SO - pH 3.75 0.65 2 2 4 TiO2 (8 layers) - Na2SO4- pH 3.75 0.60 TiO2 (8 layers) - Na2SO4- pH 3.75 + CoCl2 TiO2 (8 layers) - Na2SO4- pH 7.3 + CoCl2 0.00 0 1 2 3 4 5 6
Time (hr)
Figure 4.7 Electrochemical reduction of chlorate with Ni electrode spray-coated with TiO2
111 cathode during reduction, it was decided to test its effectiveness as a catalyst during the
electrochemical reduction of chlorate. Experiments were conducted in the presence and absence of the catalyst and with the cathodic electrolyte at two different ionic strengths with dip coated nickel electrodes. Chlorate degradation with time, at selected catalysts and electrolyte concentrations is shown in Figure 4.8. For the 0.027 M ionic strength, the presence of catalyst increased perchlorate removal from 17 to 22% after 4 hours. With CoCl2 as catalyst a removal of
20% was observed after a 4 hour run. Table 4.1 compares the first order reaction rates for the
experimental results for electrochemical reduction of chlorate. At the lower ionic strength of
0.027 M, the presence of catalyst increased the reduction rate from 0.0445 to 0.0548 hr-1, whereas at the higher ionic strength, the rate increased from 0.0482 to 0.0805 hr-1. The ionic
strength of the cathodic electrolyte had an effect of the reaction rate values and the effect was
more pronounced in the presence of the catalyst (~ 50% increase). An experiment was conducted
with a ferric oxide coated electrode and a 22% reduction was observed after a 4 hour run with a
first order reaction rate of 0.0653 hr-1. Based on the results obtained with electrochemical
perchlorate reduction, no effort was made to use Fe0 as a cathode for electrochemical chlorate
reduction. However, a test integrating electrochemical and Fe0 reduction, where a column packed
with Fe0 was added downstream to the electrochemical reactor, showed no increase in chlorate
removal. This could be attributed to the high pH effluent of the electrochemical reactor that
contained substantial amounts of hydrogen gas, adversely affecting the iron.
112 1.0
0.9
0.8
C/Co 0.7
0.027 M Electrolyte Ionic Strength, No catalyst 0.054 M Electrolyte Ionic Strength, No catalyst 0.6 0.027 M Electrolyte Ionic Strength, Pd/C-Re catalyst 0.054 M Electrolyte Ionic Strength, Pd/C catalyst 0.027 M Electrolyte Ionic Strength, CoCl2 catalyst 0.5 0.054 M Electrolyte Ionic Strength, CoCl2 catalyst, Fe3 coating 0.0 0 1 2 3 4 5 6 7 8
Time (hr)
Figure 4.8 Electrochemical reduction of chlorate with Ni electrode dip-coated with TiO2
4.4.3 XRD Analysis
The X-ray diffraction (XRD) patterns for five different iron samples, is shown in Figure
4.9. The effects of acid washing of Fe0 are shown in Figures 4.9a and 4.9b that compare the
XRD results of Fe0 before and after acid washing. Huang observed the reaction rates achieved
with untreated, raw Fe0 were significantly lower compared to that with acid washed iron [35].
This could be attributed to the higher activation energy associated with untreated iron containing
surface oxidation and other impurities [44]. Unwashed iron is covered by a passive layer of
0 Maghemite (Fe2O3), which inhibits any electron transfer and renders the Fe ineffective. Yu et al.
[45] looked at the surface properties of zero valent iron that influence its arsenite reduction
capacity. Iron oxides exposed to water are reduced by the following reactions.
113
2+ Fe2O3 + Fe + 6H+→3Fe + 3H2O (4.1)
12Fe2O3 + 3Fe→9Fe3O4 (4.2)
Acid washing reduces the maghemite content of the iron surface by reducing it to magnetite (Fe3O4) (reactions 1, 2). Even though magnetite forms passive layers, its semiconductors properties are such that it allows for electron transfer to take place [46].
At the same time the Fe0 undergoes a reaction with water to form ferrous ions.
2+ − Fe + 2H2O→Fe + 2OH + H2 (4.3)
The amorphous iron oxides formed during the initial dissolution and reprecipitation of the original rust contribute to the bulk of removal that occurs in the first few hours [45]. The presence of various oxides can be seen in Figures 4.9a and 4.90b, where the oxide peaks
(magnetite/maghemite) are more intensified in the acid washed sample compared to the untreated iron sample. These peaks signify the presence of more reactive sites available for sorption on the iron surface. Acid washing accelerates the above mentioned reduction reactions thereby generating more adsorption sites. A peak for ferrous oxide (FeO) indicates the presence of reduced ferrous ions that provide Fe0 with the ability to reduce contaminants. The ferrous iron species result in the formation of surface oxides and oxyhdroxides that are responsible for the reduction of the contaminant by the Fe0.
114 2+ − + Fe + 2OH →Fe(OH)2→FeOH →various iron oxides (4.4)
Lie et al. found that acid washing increased the oxyhydroxide and magnetite content of the Fe0 used for TCE dechlorination while reducing the maghemite content [46]. By increasing the content of the oxyhydroxide species, the availability of reactive sites on the iron surface is significantly altered. The large peak for iron oxide hydrate (FeOOH) in the acid washed iron verifies the presence of iron oxyhdroxide species. Iron corrosion is directly related to the reaction kinetics as the iron corrosion products, such as iron oxyhydroxides, provide new reactive sorption sites that eventually adsorb and reduce the perchlorate, chlorate and arsenic ions [45,
47]. The iron oxyhydroxides could present surface sites at which the perchlorate/chlorate ions form complexes resulting in their subsequent reduction.
115 1000 a. Untreated iron filings I 800
600 I Counts 400 I 200 M/H M/H M/H OH M/H 8000 b. Acid washed iron filings I 600 OH
400 I
Counts M/H M/H I 200 M/H M/H M/H M/H
5000 c. Iron filings - electrochemical reactor 400 M/H M/H I 300 M/H I Counts 200 M/H M M/H M I/S M M 100 M M C M C S 3500 d. Iron filings - column experiment 300 M/H I 250 M/H 200 I 150 M Counts M/H M/H M I M 100 M M OH 50
0 1200 e. Iron filings - H2 column 1000 M/H 800 M/H 600 M/H Counts 400 M/H I M/H M M I M 200 M M M 0 30 40 50 60 70 80 90 Degrees (2θ)
Figure 4.9 X-Ray Diffraction spectrum of zero-valent iron samples
0 [I: Iron (Fe ); M: Magnetite (Fe3O4); H: Maghemite (Fe2O3); OH: Iron oxide hydrate
(FeOOH); S: Siderite (FeCO3); C: Carbonyl Iron (FeC)]
116 XRD spectrum for iron sample after an electrochemical experiment shows the peaks for
the oxides were intensified as a result of exposure to water. Peaks for carbonyl iron (FeC) and
siderite (FeCO3) are also observed (Figure 4.9c). The two most important factors responsible for
diminishing the capacity of Fe0 to reduce contaminants are surface passivation and precipitation
of certain minerals [43, 48]. Precipitates from oxides, hydroxides and carbonates of iron cause cementation resulting in decrease in permeability leading to flow restriction and eventually clogging of the PRB. Precipitates of FeS from the sodium sulfate used as the cathodic electrolyte and sulfuric acid used to adjust the pH were possible, although because of its amorphous nature,
FeS is normally not detected by XRD [49]. During the electrochemical experiment, the pH in the cathode was elevated to (11-12) and the iron was exposed to this solution for a prolonged period of time. High pH has been found to negatively impact trichloroethylene removal by Fe0 [41]. At elevated pH, the dissolved ferrous and hydroxyl species precipitate as ferrous hydroxide on the iron surface, thereby blocking access to reactive sites and reducing the reduction rate [43]. This might have resulted in decreased reduction of perchlorate after the electrochemical experiment.
Low pH prevents the formation of passive oxide layers by keeping more of the iron dissolved in solution, thereby resulting in better perchlorate removal [45, 50]. The generation of hydrogen gas in the cathode during the electrochemical experiment might have affected the surface properties of the iron.
Figures 4.9d and 4.9e compare the XRD patters for the iron from the column experiments with and without hydrogen gas. The effects of hydrogen gas on surface oxides were not apparent since the difference in the XRD spectrums of the samples were not significant. However, other studies have shown that presence of hydrogen gas reduces the capacity of Fe0. Chen et al. [50]
117 found that Fe0 has high reactivity towards hydrogen ions to generate hydrogen gas, which has the
ability to reduce the lifetime of the Fe0, especially at low pH.
0 2+ Fe + 2H+ → Fe + H2 (4.5)
In another study, Chen et al. [41] found that presence of hydrogen gas was significantly
reducing the reaction rates for TCE degradation by Fe0. They attributed this to the formation of hydrogen gas micro-bubbles on the iron surface that bonded with the iron lattice, accumulated and eventually even adsorbed by the Fe0. These hydrogen bubbles could be hindering the
transport and diffusion of the perchlorate ions from the bulk solution to the Fe0 surface and vice
versa for the perchlorate reduction products.
4.5 Conclusions
• Use of iron filings as cathode in the electrochemical reduction of perchlorate at pH > 11
seemed to significantly affecting the surface properties of the zero valent iron deeming it
ineffective, and the result was confirmed by the X-ray diffraction spectrum data.
• Presence of hydrogen gas and its interaction with the iron during the electrochemical
experiment significantly diminished the perchlorate reduction capacity of Fe0. This was
confirmed by the Fe0 column experiment run in the presence of hydrogen gas.
• Use of nickel and platinum as electrodes resulted in very low perchlorate removal (~5%).
However coating the electrodes with thin oxide films resulted in reduction rates 3-5 times
118 higher than that obtained with Fe0. However these rates were still extremely low for any
practical consideration.
• Batch experiments with Fe0 for chlorate reduction produced rates that were significantly
higher when compared to perchlorate.
• Electrochemical experiments for chlorate resulted in rates much higher when compared
to perchlorate. Electrochemical reduction of chlorate was significantly affected by pH,
electrolyte ionic strength, coatings of thin oxide films on the electrodes and the presence
of catalysts. The reduction of chlorate obtained with electrochemical method was still
significantly slower when compared to that obtained with zero valent iron filings.
However with the use of thin oxide films and catalysts, the rates were comparable to that
obtained with Fe0.
4.6 Acknowledgements
This research study was supported by the U.S. Environmental Protection Agency (U.S.
EPA) under Contract No. EP-C-05-056. Although this work was reviewed by the EPA and approved for publication, it may not necessarily reflect official agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
119 4.7 References
1 USEPA (2007) Perchlorate. (Reuse, F. F. R. a., ed.), Federal Facilities Restoration and
Reuse
2 Clark, J. J. J. (2000) Toxicology of Perchlorate. In Perchlorate in the Environment
(Urbansky, E. T., ed.), pp. 15-29, Kluwer Academic/Plenum, New York
3 Urbansky, E. T. (1998) Perchlorate Chemistry: Implications for Analysis and
Remediation. Bioremediation Journal 2, 81-95
4 Gu, B., Ku, Y.-K. and Brown, G. M. (2002) Treatment of Perchlorate-Contaminated
Groundwater Using Highly Selective, Regenerable Ion-Exchange Technology: A Pilot-
Scale Demonstration. Remediation Journal 12, 51-68
5 Gu, B., Ku, Y. K. and Brown, G. M. (2005) Sorption and Desorption of Perchlorate and
U(VI) by Strong-Base Anion-Exchange Resins. Environmental Science & Technology
39, 901-907
6 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. I: Characterization of NH3 thermally tailored GACs.
Carbon 43, 573-580
7 Chen, W., Cannon, F. S. and Rangel-Mendez, J. R. (2005) Ammonia-tailoring of GAC to
enhance perchlorate removal. II: Perchlorate adsorption. Carbon 43, 581 - 590
8 Parette, R. and Cannon, F. S. (2005) The removal of perchlorate from groundwater by
activated carbon tailored with cationic surfactants. Water Research 39, 4020-4028
9 Yoon, Y., Amy, G., Cho, J., Her, N. and Pellegrino, J. (2002) Transport of perchlorate
- (ClO4 ) through NF and UF membranes. Desalination 147, 11-17
120 10 Yoon, J., Yoon, Y., Amy, G., Cho, J., Foss, D. and Kim, T.-H. (2003) Use of surfactant
modified ultrafiltration for perchlorate (ClO4-) removal. Water Research 37, 2001-2012
11 Xu, J., Trimble, J. J., Steinberg, L. and Logan, B. E. (2004) Chlorate and nitrate reduction
pathways are separately induced in the perchlorate-respiring bacterium Dechlorosoma sp.
KJ and the chlorate-respiring bacterium Pseudomonas sp. PDA. Water Research 38, 673-
680
12 Xiaohua, L., Roberts, D. J., Hiremath, T., Clifford, D. A., Gillogly, T. and Lehman, S. G.
(2007) Divalent Cation Addition (Ca2+ or Mg2+) Stabilizes Biological Treatment of
Perchlorate and Nitrate In Ion-Exchange Spent Brine. Environmental Engineering
Science 24, 725-735
13 Wallace, W., Beshear, S., Williams, D., Hospadar, S. and Owens, M. (1998) Perchlorate
reduction by a mixed culture in an up-flow anaerobic fixed bed reactor. Journal of
Industrial Microbiology and Biotechnology 20, 126-131
14 Steinberg, L. M., Trimble, J. J. and Logan, B. E. (2005) Enzymes responsible for chlorate
reduction by Pseudomonas sp. are different from those used for perchlorate reduction by
Azospira sp. FEMS Microbiology Letters 247, 153-159
15 Okeke, B. C., Giblin, T. and Frankenberger, W. T. (2002) Reduction of perchlorate and
nitrate by salt tolerant bacteria. Environmental Pollution 118, 357-363
16 Nerenberg, R., Kawagoshi, Y. and Rittmann, B. E. (2006) Kinetics of a hydrogen-
oxidizing, perchlorate-reducing bacterium. Water Research 40, 3290-3296
17 Kim, K. and Logan, B. E. (2001) Microbial Reduction of Perchlorate in Pure and Mixed
Culture Packed-Bed Bioreactors. Water Research 35, 3071-3076
121 18 Herman, D. C. and Frankenberger, W. T., Jr. (1999) Bacterial reduction of perchlorate
and nitrate in water. Journal of Environmental Quality 28, 1018-1024
19 Gingras, T. M. and Batista, J. R. (2002) Biological reduction of perchlorate in ion
exchange regenerant solutions containing high salinity and ammonium levels. Journal of
Environmental Monitoring 4, 96-101
20 Giblin, T., Herman, D., Deshusses, M. A. and Frankenberger, W. T., Jr. (2000) Removal
of perchlorate in ground water with a flow-through bioreactor. Journal of Environmental
Quality 29, 578-583
21 Coates, J. D., Michaelidou, U., Bruce, R. A., O'Connor, S. M., Crespi, J. N. and
Achenbach, L. A. (1999) Ubiquity and Diversity of Dissimilatory (Per)chlorate-Reducing
Bacteria. Applied & Environmental Microbiology 65, 5234-5241
22 Almeida, C. M. V. B., Gianetti, B. F. and Rabockai, T. (1997) Electrochemical study of
perchlorate reduction at tin electrodes. Journal of Electroanalytical Chemistry 422, 185-
189
23 Gu, B., Dong, W., Brown, G. M. and Cole, D. R. (2003) Complete Degradation of
Perchlorate in Ferric Chloride and Hydrochloric Acid under Controlled Temperature and
Pressure. Environmental Science & Technology 37, 2291-2295
24 Lang, G. G. and Horanyi, G. (2003) Some interesting aspects of the catalytic and
electrocatalytic reduction of perchlorate ions. Journal of Electroanalytical Chemistry 552,
197-211
25 Espenson, J. H. (2000) The Problem and Perversity of Perchlorate. In Perchlorate in the
Environment (Urbansky, E. T., ed.), pp. 1-7, Kluwer Academic/Plenum, New York
122 26 Horanyi, G. and Bakos, I. (1993) Study of the relationship between voltammetric
behavior and electroanalytic activity in the reduction of ClO4- ions at platinized platinum
electrodes. Journal of Electroanalytical Chemistry 347, 383 - 391
27 Horanyi, G., Bakos, I., Szabo, S. and Rizmayer, E., M., (1992) New observations in the
field of electrochemistry of technetium and rhenium: electrocatalytic reduction of
perchlorate ions at electrosorbed and electrodepositied Tc and Re layers in acid medium.
Journal of Electroanalytical Chemistry 337, 365-369
28 Colom, F. and Gonzalez - Tejera, M. J. (1985) Reduction of Perchlorate Ion on
Ruthenium Electrodes in Aqueous Solutions. Journal of Electroanalytical Chemistry 190,
243 - 255
29 Brown, G. M. (1986) The Reduction of Chlorate and Perchlorate at an Active Titanium
Electrode. Journal of Electroanalytical Chemistry 198, 319-330
30 Wasberg, M. and Horanyi, G. (1995) The reduction of ClO4- ions on Rh electrodes.
Journal of Electroanalytical Chemistry 385, 63-70
31 Theis, T. L., Zander, A. K., Xiang, L., Jeosadaque, S. and Anderson, M. A. (2002)
Electrochemical and Photocatalytic Reduction of Perchlorate Ion. Journal of Water
Supply: Research and Technology - AQUA 51, 367 - 374
32 Wehrli, B. (1990) Redox Reactions of Metal Ions at Mineral Surfaces. In Aquatic
Chemical Kinetics (Stumm, W., ed.), pp. 311-337, Wiley Interscience, New York
33 Gurol, M. G. and Kim, K. (2000) Investigation of Perchlorate Removal in Drinking
Water Sources by Chemical methods. In Perchlorate in the Environment (Urbansky, E.
T., ed.), pp. 99-107, Kluwer Academic/Plenum, New York
123 34 Moore, A. M., DeLeon, C. H. and Young, T. M. (2003) Rate and Extent of Aqueous
Perchlorate Removal by Iron Surfaces. Environ. Sci. Technol. 37, 3189-3198
35 Huang, H. and Sorial, G. A. (2007) Perchlorate Remediation in Aquatic Systems by Zero
Valent Iron. Environmental Engineering Science 24, 917-926
36 Siddiqui, M. S. (1996) Chlorine-ozone interactions: Formation of chlorate. Water
Research 30, 2160-2170
37 Westerhoff, P. (2003) Reduction of nitrate, bromate, and chlorate by zero valent iron
(Fe0). Journal of Environmental Engineering 129, 10-16
38 Chen, Y. and Dionysiou, D. D. (2006) Effect of calcination temperature on the
photocatalytic activity and adhesion of TiO2 films prepared by the P-25 powder-modified
sol-gel method. Journal of Molecular Catalysis. A, Chemical 244, 73-82
39 Gash, A. E., Tillotson, T. M., Satcher, J. H., Poco, J. F., Hrubesh, L. W. and Simpson, R.
L. (2001) Use of Epoxides in the Sol-Gel Synthesis of Porous Iron(III) Oxide Monoliths
from Fe(III) Salts. Chemical Materials 13, 999-1007
40 Hurley, K. D. and Shapley, J. R. (2007) Efficient heterogeneous catalytic reduction of
perchlorate in water. Environmental Science & Technology 41, 2044
41 Chen, J.-L., Al-Abed, S. R., Ryan, J. A. and Li, Z. (2001) Effects of pH on dechlorination
of trichloroethylene by zero-valent iron. Journal of Hazardous Materials 83, 243-254
42 Scherer, M. M. B., B. A. & Tratnyek, P. G. (1998) The Role of Oxides in Reduction
Reactions at the Metal-Water Interface. In Mineral-Water Interfacial Reactions: Kinetics
and Mechanisms (Grundl, D. L. S. T. J., ed.), pp. 301-322, American Chemical Society.,
Washington DC
124 43 Westerhoff, P. and James, J. (2003) Nitrate removal in zero-valent iron packed columns.
Water Research 37, 1818-1830
44 Slater, G. F., Lollar, B. S., King, R. A. and O'Hannesin, S. (2002) Isotopic fractionation
during reductive dechlorination of trichloroethene by zero-valent iron: influence of
surface treatment. Chemosphere 49, 587-596
45 Yu, X., Amrhein, C., Zhang, Y. and Matsumoto, M. R. (2006) Factors Influencing
Arsenite Removal by Zero-Valent Iron. Journal of Environmental Engineering 132, 1459-
1469
46 Liu, C.-C., Tseng, D.-H. and Wang, C.-Y. (2006) Effects of ferrous ions on the reductive
dechlorination of trichloroethylene by zero-valent iron. Journal of Hazardous Materials
136, 706-713
47 Lien, H.-L. and Wilkin, R. T. (2005) High-level arsenite removal from groundwater by
zero-valent iron. Chemosphere 59, 377-386
48 Phillips, D. H., Gu, B., Watson, D. B. and Roh, Y. (2003) Impact of Sample Preparation
on Mineralogical Analysis of Zero-Valent Iron Reactive Barrier Materials. J Environ
Qual 32, 1299-1305
49 McGeough, K. L., Kalin, R. M. and Myles, P. (2007) Carbon Disulfide Removal by Zero
Valent Iron. Environmental Science & Technology 41, 4607-4612
50 Chen, S.-S., Cheng, C.-Y., Li, C.-W., Chai, P.-H. and Chang, Y.-M. (2007) Reduction of
chromate from electroplating wastewater from pH 1 to 2 using fluidized zero valent iron
process. Journal of Hazardous Materials 142, 362-367
125 CHAPTER 5
Treatment of Taste and Odor Causing Compounds in Drinking Water:
A Review
5.1 Abstract
Problems due to the taste and odor in drinking water are common in treatment facilities
around the world. Taste and odor are primary indicators of the safely and acceptability of
drinking water and are mainly caused by the presence of two semi-volatile compounds – 2-
methyl isoborneol (MIB) and geosmin. A review of these two taste and odor causing compounds in drinking water is presented. The sources for the formation of these compounds in water are discussed alongwith the health and regulatory implications. The recent developments in the analysis of MIB/geosmin in water which have made measurements in the nanogram per liter concentrations feasible are also discussed. The relevant treatment alternatives are described in detail with emphasis on their respective advantages and problems associated with their implementation in a full-scale facility. Conventional treatment processes in water treatment plants, such as coagulation, sedimentation and chlorination have been found to be ineffective for removal of MIB/geosmin. Studies have shown powdered activated carbon (PAC), ozonation and biofiltration to be effective in treatment of these two compounds. Although some of these technologies are more effective and show more promise than the others, much work remains to be done in order to optimize these technologies so that they can be retrofitted or installed with minimal impact on the overall operation and effectiveness of the treatment system.
126 5.2 Introduction
Drinking water industries, even in most industrialized and developed nations, are facing
the necessity for more innovative and cost-effective technologies for water treatment and
purification. Taste and odor have long been associated with the suitability and safety of our
drinking water. It is not uncommon for water utilities to be flooded with complaints from
consumers about taste and odor in their drinking water, especially during warmer weather. For an
average consumer, taste and odor is the only way of determining the safety of tap water [1].
Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) and MIB (2-methyl isoborneol–
C11H20O) have been identified to be the major taste and odor-causing compounds in drinking
water obtained from surface water (Figure 5.1) [2]. Geosmin and MIB in surface water mainly
result from the biodegradation of certain types of cyanobacteria that normally bloom in the
presence of nutrients at warmer temperatures. There are currently no regulations for these two
compounds as they have not been associated with any health effects [3]. Presence of taste and
odor in drinking water may result in decreased consumer trust and subsequently, decreased water
consumption and could eventually cause the public to switch to alternate sources of drinking
water such as bottled water and in-home treatment systems. The main problem with the presence
of Geosmin and MIB is associated with their extremely low odor threshold concentrations (OTC)
and their persistence to elimination in a conventional water treatment process such as
coagulation, sedimentation, filtration and chlorination [4]. The OTC for Geosmin and MIB is 4
ng/L and 9 ng/L, respectively [2]. Another factor that adds to the challenge faced by drinking
water facilities in the removal of these contaminants is the presence of natural organic matter
(NOM). NOM, a complex mixture of organic compounds derived primarily from the decay of
127 plant and animal materials, is invariably present in all water sources and at much higher
concentrations than geosmin or MIB [2].
2 Methylisoborneol (MIB) Geosmin
Figure 5.1 Molecular structure of MIB and geosmin
The only treatment methods that have been successfully employed by water treatment plants to
remove MIB and Geosmin are adsorption by activated carbon or oxidation by strong oxidants
such as ozone. Ferguson et al. [5], Glaze et al. [6] and Bruce et al. [4] studied and demonstrated
MIB and Geosmin removal using oxidants such as ozone, hydrogen peroxide and UV. Addition of chemicals however is expensive and can result in formation of disinfection byproducts
(DBPs), which are unacceptable due to health and regulatory concerns. Adsorption by activated
carbon, either granular activated carbon (GAC) or powdered activated carbon (PAC) is
considered as one of the best available technologies for removal of organic contaminants from
water. Numerous studies have looked into GAC and PAC adsorption of MIB and Geosmin [2, 7-
10]. However, NOM levels of 3–10 mg/L competitively reduce activated carbon adsorption
capacity for MIB/Geosmin [2, 11, 12]. In natural waters, the size and concentration of NOM
particles is many folds higher than that of MIB/Geosmin, and as a result a large volume of the
GAC is not utilized for MIB/Geosmin adsorption, thereby significantly reducing the GAC
adsorption capacity.
128 There have been some good reviews of aspects related to MIB/geosmin in drinking water
– scale of the problem, sources and prediction of these two odorants, that are essential for
understanding and managing taste and odor in drinking water [13-15]. In addition to covering the
above issues, this review takes a more holistic approach and also looks at current treatment
technologies and the challenges faced by the utilities in their application. The review starts with a discussion of the sources followed by a look at the associated health effects and the current regulatory status. It then looks at recent developments in the analysis of these contaminants in drinking water. Finally the available treatment technologies are discussed in detail along with the issues related to the application of these technologies in the field. From this review, it is evident that although some technologies are more effective and applicable than the others, a completely accepted technology that could be used in any drinking water treatment facility still does not exist. More research still needs to be performed in order to arrive at a treatment control system that would have unrestricted application potential for removal of these two odor causing compounds.
5.3 Sources and Contamination of MIB/Geosmin in Water
It is not uncommon for water utilities to be overwhelmed with consumer complaints
about taste and odor when the concentrations of MIB/geosmin exceed the odor threshold,
especially during summer months. Studies indicate the main source of MIB/geosmin in water are
cyanobacteria (blue-green algae) [14, 15] . However, some studies have also shown the
MIB/geosmin in surface water can be attributed to the presence of certain type of filamentous
bacteria or actinomycetes [16]. These cyanobacteria synthesize MIB/geosmin during growth and
these algal cells release or store these odorants depending on the growth phase and also based on
129 environmental factors. Most of the MIB/geosmin is released during the death and biodegradation
of these cells (Figure 5.2). The taste and odor outbreaks are more prominent during eutrophic
conditions, when there is an overabundance of nutrients and warmer temperatures. These
conditions lead to cyanobacterial blooms in the surface water resulting in significant
MIB/geosmin production. Jutner et al. identified the various strains of cyanobacteria responsible
for the production of MIB/geosmin in water [13]. The study also identified the possible pathways
for biosynthesis of these two odorants. They summarized the MIB/geosmin concentrations in
waters around the world, alongwith the habitat supporting the cyanobacteria source.
Nutrients Biodegradation Cyanobacteria Release of MIB II, Bloom Geosmin in water Warmer Temperatures Cyanobacteria
Figure 5.2 Pathway of MIB/ Geosmin Formation
Studies have detected high concentrations of MIB/geosmin in various sources of drinking
water as well as certain types of wastewater. Watson et al. investigated geosmin and MIB in pulp
and paper mill effluent during secondary treatment [17]. They detected both compounds at
concentrations 2000-9000 times their OTCs resulting in a significant local odor in the bioreactor.
As expected, the concentrations were highest in the summer months during warmer temperatures and cyanobacteria were reported to be the likely source. Taste and odor in drinking water is not
restricted to certain geographies and is a problem faced by water treatment facilities around the
world. Lin et al. studied the correlation between musty odors and MIB concentration in two
drinking water plants in Taiwan [18]. The authors employed flavor profile analysis (FPA), where
a trained panel was asked to rate the intensity of the odor on a scale. They also measured the
130 MIB concentrations using solid phase micro-extraction (SPME) and found a good correlation between the FPA and the analytical results. Although the conventional treatment process removed upto 50% of the MIB only about 30% reduction was observed based on the FPA scale suggesting that the concentrations were still higher than the OTC.
Studies have shown MIB/geosmin accumulate in surface water reservoirs as well [19]. A
recent study found that the taste/odor impairments were widespread in the Great Lakes region
with a significant number of utilities reporting annual outbreaks during the summer months with
even a higher number experiencing erratic episodes [15]. Studies have shown that algal growth
occurring in water treatment plants in locations such as sedimentation and coagulation basins can
also result in the production of MIB/geosmin [4].
5.4 Development in Analysis of MIB/Geosmin in Water Samples
Analysis of geosmin and MIB was performed by conventional analytical techniques in
the late eighties and early nineties. Although methods such as purge and trap (P&T), liquid-
liquid extraction were effective, they were expensive, time consuming and highly labor intensive
[20]. This led to development of membrane based methods for more accurate measurements of
these taste and odor compounds. Hollow fiber stripping analysis (HFSA) used microporous
hydrophobic hollow fiber membranes for the analysis and could measure parts per trillion (ppt)
concentrations in water [21]. However the apparatus setup was complex and involved a large
number of equipment. Although, solid phase extraction (SPE) was also able to measure very low
concentrations, it also suffered from similar drawbacks. In 1996, a new method called solid
phase micro-extraction (SPME) was used to measure organic micropollutants, especially volatile
131 organic compounds (VOCs) such as benzene and toluene [22]. This method employed a fused silica fiber for extraction of the contaminants from the sample headspace followed by injection into the headspace of a Gas Chromatography – Mass Spectrophotometry (GC/GC-MS) for analysis. Like SPE, this method did not require solvent extraction and the analytical procedure was simple and quick. Unlike SPE, the fiber could also be used repeatedly. Since that time, a lot of research has been done to optimize SPME for analysis of MIB and geosmin in water and now it has become the standard method [20, 23-25].
Lloyd et al. compared SPME-GC to P&T-GC for analysis of geosmin and MIB [20]. It was seen that the precisions and limits achieved with SPME were comparable to P&T, with
SPME offering faster analysis with smaller sample size. The authors ran a series of analysis to optimize the analytical method for MIB and geosmin analysis using SPME-GC. They varied the sampling temperature, vial size, sample exposure time to the SPME fiber, percentage headspace to arrive at a method that permitted analysis at concentrations as low as10 ng/L. In the last few years, there have been significant improvements in the quality of the SPME fiber, with more robust and sensitive fibers being developed. In a more recent study, an SPME-GC-MS based method was presented where method detection limits (MDL) as low as 2 ng/L were demonstrated for both MIB and geosmin [23]. The study also details a quality-control analysis to support the method performance. Further optimization of this method has resulted in detection limits being lowered to 0.4 ng/L [25]. Analysis using headspace sampling, known as HS-SPME-
GC-MS has been further optimized for measuring MIB/geosmin with detection limits under 1 pg/L in a very recent study [24]. Because the OTCs for these compounds are in the nano gram
132 per liter range, it is imperative to have analytical methods that can measure very low concentrations with high accuracy and precision.
5.5 Health Effects/Regulation
Numerous studies have shown that the presence of these taste and odor causing compounds in water is mainly an aesthetic concern and has not been associated with any health effects [26]. MIB and geosmin have also not been correlated to presence of cyanobacterial toxins which are extremely toxic at even low concentrations [23]. As a result, there is no maximum contaminant level (MCL) or maximum contaminant level goal (MCLG) for either geosmin or MIB. Studies have detected MIB/geosmin in various species of fish but have concluded that they do not result in any toxicity to either the fish or to humans through consumption of the fish [27, 28]. Although taste and odor in drinking water is rarely associated with toxic contaminants, for consumers, it is a primary measure of the safety of the drinking water. Not only can it undermine consumer trust in the water quality but it can also result in the use of alternative supplies of drinking water, such as bottled water [1, 29]. Watson et al summarized the various types of taste and odors encountered by drinking water consumers and their possible sources [14]. Earthy and musty odors, which are the most frequently type of taste/odor in drinking water, are due to the presence of MIB/geosmin in water. Whelton et al found that the water treatment/regulation strategy for these odor causing compounds should consider perceived OTCs, values for which show considerable variation among consumers, and is also influenced by other factors such as water temperature [30]. The results showed that reducing the water temperature from 45 °C to 25 °C reduced the perceived odor intensity significantly.
133 5.6 Treatment Technologies
Studies have shown that MIB/geosmin are extremely resistant to removal by
conventional water treatment processes such as coagulation, sedimentation and filtration. Bruce
et al investigated coagulation for removal of these taste and odor compounds and found that alum coagulation could not be optimized for MIB/geosmin removal [4]. No removal was observed under a range of pH and coagulation conditions including different alum dosages. It has been seen that common oxidants such as chlorine (Cl2), chlorine dioxide (ClO2) and potassium
permanganate (KMnO4) are not very effective for removal of these compounds [6, 31]. In a pilot
plant study comparing MIB/geosmin removal with different oxidants, removal efficiencies with
chlorine (Cl2) and Chlorine dioxide (ClO2) were very low and only ozone (O3) showed any
appreciable removal of MIB/geosmin (85% for 3.8 mg/L dosage at a contact time (CT) of 6.4
mins) [9]. KMnO4 has been found to have low removal even at higher dosages and chlorine
residuals in some cases have been found to enhance and even mask the musty/earthy odors rather
than removing them. Currently PAC is the most commonly used treatment technology for
removal of seasonal tastes and odors in water. However its effectiveness for removal of
MIB/geosmin is less when compared to some other contaminants. Also, the presence of NOM or
oxidants such as chlorines or chloramines reduces its capacity further [32]. Use of advanced
oxidation technologies such as ozone or UV with H2O2 has been found to be effective in
destroying MIB/geosmin. A lot of research has also been performed in the area of
bioremediation for treatment of MIB/geosmin. The three main technologies, GAC/PAC
adsorption, AOPs and biological treatment are discussed in further detail in this section with a
focus on advantages and challenges faced in implementation of these treatment alternatives.
134 5.6.1 GAC Adsorption
Adsorption with GAC/PAC is being widely used in drinking water treatment plants
mainly for removal of organic pollutants. It would be sensible to use this already existing
technology for removal of taste and odor compounds as well. Water treatment plants invariably
use GAC in the form of filtration beds. Various researchers have looked into use of activated
carbon for MIB/geosmin remediation, with their results discussed in detail in this section and
some of the key studies summarized in Table 5.1.
Ridal et al. investigated the long-term performance of GAC filter beds in a water
treatment plant in Canada for removal of MIB/geosmin [33]. When monitored after 2 months in
place, MIB/geosmin were removed to at or below the OTCs, thereby performing effectively.
They also found that the removal efficiencies were directly related to Cl2 residual. In fact, the
practice at the plant was to increase the Cl2 residual during taste/odor episodes. It was also
observed that increasing the contact time did not necessarily result in an increase in the removal
efficiencies. When the performance of the filters was monitored after 1 and 2 year periods, the
performance had dropped significantly and the effluent concentrations were higher than the
OTCs. The authors suggested reasons for this, including flow channeling, media mixing and
coating of the GAC with dissolved organic carbon (DOC).
MacKenzie et al. compared various commercially available carbons for adsorption of
MIB [34]. Through rapid small scale column tests (RSSCTs) on water obtained from two
treatment plants, they found that although wood-based GACs had better breakthrough behavior when compared to coal-based GACs, their affinity for MIB after the odor episode was
135 significantly lower. They compared GAC regenerated with steam-curing reactivation under various conditions with conventional reactivation. The results showed that the breakthrough behavior of the steam-curing reactivated GAC was almost comparable to the virgin GAC and significantly better than conventionally reactivated GAC. Not only were the volume and mass loss with this new reactivation comparable to the conventional method, but it also resulted in a higher BET surface area.
Studies have been undertaken to tailor carbon to enhance its MIB/geosmin removal capacity. Rangel-Mendez et al. investigated MIB adsorption with tailored GAC. Tailoring the
GAC by thermal treatment with steam or steam + methane resulted in a significant increase in both micropore and mesopore volumes and treated 5 times more BV when compared to virgin
GAC [35]. Although steam treatment showed better results, upto 16% of the carbon mass was also lost during the process. Thermal treatment with steam + methane achieved similar results but with lesser mass loss.
PAC is another way of using activated carbon for water treatment during taste and odor episodes and in normally used prior to filtration. It is currently the most commonly practiced technology for MIB/geosmin removal. Cook et al. studied PAC adsorption of MIB/geosmin in four raw waters [7]. They conducted adsorption isotherm experiments and found geosmin showed better adsorption than MIB in all four waters. This was attributed to the lower solubility and the flatter structure in geosmin making it more amenable for adsorption. Based on the adsorption isotherms and kinetic experiments, they generated parameters to be used as input for
136 the homogenous surface diffusion model (HSDM) that was used to predict the PAC dosages
required. The model results showed good prediction for three of the four waters.
Ng et al. investigated geosmin adsorption on PAC obtained from activation of various
agricultural by-products and studied the effect of activation method and pore characteristics of
these carbon on geosmin adsorption [36]. They found that some pecan shells based PAC showed
better adsorption than a commercially available carbon, especially at very low concentrations [8].
Several factors including the adsorbent properties such as pore size distribution and
surface characteristics, and also water quality parameters such as presence of NOM, influence
the effectiveness of carbon adsorption. Yu et al. investigated adsorption of MIB/geosmin on
PAC obtained from five different sources to better understand the effect of various factors [37].
Among the different pore characteristics studied, only the micropore volume showed a
significant correlation with adsorption, with the best adsorption capacity exhibited by carbons
with the highest micropore volumes. Ho et al found that NOM, especially lower MW fractions,
reduced MIB adsorption by PAC during alum coagulation due to competitive adsorption [10].
Increase in turbidity and alum dose resulted in larger floc size which in turn reduced MIB
adsorption further due to incorporation of the PAC particles into these flocs.
137 Table 5.1 Activated carbon adsorption of MIB/geosmin
Key References Findings Newcombe et al. [11, 12] Simultaneous adsorption between NOM and MIB on GAC was mainly influenced by pore size/volume distribution NOM with size similar to MIB resulted in the majority of the competition Smaller NOM particles compete by direct and strong adsorption for the available adsorption sites Larger NOM compounds reduce equilibrium adsorption capacity by adsorbing closer to external surfaces and blocking access to pores To be effective in natural water, carbon should have a bimodal pore distribution Cook et al. [7, 38] Geosmin showed better adsorption than MIB on PAC for all waters studied This was attributed to the lower MW and solubility of geosmin Presence of NOM significantly reduced adsorption capacity for MIB/geosmin HSDM predicted successfully PAC doses required to reduce MIB/geosmin to the required levels in waters from three different treatment plants Pirbazari et al. [2] Developed a model to predict the adsorption of MIB/geosmin in a fixed bed GAC adsorber Experimental data, specifically adsorption and kinetic parameters were inputs for the model The model predicted breakthrough behavior well for both MIB/geosmin
Newcombe et al. investigated simultaneous adsorption of NOM and MIB on PAC [11,
12]. By studying NOM adsorption on six different types of carbon, they were able to conclude that NOM adsorption was mainly controlled by the MW distribution and the pore size distribution of the carbon. Competitive adsorption on PAC was done by studying MIB adsorption in six different NOM solutions, with different fractionation and properties. The adsorption isotherms showed significant reduction in the adsorption capacity in the presence of
NOM. The authors postulated that only the presence of low MW NOM fractions, comparable in size to the MIB molecule, resulted in competition for the available adsorption sites. It was concluded that for the carbon to be effective, it needs to have a bimodal pore distribution that
138 would provide MIB rapid access to the adsorption sites and at the same time minimize pore
blockage by the low MW NOM particles.
Pirbazari et al. investigated the effectiveness of the dispersed flow homogeneous surface
diffusion model (DFHSDM) to predict MIB/geosmin adsorption by a fixed-bed GAC adsorber
[2]. They conducted adsorption isotherms and kinetic rate studies to determine the parameters
required for model generations. They also conducted these experiments in the presence of NOM
to study their impact on GAC adsorption. The DFHSDM adsorber model predicted well the
column experimental data. The authors also used the model predictions to estimate the system
costs for the removal of these odorants in a full-scale GAC adsorber.
Compared with GAC, activated carbon fibers (ACFs) have attracted increasing attention due to their excellent surface properties, high adsorption capacity and are an ideal adsorbent for
targeting the impact of pore size [39]. Several studies have looked at adsorption of
micropollutants on activated carbon fibers (ACFs) to better understand the effect of pore size on
both single and multicomponent adsorption [39-45]. Multicomponent adsorption has been
predicted using the Ideals adsorption solute theory (IAST) in some of these studies. Srinivasan et
al looked at single solute and multicomponent adsorption of MIB/geosmin on three ACFs with different pore size distributions and compared the results with adsorption in the presence of
NOM and adsorption on GAC [46, 47]. The results indicated that the surface pore characteristics and specifically, the micropore distribution was the single most important factor influencing adsorption. It was seen that presence of NOM affected the adsorption capacity of the ACF, and this effect was more pronounced in ACF with the least microporosity. The IAST model predicted
139 the binary adsorption of MIB/geosmin well and the results indicated that adsorption of these compounds on ACFs was purely through physical adsorption with no oligomerization occurring.
Although ACFs seem to be effective for adsorption of these two odorants, more research is
required for any practical implementation of this technology.
As seen from the above discussion, various factors have to be considered for successful
application of GAC/PAC adsorption for MIB/geosmin removal. Bruce et al. looked at PAC
adsorption of MIB/geosmin in a drinking water treatment plant in Arizona [4] and came up with
practical recommendations. Prior to selection of a PAC, it is important to understand its
adsorption characteristics in the presence of NOM commonly found in a particular water.
Frequent monitoring of the influent MIB/gesomin levels is necessary and dosages should be
based on MIB concentrations since it is normally found at higher concentrations and is more
challenging to remove. Optimizing the PAC dosage is important since it can result in excessive
sludge production, reduced filter performance and larger operating costs.
5.6.2 Advanced Oxidation Processes (AOP)
A number of studies have looked into using AOPs such as ozone, UV and H2O2 for removal of MIB/geosmin. These methods are normally used in disinfection and are known to completely destroy the target contaminants. Studies looking into removal of MIB/geosmin using
AOPs are discussed in detail in this section with some of the key references summarized in Table
5.2. Collivignarelli et al. investigated MIB/geosmin removal using ozone and UV [48]. They used raw river water from a water treatment plant in Italy, spiked with MIB/geosmin and a metachlor (pesticide contaminant). Both MIB/geosmin were persistent to reaction with ozone
140 and showed low (~50%) removal rates. However, ozone followed by exposure to UV increased
the removal close to 90%. Molecular ozone had limited reaction with these two compounds and
UV rays were required for decomposition of the ozone molecule for generation of hydroxyl radicals (·OH) which subsequently reacted with MIB/geosmin. Also when compared to metachlor, MIB/geosmin required longer contact times with ozone (2-3 minutes at 2-3 mg/L) and higher UV doses (5000-6000 J/m2) for the same degree of removal. It is worth noting that this
process did not result in complete removal of MIB/geosmin and the resulting effluent
concentrations were higher than the OTCs.
Another study determined the reaction rates for ozone oxidation of several taste and odor
compounds in drinking water [49]. Similar to the results from the previous study, the second
- order reaction rates for oxidation with hydroxyl radical generated from UV/H2O2 (MIB: 0.35 M
1s-1; geosmin: 0.1 M-1s-1) were significantly higher than the rates obtained with molecular ozone
(MIB: 5.1 x 109 M-1s-1; geosmin: 7.8 x 109 M-1s-1 ). Removal of MIB/geosmin was studied in
water from two different lakes in Switzerland, and the oxidation efficiency for both compounds
was observed to be in the 50-70% range.
Similar to GAC adsorption, presence of NOM in water can influence AOPs as well. Ho et
al. investigated the effect of NOM characterization on the ozonation of MIB/geosmin [50].
Ozonation of fractionated NOM was studied and it was found that the NOM containing higher
molecular weight (MW) fractions had higher specific UV absorbance (SUVA) and also showed
the highest O3 demand, which inturn translated into a higher hydroxyl radical generation during
ozonation for a given contact time. This was confirmed with the highest MIB/geosmin removal
141 (98%) being observed for the lowest CT, which corresponded to the NOM fraction with the highest MW. However when the same ozonation was performed on two surface waters, MIB removal was higher in the water containing lower MW/SUVA NOM, thereby contradicting the experimental results. However, this was attributed to higher hydroxyl radical consumption due to the higher concentration of dissolved organic carbon (DOC). However, the MIB removal observed was similar when the same O3 dose to DOC ratio was compared. Parl et al also reached similar conclusion when they looked at kinetics of MIB/geosmin oxidation using H2O2 [51].
Table 5.2 Removal of MIB/geosmin by AOPs
Key References Findings Ho et al. [50] Studied NOM fractionations in a reservoir supplying drinking water NOM with higher SUVA characteristics had lower CTs and faster reaction with ozone Higher MIB/geosmin degradation in fractions containing higher SUVA NOMs Hydroxy radical was main mechanism of destruction of MIB/geosmin However, some NOM fractions competed with MIB/geosmin for the hydroxyl radicals generated during ozonation Rosenfeldt et al. [52] UV/H2O2 successfully destroyed MIB/geosmin although at doses higher than that required for disinfection Water quality parameters such as turbidity significantly influence UV oxidation of MIB/goesmin A steady state model to predict MIB/geosmin destruction was developed The model used inputs such as pH, alkalinity, TOC, UV absorbance to predict well the reaction rates Westerhoff et al. [53] Hydroxy radicals mediated reactions dominated MIB/geosmin oxidation during ozonation Second order reaction kinetics better for geosmin when compared to MIB MIB/geosmin removal directly related to pH, temperature, ozone and /H2O2 doses Developed an empirical model that combined effects of interconnected parameters such as bromate formation, ozone residual, CT inactivation on MIB/geosmin oxidation with ozone
142 Westerhoff et al. conducted batch ozonation experiments to study the effect of ozone oxidation parameters such as pH, ozone and H2O2 dosage and water quality parameters such as temperature and initial concentrations on the removal of MIB/geosmin [53]. Like the previous studies, hydroxyl radical dominated the oxidation reaction when compared to ozone and also geosmin showed better reaction kinetics when compared to MIB. Introduction of a hydroxyl radical scavenger during the reaction significantly reduced the removal efficiency. The results showed that removal efficiencies for both MIB/geosmin increased with increase in temperature, ozone dosage, pH and H2O2 concentration. An empirical model was developed to predict CT requirement, bromate formation and odorant oxidation, which are all interconnected. The CT ratio of hydroxyl radical to ozone was found to be the most critical factor for oxidation of
MIB/geosmin. Liang et al. also found similar results where they found that pH is a significant factor influencing oxidation as it is directly related to the hydroxyl radical concentrations [54].
Contrary to some earler results, it was seen that presence of background organics did not have a significant effect on ozonation of MIB/geosmin. Rosenfeldt et al. looked at UV oxidation of
MIB/geosmin and found that dosages higher than normally used for disinfection are required for complete removal [52]. They found that UV combined with H2O2 significantly increased the oxidation rates. It was also found that water quality parameters such as turbidity from NOM also influence the removal rates, as the destruction in clearwell water was much higher than with raw river water. The authors were also able to effectively predict removal by UV/ H2O2 through a steady state model.
Removal of MIB/geosmin by AOPs is also dependent on various water quality parameters such as pH and NOM concentrations. The capital and energy costs associated with
143 these technologies can be significantly high, especially for large scale applications. There is also the risk of formation of harmful disinfection by-products through these processes. However these technologies are being used more commonly now and could be retrofitted or optimized for effective removal of these odorants.
5.6.3 Bioremediation
In the last few years there has been an extensive use of biological treatment for removal of certain contaminants in water and wastewater. Unlike wastewater, biological methods have limited application in drinking water and are mainly used with filtration, or biofiltration. Results from various studies dealing with biological removal of MIB/geosmin are discussed with the key results summarized in Table 5.3. Huck et al. were one of the first researchers to study biological removal of odor causing compounds in drinking water [55]. They studied microbial geosmin removal in a lab scale bioreactor. The results demonstrated very low geosmin removal and the authors concluded that biodegradation was not the most effective technology for removal of T/O compounds in drinking water. However subsequent research has shown results to the contrary.
Ho et al. demonstrated removal of MIB and geosmin in a biologically active sand filter
[56]. They used water from a river in Australia known for episodes of significant odor outbreaks due to MIB and geosmin. However the produced water in a treatment plant using conventional treatment is surprisingly free of any MIB/geosmin. The results from this study corroborate the removal of these compounds by biological sand filtration. The authors also determined the pseudo first order reaction kinetics by running batch experiments with the biofilm obtained from one of the sand filters with rates as high as 0.6 day-1. Four different bacteria thought to be
144 responsible for this biodegradation were also identified. In a related study, Hoefel et al. identified three gram-negative bacteria that coordinated the biological degradation of geosmin and interestingly degradation did not occur even if one of the three isolates was absent [57].
Table 5.3 Biological removal of MIB/geosmin
Key References Findings Ho et al. [56] Rapid biodegradation of MIB/geosmin observed in sand filters Reaction rates depended on the initial concentration of the inoculum but not the contaminants Rates increased upon reexposure of the biofilm to the compounds Four strains of bacteria responsible for this biodegradation were identified Elhadi et al. [58, 59] Temperature, media type and initial MIB/geosmin concentrations were factors affecting biofilter performance Higher removals were observed at the higher temperature and when GAC was used as the support media Simulated ozonation resulted in biodegradable organic matter which resulted in higher biomass concentration and eventually better performance The biofilter showed satisfactory performance even with the transient presence of MIB/geosmin
Elhadi et al. investigated MIB/geomin removal in a dual media filter in a bench scale study [59]. They conducted a factorial design experiment to study the effects of different factors
on removal efficiency – temperature, media type, presence of biodegradable organic matter
(BOM) and influent concentrations. Typical ozonation byproducts were used as BOM to
simulate filtration following ozonation. The results showed that removal for both compounds
was higher at the higher temperature of 20 °C when compared to 8 °C. Also removal was better
when GAC-sand was used as the media when compared to the anthracite-sand media. The
removal was also found to be higher at the higher influent concentrations and at higher BOM
concentrations. This was mainly due to higher density of biomass at the higher BOM
concentrations. They also demonstrated removal at very low influent concentrations of 25 ng/L.
145 The optimum levels for each of the factors for maximum MIB/geosmin removal were also listed.
Although the study demonstrated a maximum removal of around 60% at an influent
concentration of 100 ng/L, the resulting effluent concentrations would still be significantly
higher than the OTCs for these compounds, requiring further polishing prior to distribution.
5.6.4 Integrated Technologies
Various studies have also looked into the possibility of combining different technologies
to enhance removal of MIB/geosmin in water. The combination of methods could be used either
as a polishing step or for complete removal of contaminants, which may have not been possible
with just one technology. Elhadi et al. compared MIB/geosmin in fresh GAC/sand biolfilter with
an exhausted GAC/sand filter [58]. The results show that as expected, initial removal was higher
in the fresh GAC/sand filter when compared to the exhausted GAC/sand filter because of better
adsorption. However with time, as a stable biomass developed on both filters, the removal rates
were comparable. Removal rates after a two month run were close to 87% for geosmin and 52%
for MIB. This phenomenon of higher removal of geosmin when compared to MIB has been
observed in other studies as well confirming that geosmin displays better reaction kinetics [7,
38]. The authors also simulated biofilter startup and transient presence of MIB/geosmin by
spiking the influent water accordingly and the biofilter was able to perform consistently under these conditions.
Matsui et al demonstrated geosmin removal with a combination of super-PAC (S-PAC)
and microfiltration (MF) in a small-scale pilot sudy [60]. S-PAC was obtained by pulverization
of PAC to sub-micron particle size. S-PAC with MF showed better efficiency than PAC at a
146 significantly low dosage. The authors claimed that up to 90% savings in dosage and a better
removal could be achieved with S-PAC. However the cost-effectiveness of this process along
with the material loss during the sieving process were not discussed in detail.
Nerenberg et al. demonstrated ozonation followed by biofiltration in a water treatment facility
for removal of MIB/geosmin [32]. They found that ozonation and biofiltration show a synergistic
effect. Ozonation resulted in partial destruction of the odorant and also transformed some of the
nonbiodegradable NOM into smaller compounds that can be used by the bacteria as substrate.
This enhances the ability of the biofilter to completely remove the remaining MIB/geosmin in
the water. Park et al. compared oxidation of MIB/geosmin with ferrate (Fe VI) to ozone and
found that due to low selectivity of the ferrate ion, the removal rates were significantly lower
when compared to ozone [61].
It is clear that some of the integrated technologies like GAC + biofiltration and ozonation
+ biofiltration have shown improved removal of MIB/geosmin and have the potential to be
further developed for implementation in the field.
5.6.5 Other Novel Treatment Methods
The above sections discussed technologies that are conventionally used in water
treatment facilities. However, MIB/geosmin removal in water has been demonstrated by other
novel methods on a bench/laboratory scale. Lawton et al investigated MIB/geosmin removal using titanium dioxide (TiO2) photocatalysis [62]. The results showed rapid degradation of both
MIB/geosmin with more than 99% removal within 60 minutes. The pseudo first order reaction
147 rates were also determined. The authors acknowledge the challenges involved with field
application of this technology and also the uncertainty regarding the process efficiency with the
use of natural water containing NOM. Song et al. investigated MIB/geosmin removal in water by ultrasonic irradiation [63]. When water is subjected to ultrasonic radiation, it is associated with generation of heat, and the resulting pyrolysis causes degradation of the MIB/ geosmin. The
study found that with this method, complete removal could be achieved in a matter of minutes.
The authors also developed degradation kinetics and suggested possible degradation pathways.
Since the method is not based on chemical removal, the presence of hydroxyl radical scavengers
in the water did not influence the MIB/geosmin removal efficiency. Although this technology
might be applicable for the aquaculture industry, there are several reasons that would prohibit its
large-scale application in a drinking water treatment facility.
5.7 Current Status
Although some of the conventional technologies discussed in the previous section are
effective for removal of MIB/geosmin in water, it is not necessary for treatment plants to install
any one technology exclusively for treating taste and odor. The current practice most commonly
followed is application of PAC during severe taste and odor outbreaks. There are several factors
that affect PAC dosage as well. Westerhoff et al. looked at the mechanisms affecting
MIB/geosmin concentrations in water supply reservoirs and found that thermal stratification
significantly influenced the MIB/geosmin concentrations released into the supplied water [19].
The mechanism of water intake from a reservoir could also influence its MIB/geosmin
concentration. Based on the thermal destratification, the amount of MIB/geosmin released and
148 the duration of the episode could be estimated. This data could then be used by the water utility to set the PAC dosage.
Another way of controlling these compounds in water treatment plants is periodic chlorination to prevent algal growth. Although chlorination is effective, its use with PAC results in increased chlorine demand from PAC, reduction in the sorption capacity due to oxidation and has also been found to result in desorption of the adsorbed odorant.
Based on this discussion, it is clear that it would not be economical or practical for water treatment facilities to install a technology exclusively for treatment of MIB/geosmin. It would be preferable to have a conventional technology that is optimized for MIB/geosmin removal. An even better alternative would be a system that has been installed for treating other contaminants and that would be effective for these odorants also during severe taste or odor episodes.
5.8 Conclusion
MIB and geosmin have been identified as the main taste and odor causing compounds in drinking water. Although these two compounds have been not been associated with any serious health affect, the resulting taste and odor from their presence is perceived as unsafe by consumers. The main challenge faced by drinking water utilities in their removal is that they can be detected at extremely low concentrations, down to nanograms per liter. This taste and odor problem is acute in summer months when cyanobacterial blooms are common due to the warmer, nutrient rich conditions. Studies have found that it is extremely difficult to remove these two odorants by conventional water treatment methods. Coagulation, sedimentation and chlorination
149 have been found to be ineffective for their treatment. Currently adsorption by powdered activated
carbon (PAC) is the most effective technology and is routinely used in water treatment plants to
treat these odorants during severe outbreaks. Another factor that adds to the challenge faced by drinking water facilities in the removal of these contaminants is the presence of natural organic matter (NOM). NOM, a complex mixture of organic compounds derived primarily from the decay of plant and animal materials, is invariably present in all water sources and at much higher concentrations than geosmin or MIB. Although PAC is practical for intermittent use, it can result in formation of excessive sludge and increases operating costs. GAC filtration is more effective when the carbon is tailored and also its performance is significantly influenced by the presence of NOM. Competitive adsorption reduces the capacity of the carbon to adsorb MIB/geosmin in the presence of other organic contaminants. Studies have shown that advanced oxidation processes (AOPs) such as ozone and UV/H2O2 are effective for removal of these compounds in
water where the hydroxyl radical is the main mechanism of oxidation. However, the capital and
operating costs associated with these AOPs can be significantly high, especially at the higher dosages required for MIB/geosmin. Also they could result in the formation of disinfection by- products which could be of health or regulatory concerns. Another treatment alternative that has received a lot of attention is biological treatment, specifically biofiltation. Sand filtration is invariably used in most water treatment plants and in most cases these filters support biological activity which enhances filtration. Sand biofiltration has been found to be effective for removal of MIB/geosmin. However complete removal has been observed mainly where biofilter is
supported on GAC or is followed by ozonation. Although some bacteria have been identified for
removal of MIB/geosmin, it would be difficult to introduce these microorganisms into the
treatment system, before knowing if they have any associated health effects. All these major
150 technologies along with their advantages and challenges associated with their application have been discussed in detail in this review. Because taste and odor events due to MIB/geosmin in water are seasonal, their treatment may not be required at all times. Consequently, a treatment system with PAC, where the dosage can be adjusted easily based on the influent odorant concentrations is preferable. A lot of research has been focused into optimizing PAC dosages and trying to resolve some of the challenges faced by utilities in its application. Future work in this area would enable drinking water treatment facilities to implement these technologies with minimal operational issues and ensure clean wholesome water to the consumers.
151 5.9 References
1 McGuire, M. J. (1995) Off-flavor as the consumer's measure of drinking water
safety. Water Science & Technology 31, 1-8
2 Pirbazari, M., Ravindran, V., Badriyha, B. N., Craig, S. and McGuire, M. J.
(1993) GAC adsorber design protocol for the removal of off-flavors. Water
Research 27, 1153-1166
3 (OWWRC), O. W. W. R. C. (07/2008) T&O and Health,
http://www.owwrc.com/TOhealth.htm.
4 Bruce, D., Westerhoff, P. and Brawley-Chesworth, A. (2002) Removal of 2-
methylisoborneol and geosmin in surface water treatment plants in Arizona.
Journal of Water Supply: Research and Technology - AQUA 51, 183-197
5 Ferguson, D. W., McGuire, M. J., Koch, B., Wolfe, R. L. and Aieta, E. (1990)
Comparing peroxone and ozone for controlling taste and odor compound,
disinfection by-products, and microorganisms. Journal American Water Works
Association 82, 181-191
6 Glaze, W. H., Zarnoch, J. J., Ruth, E. C., Chauncey, W. and Schep, R. (1990)
Evaluating Oxidants for the Removal of Model Taste and Odor Compounds from
a Municipal Water Supply. Journal of the American Water Works Association 82,
79-84
7 Cook, D., Newcombe, G. and Sztajnbok, P. (2001) The application of powdered
activated carbon for MIB and geosmin removal: predicting PAC doses in four raw
waters. Water Research 35, 1325-1333
152 8 Ng, C., Losso, J. N., Marshall, W. E. and Rao, R. M. (2002) Freundlich
adsorption isotherms of agricultural by-product-based powdered activated carbons
in a geosmin - water system. Agricultural Wastes 85, 131-135
9 Jung, S. W., Baek, K. H. and Yu, M. J. (2004) Treatment of taste and odor
material by oxidation and adsorption. Water science and technology : a journal of
the International Association on Water Pollution Research and Control 49, 289-
295
10 Ho, L. and Newcombe, G. (2005) Effect of NOM, turbidity and floc size on the
PAC adsorption of MIB during alum coagulation. Water Research 39, 3668-3674
11 Newcombe, G., Morrison, J., Hepplewhite, C. and Knappe, D. R. U. (2002)
Simultaneous adsorption of MIB and NOM onto activated carbon-II. Competitive
effects. Carbon 40, 2147-2156
12 Newcombe, G., Morrison, J. and Hepplewhite, C. (2002) Simultaneous adsorption
of MIB and NOM onto activated carbon. I. Characterisation of the system and
NOM adsorption. Carbon 40, 2135-2146
13 Juttner, F. and Watson, S. B. (2007) Biochemical and Ecological Control of
Geosmin and 2-Methylisoborneol in Source Waters. Applied and Environmental
Microbiology 73, 4395-4406
14 Watson, S. B. (2004) Aquatic Taste and Odor: A Primary Signal of Drinking-
Water Integrity. Journal of Toxicology & Environmental Health: Part A 67, 1779-
1795
153 15 Watson, S. B., Ridal, J. and Boyer, G. L. (2008) Taste and odour and
cyanobacterial toxins: impairment, prediction, and management in the Great
Lakes. Canadian Journal of Fisheries & Aquatic Sciences 65, 1779-1796
16 Zaitlin, B. and Watson, S. B. (2006) Actinomycetes in relation to taste and odour
in drinking water: Myths, tenets and truths. Water Research 40, 1741-1753
17 Watson, S. B., Ridal, J., Zaitlin, B. and Lo, A. (2003) Odours from pulp mill
effluent treatment ponds: the origin of significant levels of geosmin and 2-
methylisoborneol (MIB). Chemosphere 51, 765
18 Lin, T.-F., Wong, J.-Y. and Kao, H.-P. (2002) Correlation of musty odor and 2-
MIB in two drinking water treatment plants in South Taiwan. Science of the Total
Environment 289, 225
19 Westerhoff, P., Rodriguez-Hernandez, M., Baker, L. and Sommerfeld, M. (2005)
Seasonal occurrence and degradation of 2-methylisoborneol in water supply
reservoirs. Water Research 39, 4899-4912
20 Lloyd, S. W., Lea, J. M., Zimba, P. V. and Grimm, C. C. (1998) Rapid analysis of
geosmin and 2-methylisoborneol in water using solid phase micro extraction
procedures. Water Research 32, 2140-2146
21 Zander, A. K. and Pingert, P. (1997) Membrane-based extraction for detection of
tastes and odors in water. Water Research 31, 301-309
22 Eisert, R. and Levsen, K. (1996) Solid-phase microextraction coupled to gas
chromatography: a new method for the analysis of organics in water. Journal of
Chromatography A 733, 143-157
154 23 Zimmerman, L. R., Ziegler, A. C. and Thurman, E. M. (2002) Method of Analysis
and Quality-Assurance Practices by U.S. Geological Survey Organic
Geochemistry Research Group--Determination of Geosmin and Methylisoborneol
in Water Using Solid-Phase Microextraction and Gas Chromatography/Mass
Spectrometry. 1-12
24 Saito, K., Okamura, K. and Kataoka, H. (2008) Determination of musty odorants,
2-methylisoborneol and geosmin, in environmental water by headspace solid-
phase microextraction and gas chromatography–mass spectrometry. Journal of
Chromatography A 1186, 434-437
25 Chang, J., Biniakewitz, R. and Harkey, G. (2008) Determination of Geosmin and
2-MIB in Drinking Water by SPME-PTV-GC-MS. In Environmental: The
applications book, pp. 1-3
26 Dionigi, C. P., Lawlor, T. E., McFarland, J. E. and Johnsen, P. B. (1993)
Evaluation of geosmin and 2-methylisoborneol on the histidine dependence of
TA98 and TA100 Salmonella typhimurium tester strains Water Research 27,
1615-1618
27 Robedson, R. F., Hammond, A., Jauncey, K., Beveridge, M. C. M. and Lawton,
L. A. (2006) An investigation into the occurrence of geosmin responsible for
earthy-musty taints in UK farmed rainbow trout, Onchorhynchus mykiss.
Aquaculture 259, 153-163
28 Schulz, S., Fuhlendorff, J. and Reichenbach, H. (2004) Identification and
synthesis of volatiles released by the myxobacterium Chondromyces crocatus.
Tetrahedron 60, 3863
155 29 Watson (2000) Quantitative analysis of trace levels of geosmin and MIB in source
and drinking water using headspace SPME. Water Research 34, 2818-2828
30 Whelton, A. J. and Dietrich, A. M. (2004) Relationship between intensity,
concentration, and temperature for drinking water odorants. Water Research 38,
1604-1614
31 Lalezary, S., Pirbazari, M. and McGuire, M. J. (1986) Oxidation of five earthy-
musty taste and odor compounds. Journal American Water Works Association 78,
62-69
32 Nerenberg, R., Rittmann, B. E. and Soucie, W. J. (2000) Ozone/biofiltration for
removing MIB and geosmin. Journal American Water Works Association 92, 85-
97
33 Ridal, J., Brownlee, B., McKenna, G. and Levac, N. (2001) Removal of Taste and
Odour Compounds by Conventional Granular Activated Carbon Filtration Water
Qual. Res. J. Canada 36, 43-54
34 MacKenzie, J. A., Tennant, M. F. and Mazyck, D. W. (2005) Tailored GAC for
the Effective Control of 2-Methylisoborneol. Journal American Water Works
Association 97, 76-88
35 Rangel-Mendez, J. R. and Cannon, F. S. (2005) Improved activated carbon by
thermal treatment in methane and steam: Physicochemical influences on MIB
sorption capacity. Carbon 43, 467-479
36 Ng, C., Losso, J. N., Marshall, W. E. and Rao, R. M. (2002) Physical and
chemical properties of selected agricultural byproduct-based activated carbons
and their ability to adsorb geosmin. Agricultural Wastes 84, 177-185
156 37 Yu, J., Yang, M., Lin, T.-F., Guo, Z., Zhang, Y., Gu, J. and Zhang, S. (2007)
Effects of surface characteristics of activated carbon on the adsorption of 2-
methylisobornel (MIB) and geosmin from natural water. Separation and
Purification Technology 56, 363-370
38 Cook, D. and Newcombe, G. (2004) Can we predict the removal of MIB and
geosmin with PAC by using water quality parameters? Water Science and
Technology: Water Supply 4 221-226
39 Lu, Q. and Sorial, G. A. (2004) The role of adsorbent pore size distribution in
multicomponent adsorption on activated carbon. Carbon 42, 3133-3142
40 Pelekani, C. and Snoeyink, V. L. (2000) Competitive adsorption between atrazine
and methylene blue on activated carbon: the importance of pore size distribution.
Carbon 38, 1423-1436
41 Pelekani, C. and Snoeyink, V. L. (1999) Competitive adsorption in natural water:
Role of activated carbon pore size. Water Research 33, 1209-1219
42 Pelekani, C. and Snoeyink, V. L. (2001) A kinetic and equilibrium study of
competitive adsorption between atrazine and Congo red dye on activated carbon:
the importance of pore size distribution. Carbon 39, 25-37
43 Lu, Q. and Sorial, G. A. (2004) Adsorption of phenolics on activated carbon--
impact of pore size and molecular oxygen. Chemosphere 55, 671-679
44 Lu, Q. and Sorial, G. A. (2007) The effect of functional groups on
oligomerization of phenolics on activated carbon. Journal of Hazardous Materials
148, 436-445
157 45 Lu, Q. and Sorial, G. A. (2005) Impact of pore size on competitive adsorption of
phenolic compounds. Water Science and Technology: Water Supply 4, 1-7
46 Srinivasan, R., Sorial, G. A., Ononye, G., Husting, C. and Jackson, E. (2008)
Elimination of persistent odorous compounds in drinking water. Water Science
and Technology: Water Supply 8, 121-127
47 Srinivasan, R. and Sorial, G. A. (2009) Adsorption of Geosmin and MIB on
Activated Carbon - Single and Binary Solute Systems. Water, Air and Soil
Pollution: Focus, submitted for publication
48 Collivignarelli, C. and Sorlini, S. (2004) AOPs with ozone and UV radiation in
drinking water: contaminants removal and effects on disinfection byproducts
formation. Water Science and Technology 49, 51-56
49 Peter, A. and Gunten, U. V. (2007) Oxidation Kinetics of Selected Taste and Odor
Compounds During Ozonation of Drinking Water. Environmental Science &
Technology 41, 626-631
50 Ho, L., Newcombe, G., Crou, eacute and Jean-Philippe (2002) Influence of the
character of NOM on the ozonation of MIB and geosmin. Water Research 36,
511-518
51 Park, G., Yu, M., Koo, J.-Y., Joe, W. H. and Kim, H. (2006) Oxidation of
geosmin and MIB in water using O3/H2O2: kinetic evaluation. Water Science and
Technology: Water Supply 6 63-69
52 Rosenfeldt, E. J., Melcher, B. and Linden, K. G. (2005) UV and UV/H2O2
treatment of methylisoborneol (MIB) and geosmin in water. Journal of Water
Supply: Research and Technology—AQUA 54, 423-434
158 53 Westerhoff, P., Nalinakumari, B. and Pei, P. (2006) Kinetics of MIB and
Geosmin Oxidation during Ozonation. Ozone: Science & Engineering 28, 277-
286
54 Liang, C., Wang, D., Chen, J., Zhu, L. and Yang, M. (2007) Kinetics Analysis on
the Ozonation of MIB and Geosmin. Ozone: Science & Engineering 29, 185-189
55 Huck, P. M., Kenefick, S. L., Hrudey, S. E. and Zhang, S. (1995) Bench Scale
Determination of the Removal of Odour Compounds with Biological Treatment.
Water Science and Technology 31, 203-209
56 Ho, L., Hoefel, D., Bock, F., Saint, C. P. and Newcombe, G. (2007)
Biodegradation rates of 2-methylisoborneol (MIB) and geosmin through sand
filters and in bioreactors. Chemosphere 66, 2210-2218
57 Hoefel, D., Ho, L., Aunkofer, W., Monis, P. T., Keegan, A., Newcombe, G. and
Saint, C. P. (2006) Cooperative biodegradation of geosmin by a consortium
comprising three gram-negative bacteria isolated from the biofilm of a sand filter
column. Letters in Applied Microbiology 43, 417-423
58 Elhadi, S. L. N., Huck, P. M. and Slawson, R. M. (2004) Removal of geosmin and
2-methylisoborneol by biological filtration. Water Science and Technology 49,
273-280
59 Elhadi, S. L. N., Huck, P. M. and Slawson, R. M. (2006) Factors affecting the
removal of geosmin and MIB in drinking water biofilters. Journal American
Water Works Association 98, 108-120
60 Matsui, Y., Aizawa, T., Kanda, F., Nigorikawa, N., Mima, S. and Kawase, Y.
(2007) Adsorptive removal of geosmin by ceramic membrane filtration with
159 super-powdered activated carbon. Journal of Water Supply: Research and
Technology—AQUA 56.6, 411-418
61 Park, G., Yu, M., Go, J., Kim, E. and Kim, H. (2007) Comparison between ozone
and ferrate in oxidising geosmin and 2-MIB in water. Water Science &
Technology 55, 117-125
62 Lawton, L. A., Robertson, P. K. J., Robertson, R. F. and Bruce, F. G. (2003) The
destruction of 2-methylisoborneol and geosmin using titanium dioxide
photocatalysis. Applied Catalysis B: Environmental 44, 9-13
63 Song, W., rsquo and Shea, K. E. (2007) Ultrasonically induced degradation of 2-
methylisoborneol and geosmin. Water Research 41, 2672-2678
160 CHAPTER 6
Single Solute Adsorption of MIB and Geosmin on Activated Carbon
Fibers
6.1 Abstract
MIB (2-methyl isoborneol - C11H20O) and Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) are the major taste and odor causing compounds in drinking water. The
objective of this study was to investigate the adsorption of these odor causing compounds
by three types of activated carbon fibers (ACF), namely ACC-25, ACC-20 and ACC-15
with different micropore size distributions. Adsorption for both MIB and geosmin was well defined by Freundlich isotherms. For MIB, the adsorption capacity increased with increase in micropore volume of the adsorbents whereas in the case of geosmin, the adsorption was much more complex. Isotherms were also conducted for ACF adsorption for MIB and geosmin in the presence of natural organic matter (NOM) to better understand competitive adsorption. Overall the ACF ACC-25 was the most effective adsorbent for both MIB and geosmin and was least affected by the competitive effect of the NOM because of its wider distribution of micropores. Adsorption isotherms were also conducted with a granular activated carbon F-400 to compare adsorption with the ACFs.
F-400 was found to have significantly lower adsorption capacity for both MIB and geosmin when compared to the ACFs and also presence of NOM significantly reduced its adsorption capacity due to competitive adsorption.
161 6.2 Introduction
It is not uncommon for water utilities to be flooded with complaints from consumers about taste and odor in their drinking water, especially during warmer weather. For an average consumer, off-flavors are the only way of determining the safety of tap water [1]. As a result, drinking water utilities throughout the world are facing the necessity for more innovative and cost-effective technologies for removal of taste and odor removal in water treatment and purification. Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) and 2-methyl isoborneol–C11H20O (MIB) are the major taste-and-odor-
causing compounds in drinking water obtained from surface water and are associated
with earthy and musty odors [2]. Geosmin and MIB in surface water mainly result from
the biodegradation of certain types of cyanobacteria that normally bloom in the presence
of nutrients at warmer temperatures. There are currently no regulations for these two
compounds as they have not been associated with any health effects [1]. The main problem of the presence of geosmin, 2-MIB and other odor-causing substances is associated with their extremely low odor threshold concentrations (OTC) and their persistence to elimination in a conventional water treatment process such as coagulation,
sedimentation, filtration and chlorination. The OTC for geosmin and MIB is 4 ng/L and
9 ng/L, respectively [3]. Another factor that adds to this challenge faced by drinking
water facilities, when it comes to removing taste and odor compounds, is the presence of
natural organic matter (NOM). NOM, a complex mixture of organic compounds derived
primarily from the decay of plant and animal materials, is invariably present in all water
sources and at much higher concentrations than geosmin or MIB.
162 Adsorption by granular activated carbon (GAC) is considered as one of the best available
technologies for removal of organic contaminants from water. Numerous studies have
looked into GAC and PAC (powdered activated carbon) adsorption of MIB and Geosmin
[3-7]. However, NOM levels of 3–10 mg/L competitively reduce activated carbon
adsorption capacity for MIB or Geosmin [3, 8, 9]. Compared with GAC, activated
carbon fibers (ACFs) have attracted increasing attention due to their excellent surface
properties, high adsorption capacity and are an ideal adsorbent for targeting the impact of
pore size [10]. The low ash and surface oxide content and controlled pore structure are
key advantages of these types of adsorbents over GAC. Pelekani and Snoeyink looked at
activated carbon fiber adsorption of an organic micropollutant atrazine and the
mechanism of competitive adsorption between atrazine and another compound much
larger in size.
This study aims to study the adsorption of two taste and odor causing compounds
geosmin and MIB on a series of activated carbon fibers with different micropore size distributions. In order to better understand the effect of pore size characteristics on adsorption, activated carbon fiber adsorption was compared to adsorption with a commercially available granular activated carbon. Experiments were also conducted to look at the effect of natural organic matter on geosmin and MIB adsorption on activated carbon fibers.
163 6.3 Materials and Methods
6.3.1 Adsorbates
Geosmin (Sigma Aldrich, St Louis, MO) was obtained as a solution with concentration of
2 mg/mL in methanol. MIB (Sigma Aldrich, St Louis, MO) was also obtained as a
solution with concentration of 10 mg/mL in methanol. Humic acid (Sigma Aldrich, St
Louis, MO) was used as a representative of NOM in powder form with purity over 99%.
6.3.2 Adsorbents
Four microporous phenolic resin-based activated carbon fibers (ACFs) with increasing
degrees of activation designated as ACC-15, ACC-20, and ACC-25 (Nippon Kynol,
Japan) were used in the study. They were received as twilled-weave fabrics. F400
(Filtrasorb 400, Calgon, Pittsburgh, PA), which is a bituminous base activated carbon, was chosen as a typical GAC. All the ACFs and F-400 used in the study were dried in an oven at 105 oC overnight, and then stored in a desiccator until use. The properties of
these are listed in Table 6.1.
6.3.3 Adsorption Isotherm Procedure
The bottle point method was used for conducting the adsorption isotherms at room
temperature. Various initial concentrations of adsorbate, ranging from 50µg/l to 1000
µg/l were used. An initial concentration of 10 mg/L humic acid was used for adsorbate
solutions containing NOM. Each set of bottles was accompanied by two blanks to check
164 for any volatilization, adsorption onto the walls of the bottles, and biodegradation of adsorbate. Accurately weighed (±0.1 mg) masses of activated carbon fibers or GAC were placed in 250-ml glass-amber bottles. The bottles were tightly sealed with Teflon lined caps after purging. Adsorbate solutions were prepared in autoclaved de-ionized water buffered with potassium phosphate (KH2PO4) and the pH adjusted to 7.0 by adding sodium hydroxide. The glass amber bottles containing the different masses of carbon were then completely filled with the adsorbate solution sealed tightly with Teflon lined caps and covered by parafilm.
Table 6.1 Physical properties of the adsorbents used in the study [10]
Property F-400 ACC-15 ACC-20 ACC-25 BET surface area 993 1137 1455.6 2014.0 (m2 g-1)a Micropore area 893.0 1089.7 1379.0 1919.6 (m2 g-1)b Mesopore area 100.0 47.3 76.6 94.4 (m2 g-1)c Micropore volume 0.373 0.560 0.705 0.767 (cm3 g-1)b Critical pore diameter (Å) 4-800 12.8 17.4 19.7 Total pore volume 0.615 0.598 0.766 0.842 (cm3 g-1)d Fraction microporosity (%)e 60.7 93.6 92.0 91.0
a P/P0 ranges from 0.03 to 0.111
b.Applying BJH model
c.Mesopore area = BET surface area – Micropore area d.Single point desorption total pore volume of pores less than 2526.8 Å at P/P0 = 0.992 Micropore Volume e. Microporosity = ×100% PoreTotal Volume
165 The isotherm bottles were then placed in a rotary shaker and allowed to
equilibriate for eight days. After equilibration, all samples were filtered through 0.45-µm
nylon filters (Micron Separation, Inc.) prior to analysis in order to minimize interferences
of the carbon fines with analysis. The first 5 ml of the filtered samples were discarded
before samples were taken for analysis in order to minimize the impact of potential
adsorption of the adsorbate on the filter membrane.
6.3.4 Analytical procedure
Analysis of MIB and geosmin were carried using solid-phase microextraction/gas
chromatography (SPME-GC) (Lloyd et al. 1998; Watson et al. 2000). For Geosmin
analysis, 20 ml of sample was added to a 40 ml septum capped vial containing 6 g
sodium chloride and a magnetic stir bar. In case of MIB, 25 ml of sample was used with
7.5 g of sodium chloride. The vials were then placed in a water bath on a magnetic stir
plate heated to 65 ± 1.5 °C. A SPME fiber (Supelco, Bellefonte, PA) was introduced into the head space gas through the septum and the sample stirred for 30 min. The fiber was
removed from the vial and inserted into the GC (Agilent 6890 Series; Agilent
Technologies, Wilmington, DE) inlet port and left exposed for 10 minutes for complete desorption of the target compounds. The GC was equipped with Agilent DB-5 column
(30 m x 0.32 mm inside diameter x 0.25 µm film thickness) and flame ionization detector
(FID). The injection port and detector temperature were set at 250 °C. The GC oven temperature was maintained at 60 °C for the first 4 min, then ramped to 270 °C at 10
°C/min and kept at 270 °C for 1 min. The flow rate of the carrier gas (N2) was set at 1.5
ml/min. The detector makeup gas (N2) flow rate was set at 43.5 ml/min. The flame gases
166 hydrogen and air flow rates were set at 40 and 450 ml/min, respectively. The retention time for geosmin was 14.1 min and for MIB, 11.0 min. The fiber was then retracted into the holder, removed from the GC inlet and reused for the next sample.
6.4 Results and Discussion
Figure 6.1a compares the seven day single solute MIB adsorption isotherms for
the three ACFs, ACC-25, ACC-20, ACC-15 and the GAC F-400 in organic free water
(OFW). Adsorption in the entire concentration range (1 to 1000µg/L) was well defined
by the Freundlich isotherm, parameters for which are specified in Table 6.2. It is clearly
seen from the figure that the MIB adsorption capacity increases in the order of increasing
micropore volume from F-400 to ACC-25. The adsorption capacities of ACC-25 and
ACC-20 seem to very similar and significantly higher than ACC-15 and F-400. This is
reflected in the k values (Table 6.2) as well, where the values decrease as we go from
ACC-25 to F-400. Figure 6.1b shows the adsorption isotherms for MIB in the presence of
10mg/L humic acid. As in the case with organic free water, adsorption capacity for MIB
decreases with reduction in the micropore volume.
167 Table 6.2. Freundlich constants for MIB and geosmin adsorption (OFW: organic free water; HA solution: water with 10mg/L HA conc.)
Geosmin MIB
OFW HA solution OFW HA solution K 1/n K 1/n K 1/n K 1/n F-400 1434 0.6254 73 0.9357 825 0.6224 13 1.1406 ACC-15 4979 0.7388 512 0.6669 947 0.7145 89 1.0564 ACC-20 14099 0.2242 9032 0.3264 3854 0.58 1149 0.714 ACC-25 20100 0.3671 10122 0.3917 5183 0.5591 4216 0.5632
Figure 6.2 shows the effect of humic acid on MIB adsorption for all four adsorbents.
From the figure, it can be seen that humic acid reduces MIB uptake in all the four adsorbents. However the effect is least pronounced in ACC-25 and most pronounced in
F-400 with the remaining two adsorbents in the middle. This could be attributed to the difference in the pore size characteristics of these adsorbents. MIB has a molecular weight of 182.3 g/mol whereas the molecular weight of humic acid is estimated to be between 2000 to 3000 g/mol [11], thereby resulting in a ten time bigger particle size. The average micropore size in both ACC-15
168 ACC-20 1e+5 ACC-25 ACC-15 F400 g/g) μ ( e 1e+4 Q
1e+3 a. Organic Free Water
ACC-25 1e+5 ACC-20 ACC-15 F400 g/g) μ
( 1e+4 e Q
1e+3 b. Humic acid solution
1 10 100 1000 C (μg/L) e
Figure 6.1. MIB adsorption isotherms in organic free water and humic acid
solutions
169 Organic free water Organic free water 1e+5 Humic acid solution Humic acid solution
g/g) 1e+4 μ ( e Q
ACC - 25 ACC - 20 1e+3
Organic free water Organic free water Humic acid solution 1e+5 Humic acid solution
1e+4 g/g) μ ( e Q
ACC - 15 F 400
1e+3
1 10 100 10001 10 100 1000 Ce (μg/L) Ce (μg/L)
Figure 6.2 Comparison of MIB adsorption isotherms
and ACC-20 is around 7.8 Å [12], which are too small for the humic acid to access resulting in pore blockage. If it is only complete pore blockage, there should be no significant change in the slopes resulting in parallel isotherms, as in the case of ACC-20.
However there is a change in the slopes of the isotherms for ACC-15 and this could be a result of micropore constriction by humic acid particles. In case of ACC-25, humic acid does not reduce the adsorption capacity for MIB. In addition to a primary micropore distribution in the 7-9 Å range, ACC-25 has a small volume of secondary micropores in
170 the 14-20 Å range. It is in these larger micropores that the humic acid particles adsorb without blocking the smaller micropores and
ACC-20 1e+5 ACC-25 ACC-15 F400 g/g) μ ( e
Q 1e+4
a. Organic Free Water
1e+3
ACC-25 1e+5 ACC-20 ACC-15 F400 g/g) μ ( e 1e+4 Q
b. Humic acid solution 1e+3 0.1 1 10 100 1000 C (μg/L) e
Figure 6.3 Geosmin adsorption isotherms in organic free water and humic acid
solutions
171
thereby allowing access to the smaller MIB particles. On the contrary, F-400 is significantly impacted by the competitive adsorption between MIB and humic acid. F-
400 possesses a wide range of micropores where the humic acid particles present at a much higher concentration adsorb and compete with the MIB particles.
Figure 6.3a compares the seven day single solute geosmin adsorption isotherms for the three ACFs, ACC-25, ACC-20, ACC-15 and the GAC F-400. When compared to
MIB, geosmin adsorption on activated carbon seems to be a bit more complicated. As in the case of MIB, ACC-25 seems to have the highest adsorption capacity (Figure 6.3a).
However with ACC-15, significantly higher adsorption is observed in the 10-100 µg/L residual concentration range. In the same concentration range, ACC-20 displays a much lower adsorption capacity. Presently the reason for this is unexplained and further research is required to fully understand this behavior. As expected, F-400 adsorption is lower when compared to the ACFs. Figure 6.3b shows the isotherms for geosmin adsorption in the presence on humic acid. The order of adsorption in this case is ACC-
25>ACC-20> F-400>ACC-15. It is interesting to note that the adsorption with F-400 is in fact better than that obtained with ACC-15. Figure 6.4 shows the effect of humic acid on geosmin adsorption for all four adsorbents. It is clearly seen that the presence of humic acid reduces geosmin adsorption in all four adsorbents. With ACC-25, humic acid resulted in a small reduction in geosmin adsorption. For ACC-20, the reduction is lowered for residual concentrations below 100µg/L. Since the slopes for the two isotherms are different in case of ACC-20, the reduction in geosmin adsorption capacity
172 could be a result of a combination of pore blockage and micropore constriction. However in the ACC-15, there is a considerable reduction in Geosmin capacity resulting from significant surface pore blockage for the same reasons that reduced MIB adsorption in the presence of humic acid. This effect is less pronounced for ACC-25, where the effect of pore blockage is substantially reduced due to a wider distribution of micropores. From the isotherms for F-400, it is clear that there is a significant competitive effect between geosmin and humic acid for the available adsorption sites.
Organic free water Organic free water 1e+5 Humic acid solution Humic acid solution g/g) μ
( 1e+4 e Q
ACC - 25 ACC - 20
1e+3
Organic free water Organic free water 1e+5 Humic acid solution Humic acid solution g/g) μ
( 1e+4 e Q
ACC - 15 F 400
1e+3 0.1 1 10 100 10000.1 1 10 100 1000 C (μg/L) e Ce (μg/L)
Figure 6.4. Comparison of Geosmin adsorption isotherms
173 From Table 6.2, it is seen that for both MIB and geosmin adsorption, the value of the
Freundlich k parameter increases with increasing micropore volume of the adsorbent.
This trend holds true even in the presence of humic acid. This proves that the pore size
distribution is the most important factor influencing adsorption of these compounds on
activated carbon. Also the k values, which are significantly higher for geosmin compared
to MIB, confirm that geosmin showed better adsorption. [4] also found that geosmin was
more amenable to adsorption that MIB and attributed this to the slightly lower molecular
weight, lower solubility and flatter structure of geosmin.
6.5 Conclusion
Activated carbon adsorption of two odor causing compounds namely geosmin and
MIB was studied using three types of activated carbon fibers (ACF), namely ACC-25,
ACC-20 and ACC-15 with different micropore size distributions and one granular
activated carbon, F-400. Adsorption for both MIB and geosmin was well defined by
Freundlich isotherms. The adsorption capacity for MIB increased with increase in
micropore volume of the adsorbents i.e. in the order ACC-25 > ACC-20 > ACC-15 > F-
400. Humic acid, which was used to represent NOM affected MIB adsorption on all the
adsorbents. However this effect was least pronounced for ACC-25 owing to a significant
volume of secondary micropores and most pronounced for ACC-15 due to surface pore blockage. Among the four adsorbents F-400 had the smallest capacity for MIB adsorption and was also significantly impacted by competitive adsorption of the humic acid. In the case of geosmin adsorption, ACC-25 was the most effective adsorbent and its adsorption capacity was significantly higher when compared to F-400. As in the case of MIB,
174 presence of humic acid reduced the adsorption capacity for all the adsorbents. This was
especially significant for ACC-15 and F-400. Overall the ACF ACC-25 was the most
effective adsorbent for both MIB and geosmin and was least affected by the competitive effect of the NOM because of its wider distribution of micropores. All the adsorbents seem to have a higher capacity for adsorption of geosmin when compared to MIB. The results confirmed that the pore size characteristics, especially the distribution of micropores, were the most important factor influencing the activated carbon adsorption of these two odor causing compounds.
6.6 References
1 McGuire, M. J. (1995) Off-flavor as the consumer's measure of drinking water
safety. Water Science & Technology 31, 1-8
2 Bruce, D., Westerhoff, P. and Brawley-Chesworth, A. (2002) Removal of 2-
methylisoborneol and geosmin in surface water treatment plants in Arizona.
Journal of Water Supply: Research and Technology - AQUA 51, 183-197
3 Pirbazari, M., Ravindran, V., Badriyha, B. N., Craig, S. and McGuire, M. J.
(1993) GAC adsorber design protocol for the removal of off-flavors. Water
Research 27, 1153-1166
4 Cook, D., Newcombe, G. and Sztajnbok, P. (2001) The application of powdered
activated carbon for MIB and geosmin removal: predicting PAC doses in four raw
waters. Water Research 35, 1325-1333
175 5 Ng, C., Losso, J. N., Marshall, W. E. and Rao, R. M. (2002) Freundlich
adsorption isotherms of agricultural by-product-based powdered activated carbons
in a geosmin - water system. Agricultural Wastes 85, 131-135
6 Jung, S. W., Baek, K. H. and Yu, M. J. (2004) Treatment of taste and odor
material by oxidation and adsorption. Water science and technology : a journal of
the International Association on Water Pollution Research and Control 49, 289-
295
7 Ho, L. and Newcombe, G. (2005) Effect of NOM, turbidity and floc size on the
PAC adsorption of MIB during alum coagulation. Water Research 39, 3668-3674
8 Newcombe, G., Morrison, J., Hepplewhite, C. and Knappe, D. R. U. (2002)
Simultaneous adsorption of MIB and NOM onto activated carbon-II. Competitive
effects. Carbon 40, 2147-2156
9 Newcombe, G., Morrison, J. and Hepplewhite, C. (2002) Simultaneous adsorption
of MIB and NOM onto activated carbon. I. Characterisation of the system and
NOM adsorption. Carbon 40, 2135-2146
10 Lu, Q. and Sorial, G. A. (2004) Adsorption of phenolics on activated carbon--
impact of pore size and molecular oxygen. Chemosphere 55, 671-679
11 Matthews, J. C. (1995) Nature and extent of the interactions of humic acids with a
water treatment algicide and a fungicide. Chemosphere 30, 1565-1572
12 Pelekani, C. and Snoeyink, V. L. (2001) A kinetic and equilibrium study of
competitive adsorption between atrazine and Congo red dye on activated carbon:
the importance of pore size distribution. Carbon 39, 25-37
176
CHAPTER 7
Adsorption of Geosmin and MIB on Activated Carbon – Single and
binary solute system
7.1 Abstract
The adsorption of two taste and odor causing compounds, namely MIB (2-methyl isoborneol) and geosmin on activated carbon was investigated in this study. The impact
of adsorbent pore size distribution on adsorption of MIB and geosmin was evaluated through single solute and multicomponent adsorption of these compounds on three types of activated carbon fibers (ACFs) and one granular activated carbon (GAC). The ACFs
(ACC-15, ACC-20 and ACC-25) with different degrees of activation had narrow pore size distributions and specific critical pore diameters whereas the GAC (F-400) had a wider pore size distribution and lesser microporosity. The effect of the presence of natural organic matter (NOM) on MIB and geosmin adsorption was also studied for both the single solute and binary systems. Myers equation was used to evaluate the single solute isotherms as it converges to Henry’s law at low coverage and also serves as an input for predicting multicomponent adsorption. The single solute adsorption isotherms fit Myers equation well and pore size distribution significantly influenced adsorption on the ACFs and GAC. The ideal adsorbed solute theory (IAST), which is a well established thermodynamic model for multicomponent adsorption, was used to predict the binary
adsorption of MIB and geosmin. The IAST predicted well the binary adsorption on the
177 ACFs and GAC. Binary adsorption isotherms were also conducted in the presence of
oxygen (oxic) and absence of oxygen (anoxic). There were no significant differences in
the binary isotherm between the oxic and anoxic conditions, indicating that adsorption
was purely through physical adsorption and no oligomerization was taking place. Binary
adsorptions for the four adsorbents were also conducted in the presence of humic acid to
determine the effect of NOM and to compare with IAST predictions. The presence of
NOM interestingly resulted in deviation from IAST behavior in case of two adsorbents,
ACC-15 and F-400.
7.2 Introduction
It is not uncommon for water utilities to be flooded with complaints from consumers about taste and odor in their drinking water, especially during warmer weather. For an average consumer, off-flavors are the only way of determining the safety of tap water [1]. As a result, drinking water utilities throughout the world are facing the necessity for more innovative and cost-effective technologies for removal of taste and odor removal in water treatment and purification. Two compounds that have been identified as major taste-and-odor-causing compounds in drinking water obtained from surface water are Geosmin (trans-1, 10-dimethyl-trans-9 decalol-C12H22O) and 2-methyl
isoborneol–C11H20O (MIB) [2]. Geosmin and MIB in surface water mainly result from
the biodegradation of certain types of cyanobacteria that normally bloom in the presence
of nutrients at warmer temperatures. There are currently no regulations for these two
compounds as they have not been associated with any health effects [3]. The main challenge faced by the utilities in treatment of geosmin, 2-MIB and other odor-causing
178 substances is associated with their extremely low odor threshold concentrations (OTC).
Also, they are extremely persistent to elimination by conventional water treatment processes such as coagulation, sedimentation, filtration and chlorination. The OTC for geosmin and MIB is 4 ng/L and 9 ng/L, respectively [4]. Another factor that adds to the challenge faced by drinking water facilities in the removal of these contaminants is the presence of natural organic matter (NOM). NOM, a complex mixture of organic compounds derived primarily from the decay of plant and animal materials, is invariably present in all water sources and at much higher concentrations than geosmin or MIB [3].
The only treatment methods that have been successfully employed by water treatment plants to remove MIB and Geosmin are adsorption by activated carbon or oxidation by strong oxidants such as ozone. Ferguson et al.[5], Glaze et al. [6] and Bruce et al. [2] studied and demonstrated MIB and Geosmin removal using oxidants such as ozone, hydrogen peroxide and UV. Addition of chemicals however is expensive and can result in formation of disinfection byproducts (DBPs), which are unacceptable due to health and regulatory concerns. Studies have shown that conventional water treatment processes such as coagulation, sedimentation and filtration are unable to achieve any significant removal of MIB and Geosmin [2]. Adsorption by activated carbon, either granular activated carbon (GAC) or powdered activated carbon (PAC) is considered as one of the best available technologies for removal of organic contaminants from water.
Numerous studies have looked into GAC and PAC adsorption of MIB and Geosmin [4,
7-10]. However, NOM levels of 3–10 mg/L competitively reduce activated carbon adsorption capacity for MIB or Geosmin [4, 11, 12]. In natural waters, the size and
179 concentration of NOM particles is many folds higher than that of MIB or Geosmin, and as a result a large volume of the GAC is not utilized for MIB/Geosmin adsorption, thereby significantly reducing the GAC adsorption capacity. Several studies have looked at competitive adsorption of a micropollutant on activated carbon in the presence of
NOM [11-14]. Compared with GAC, activated carbon fibers (ACFs) have attracted increasing attention due to their excellent surface properties, high adsorption capacity and are an ideal adsorbent for targeting the impact of pore size [13, 15].
This study aims to understand the adsorption of two taste and odor causing
compounds, geosmin and MIB on a series of ACFs with different micropore size distributions. In order to study the impact of adsorbent pore size distributions, three different ACFs with different pore sizes were selected. For comparison basis, one bituminous base GAC was also used in the study. Isotherm experiments were conducted in the presence and absence of NOM in order to study the impact of competitive adsorption. The investigations were conducted for both single and binary solute systems.
Binary solute adsorption isotherms were conducted to simulate practical usage. In practice, fixed beds of activated carbon are invariably used to treat water/wastewater by adsorption of more than one contaminant. It is also necessary to note that single solute activated carbon adsorption data could be used to predict multicomponent adsorption. In order to predict multicomponent adsorption, ideal adsorbed solute theory (IAST) model was utilized. Using IAST, single solute parameters were used to generate predictions for the binary solute system and the experimental data was compared with model predictions.
180 7.3 Experimental
7.3.1 Materials
7.3.1.1 Adsorbates
Geosmin (Sigma Aldrich, St Louis, MO) was obtained as a solution with
concentration of 2 mg/mL in methanol. MIB (Sigma Aldrich, St Louis, MO) was also
obtained as a solution with concentration of 10 mg/mL in methanol. Humic acid (Sigma
Aldrich, St Louis, MO) was used as a representative of NOM in powder form with purity
over 99%.
7.3.1.2 Adsorbents
Three microporous phenolic resin-based activated carbon fibers (ACFs) with
increasing degrees of activation designated as ACC-15, ACC-20, and ACC-25 (Nippon
Kynol, Japan) were used in the study. They were received as twilled-weave fabrics. F400
(Filtrasorb 400, Calgon, Pittsburgh, PA), which is a bituminous base activated carbon, was chosen as a typical GAC. All the ACFs and F-400 used in the study were dried in an oven at 105 °C overnight, and then stored in a desiccator until use.
7.3.2 Methods
7.3.2.1 Adsorption Isotherm Procedure
The bottle point method was used for conducting the adsorption isotherms at
room temperature. An initial adsorbate concentration of 200 µg/l was used for Geosmin
181 and MIB. An initial concentration of 10 mg/L humic acid was used for adsorbate
solutions containing NOM. Each set of bottles was accompanied by two blanks to check
for any volatilization, adsorption onto the walls of the bottles, and biodegradation of
adsorbate. For the anoxic isotherm, dissolved oxygen (DO) was removed from the
isotherm bottles and adsorbents and was maintained under 1 mg/L. Accurately weighed
(±0.1 mg) masses of ACFs or GAC were placed in 250-ml glass-amber bottles.
Adsorbate solutions were prepared in autoclaved de-ionized water buffered with
potassium phosphate (KH2PO4) and the pH adjusted to 7.0 by adding sodium hydroxide.
The glass amber bottles containing the different masses of carbon were then completely
filled with the adsorbate solution sealed tightly with Teflon lined caps and covered by
parafilm. The isotherm bottles were then placed in a rotary shaker and allowed to
equilibrate for eight days. After equilibration, all samples were filtered through 0.45-µm
nylon filters (Micron Separation, Inc.) prior to analysis in order to minimize interferences
of the carbon fines with analysis. The first 5 ml of the filtered samples were discarded
before samples were taken for analysis in order to minimize the impact of potential
adsorption of the adsorbate on the filter membrane.
7.3.3 Analytical procedure
Analysis of MIB and geosmin were carried using solid-phase microextraction/gas
chromatography (SPME-GC) [16, 17]. For the analysis, 25 ml of sample was saturated by adding 7.5 g of sodium chloride in a 40 ml septum capped vial. The vials were then
placed in a water bath on a magnetic stir plate heated to 65 ± 1.5 °C. A SPME fiber
(Supelco, Bellefonte, PA) was introduced into the head space gas through the septum and
182 the sample stirred for 30 min. The fiber was removed from the vial and inserted into the
Gas Chromatograph (GC) (Agilent 6890 Series; Agilent Technologies, Wilmington, DE)
inlet port and left exposed for 10 minutes for complete desorption of the target
compounds. The GC was equipped with Agilent DB-5 column (30 m x 0.32 mm inside
diameter x 0.25 µm film thickness) and flame ionization detector (FID). The injection port and detector temperature were set at 250 °C. The GC oven temperature was maintained at 60 °C for the first 4 min, then ramped to 270 °C at 10 °C/min and kept at
270 °C for 1 min. The flow rate of the carrier gas (N2) was set at 1.5 ml/min. The detector
makeup gas (N2) flow rate was set at 43.5 ml/min. The flame gases hydrogen and air flow
rates were set at 40 and 450 ml/min, respectively. The retention time for geosmin was
14.1 min and for MIB, 11.0 min. The fiber was then retracted into the holder, removed
from the GC inlet and reused for the next sample.
7.4 Results and Discussion
7.4.1 Single solute adsorption
Single solute adsorption isotherms were conducted for MIB and Geosmin on three
activated carbon fibers with different degrees of activation, namely ACC-15, ACC-20
and ACC-25. Isotherm experiments were also conducted with F-400 (a bituminous based
GAC) in order to compare ACF and GAC adsorption of MIB/ Geosmin. Table 6.1 in the
previous chapter summarized the physical properties of the adsorbents including surface
area, pore volume and microporosity [15]. It can be seen that F-400 has a wide range of
critical pore diameter, which is the diameter beyond which no adsorption takes place.
Contrarily, the three ACFs have significantly higher proportion of micropores and also
183 have specific critical pore diameters, which increase with the degree of activation (ACC-
15< ACC-20< ACC-25). In order to better understand the effect of NOM on activated
carbon adsorption of MIB and Geosmin, adsorption experiments were also carried out in
the presence of humic acid, which is commonly used to represent NOM.
Myers equation was used to model the single solute adsorption isotherms. The
Myers equation is given by
q ,ie pi C ,ie ×= qK ,iei )exp( (7.1) Hi
where Ce,i is the equilibrium liquid phase concentration, qe,i is the equilibrium surface
loading, Hi, Ki and pi are regression parameters. Myers equation was chosen because at low concentrations, it becomes linear and follows Henry’s law equation.
dC ,ie 1 q ,ie →= 0; (7.2) , Hdq iie
This phenomenon is very important to note when the single solute data are used to
predict multicomponent data by using a thermodynamic model (See Section 7.4.2).
The mass balance around the isotherm bottle is used to calculate the equilibrium surface
loading (qe,i) and is given by
( − ,, ieio ).VCC q = (7.3) ,ie m
where C0,i is the initial liquid phase concentration of MIB/Geosmin, m is the mass of the
adsorbent and V is the adsorbate volume. An adsorbate volume of 250 mL was used for
all the isotherm experiments. Figures 7.1 and 7.2 show the adsorption isotherm
184 experimental data fit to the Myers adsorption equation for Geosmin and MIB
respectively. The figures also include adsorption isotherms for MIB and Geosmin in the
presence of humic acid. Correlation of the experimental isotherm data to the Myers
equation was done using a nonlinear least-squares regression algorithm supplied by the
International Mathematics and Statistics Library for IBM computers. This algorithm used
an optimization tool to arrive at the value of the regression parameters by minimizing the
sum of the squares of the residuals. The sum of squares of relative error (SSRE) is given
by
2 ⎛ observedq − q predicted ⎞ ⎜ ,ie ,ie ⎟ SSRE = ∑ ⎜ ⎟ (7.4) ⎝ ,ie observedq ⎠
Table 7.1 shows the regression parameters for Myers fit for the adsorption
isotherms for MIB and Geosmin on the three ACFs and one GAC. It can be seen from
Figures 7.1 and 7.2 and from the low SSRE values in Table 7.1 that all adsorption isotherms were well defined by the Myers equation.
185 10 10 ACC-25 ACC-20
MIB MIB MIB/HA MIB/HA MIB - Myers fit MIB- Myers fit MIB/HA - Myers fit MIB/HA - Myers fit g/g) 1 1 μ
( 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 e 10 10 q ACC-15 F-400
MIB MIB MIB/HA MIB/HA MIB - Myers fit MIB - Myers fit MIB/HA - Myers fit MIB/HA - Myers fit 1 1 0.0 0.5 1.0 1.5 2.0 2.5 0.00.51.01.52.02.53.0 log C (μg/L) e
Figure 7.1 Single solute adsorption isotherm of MIB
186 10 10 ACC-25 ACC-20
Geosmin Geosmin Geosmin/HA Geosmin/HA Geosmin - Myers fit Geosmin - Myers fit Geosmin/HA - Myers fit Geosmin/HA - Myers fit
g/g) 1 1 μ
( -0.5 0.0 0.5 1.0 1.5 2.0 2.5 -10123
e 10 10
q ACC-15 F-400
Geosmin Geosmin Geosmin/HA Geosmin/HA Geosmin - Myers fit Geosmin - Myers fit Geosmin/HA - Myers fit Geosmin/HA - Myers fit 1 1 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 log C (μg/L) e
Figure 7.2 Single solute adsorption isotherm of geosmin
It can also be clearly seen that the presence of humic acid influences the adsorption capacity for all the adsorbents. The effect of the adsorbent pore size characteristics on adsorption of MIB and Geosmin and the competitive adsorption of humic acid has been discussed in detail elsewhere [18]. The effect of humic acid on the adsorption of MIB and geosmin was least pronounced for ACC-25 owing to a significant volume of secondary micropores and most pronounced for ACC-15 due to surface pore blockage The results confirmed that the pore size characteristics, especially the distribution of micropores, were the most important factor influencing the activated carbon adsorption of these two odor causing compounds.
187
Table 7.1 Myers isotherm equation parameters
Adsorbent H (lg-1) K (mmol g-1)-p P SSRE MIB F-400 3.73 x 107 9.528 .0904 .0011 ACC-15 1.2 x 103 1.263 .0562 .000067 ACC-20 2.23 x 108 9.104 .0558 .0047 ACC-25 9.72 x 1010 14.459 .0434 .00325 F-400 + HA 2.66 x 101 1.37 x 10-12 1.00 x 10-13 .2320 ACC-15 + HA 1.18 x 102 4.06 x 10-11 1.00 x 10-12 .032 ACC-20 + HA 5.47 x 105 5.620 .055 .000315 ACC-25 + HA 7.98 x 109 12.34 .049 .00242 Geosmin F-400 1.10 x 106 5.885 .071 .00084 ACC-15 5.65 x 104 2.088 .0867 .00021 ACC-20 1.09 x 107 40.12 .0628 .0076 ACC-25 1.72 x 1023 36.40 .0383 .00904 F-400 + HA 1.52 x 102 1.00 x 10-10 .327 .1549 ACC-15 + HA 1.87 x 106 5.63 .0675 .00032 ACC-20 + HA 1.47 x 1019 28.02 .057 .00998 ACC-25 + HA 4.52 x 1015 21.51 .055 .0044
7.4.2 Binary adsorption
In water treatment plants, activated carbon based adsorber beds are designed for a mixture of contaminant and not just for one particular contaminant. It is very likely that surface water containing taste and odor causing compounds contains certain concentrations of both MIB and Geosmin. In order to simulate practical usage, binary adsorption isotherms were conducted for solution containing both MIB and Geosmin.
Additional isotherms were also conducted in the presence of humic acid to better understand the effect of NOM on binary adsorption of these two contaminants. In order to better understand competitive adsorption, a mathematical model must be selected that can accurately predict adsorption equilibrium in a multicomponent system. Invariably, fixed bed adsorber models for multicomponent systems are based on parameters obtained
188 from single solute adsorption systems. Over the years, many adsorption models have been developed [15]. However, the Ideal adsorbed Solute Theory (IAST) model has been found to be extremely effective in accurately predicting multicomponent adsorption and has become well established as a result of its simplicity of calculation and thermodynamic soundness. Lu and Sorial [19] discussed in detail the various mathematical equations used in the development of the IAST model. The IAST model is based on the theory that the adsorbed mixture forms an ideal solution at a constant spreading pressure. This model involves adsorption at low coverage and hence Myers equation used for predicting single solute isotherms is ideal as it converges to Henry’s law at low coverage. Another important aspect of the IAST is that it can be applied only to systems where physical adsorption is the predominant adsorption mechanism [20]. The effectiveness of IAST for predicting the competitive adsorption between MIB and
Geosmin on ACFs and GAC was evaluated in the presence and absence of NOM in this study.
Studies have also shown that certain organic compounds such as phenols undergo
chemical transformations, namely oligomerization on the surface of the activated carbon
in the presence of molecular oxygen [15, 19, 21, 22]. This oxidative coupling on the
surface of the adsorbent resulted in the formation of very stable molecular products that
significantly affected the regeneration efficiency of the carbon. In order to determine if
any oligomerization was occurring, adsorption isotherms were conducted under oxic
(presence of molecular oxygen) and anoxic (absence of oxygen) conditions. The
procedure for anoxic and oxic conditions is presented elsewhere [19]. Binary adsorption
189 isotherms were also conducted at three different initial concentration combinations for all four adsorbents in order to determine if there was any significant adsorbate/adsorbate
interaction taking place that would result in deviations from the ideal adsorbed solution
behavior.
It can be seen from Figure 7.3a that for ACC-15, the anoxic and oxic
experimental adsorption isotherms for both MIB and Geosmin overlapped, indicating no
significant differences between both isotherm conditions. The same phenomenon was observed in case of F-400 as well. This confirmed that the presence of molecular oxygen had little or no effect on adsorption of MIB and Geosmin by activated carbon and no oligomerization was taking place. All subsequent adsorption isotherms were hence conducted under oxic conditions. From Figures 7.3, 7.4, 7.5 and 7.6, it can be seen that the IAST model predicted reasonably well the competitive adsorption of MIB and geosmin on all four adsorbents. This confirms that the adsorption is predominantly through physical adsorption with no chemisorption or oligomerization taking place. In case of ACC-15, although both isotherms show an increasing trend of adsorbed-phase capacity with increasing equilibrium liquid phase concentration (Figure 7.3), the adsorptive capacity for geosmin is marginally higher when compared to MIB. Because of high microporosity in ACC-15, both adsorbates are able to adsorb onto micropores without significant competition. In case of F-400, MIB isotherm manifests a decreasing trend signifying lesser available sites on the carbon at low carbon dosage due to competitive adsorption (Figure 7.4). F-400 has the least microporosity among the four adsorbents and hence the increased competition for the available sites. From Figures 7.5
190 and 7.6, it is seen that for ACC-20 and ACC-25, there is a very small difference between the two isotherms and both isotherms display an increasing trend of adsorbed phase capacity. It is interesting to note that for ACC-20, the adsorptive capacity for MIB is higher than for geosmin at the lower carbon dosage.
191 10 a. Concentration ratio 1:1
Geosmin IAST fit Geosmin (C0: 200 mg/L) MIB IAST fit MIB (C0: 200 mg/L) Anoxic Geosmin Anoxic MIB 1 0.5 0.6 0.7 0.8 0.9 1 2 3 10 b. Concentration ratio 2:1 g/g) μ e( q Geosmin IAST fit
Geosmin (C0: 200 μg/L) MIB IAST fit
MIB (C0: 100 μg/L) 1 0.5 0.6 0.7 0.8 0.9 1 2 3 10 c. Concentration ratio 1:2
Geosmin IAST fit
Geosmin (C0: 200 mg/L) MIB IAST fit
MIB (C0: 400 mg/L) 1 0.5 0.6 0.7 0.8 0.9 1 2 3 C (μg/L) e
Figure 7.3 Binary adsorption isotherms of MIB and Geosmin on ACC-15
192 10 a. Concentration ratio 1:1
Geosmin IAST fit Geosmin (C0: 200 mg/L) MIB IAST fit MIB (C0: 200 mg/L) Geosmin anoxic MIB anoxic 1 1.5 2 2.5 10 b. Concentration ratio 2:1 g/g) μ ( e q Geosmin IAST fit
Geosmin (C0: 200 μg/L) MIB IAST fit
MIB (C0: 100 μg/L) 1 1.5 2 2.5 10 c. Concentration ratio 1:2
Geosmin IAST fit
Geosmin (C0: 200 mg/L) MIB IAST fit
MIB (C0: 400 mg/L) 1 1 23 C (μg/L) e
Figure 7.4 Binary adsorption isotherms of MIB and Geosmin on F-400
193 10 a. Concentration ratio 1:1
Geosmin IAST fit
Geosmin (C0: 200 μg/L) MIB IAST fit
MIB (C0: 200 μg/L) 1 0.5 0.6 0.7 0.8 0.9 1 2 10 b. Concentration ratio 2:1 g/g) g/g) μ ( e q Geosmin IAST fit
Geosmin (C0: 200 μg/L) MIB IAST fit
MIB (C0: 100 μg/L) 1 0.5 0.6 0.7 0.8 0.9 1 2 10 c. Concentration ratio 1:2
Geosmin IAST fit
Geosmin (C0: 200 mg/L) MIB IAST fit
MIB (C0: 400 mg/L) 1 0.4 0.5 0.6 0.7 0.8 0.9 1 2 3 C (μg/L) e
Figure 7.5 Binary adsorption isotherms of MIB and Geosmin on ACC-20
194 10 a. Concentration ratio 1:1
Geosmin IAST fit
Geosmin (C0: 200 μg/L) MIB IAST fit
MIB (C0: 200 μg/L) 1 0.5 0.75 1 2.5 10 b. Concentration ratio 2:1 g/g) g/g) μ ( e q Geosmin (C0: 200 mg/L) MIB (C0: 100 mg/L) MIB IAST fit Geosmin IAST fit 1 0.05 0.075 0.1 0.25 0.5 0.75 1 2.5 10 c. Concentration ratio 1:2
Geosmin IAST fit
Geosmin (C0: 200 mg/L) MIB IAST fit
MIB (C0: 400 mg/L) 1 0.1 0.25 0.5 0.75 1 2.5 C (μg/L) e
Figure 7.6 Binary adsorption isotherms of MIB and Geosmin on ACC-25
195 10 10 a. ACC-25 b. ACC-20
Geosmin IAST fit Geosmin (C0: 200 mg/L) Geosmin (C0: 200 μg/L) MIB (C0: 100 mg/L) MIB IAST fit MIB IAST fit MIB (C 0: 200 μg/L) Geosmin IAST fit
g/g) 1 1 μ (
e 0.25 0.5 0.75 1 2.5 1 1.5 2 2.5
q 10 10 c. ACC-15 d. F-400
Geosmin IAST fit Geosmin IAST fit Geosmin (C0: 200 mg/L) Geosmin (C0: 200 mg/L) MIB IAST fit MIB IAST fit MIB (C0: 400 mg/L) MIB (C0: 400 mg/L) 1 1 1.5 2 2.5 1.5 2 2.5 Ce(μg/L)
Figure 7.7 Binary adsorption isotherms of MIB/Geosmin in the presence of humic
acid (10 mg/L) on F-400, ACC-15, ACC-20 and ACC-25
Binary adsorption isotherms were also conducted with 10 mg/L humic acid to study the effect of the presence of NOM on MIB and geosmin adsorption. The binary adsorption data was also compared with IAST model predictions. Myers isotherm parameters for single solute adsorption in the presence humic acid were used to generate the IAST model predictions. From Figure 7.7, it is seen that the IAST predicted well the binary adsorption in the presence of humic acid for ACC-20 and ACC-25. However, the
IAST predictions were less effective in case of ACC-15 and F-400. Although this ineffectiveness of the IAST model is clearly seen for F-400 (Figure 7.7d), it is not so
196 apparent in case of ACC-15 at the current scale of the figure (Figure 7.7c). At a closer inspection, it can be seen the experimental data for ACC-15 follow a trend that is parallel to the x-axis whereas the IAST model predictions are at an angle with respect to the x- axis. This can be attributed to blockage of micropores. Presence of NOM shifted the competitive adsorption from direct competition to surface pore blockage, especially of the micropores. Constriction of micropores in turn reduced the rate of diffusion of the target contaminants, in this case MIB and geosmin. Studies have shown that the IAST model is not very effective for predicting competitive adsorption in the presence of
NOM, as it does not consider pore blockage as a potential competitive mechanism [11,
23]. In case of NOM adsorption on GAC, studies have shown that presence of dissolved oxygen (DO) significantly influences adsorption [24]. Hence, the inability of the IAST model to predict accurately binary adsorption in the presence of humic acid on F-400 could be a result of oligomerization taking place. The order of effectiveness of the IAST model to predict binary adsorption of MIB and geosmin in the presence of humic acid is
ACC-25 > ACC-20 > ACC-15 > F-400. From this, it can be concluded that binary adsorption of geosmin and MIB on ACC-25 is least affected by the presence of humic acid. This confirms the results of the previous study [18] dealing with single solute adsorption of these two compounds on the three ACFs and F-400, where again ACC-25 was least affected by the presence of NOM in the form of humic acid.
7.5 Conclusion
Adosrption of MIB and geosmin on three activated carbon fibers (ACF) and one
GAC was studied through single and multicomponent adsorption isotherms. The single
197 solute adsorption isotherms were well defined by Myers equation and pore size properties
were the single most important factor influencing adsorption. Presence of humic acid as
NOM significantly influenced MIB and geosmin adsorption through competitive
adsorption and pore blockage mechanisms. However, the influence is less pronounced on
ACC25. Binary adsorption isotherms were also conducted for MIB and geosmin in the
presence and absence of NOM in order to simulate practical usage. The IAST model
using the Myers single solute isotherm parameters as input was used to predict binary
adsorption. The binary adsorption isotherms fit the IAST model predictions reasonably
well. It was also found that the presence of molecular oxygen did not significantly impact
the adsorptive capacity of the adsorbents for the binary system. The results reveal that
pure physical adsorption is taking place on the surface of the ACFs and GAC and there
were no signs of any oligomerization occurring. Presence of NOM in case of binary
isotherms resulted in some deviations from the IAST predictions especially for F400 and
to a certain extent ACC15. The results from binary adsorption confirm the results of the
single solute adsorption studied previously as in both cases, among the four adsorbents, the adsorption on ACC-25 was least influenced by the presence of humic acid.
7.6 References
1 McGuire, M. J. (1995) Off-flavor as the consumer's measure of drinking water
safety. Water Science & Technology 31, 1-8
2 Bruce, D., Westerhoff, P. and Brawley-Chesworth, A. (2002) Removal of 2-
methylisoborneol and geosmin in surface water treatment plants in Arizona.
Journal of Water Supply: Research and Technology - AQUA 51, 183-197
198 3 http://www.owwrc.com/TOhealth.htm
4 Pirbazari, M., Ravindran, V., Badriyha, B. N., Craig, S. and McGuire, M. J.
(1993) GAC adsorber design protocol for the removal of off-flavors. Water
Research 27, 1153-1166
5 Ferguson, D. W., McGuire, M. J., Koch, B., Wolfe, R. L. and Aieta, E. (1990)
Comparing PEROXONE and Ozone for Controlling Taste and Odor Compounds,
Disinfection By-Products, and Microorganisms. Journal of the American Water
Works Association 82, 181-191
6 Glaze, W. H., Zarnoch, J. J., Ruth, E. C., Chauncey, W. and Schep, R. (1990)
Evaluating Oxidants for the Removal of Model Taste and Odor Compounds from
a Municipal Water Supply. Journal of the American Water Works Association 82,
79-84
7 Cook, D., Newcombe, G. and Sztajnbok, P. (2001) The application of powdered
activated carbon for MIB and geosmin removal: predicting PAC doses in four raw
waters. Water Research 35, 1325-1333
8 Ng, C., Losso, J. N., Marshall, W. E. and Rao, R. M. (2002) Freundlich
adsorption isotherms of agricultural by-product-based powdered activated carbons
in a geosmin - water system. Agricultural Wastes 85, 131-135
9 Jung, S. W., Baek, K. H. and Yu, M. J. (2004) Treatment of taste and odor
material by oxidation and adsorption. Water science and technology : a journal of
the International Association on Water Pollution Research and Control 49, 289-
295
199 10 Ho, L. and Newcombe, G. (2005) Effect of NOM, turbidity and floc size on the
PAC adsorption of MIB during alum coagulation. Water Research 39, 3668-3674
11 Newcombe, G., Morrison, J., Hepplewhite, C. and Knappe, D. R. U. (2002)
Simultaneous adsorption of MIB and NOM onto activated carbon-II. Competitive
effects. Carbon 40, 2147-2156
12 Newcombe, G., Morrison, J. and Hepplewhite, C. (2002) Simultaneous adsorption
of MIB and NOM onto activated carbon. I. Characterisation of the system and
NOM adsorption. Carbon 40, 2135-2146
13 Pelekani, C. and Snoeyink, V. L. (1999) Competitive adsorption in natural water:
Role of activated carbon pore size. Water Research 33, 1209-1219
14 Pelekani, C. and Snoeyink, V. L. (2000) Competitive adsorption between atrazine
and methylene blue on activated carbon: the importance of pore size distribution.
Carbon 38, 1423-1436
15 Lu, Q. and Sorial, G. A. (2004) The role of adsorbent pore size distribution in
multicomponent adsorption on activated carbon. Carbon 42, 3133-3142
16 Lloyd, S. W., Lea, J. M., Zimba, P. V. and Grimm, C. C. (1998) Rapid analysis of
geosmin and 2-methylisoborneol in water using solid phase micro extraction
procedures. Water Research 32, 2140-2146
17 Watson (2000) Quantitative analysis of trace levels of geosmin and MIB in source
and drinking water using headspace SPME. Water Research 34, 2818-2828
18 Srinivasan, R., Sorial, G. A., Ononye, G., Husting, C. and Jackson, E. (2008)
Elimination of persistent odorous compounds in drinking water. Water Science
and Technology: Water Supply 8, 121-127
200 19 Lu, Q. and Sorial, G. A. (2004) Adsorption of phenolics on activated carbon--
impact of pore size and molecular oxygen. Chemosphere 55, 671-679
20 Noroozi, B., Sorial, G. A., Bahrami, H. and Arami, M. (2008) Adsorption of
binary mixtures of cationic dyes. Dyes and Pigments 76, 784
21 Lu, Q. and Sorial, G. A. (2005) Impact of pore size on competitive adsorption of
phenolic compounds. Water Science and Technology: Water Supply 4, 1-7
22 Lu, Q. and Sorial, G. A. (2007) The effect of functional groups on
oligomerization of phenolics on activated carbon. Journal of Hazardous Materials
148, 436-445
23 Ebie, K., Li, F., Azuma, Y., Yuasa, A. and Hagishita, T. (2001) Pore distribution
effect of activated carbon in adsorbing organic micropollutants from natural
water. Water Research 35, 167-179
24 Karanfil, T., Schlautman, M., Kilduff, J. and Weber, W. J. (1996) Adsorption of
Organic Macromolecules by Granular Activated Carbon. 2. Influence of
Dissolved Oxygen. Environmental Science & Technology 30, 2195-2201
201 CHAPTER 8
CONCLUSIONS AND RECOMMENDATIONS
8.1 Conclusions
8.1.1 Perchlorate
Through an extensive literature review, this research looked at the challenges
associated with perchlorate remediation and the various treatment technologies available
for perchlorate treatment. This study looked at the potential to develop an integrated
method for removal of perchlorate in water by combining zero-valent iron (Fe0) and electrochemical reduction.
The main conclusions drawn from the study were:
1. Perchlorate has been the most challenging contaminant to treat and regulate for
the policy makers and the environmental technology community alike in recent
times.
2. Among all the available technologies for perchlorate removal, only ion exchange
and biological treatment have been found to be feasible and have been
implemented on remediation sites. However, both technologies have limitations
when it comes to drinking water treatment.
3. Although a plethora of new technologies have been identified and have shown
promise at a bench or pilot scale, none of these technologies are currently at a
202 stage were they can be implemented in the field. A lot of research still requires to
be done to make them cost effective and practical for field application purposes.
4. Integrating zero-valent iron reduction with electrochemical reduction of
perchlorate did not result in the expected synergistic effect. High pH and
hydrogen gas generated during electrochemical reduction significantly influenced
the surface properties of the zero-valent iron and adversely affected the reaction
kinetics.
5. The reaction rates obtained with electrochemical reduction of perchlorate were
extremely low. Coating the electrodes with thin oxide films resulted in reduction
rates 3-5 times higher than that obtained with Fe0. However these rates were still
extremely low for any practical consideration.
8.1.2 MIB/ Geosmin
A critical review was conducted to better understand treatment of these taste and odor causing compounds in water. Since activated carbon adsorption is the best available technology for treatment of MIB/ Geosmin, single solute and binary adsorptions for MIB and Geosmin were studied on activated carbon fibers (ACFs) to evaluate the effect of pore size characteristics and distribution. The main conclusions are listed below.
1. A large percentage of water treatment plants are plagued with taste and odor
episodes due to presence of these compounds especially during summer months.
2. Powdered activated carbon (PAC) is currently the most commonly practiced
approach for treatment of these compounds. For MIB/ Geosmin removal,
203 activated carbon adsorption is significantly better than advanced oxidation
processes (AOPs) both in terms of effectiveness and cost feasibility.
3. Single solute adsorption isotherms of MIB and geosmin on the activated carbon
fibers (ACFs) were well defined by Myers equation and pore size properties were
the single most important factor influencing adsorption.
4. Binary adsorption isotherms were also conducted for MIB and geosmin in the
presence and absence of natural organic matter (NOM) in order to simulate
practical usage and fit the IAST model predictions reasonably well.
5. Presence of NOM significantly influenced MIB and geosmin adsorption through
competitive adsorption and pore blockage mechanisms and resulted in some
deviations from the IAST predictions in case of binary isotherms.
6. The results reveal that pure physical adsorption is taking place on the surface of
the ACFs and GAC and there were no signs of any oligomerization occurring.
8.2 Recommendations for future research
8.2.1 Perchlorate
Based on the literature review for perchlorate treatment, it is highly likely that an
integrated technology that combines two or more different technologies might be
necessary to overcome the drawbacks of one technology. A perchlorate removal technology that has high reaction rates can be combined with a destruction technology that completely destroys the perchlorate in the concentrated brine by reducing it to chloride. Ion exchange followed by catalytic chemical reduction or biological treatment would be one such alternative. Future work should look at a combination or two or more
204 technologies that have a synergistic effect and that compensate for the drawbacks.
Making these technologies adaptable to drinking water environment is equally challenging and would require extensive research. In the end, it will be the regulatory decisions by the USEPA that will dictate the extent and direction of perchlorate research.
Although reaction kinetics associated with electrochemical reduction of perchlorate are too slow, the major advantage is that it is a destruction technology that results in complete removal of perchlorate. Future work is required in determining the right combination of electrodes, thin oxide films on the electrodes and heterogeneous catalysts that can result in reaction kinetics that prove to be feasible. Research on electrochemical treatment could be for direct treatment of perchlorate in water or brines concentrated in perchlorate through other treatment technologies.
8.2.2 MIB/ Geosmin
Since activated carbon adsorption has been found to be the most feasible technology for treatment of MIB/ Geosmin in water, most work will have to be concentrated in optimizing PAC dosages in water treatment plants during taste and odor episodes. Research should focus on better understanding the PAC pore size characteristics in order to tailor the carbon to treat MIB/ Geosmin in the presence of natural organic matter (NOM). ACFs have been found to be ideal adsorbents for targeting the impact of pore size on single and multicomponent adsorption.
205 Adsorption on ACF could determine the degree of tailoring the PAC would have
to undergo to support both microporous and macroporous adsorption. This would enable
simultaneous adsorption of the smaller MIB/ Geosmin molecules along with the larger
NOM molecules. Future work should look into novel methods of producing and tailoring
PAC so that it could approach the performance of ACF. Research should focus on optimizing factors such as activation temperature and duration that would result in the
ideal pore size distribution for effective and economical removal of MIB and geosmin.
206