Taste and odor in drinking water are one of the main problems of the water supply and treatment sector. Peroxone is an effective advanced oxidation process, which combines ozone with hydrogen peroxide to create hydroxyl radicals that decompose organic compounds. 2-Methylisoborneol (2-MIB) is one of the significant taste and odor causing compounds, which can be removed with the peroxone process. In this study, removal of a 2-MIB compound by peroxone process was investigated in a pilot-scale treatment plant and optimum operational conditions were determined. For safety reasons, it is important that residual O3 and H2O2 concentrations in the water leaving the reactor should not exceed 0.1 and 0.5 mg/L, respectively. Results indicate that while dissolved ozone concentration was below the indicated limit for all experiments, concentrations over 0.5 mg/L residual H2O2 were observed during the experiments with an H2O2:O3 ratio of 0.5. This limit exceedance affected the decision on the ideal peroxone ratio along with the 2-MIB removal results. Therefore; optimum H2O2:O3 ratio was determined as 0.3. 2-MIB removal efficiency of 81% was achieved at the optimum H2O2:O3 ratio with a contact time of 15 min. According to the results, 2-MIB removal rate had a linear correlation with the contact time.

  • 2-MIB and geosmin are two major taste and odor causing compounds.

  • Peroxone is an effective advanced oxidation process for micropollutant removal.

  • Optimizing operational parameters plays a key role in water treatment.

  • Results obtained from pilot plant provide major guidance for full-scale applications.

Graphical Abstract

Graphical Abstract
Graphical Abstract

Taste and odor in drinking water affect the perception of customers in terms of both safety and aesthetics. Two of the most common taste and odor compounds, 2-methylisoborneol (2-MIB) and geosmin, are detectable even at concentrations below 10 ng/L (Fakioglu 2017). Water authorities not only search for the cause of taste and odor problems, but also try to find an effective process that can be integrated into existing water treatment plants.

Previous studies have shown that conventional water treatment processes such as coagulation, flocculation, sedimentation and chlorination are ineffective for removal of extracellular taste and odor compounds (Lalezary et al. 1986; Glaze et al. 1990; Bruce et al. 2002; Srinivasan & Sorial 2010). However, processes including powdered activated carbon, ozonation, and biofiltration have been found to be successful for the removal of these compounds where they are present dissolved in the water (Cook et al. 2001; Ho 2009; Hsieh et al. 2010; Srinivasan & Sorial 2010; Wang 2014).

Removal of taste and odor compounds by the peroxone process; that is, a combination of ozone (O3) and hydrogen peroxide (H2O2), was proven to be effective by several studies (Srinivasan & Sorial 2010; Fakioglu 2017). These studies have shown that removal of 2-MIB and geosmin compounds was quite higher with the peroxone process in comparison to sole O3 application at the same O3 dosage. The findings of Koch et al. (1992) showed that 80–90% removal of 2-MIB and geosmin could be achieved with only 2 mg/L ozone dosage and 0.2 ratio of H2O2/O3, which is the rate of removal that can only be achieved with a high sole ozone dosage of 4 mg/L. Determining an ideal peroxone ratio (H2O2:O3 by mole) was the main objective of these studies. According to several studies, addition of H2O2 to O3 has decreased the formation of potentially carcinogenic bromate, which is a byproduct of ozonation (Mizuno et al. 2011; Wang 2014). The operating temperature was set to a certain value in most of the studies, since the solubility of ozone is highly affected by the temperature (von Sonntag & von Gunten 2012).

In this study, optimum operational conditions such as peroxone ratio and contact time were determined for removal of 2-MIB in a pilot-scale peroxone system. Water samples used in all experiments were taken from one of the main drinking water sources of a mega city, Istanbul (Turkey). Other important parameters such as geosmin, total organic carbon and bromate were also determined along with the 2-MIB. During the experimental study, residual (dissolved) O3 and residual H2O2 were monitored in order to investigate whether the applied chemical dosages were reasonable or not.

Characteristics of raw water and reagents

Raw water was obtained from the outlet of an aeration unit in a full-scale drinking water treatment plant that was having taste and odor problems during the summer seasons. Characterization of the raw water, which was originally received from a surface water (a lake in Istanbul, Turkey), is provided in Table 1. All analyses in Table 1 were conducted in accordance with Standard Methods (APHA 2005). During some of the summer seasons in the previous years, the lake had experienced high 2-MIB concentrations of on average 90 ng/L. However, 2-MIB and geosmin concentrations were lower than 10 ng/L during the sampling period, which is known as the limit concentrations to be perceived. Therefore, 2-MIB was spiked into the water samples to obtain a concentration of 90 ng/L, which also simulated the average 2-MIB concentration of the investigated water source during the peak season when taste and odor problem was experienced. Chromatographic grade 2-MIB standard (Supelco, USA) (10 mg/ml) was used for maintaining the required concentration of 2-MIB in the water samples. Since 2-MIB is more difficult to be removed with O3 in comparison to geosmin (WQTS 2009; Fakioglu 2017), operational conditions for 2-MIB removal were investigated to set the process design requirements. All reagents and solvents for H2O2 analysis were at least of analytical grade. All solutions were prepared with deionized water.

Table 1

Raw water quality

ParameterUnitValue
Chloride (mg/L) 20.4 
Color (Pt-Co) 8.0 
Conductivity (25 °C) (μS/cm) 320 
pH – 7.71 
Iron (mg/L) 0.213 
Manganese (mg/L) 0.058 
Sulfate (mg/L) 19.9 
Sodium (mg/L) 16.87 
Potassium (mg/L) 3.11 
Total organic carbon (TOC) (mg/L) 2.7 
Turbidity (NTU) 3.6 
UV254 (cm−10.07 
2-MIB (ng/L) <0.5 
Geosmin (ng/L) 0.28 
Bromide (mg/L) 0.01 
Total dissolved solids (TDS) (mg/L) 192 
Total Hardness (mg CaCO3/L) 144 
Total Alkalinity (mg CaCO3/L) 116.4 
ParameterUnitValue
Chloride (mg/L) 20.4 
Color (Pt-Co) 8.0 
Conductivity (25 °C) (μS/cm) 320 
pH – 7.71 
Iron (mg/L) 0.213 
Manganese (mg/L) 0.058 
Sulfate (mg/L) 19.9 
Sodium (mg/L) 16.87 
Potassium (mg/L) 3.11 
Total organic carbon (TOC) (mg/L) 2.7 
Turbidity (NTU) 3.6 
UV254 (cm−10.07 
2-MIB (ng/L) <0.5 
Geosmin (ng/L) 0.28 
Bromide (mg/L) 0.01 
Total dissolved solids (TDS) (mg/L) 192 
Total Hardness (mg CaCO3/L) 144 
Total Alkalinity (mg CaCO3/L) 116.4 

Pilot plant set-up

Pilot plant consisted of a raw water storage tank, reactor column (ozone contact column, of 2.4 m diameter and 5 m height), hydrogen peroxide (H2O2) dosing system and ozone generator along with the required pipes and measuring equipment (Figure 1). Raw water storage tank, which was fed with water from the aeration unit outlet of a full-scale water treatment plant, was made of polyethylene (PE) (of 2.8 m diameter, 2.4 m height and 15 m3 total volume) and contained a low speed mixer (1.5 kW, 6–8 rev/min) along with an ultrasonic water level sensor (Sick AG, Germany). The ozone contact reactor was made of plexiglass (with 0.24 m diameter, 5 m height and 0.226 m3 volume). Raw water was fed with a centrifuge feed pump (Ebara, Japan) to the reactor. The ozonation reactor contained an ultrasonic water level measuring device (Keller, Switzerland), a gas phase ozone measuring device (2B Technologies Model 106-M, USA), a dissolved ozone measuring device (Chemitec 42 Series, China), a thermometer (Chemitec 42 Series, China) and a pH meter (Chemitec 42 Series, China). Ozone was produced by the ozone generator (BNP, China), which was fed with an air compressor. Ozone was then transferred to the reactor by a ceramic diffuser (with 0.185 m diameter). H2O2 solution was dosed by a solenoid dosage pump (Sisdoz, Turkey) to a pipe equipped with a static mixer (Polyvinylchloride, PVC), which delivered water to the reactor.

Figure 1

Layout of the pilot-scale set-up.

Figure 1

Layout of the pilot-scale set-up.

Close modal
Figure 2

Consumed ozone ratio depending on reactor height.

Figure 2

Consumed ozone ratio depending on reactor height.

Close modal

Gas phase ozone at the inlet and at the outlet of the reactor was measured with ozone measurement devices (2B Technologies, USA). Surplus (off-gas) ozone was destroyed by an ozone destruction unit (BNP, China) before reaching to the atmosphere.

Experimental plan

Relation between ozone consumption and reactor height

Before starting the experiments to determine the ideal peroxone ratio and contact time experiments, ozone consumption versus reactor height was investigated through running the system in batch mode. Experiments were performed with different reactor heights (1 m, 2 m, 3 m, 4 m and 4.5 m), while initial ozone concentration and contact time were set to 4 mg O3/L and 10 minutes, respectively. Ozone consumption was calculated by using the inlet and the off-gas ozone concentrations. All experiments were conducted in duplicate and average values were used when presenting the results. Results obtained from the experiments conducted in the pilot treatment plant are given in Table 3 and Figure 2.

Determination of ideal peroxone ratio and contact time

Different H2O2:O3 ratios were tested in order to find the ideal ratio for the removal of 2-MIB (Table 2). Inlet O3 dose was set to 4 mg/L and the reactor was continuously fed with 2-MIB spiked water. Average water temperature in the pilot plant was recorded as around 16 °C. After the determination of the optimum H2O2:O3 ratio, the reactor was operated at different contact times including 5, 10 and 15 min (Table 2). TOC and geosmin removal efficiencies were also determined throughout the experimental study. Additionally, variations in residual H2O2 concentration were measured during each experiment.

Table 2

Experimental plan applied to determine the optimum H2O2:O3 ratio and contact time

(a) Determination of optimum H2O2:O3 ratio
(b) Determination of optimum contact time
H2O2:O3Inlet O3 concentrationContact timeContact timeInlet O3 concentrationH2O2:O3
4 mg/L 10 min    
0.1 5 min 4 mg/L Selected H2O2:O3 ratio 
0.3 10 min 
0.5 15 min 
(a) Determination of optimum H2O2:O3 ratio
(b) Determination of optimum contact time
H2O2:O3Inlet O3 concentrationContact timeContact timeInlet O3 concentrationH2O2:O3
4 mg/L 10 min    
0.1 5 min 4 mg/L Selected H2O2:O3 ratio 
0.3 10 min 
0.5 15 min 

Analytical and instrumental procedures

Residual H2O2 concentration in the liquid phase was analysed following the iodide/molybdate method (HP-02) (Klassen et al. 1994). Concentrations of 2-MIB and geosmin were measured by gas chromatography mass spectrometry (GC/MS) (Agilent Technologies 7890B, USA) using Method 8270D (USEPA 1998). The solid phase micro extraction (SPME) method was used for preparing the samples for GC/MS. TOC analysis was conducted by a TOC analyzer (Shimadzu TOC-L Series, Japan). Bromate analysis was conducted by following Method 300.1 (USEPA 1999).

Relation between ozone consumption and reactor height

Results obtained from the experiments conducted in the pilot treatment plant are given in Table 3 and Figure 1. Ozone consumption ratio was found to be 51% at 1 m reactor height and 97% at 4.5 m reactor height by applying 4 mg/L inlet ozone concentration. In order to reach an ozone consumption ratio above 75%, at least 3 m reactor height was required.

Table 3

Ozone consumption and TOC removal at different reactor heights

Inlet ozone concentration (mg/m3)4,000
Reactor height (m) 4.5 
Inlet ozone concentration (mg/m34,031 4,072 4,057 3,968 3,984 
Off-gas ozone concentration (mg/m31,976 1,492 953 378 114 
Ozone consumption (mg/m32,055 2,580 3,104 3,590 3,870 
Ozone consumption ratio (%) 51 63 77 90 97 
TOC removal (%) 
Inlet ozone concentration (mg/m3)4,000
Reactor height (m) 4.5 
Inlet ozone concentration (mg/m34,031 4,072 4,057 3,968 3,984 
Off-gas ozone concentration (mg/m31,976 1,492 953 378 114 
Ozone consumption (mg/m32,055 2,580 3,104 3,590 3,870 
Ozone consumption ratio (%) 51 63 77 90 97 
TOC removal (%) 

Along with the ozone consumption ratio, TOC removal was investigated during study. As seen in Table 3, TOC removal was lower than 6% for all operating conditions. Based on the results obtained from this study, ozone cannot remove TOC effectively alone. However, ozone is proven to increase the colloidal destabilization, which dramatically increases the removal efficiencies of organic compounds by coagulation, flocculation and filtration units in water treatment plants (Yuksel et al. 2002).

Results indicated that consumed ozone increased with increasing reactor height. With increasing heights, ozone contact time can be extended since water will stay longer in the reactor. Therefore, ozone fed at the bottom of the reactor could have more time to react with the compounds in the water and could be consumed in higher ratios.

Determination of ideal peroxone ratio and contact time

Variations in dissolved O3 and residual H2O2 concentrations

Dissolved and consumed O3 concentrations were measured for each H2O2:O3 ratio at a contact time of 10 minutes (Figures 3 and 4). It is essential that the residual O3 concentration in the water leaving the reactor should not exceed 0.1 mg/L for safety reasons (WQTS 2009). Considering Figure 3, dissolved (residual) O3 concentration was around 0.05 mg/L, below 0.1 mg/L for each H2O2:O3 ratio, while consumed O3 concentration was around 4 mg/L. Dissolved ozone concentrations of each ratio (0.0–0.5) at the end of the 10th minute showed that the highest dissolved ozone concentration was reached with sole ozonation, which is as expected since the ozone was consumed in the peroxone reaction (Figure 3). On the other hand, the consumed ozone concentrations at the beginning showed that the most stable and unstable consumptions of ozone were in the case of having an H2O2:O3 ratio of 0.5 and sole ozonation. However, overall ozone consumptions obtained for each ratio including the H2O2:O3 ratio of zero showed that the consumptions were very close to each other.

Figure 3

Dissolved O3 concentrations for various peroxone ratios.

Figure 3

Dissolved O3 concentrations for various peroxone ratios.

Close modal
Figure 4

Consumed O3 concentrations at various peroxone ratios.

Figure 4

Consumed O3 concentrations at various peroxone ratios.

Close modal

Residual H2O2 concentration in the peroxone process was recommended not to exceed 0.5 mg/L (Oturan & Aaron 2014), since higher concentrations might cause operational problems such as residual chlorine consumption in the water treatment plants. Figure 5 shows that limit residual H2O2 concentration was not exceeded during the experiments at a contact time of 10 min and H2O2:O3 ratio of 0.1. On the other hand, the average residual H2O2 concentration for H2O2:O3 ratio of 0.3 peroxone ratio was around 0.5 mg/L. However, a residual H2O2 concentration of over 0.5 mg/L was observed even at the 4th min at a H2O2:O3 ratio of 0.5.

Figure 5

Residual H2O2 concentrations along experiments for different peroxone ratios.

Figure 5

Residual H2O2 concentrations along experiments for different peroxone ratios.

Close modal

Impact of H2O2:O3 ratio on treatment performance

The experiments for determining the optimum peroxone ratio were conducted with 10 minutes contact time as depicted in Table 2. TOC removal efficiencies were below 10% during the experiments at each peroxone ratio. Geosmin removal efficiencies could not be evaluated depending on the quite low initial geosmin concentrations, which were close to the detection limit. Bromate concentrations were below detection limits for all the experiments (<2 μg/L). 2-MIB concentrations in the influent and effluent were between 75–83 and 32–52 ng/L, respectively, having a 2-MIB removal efficiency between 37.5 and 57.3% (Figure 6). High residual H2O2 concentration was observed in the reactor effluent during the experiments with an H2O2:O3 ratio of 0.5. According to the results obtained from the pilot plant study, the ideal H2O2:O3 ratio was determined as 0.3. The optimum peroxone ratio achieved 52.3% removal of 2-MIB, while sole ozonation only reached 37.5% removal of 2-MIB. The highest peroxone ratio (0.5) removed the compound by 57.3%, which shows the removal could be improved by 20% compared to sole ozonation.

Figure 6

2-MIB removal efficiency at different peroxone ratios.

Figure 6

2-MIB removal efficiency at different peroxone ratios.

Close modal

Impact of contact time on treatment performance

Optimum contact time at the ideal H2O2:O3 ratio was determined regarding 2-MIB removal efficiency (Figure 7). TOC removal efficiencies were below 9% during the experiments conducted at different peroxone ratios. According to the results, 2-MIB removal efficiency had a linear correlation with the contact time. Thus, the best removal efficiency (81%) was obtained at a contact time of 15 min by applying an O3 dose of 4 mg/L. Overall results confirmed the significance of contact time on the 2-MIB removal efficiency.

Figure 7

2-MIB removal efficiency at different contact times (H2O2:O3 ratio: 0.3).

Figure 7

2-MIB removal efficiency at different contact times (H2O2:O3 ratio: 0.3).

Close modal

Relation between the ozone consumption and reactor height was determined in this study. Results indicated that consumed ozone increased with increasing reactor height. In order to reach an ozone consumption ratio above 75%, at least 3 m reactor height was needed. TOC could not be removed by only ozone application.

Optimum operational conditions were investigated for the removal of 2-MIB by peroxone process. It is recommended that the residual O3 concentration in the outlet of the reactor should not exceed 0.1 mg/L for operator safety reasons. Dissolved (residual) ozone concentration was below the limit residual O3 concentration at all H2O2:O3 ratios. Results show that addition of an ozone quenching chemical to the reactor is not necessary for any of the cases in this study. Besides, it is important not to exceed 0.5 mg/L of residual H2O2 concentration in the peroxone process. However, a residual H2O2 concentration of over 0.5 mg/L was measured at an H2O2:O3 ratio of 0.5. Although the average residual H2O2 concentration with an H2O2:O3 ratio of 0.3 was around 0.5 mg/L, the removal rate at H2O2:O3 ratio of 0.3 showed a slight increase in comparison to an H2O2:O3 ratio of 0.1. Therefore, the optimum H2O2:O3 ratio was determined as 0.3 for the peroxone process in the tested conditions. A 2-MIB removal efficiency of 81% was achieved at the optimum H2O2:O3 ratio with a contact time of 15 min. According to the results, the 2-MIB removal rate had a linear correlation with the contact time.

Authors suggest that further studies should be conducted for both ozone and peroxone processes regarding the formation of disinfection by-products. In addition to this, if the water treatment plant has a disinfection concern other than the taste and odor problem, virus and bacteria inactivation rates should be investigated with the peroxone process. Last but not least, the effect of temperature for the removal of taste and odor compounds in a full-scale water treatment should be taken into consideration.

This research was funded by Istanbul Water and Sewerage Administration (ISKI) with the project titled as ‘Evaluation of Operation Results with Hydrogen Peroxide and Ozone Oxidation (Peroxone) Process’. The authors express their gratitude to Fatih Turan, General Director of ISKI. The authors would like to thank the Head and all staff of the Water Treatment Department of ISKI for their support.

All relevant data are included in the paper or its Supplementary Information.

APHA
2005
Standard Methods for the Examination of Water and Wastewater
, 21st edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
.
Bruce
D.
Westerhoff
P.
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
(
4
),
183
198
.
Fakioglu
M.
2017
Removal of Taste and Odor Causing Compounds in Drinking Water by Peroxone Process
.
Master's Thesis
,
Istanbul Technical University
,
Istanbul, Turkey
.
Glaze
W. H.
Zarnoch
J. J.
Ruth
E. C.
Chauncey
W.
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
.
Hsieh
S. T.
Lin
T. F.
Wang
G. S.
2010
Biodegradation of MIB and geosmin with slow sand filters
.
Journal of Environmental Science and Health
45
(
8
),
951
957
.
Klassen
N. V.
Marchington
D.
McGowan
C. E.
1994
Method HP-02 analysis of hydrogen peroxide
.
Analytical Chemistry
66
,
2921
2925
.
Koch
B.
Gramith
J. T.
Dale
M. S.
Ferguson
D. W.
1992
Control of 2-methylisoborneol and geosmin by ozone and peroxone – a pilot-study
.
Water Science and Technology
25
(
2
),
291
298
.
Lalezary
S.
Pirbazari
M.
McGuire
M. J.
1986
Oxidation of five earthy-musty taste and odor compounds
.
Journal of the American Water Works Association
78
,
62
69
.
Mizuno
T.
Ohara
S.
Nishimura
F.
Tsuno
H.
2011
O3/H2O2 process for both removal of odorous algal-derived compounds and control of bromate ion formation
.
Ozone: Science & Engineering
33
(
2
),
121
135
.
Oturan
M. A.
Aaron
J. J.
2014
Advanced oxidation processes in water/wastewater treatment: principles and applications. A review
.
Critical Reviews in Environmental Science and Technology
44
(
23
),
2577
2641
.
USEPA
1998
EPA Method 8270D
.
U.S. Environmental Protection Agency
,
Washington, DC
.
USEPA
1999
Determination of Inorganic Anions in Drinking Water by ion Chromatography
.
EPA Method 300.1
,
U.S. Environmental Protection Agency
,
Washington, DC
.
von Sonntag
C.
von Gunten
U.
2012
Chemistry of Ozone in Water and Wastewater Treatment
.
IWA Publishing
,
London, England
.
WQTS
2009
Evaluation of Ozone and Peroxone for Water Quality Enhancement at the del Valle and Patterson Pass Water Treatment Plants
.
Water Quality and Solutions Inc., Los Angeles, CA, Project Report
.
Yuksel
E.
Akgiray
O.
Saatci
A. M.
Sarikaya
H. Z.
Koyuncu
I.
2002
Effect of ozone injection location on filter performance in direct filtration
.
Water Science and Technology
46
(
9
),
171
177
.