Abstract

The removal efficiencies of geosmin/2-MIB by conventional treatment (flocculation, sedimentation and sand filtration) combined with advanced treatment (ozonation and granular activated carbon (GAC) filtration) in a pilot-scale experiment were investigated in a comprehensive manner. The objective of this study is to provide useful information for practical applications to solve the taste and odor problem during algal blooms in Lake Taihu. Results showed that the conventional treatment removed 38–59% and 36–64% of 2-MIB and geosmin, respectively, with ozone dosage from 0 to 1.0 mg/L. In particular, the increase in ozone concentration promoted the removal of 2-MIB/geosmin by sand filtration, meanwhile lowering the odorant removal efficiency by GAC filtration, with the key reason being the higher biomass produced in the sand filter through the strengthening effect of the ozonation. The organics with higher molecular weight (MW) showed the most significant decline in the pre-ozonation process with 1.0 mg/L ozone dosage, resulting in an enhanced removal efficiency of 2-MIB/geosmin by pre-ozonation. With ozone dosage of 1.0 mg/L, geosmin and 2-MIB in the treated water were 0.1 and 3.85 ng/L, respectively, which were below their odor threshold concentrations (OTCs) with the preliminary concentration of ∼200 ng/L of 2-MIB/geosmin.

INTRODUCTION

Taste and odor issues have become significant challenges to the sustainability and safety of drinking water. Every spring and summer, Lake Taihu is threatened by the taste and odor compounds produced by cyanobacteria in surface water with 2-methylisoborneol (2-MIB) and trans-1,10-dimethys-trans-9-decalol (geosmin) as two typically identified taste and odor compounds (Su et al. 2015). Conventional treatment could not completely remove 2-MIB and geosmin in the surface water. Therefore, a subsequent advanced treatment process was required for better removal of these taste and odor compounds. Currently, ozone/granular activated carbon (O3/GAC) is widely used in water utilities as an advanced technology for taste and odor removal for drinking water treatment.

Ozone is a strong oxidant capable of oxidizing various organic compounds, however it cannot effectively destroy 2-MIB and geosmin below their odor threshold levels if used alone (Collivignarelli & Sorlini 2004). The resistance of 2-MIB/geosmin by ozone oxidation may be due to the fact that they have the structures of saturated cycled tertiary alcohols (Antonopoulou et al. 2014). Ozone dosage is an important factor in maximizing the removal efficiency of odorants during ozonation. He et al. (2017) investigated the odor removal performance by O3 pre-oxidation and obtained total odor removal rates ranging from 77.8% to 86.7% with ozone dosage from 1 to 4 mg/L.

GAC has been considered a useful method for the removal of taste and odor compounds in surface water. Zamyadi et al. (2015) figured out that 80% of geosmin and 60% of 2-MIB were removed with the use of aged GAC filters in a two-year full-scale plant operation; better geosmin removal by GAC filter as compared with 2-MIB was also observed. In the GAC filter, biodegradation and adsorption play equally significant roles for the removal of organic compounds; biodegradation exhibited dominance in the GAC filter in long-term operation where stable biomass had been produced on the surface of the adsorptive media (De Waters & DiGiano 1990; Takeuchi et al. 1997).

Owing to the fact that ozone is able to encourage biological activities on the GAC absorptive surface, an enhanced removal efficiency of organic materials is obtained in the GAC filter (Takeuchi et al. 1997; Simpson 2008). Several studies have proved that ozonation and GAC filtration are a well-suited combination for effective removal of 2-MIB and geosmin in surface water. Elhadi et al. (2004) discovered that, in a bench-scale study employing O3/GAC, geosmin removals were accounted to be from 76% to 100% while 2-MIB removals were from 47% to 100%. Park et al. (2015) figured out that the O3/GAC process was capable of removing 94% to 100% of geosmin at 50–1,000 ng/L and 87% to 100% of 2-MIB at 50–300 ng/L. Instead of being used alone, O3/GAC was typically employed as an advanced treatment process other than the conventional treatment in numerous water plants. However, there are rare reports of the removal performance of geosmin and 2-MIB by O3/GAC in the overall treatment processes.

This study aims to evaluate the effectiveness of the O3/GAC and conventional treatment process for the removal of 2-MIB/geosmin in a pilot-scale plant to provide useful information for practical applications to solve the taste and odor problem during algal blooms in Lake Taihu. The effect of ozone dosage, together with initial concentrations of geosmin/2-MIB on the removal of geosmin/2-MIB by both the conventional processes and O3/GAC treatment were investigated. The relationship between the removal performance of dissolved organic compound (DOC) and geosmin/2-MIB was also discussed.

MATERIALS AND METHODS

Materials

The mixed stock solutions of 2-MIB and geosmin (100 μg/mL dissolved in methanol, Supelco, PA) were diluted to the desirable concentration with Milli-Q water as per requirement. Raw water was prepared by dosing 2-MIB and geosmin mixed solution at a specific concentration as taste and odor compounds. The coagulant, poly aluminium chloride (supplied by the water plant) was also prepared.

Analytical methods of geosmin and 2-MIB

Geosmin and 2-MIB were analysed using solid-phase micro-extraction gas chromatography/mass spectrometry (SPME-GC/MS). Based on the methodology from Watson et al. (2000), the SPME analysis was performed using a 30 mL sample after a 0.45 μm filter and 9.0 g NaCl in a 40 mL septum capped vial. An internal standard of 3-isobutyl-2-methoxypyrazine (IBMP of 30 ng/L, Supelco, PA), together with a small PTFE-coated stirring bar, was added in the vial. The extraction was conducted in a water bath with a magnetic stirrer maintained at 65°C and a constant speed of 800 RPM. DVB/CAR/PDMS fibre (Supelco #57328 U) was incorporated into the headspace of the vial followed by exposure of 30 minutes. Following the extraction, the fibre was injected to the gas chromatography/mass spectrometer (GC/MS, Trace DSQ, and Thermo Fisher, USA) and desorbed in splitless mode at 250°C with a 2-min period. The gas chromatograph column (HP-5MS, 30 m × 0.25 mm × 0.25 μm, Supelco, PA) was maintained at 40°C for a 2-min duration, followed by programming with 10°C/min to 250°C and remained for 2 min. As regards the selected ion monitoring (SIM) mode, three ions were monitored (m/z 95, for 2-MIB; m/z 112, for geosmin; and m/z 167, for IBMP). The threshold concentrations for 2-MIB and geosmin were reported to be 6.3 and 1.3 ng/L, respectively (Young et al. 1996).

High-performance size-exclusion chromatography (HPSEC) analysis

DOC fraction characteristics were examined using HPSEC analysis (Alliance HPLC system, Waters, USA) with the methodology adopted from Chin et al. (1994). Analysis of DOC was carried out by HPSEC using a Sievers 900 online total organic carbon (TOC) analyser (GE, USA). UV absorbance was measured at a wavelength of 254 nm for HPSEC detection (2489 UV/VIS detector, Waters, USA).

Pilot system

The pilot unit for this study was developed in Suzhou New District Water Treatment Plant (WTP), China, which was comprised of pre-ozonation, flocculation, sedimentation, sand filtration, post-ozonation and GAC filtration (Figure 1). The capacity of the pilot plant is around 3.0 m³/day. The pilot system was operated with the influent water containing geosmin and 2-MIB. Poly aluminium chloride was used as a coagulant with a feed Al concentration of approximately 40 mg/L according to past tests. Sedimentation was implemented with a hydraulic retention time of 40 minutes and a sludge retention time of 20 hours. The hydraulic retention times for sand filtration and GAC filtration were 17 and 40 minutes, respectively. Ozone was produced using an ozone generator (Pacific Ozone Technology, G11), which was fed with pure and dry oxygen, followed by injection into the main ozone contact reactor with a volume of 1,500 L for a contact duration of 7 min. An ozone analyser (BMT-964, Germany) was used for the measurement of the ozone concentration in both the pre-ozone and post-ozone reactors at the fixed time over the reaction period. In the current study, the ozone concentration for pre-ozonation and post-ozonation was maintained the same.

Figure 1

A schematic diagram of the pilot-plant system.

Figure 1

A schematic diagram of the pilot-plant system.

Raw water characteristics

The pilot plant was located at Suzhou New District WTP with the influent taken from Lake Taihu. Water quality characteristics during the test period are shown in Table 1.

Table 1

Raw water quality characteristics and analysis methods

ParameterMonthly average valueMethod
Temperature (°C) 31.6 Thermometer 
pH 7.3 METTLER-TOLEDO SevenEasy S20 K (Switzerland) 
Turbidity (NTU) 15.23 Hach 2100Q (USA) 
TOC (mg/L) 3.59 Shimadzu TOC-L (Japan) 
UV254 (cm-10.050 UV Spectrophotometer Hach DR5000 (USA) 
DO (mg/L) 21.22 HQ10 Portable LDOTM Dissolved Oxygen Meter, Hach (USA) 
ParameterMonthly average valueMethod
Temperature (°C) 31.6 Thermometer 
pH 7.3 METTLER-TOLEDO SevenEasy S20 K (Switzerland) 
Turbidity (NTU) 15.23 Hach 2100Q (USA) 
TOC (mg/L) 3.59 Shimadzu TOC-L (Japan) 
UV254 (cm-10.050 UV Spectrophotometer Hach DR5000 (USA) 
DO (mg/L) 21.22 HQ10 Portable LDOTM Dissolved Oxygen Meter, Hach (USA) 

RESULTS AND DISCUSSION

Effects of ozone dosages on 2-MIB/geosmin removal

Figure 2 shows the removal efficiencies of geosmin and 2-MIB in the ozonation process with different ozone concentrations varying from 0 to 1.0 mg/L. The initial concentrations for both chemicals were ∼200 ng/L.

Figure 2

Variation and removal efficiencies of (a) geosmin and (b) 2-MIB in each process for different ozone dosages.

Figure 2

Variation and removal efficiencies of (a) geosmin and (b) 2-MIB in each process for different ozone dosages.

Figure 3

Representative HPSEC chromatograms highlighting the MW distribution of TOC for raw water with different ozone dosages.

Figure 3

Representative HPSEC chromatograms highlighting the MW distribution of TOC for raw water with different ozone dosages.

With the increase of the ozone dosage from 0 to 0.5 mg/L, removal of 2-MIB by pre-ozonation did not exhibit substantial change. With the higher ozone dose of 1.0 mg/L, a removal percentage of 32% for 2-MIB was observed, which almost reached four times higher than that with 0.5 mg/L O3. Compared with 2-MIB, removal of geosmin by pre-ozonation exhibited more rapid increase with ozone dosage and achieved 30.99% with 1.0 mg/L O3. Likewise, Ho et al. (2004) found geosmin was oxidized faster than 2-MIB with 2 mg/L O3 dosage, which was attributed to the greater steric hindrance in the structure of 2-MIB.

With no ozone dosage, filter efficiencies were examined to be 19% and 16% for 2-MIB and geosmin and were substantially increased to 34% and 37% with 1.0 mg/L O3, respectively. Ho et al. (2007) demonstrated that 2-MIB and geosmin were removed in bench-scale sand filters predominantly through a biodegradation process. Ozone promoted the establishment of biofilm on the sand media, subsequently enhancing the filter's ability to remove 2-MIB and geosmin (De Waters & DiGiano 1990). Conversely, removal of 2-MIB/geosmin by GAC filter was observed to decrease from 37% to 4% and from 38% to 6%, respectively with O3 dosage increasing from 0 to 1.0 mg/L. This was probably due to the competitive impact of the sand filter, which has higher biomass densities than the GAC filter. A similar result was reported by Westerhoff et al. (2005), suggesting that 2-MIB removals achieved higher removal efficiencies in slow sand filters than in GAC filters.

Generally, the conventional treatment process removed 38% to 59% and 36% to 64% of 2-MIB and geosmin, respectively, with ozone dosages increasing from 0 to 1.0 mg/L. At ∼200 ng/L of 2-MIB and geosmin in raw water, a 1.0 mg/L O3 dose was sufficient for the destruction of 2-MIB/geosmin below their OTCs in the treated water.

Effect of ozone dosages on molecular weight (MW) distribution of DOC

Figures 38 show the HPSEC chromatograms in terms of DOC concentrations in the raw, pre-ozonated, sedimentation, sand filtration, post-ozonated and GAC filtration water with regard to different ozone dosages. In this study, TOC in water samples was fractionated and characterized into three types: Biopolymer (BP), Middle Molecular Weight (MMW) and Low Molecular Weight (LMW) matter in accordance with its MW and UV absorbance based on the methodology suggested by Gibert et al. (2013).

It can be seen that pre-ozonation induced a substantial reduction in BP fraction with a MW from 104 to 106 Da with the increase of ozone dosage. This fraction primarily represents the natural organic matter (NOM) of high MW including polysaccharides, proteins and amino sugars that are typically oxidized by ozone into lower molecular weight biodegradable organics (Figure 4). This could be the reason for the results in Figure 2 that better 2-MIB/geosmin removal was achieved with higher ozone dosage. As suggested by previous studies, the NOM with higher molecular weight and specific UV absorbance (SUVA) improved the oxidation efficiency of 2-MIB/geosmin through the promotion of the production of ·OH radicals, which was identified as the dominant process over the ozonation of 2-MIB/geosmin (Westerhoff et al. 1999).

Figure 4

Representative HPSEC chromatograms highlighting the MW distribution of TOC for pre-ozonated water with different ozone dosages.

Figure 4

Representative HPSEC chromatograms highlighting the MW distribution of TOC for pre-ozonated water with different ozone dosages.

In the sedimentation effluent, the peak value of BP matter was lowered to 0 with 1.0 mg/L O3, indicating that the BP matter had been fully removed (Figure 5). Moreover, the peak value of LMW matter exhibited a substantial decrease in sand filtration effluent and increased significantly in post-ozonated effluent with 1.0 mg/L O3 (Figures 6 and 7). In the GAC filtered sample (Figure 8), organics with MW of 104 Da exhibited a small increase with 0.5 mg/L O3, indicating that the existence of these organics diminished the removal efficiency of 2-MIB/geosmin in the GAC filter with 0.5 mg/L O3 (Figure 2). It has been reported that NOM with similar particle size as 2-MIB is capable of competing with 2-MIB/geosmin for adsorption sites on carbon surfaces (Newcombe et al. 2002).

Figure 5

Representative HPSEC chromatograms highlighting the MW distribution of TOC for sedimentation water with different ozone dosages.

Figure 5

Representative HPSEC chromatograms highlighting the MW distribution of TOC for sedimentation water with different ozone dosages.

Figure 6

Representative HPSEC chromatograms highlighting the MW distribution of TOC for sand filtration water with different ozone dosages.

Figure 6

Representative HPSEC chromatograms highlighting the MW distribution of TOC for sand filtration water with different ozone dosages.

Figure 7

Representative HPSEC chromatograms highlighting the MW distribution of TOC for post-ozonated water with different ozone dosages.

Figure 7

Representative HPSEC chromatograms highlighting the MW distribution of TOC for post-ozonated water with different ozone dosages.

Figure 8

Representative HPSEC chromatograms highlighting the MW distribution of TOC for GAC filtration water with different ozone dosages.

Figure 8

Representative HPSEC chromatograms highlighting the MW distribution of TOC for GAC filtration water with different ozone dosages.

The DOC removal efficiencies in all treatment processes under different ozone dosages are also examined and the result is shown in Figure 9. With ozone increasing from 0 to 1.0 mg/L, the removal of DOC by pre-ozonation decreased from 4% to −7%, which is probably due to the release of intracellular organic matter from the broken algal cells during pre-ozonation. In the present experiment, the effect of ozone as a coagulation aid on DOC removal was observed with ozone dosage of 1.0 mg/L. In general, the highest DOC removal was achieved in coagulation and sedimentation with the value from 13% to 21% with ozone dosage increasing from 0 to 1.0 mg/L, while the removal efficiencies of DOC were the weakest in pre-ozonation and post-ozonation (−7% to 4% and −3% to 6%, respectively).

Figure 9

DOC removal efficiencies in treatment process units with different ozone dosages.

Figure 9

DOC removal efficiencies in treatment process units with different ozone dosages.

Effects of initial concentrations on 2-MIB/geosmin removal

Figure 10 shows the residual concentrations and removal efficiencies of geosmin and 2-MIB in the pilot-scale treatment processes with variation of their initial concentrations of ∼50, ∼70, and ∼200 ng/L with a constant O3 dosage of 1.0 mg/L.

Figure 10

Variation and removal efficiencies of (a) geosmin and (b) 2-MIB with both conventional process and advanced treatment with different initial concentrations.

Figure 10

Variation and removal efficiencies of (a) geosmin and (b) 2-MIB with both conventional process and advanced treatment with different initial concentrations.

Regarding the effects of initial concentrations on geosmin and 2-MIB oxidation, it was found that the initial concentration increase did not promote 2-MIB removal by pre-ozonation, conversely the geosmin removal efficiency increased from 12.3% to 31.0%. Moreover, a decline of approximately 12% in the removal efficiencies of both compounds was noticed in the sand filtration with the increase of initial concentrations from 50 ng/L to 200 ng/L. Likewise, the efficiencies of GAC filtration in removing 2-MIB and geosmin exhibited a decline from 19% to 4% and from 11% to 5%, respectively, with the increase of the initial concentrations from 50 ng/L to 200 ng/L. With lower initial concentrations of 50 ng/L and 70 ng/L, the removal of geosmin by GAC filtration was higher than that of 2-MIB, which was probably due to the slightly lower solubility and molecular weight, in addition to the flatter structure of geosmin (Cook et al. 2001). Nevertheless, at the higher initial concentration of 200 ng/L, the discrepancy was lowered.

Overall, sand filtration was the most effective for 2-MIB/geosmin removal (37% geosmin and 34% MIB with initial concentrations of ∼200 ng/L). Pre-ozonation contributed 20% to 30% removal of 2-MIB/geosmin. With subsequent GAC filtration, 2-MIB and geosmin were successfully reduced below their threshold odor levels.

CONCLUSION

Pilot-plant studies were carried out for the evaluation of the removal efficiencies of 2-MIB/geosmin by O3/GAC advanced treatment process. The removal efficiency of 2-MIB/geosmin by pre-ozonation increased in proportion to the ozone dosages and the initial concentrations of both chemicals. Additionally, geosmin was found to have a higher removal efficiency by pre-ozonation as compared with 2-MIB. The sand filtration enhanced the removal of 2-MIB/geosmin with the increase of ozone dosages, meanwhile exhibiting an opposite trend in proportion to the initial concentrations. As compared with the sand filtration, the removal of 2-MIB/geosmin by the GAC filter exhibited a decline with increase in the O3 dosage. Geosmin showed better adsorption efficiency than 2-MIB by GAC filtration. The conventional treatment process removed a substantial part of 2-MIB/geosmin; furthermore, the removals showed increased values of 38% to 59% and 36% to 64% for 2-MIB and geosmin, respectively, with increased ozone dosages. An O3 dosage of 1.0 mg/L was sufficient for 200 ng/L 2-MIB and geosmin in the influent water to be lowered below their threshold levels. The TOC removal of the pre-ozonation, together with conventional treatment, exhibited increase with ozone dosages; conversely, with the GAC filtration an opposing tendency was observed. Pre-ozonation lowered the response of organic compounds with MW from 104 to 106 Da. With 1.0 mg/L O3 dosage, low molecular weight organics were substantially removed by the sand filtration. In the GAC filtrated water, organics of higher MW were basically removed, while the organic matter with MW of 104 Da increased with the 0.5 mg/L O3 dose. An O3 dose of 1.0 mg/L was sufficient for 200 ng/L 2-MIB and geosmin in the influent water to be below their threshold levels. The O3/GAC advanced treatment was examined to be an effective method for the removal of 2-MIB and geosmin in drinking water treatment.

ACKNOWLEDGEMENTS

This work was supported by the International Science and Technology Cooperation Program of China (2016YFE0123700), the Major Science and Technology Program for Water Pollution Control and Treatment (2017ZX07201001), the National Natural Science Foundation of China (No. 51708130) and the Key Laboratory of Yangtze River Water Environment, Ministry of Education.

REFERENCES

REFERENCES
Antonopoulou
M.
,
Evgenidou
E.
,
Lambropoulou
D.
&
Konstantinou
I
, .
2014
A review on advanced oxidation processes for the removal of taste and odor compounds from aqueous media
.
Water Research
53
,
215
234
.
Chin
Y. P.
,
Aiken
G.
&
O'Loughlin
E
, .
1994
Molecular weight, polydispersity, and spectroscopic properties of aquatic humic substances
.
Environmental Science & Technology
28
(
11
),
1853
1858
.
De Waters
J. E.
&
DiGiano
F. A
, .
1990
The influence of ozonated natural organic matter on the biodegradation of a micropollutant in a GAC bed
.
Journal (American Water Works Association
)
82
(
8
),
69
75
.
Elhadi
S. L. N.
,
Huck
P. M.
&
Slawson
R. M
, .
2004
Removal of geosmin and 2-methylisoborneol by biological filtration
.
Water Science & Technology
49
(
9
),
273
280
.
Gibert
O.
,
Lefèvre
B.
,
Fernández
M.
,
Bernat
X.
,
Paraira
M.
&
Pons
M
, .
2013
Fractionation and removal of dissolved organic carbon in a full-scale granular activated carbon filter used for drinking water production
.
Water Research
47
(
8
),
2821
2829
.
He
Q.
,
Wang
H.
,
Zhong
L.
,
Yang
K.
&
Zou
Z
, .
2017
The use of potassium permanganate, ozone and associated coupled processes for odor removal in drinking water: bench and pilot scale tests
.
Journal of Water Supply: Research and Technology - AQUA
66
(
4
),
249
256
.
Ho
L.
,
Croué
J.-P.
&
Newcombe
G
, .
2004
The effect of water quality and NOM character on the ozonation of 2-MIB and geosmin
.
Water Science & Technology
49
(
9
),
249
255
.
Ho
L.
,
Hoefel
D.
,
Bock
F.
,
Saint
C. P.
&
Newcombe
G
, .
2007
Biodegradation rates of 2-methylisoborneol (2-MIB) and geosmin through sand filters and in bioreactors
.
Chemosphere
66
(
11
),
2210
2218
.
Newcombe
G.
,
Morrison
J.
,
Hepplewhite
C.
&
Knappe
D. R. U
, .
2002
Simultaneous adsorption of MIB and NOM onto activated carbon: II. Competitive effects
.
Carbon
40
(
12
),
2147
2156
.
Su
M.
,
Yu
J.
,
Zhang
J.
,
Chen
H.
,
An
W.
,
Vogt
R. D.
,
Andersen
T.
,
Jia
D.
,
Wang
J.
&
Yang
M
, .
2015
MIB-producing cyanobacteria (Planktothrix sp.) in a drinking water reservoir: distribution and odor producing potential
.
Water Research
68
,
444
453
.
Takeuchi
Y.
,
Mochidzuki
K.
,
Matsunobu
N.
,
Kojima
R.
,
Motohashi
H.
&
Yoshimoto
S
, .
1997
Removal of organic substances from water by ozone treatment followed by biological activated carbon treatment
.
Water Science & Technology
35
(
7
),
171
178
.
Watson
S. B.
,
Brownlee
B.
,
Satchwill
T.
&
Hargesheimer
E. E
, .
2000
Quantitative analysis of trace levels of geosmin and MIB in source and drinking water using headspace SPME
.
Water Research
34
(
10
),
2818
2828
.
Westerhoff
P.
,
Summers
R. S.
,
Chowdhury
Z.
&
Kommineni
S.
2005
Ozone-Enhanced Biofiltration for Geosmin and MIB Removal
.
Report No. 91075
,
AWWA Research Foundation and American Water Works Association
,
Denver, CO
,
USA
.
Young
W. F.
,
Horth
H.
,
Crane
R. A.
,
Ogden
T.
&
Arnott
M
, .
1996
Taste and odour threshold concentrations of potential potable water contaminants
.
Water Research
30
(
2
),
331
340
.
Zamyadi
A.
,
Henderson
R.
,
Stuetz
R.
,
Hofmann
R.
,
Ho
L.
&
Newcombe
G
, .
2015
Fate of geosmin and 2-methylisoborneol in full-scale water treatment plants
.
Water Research
83
,
171
183
.