Removal of N-nitrosamines in a membrane bioreactor and nanofiltration hybrid system for municipal wastewater reclamation: Process efficiency and mechanisms

Accepted Manuscript Removal of N-nitrosamines in a membrane bioreactor and nanofiltration hybrid system for municipal wastewater reclamation: Process efficiency and mechanisms Kangmin Chon, Sung Hyun Kim, Jaeweon Cho PII: DOI: Reference:

S0960-8524(15)00266-7 http://dx.doi.org/10.1016/j.biortech.2015.02.080 BITE 14653

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

31 December 2014 19 February 2015 20 February 2015

Please cite this article as: Chon, K., Kim, S.H., Cho, J., Removal of N-nitrosamines in a membrane bioreactor and nanofiltration hybrid system for municipal wastewater reclamation: Process efficiency and mechanisms, Bioresource Technology (2015), doi: http://dx.doi.org/10.1016/j.biortech.2015.02.080

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Removal of N-nitrosamines in a membrane bioreactor and nanofiltration hybrid system 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

for municipal wastewater reclamation: Process efficiency and mechanisms

Kangmin Chon a,b, Sung Hyun Kim c, Jaeweon Cho d,*

a

Jeju Global Research Center (JGRC), Korea Institute of Energy Research (KIER), 200 Haemajihaean-

ro, Gujwa-eup, Jeju-si, Jeju-do 695-971, Republic of Korea b

School of Environmental Science and Engineering, Gwangju Institute of Science and Technology

(GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712, Republic of Korea c

Industry Service Research Center, World Institute of Kimchi (WiKim), 86 kimchiro, Namgu, Gwangju

503-360, Republic of Korea d

Department of Civil and Environmental Engineering, College of Engineering, Yonsei University, 50

Yonsei-ro, Seodaemun-gu, Seoul 120-749, Republic of Korea

* Corresponding author. Tel.: +82-2-2123-8296; Fax: +82-2-364-5300. E-mail address: [email protected] (J. Cho).

ABSTRACT This study investigated the removal efficiency and mechanisms of water contaminants (mainly Nnitrosamines) during municipal wastewater reclamation by a membrane bioreactor (MBR) and nanofiltration (NF) hybrid system. The removal of bulk water contaminants was governed by the microbial activities in the MBR and molecular weight cut-off (MWCO) of the NF membranes. The removal of N-nitrosamines by the MBR was primarily attributed to biodegradation by aerobic bacteria, which can be determined by the reactivity of the amine functional groups with the catabolic enzymes (removal efficiency = 45 – 84 %). Adsorption and formation of membrane fouling can enhance the 1

removal of N-nitrosamines by the NF membranes. However, size-exclusion is found to play a major role 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

in the removal of N-nitrosamines by the NF membranes since the removal efficiencies of N-nitrosamines varied significantly depending on molecular weight of the N-nitrosamines and MWCO of the NF membranes (removal efficiency: NE90 > NE70).

Keywords: Biodegradation; Membrane bioreactor; Nanofiltration; N-nitrosamines; Size-exclusion

1. Introduction Municipal wastewater reclamation is an important strategy for securing a sustainable water resource to solve the water scarcity crisis induced by the combination of climate changes and rapid population growth from urbanization and industrialization (Asano et al., 2007). Nevertheless, the qualities of reclaimed wastewater are strictly regulated by many water authorities since municipal wastewater is a primary source of waterborne pathogens, organic materials, heavy metals, metalloids, nutrients (i.e., nitrogen and phosphorous species), and trace organic contaminants (i.e., pharmaceuticals and personal care products (PPCPs), endocrine disrupting chemicals, and disinfection byproducts) (Kolpin et al., 2002; Australian Capital Territory, 2004; California Department of Public Health (CDPH), 2013; Chon et al., 2014). During the past decades, wastewater treatment technologies have mainly focused on the removal of waterborne pathogens, and nutrients (i.e., carbon, nitrogen and phosphorous species) for both protection of public health and prevention of environmental pollution (Asano et al., 2007). In recent years, the occurrence and removal of trace organic contaminants have gained great attention in municipal wastewater reclamation as they can cause adverse effects on human health and aquatic ecosystems (Ternes and Joss, 2006; Asano et al., 2007; Chon et al., 2013a). Although trace organic contaminants are present at very low concentrations in municipal wastewater (i.e., low ng/L to µg/L range), long-term exposure to them may lead chronic health problems on humans and animals (Kolpin et al., 2002; Ternes and Joss, 2006). Furthermore, conventional wastewater treatment processes,

2

including activated sludge, coagulation and floatation are not effective for the removal of trace organic 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

contaminants (Carballa et al., 2005). Therefore, integrated membrane systems consisting of microfiltration (MF)/ultrafiltration (UF) and reverse osmosis (RO) membranes or membrane bioreactor (MBR) and RO membranes have been widely used for municipal wastewater reclamation due to their economy feasibility and high quality effluents (Dolar et al., 2012; Chon et al., 2013a). Among various membrane processes, MBR systems which combine an activated sludge process with membrane filtration separation are considered as a promising option for municipal wastewater reclamation since they produce much less amount of sludge and require a smaller footprint than the conventional biological wastewater treatment consisting of an activated sludge process and a secondary sedimentation (Chon et al., 2011;; Chon et al., 2012a). Most of water contaminants, including particulates, colloids, suspended solids, organic materials, and ammonium, can be more efficiently removed by MBR systems than conventional activated sludge (CAS) processes (Xing et al., 2000). In addition, relatively higher mixed liquor suspended solids (MLSS) concentration and longer sludge retention time in MBR systems compared to CAS processes may intensify biodegradation of trace organic contaminants by activated sludge (Clara et al., 2005). However, some of trace organic contaminants (e.g., carbamazepine, diclofenac, nonylphenoxyacetic and nonylphenoxyethoxy acetic acids) cannot be effectively degraded using MBR processes as they are recalcitrant to biological wastewater treatment techniques (Clara et al., 2005; Chon et al., 2012a). Based on these reasons, the permeates from MBR processes have to be further treated by high pressure membranes (i.e., nanofiltration (NF) and RO) for ensuring removals of both biodegradable and non-biodegradable trace organic contaminants (Clara et al., 2005; Asano et al., 2007; Chon et al., 2012a). A wide range of dissolved water contaminants, including heavy metals, metalloids, and most trace organic contaminants, can be successfully removed from municipal wastewater by NF and RO membranes (Chon et al., 2012a; Chon et al., 2013a). Nevertheless, N-nitrosamines are insufficiently removed by high pressure membranes since they are small, uncharged and hydrophilic compounds

3

(Asano et al., 2007; Fujioka et al., 2013a; Fujioka et al., 2013b). In addition, the use of disinfection 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

processes (e.g., chlorination and chloramination) in municipal wastewater treatment plants may increase concentrations of N-nitrosamines in municipal wastewater effluents (Mitch and Sedlak, 2002; Wijekoon et al., 2013; Chon et al., 2013b). N-nitrosamines, including N-nitrosodimethylamine (NDMA), Nnitrosopyrrolidine

(NPYR),

N-nitrosodiethylamine

(NDEA),

N-nitrosopiperidine

(NPIP),

N-

nitrosomorpholine (NMOR), N-nitrosodibutylamine (NDBA), and N-nitrosodiphenylamine (NDPHA), have been categorized as an emerging class of trace organic contaminants by the United States Environmental Protection Agency (U.S. EPA) due to their high potential carcinogenic risks to human (U.S. EPA, 1993). Their carcinogenic effects have observed even at a few ng/L levels (Sacher et al., 2008). Hence, the CDPH has set a 10 ng/L as maximum admissible levels in drinking water for NDMA, NDEA, and NDPA (CDPH, 2013) and the Australian Guidelines for Water Recycling has recommended that maximum concentrations of NDMA, NDEA, and NMOR in recycled water for potable uses do not exceed 10 ng/L, 10 ng/L, and 1 ng/L, respectively (NRMMC et al., 2009). Despite of a high mutagenic and carcinogenic potential, only a few researchers have investigated the removal of N-nitrosamines from municipal wastewater by biological and/or physico-chemical processes (Krauss et al., 2009; Fujioka et al., 2013a; Wijekoon et al., 2013). Removal efficiencies of Nnitrosamines in 21 full-scale wastewater treatment plants using CAS processes varied from 40 to 60 % (Krauss et al., 2009). The removal efficiencies of N-nitrosamines by RO membranes appear to be highly variable depending on the types of N-nitrosamines. The removal efficiencies of NDMA, NDEA, and NMOR ranged from 4 – 47 %, 0 – 53 %, and 81 – 84 %, respectively (Fujioka et al., 2013a). Some researchers found that feed water characteristics (i.e.., temperature), membrane properties (i.e., pore size and thickness of active skin layers), and operating conditions (i.e., recovery rates, permeate flux, and membrane fouling) may exert considerable effects on the removal of N-nitrosamines by RO membranes (Fujioka et al., 2012; Fujioka et al., 2013a; Fujioka et al., 2013b; Fujioka et al., 2013c). Even though NF membranes may produce high quality effluents which are similar to permeates of RO membranes at

4

relatively low pressure, very little is known about the removal of N-nitrosamines by integrated 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

membrane systems consisting of a MBR and NF membranes compared to CAS/MBR and RO membrane processes as most of previous studies have mainly focused on the removal of PPCPs and identification of fouling behavior in MBR and NF hybrid systems (Chon et al., 2012a; Chon et al., 2013c). The main objective of this study is to provide deeper insights into the removal behavior of Nnitrosamines by a MBR and NF hybrid system for municipal wastewater reclamation. Therefore, the concentration and removal of N-nitrosamines in the MBR and NF hybrid system were investigated and directly correlated to the physico-chemical characteristics of N-nitrosamines and membranes to elucidate their removal mechanisms. Furthermore, the removal efficiencies of bulk water contaminants, including organic, inorganic constituents, and dissolved organic matter (DOM), were identified to evaluate the performance of the MBR and NF hybrid system for municipal wastewater reclamation.

2. Materials and methods 2.1. MBR and NF hybrid system description A laboratory-scale MBR and NF hybrid system consisted of one submerged MBR and arranged two NF units equipped with two different flat-sheet NF membranes (Fig. 1). The physico-chemical properties of the MF and NF membranes are listed in Table 1 (Chon et al., 2012a; Chon et al., 2013c). Primary effluents collected from Gwangju wastewater treatment plant (WWTP; Gwangju, Korea) was fed to the submerged MBR (effective volume = 12.0 L; hydraulic retention time = 12.5 h; temperature = 21 ± 1 °C; dissolved oxygen = 7.1 ± 1.2 mg/L) equipped with a U-shaped hollow fiber MF membrane (effective surface area = 502.4 cm2) and further treated using two different kinds of NF membranes (effective surface area of each NF membrane = 58.2 cm2), including NE70 and NE90 membranes (Toray Chemical, Seoul, Korea). Concentrations of mixed liquor suspended solids in the MBR were maintained approximately 2533±73 mg/L (sludge retention time = 30 d) to keep a stable permeate flux without chemical cleaning and the MBR was operated under continuous aeration (flow rate = 8 L/min) at and

5

an intermittent suction mode (suction = 9 min; backwashing = 1 min) to reduce fouling formation on the 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

surfaces on the submerged MF membranes. To avoid the photolytic degradation of N-nitrosamines, the MBR and reservoirs for the MBR and NF permeates were made of polyvinyl chloride and covered with aluminum foils. The operating temperature, retentate flux, and transmembrane pressure (TMP) of the NF membranes were maintained at 20±2 °C, 500 mL/min, and 350 kPa, respectively (Chon et al., 2013c).

2.2. Collection of samples Four different water samples, including primary effluent (Feed), MBR permeate (MBR), NE70 permeate (NE70), and NE90 permeate (NE90), were collected 3 times from the MBR and NF hybrid system over the testing period of 2 month to monitor changes in bulk water quality parameters and characteristics of DOM. The collection of samples for the analysis of N-nitrosamines was conducted between 3 to 6 weeks with two comprehensive sampling campaigns after the beginning of the operation.

2.3. Measurement of N-nitrosamines 2.3.1. Reagents Seven different types of N-nitrosamines, including NDMA, NPYR, NDEA, NPIP, NMOR, NDBA, and NDPHA, were provided from Supelco (Bellefonte, PA, USA). Two types of isotope-labeled standards, including NYPR-d8 (Purity = 98%) and NDEA-d10 (Purity = 98 %), were purchased from Cambridge Isotope Laboratory (Andover, MA, USA). Methanol (HPLC grade), acetonitrile (HPLC grade) and dichloromethane (HPLC grade) were obtained from Fisher Scientific (Nepean, ON, Canada), and formic acid, sodium azide, and L-ascorbic acid were provided from Fluka (Buchs, Switzerland) and Sigma-Aldrich (St Louis, MO, USA), respectively. The physico-chemical properties of the selected Nnitrosamines are provided in Table S1 in Supplementary Information (SI).

2.3.2. Preparation of standard solutions and samples.

6

The unlabeled and isotope-labeled N-nitrosamine standard stock solutions were prepared in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

methanol (concentration of each N-nitrosamine = 2 mg/L; concentration of each isotope-labeled Nnitrosamine = 2 mg/L) and stored at –20 °C. The unlabeled N-nitrosamine standard solutions were prepared using methanol (concentration = 0.00, 0.95, 1.90, 4.75, 9.50, 19.00, 47.50, and 95.00 µg/L) and each standard solution contains 100 µg/L of the isotope-labeled N-nitrosamines. Raw and treated wastewaters were collected in amber glass bottles (volume = 1 L). The collected samples were treated using 1 mL of 130 g/L sodium azide (final concentration = 2 mM/L) to prevent microbial degradation, and then filtered with 0.7 µm glass-fiber filter (GF/F, Whatman, Maid stone, Kent, UK). The filtered samples were divided into two subsamples (volume of each subsample = 500 mL) and stored at 4 °C prior to extraction. To calculate recovery rates of the N-nitrosamines, 25 µL of the isotope-labeled Nnitrosamine stock solution was spiked in two subsamples (final concentration of each isotope-labeled Nnitrosamine = 100 µg/L) and 25 µL of the unlabeled N-nitrosamine stock solution was spiked in one of two subsamples (final concentration of each N-nitrosamine = 100 µg/L). All the procedures for preparing the standard solutions and samples were performed just before high resolution gas chromatography mass spectrometry (HR-GC/MS) analysis.

2.3.3. Solid phase extraction (SPE) An Auto Trace automated SPE system (Claiper Corporation, Hopkington, MA, USA) using Super clean SPE cartridges (volume = 6 mL) packed with 2 g of 80-120 mesh coconut charcoal (Supelco, Bellefonte, PA, USA) was employed to extract N-nitrosamines from raw and treated wastewaters. The detailed procedures were provided in United StatesEnvironmental Protection Agency (EPA) Method 521 proposed by Munch and Basset (2004). Each SPE cartridge was preconditioned with 10 mL of dichloromethane and 10 mL of methanol, and residual solvents were removed under the vacuum. Nnitrosamines in the samples were extracted by the SPE cartridges at a flow rate of 5 mL/min. After loading the samples, the SPE cartridges were rinsed using 5 mL of deionized (DI) water and dried with

7

air for 3 hours. The absorbed N-nitrosamines on the SPE cartridges were eluted using 10 mL of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

dichloromethane. The eluted solutions were concentrated in a water bath at 40 °C with ultrapure nitrogen gas (purity = 99.999%, Praxair Korea, Kiheung, Gyeonggi-do, Korea) to less than 500 µL and then they were brought to final volume of 500 µL with methanol.

2.3.4. HR-GC/MS analysis Concentrations of the N-nitrosamines were analyzed by a Trace Ultra gas chromatography (Thermo Fisher Scientific, San Jose, CA, USA) equipped with a DB-5MS capillary column (30m × 0.25 mm × 0.25 mm, J&W Scientific, Folsom, CA, USA) coupled to a DFS high resolution magnetic sector mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). The temperature of the gas chromatography oven was initially kept for 3 min at 35 °C, raised to 80 °C at 15 °C/min, and increased to 260 °C at 30°C/min and then held for 3 min (injection volume of samples = 2 μL; carrier gas = helium; flow rate = 1.0 mL/min). All the target N-nitrosamines were ionized using electron ionization source in positive mode (electron energy = 35 eV) and current of the filament current was 0.5 mA. The temperatures of the source and the transfer line were 200 °C and 280 °C, respectively. Mass spectrometry was conducted at mass resolving power of 10,000 (5 % peak height) in multiple ion detection mode.

2.4. Analytical methods for bulk water quality parameters and DOM characterization Chemical oxygen demand (COD) was analyzed by a Water Analyzer (HS-3300, Humas, Daejeon, Korea) with the manganese III reactor digestion method at a wavelength of 540 nm. Concentrations of dissolved organic carbon (DOC) and total nitrogen (TN) were determined using a combustion-type total organic carbon (TOC) analyzer (TOC-VCPH, Shimadzu, Kyoto, Japan) coupled to a TNM-1 unit (Shimadzu, Kyoto, Japan). Concentration of calcium was measured by Ion chromatography (DX 120, Dionex, Sunny vale, CA, USA) and inductively coupled plasma-mass spectrometry (7500ce, Agilent, Santa Clara, CA, USA) was employed to quantify levels of metals and metalloids. The aromatic

8

moieties of DOM were quantified by a UV/Vis spectrophotometer (UV-1601, Shimadzu, Japan) at 254 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

nm and the specific UV absorbance (SUVA) values were calculated using the following equation: SUVA (L/mgC·m)= UV absorbance at 254 nm (UVA254;/cm) / DOC (mg/L) × 100. High-performance sizeexclusion chromatography (HPSEC) equipped with a Protein-Pak 125 column (Waters, Milford, MA, USA), UVA detection (SPD-10AVP, Shimadzu, Kyoto, Japan), and fluorescence detection (RF-10A XL, Shimadzu, Kyoto, Japan) was used to ascertain molecular weight (MW) distribution of aromatic and protein-like substances in the raw and treated wastewaters (Chon et al., 2013c). Fluorescence spectral properties of DOM were confirmed using a fluorescence spectrophotometer (F-2500, Hitachi, Tokyo, Japan) and the raman scattering peaks were corrected with a DI water blank (Chon et al., 2013d).

3. Results and discussion 3.1. Changes of bulk water quality parameters Selected water quality parameters (Table 2) in the feed and treated wastewaters were monitored during the operation of the MBR and NF hybrid system to evaluate the process performance for water quality improvement. The pH of the feed water was slightly reduced after the MBR since the alkalinity was consumed by nitrifying bacteria activities in the MBR (Chon et al., 2012a; Do et al., 2012). Water conductivity change was negligible during the MBR process while significant decreases of the conductivity were observed for the NF processes. The conductivity removal efficiency was higher for the NE90 membrane (74%) compared to the NE70 membrane (33 %) which could be explained by the corresponding molecular weight cut-off (MWCO) of the two membranes (i.e., NE90 = 210 Da; NE70 = 350 Da). During the MBR process, CODMn (removal efficiency = 40 %) and DOC (removal efficiency = 59 %) concentrations decreased considerably by the microbial activities. The removal of DOC was inversely proportional to MWCOs of the NF membranes (NE70 = 88 %; NE90 = 93 %) (Chon et al., 2012a). Consequently, the averaged DOC concentration in the NE90 permeates (0.40 mg/L) satisfied the

9

CDPH draft regulations on groundwater recharge (TOC < 0.5 mg/L). This demonstrates the good 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

performance of the MBR and NF hybrid system on the removal of bulk organic compounds.

3.2. Changes of DOM characteristics 3.2.1. HPSEC analysis MW distribution of DOM in the feed and treated wastewaters, in terms of aromatic and proteinlike substances, is depicted in Fig. 2. The aromatic and protein-like substances of DOM in the feed water included either low (aromatic substances = 670 – 920 daltions (Da); protein-like substances = 50 – 860 Da) or high MW fractions (aromatic substances = 35,200 Da; protein-like substances = 31,660 Da). The high MW fractions were mostly removed by the MBR whereas significant increases were observed for the low MW peaks (aromatic substances = 830 Da; protein-like substances = 930 Da), indicating that the high MW fractions were degraded into the low MW fractions by the microbial activities in the MBR (Chon et al., 2013c). Furthermore, MW distribution of DOM in the feed water was slightly shifted to lower MW after the MBR due to the enhanced size-exclusion effects by the fouling formation on the membrane surfaces which may lead to the increase in TMP and the decrease in the permeate flux of MBR processes (Banu et al., 2000; Banu et al., 2001; Chon et al., 2013a; Raj et al., 2013). Therefore, only a low MW fraction was found for the MBR permeate (aromatic substances = 830 Da; protein-like substances = 20 – 930 Da). Both the NE70 and NE90 membranes were effective for the removal of the low MW fractions. However, despite of the lower negative surface charge (NE70 = –18 mV; NE90 = – 15 mV), the low MW fractions of the aromatic substances with a negative charge were more effectively removed by the NE90 membrane than the NE70 membrane due to its lower MWCO (NE70 = 350 Da; NE90 = 210 Da) (Chon et al., 2012b; Chon et al., 2013c).

3.2.2. Spectroscopic analysis

10

Changes in the 3 dimensional fluorescence spectra of DOM using the MBR and NF hybrid 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

system are represented in Fig. S1 (SI). Three distinctive fluorescence peaks were observed for the feed water. The fluorescence peak at excitation (Ex) = 280 nm and emission (Em) = 340 nm (Feed I) was associated with protein-like fluorescence and the fluorescence peaks at Ex = 280 nm / Em = 410 nm (Feed II) and Ex = 340 nm / Em = 420 nm (Feed III) were indicative of humic-like fluorescence (Chon et al., 2011; Chon et al., 2013d). The fluorescence intensity of Feed II and III was poorly reduced after the MBR while that of Feed I was considerably decreased as protein-like substances are more easily degraded by microbial activities in the MBR than humic-like substances (Chon et al., 2013c). Proteinlike or humic-like fluorophores in the MBR permeate were mostly eliminated by both the NE70 and NE90 membranes. Nevertheless, the decreasing rates of humic-like fluorescence intensities by the NE70 membrane (MBR I = 80 %; MBR III = 83 %) were substantially lower than those of the NE90 membrane (MBR I = 90 %; MBR III = 91 %) as size-exclusion may play a critical role in the removal of DOM by the NF membranes. These removal trends were consistent with the results from previous studies (Chon et al., 2012a; Chon et al., 2013c), which ensure the adequate operation of the MBR and NF hybrid system to maintain its high effluent qualities for municipal wastewater reclamation. Although distinctive differences were observed for the SUVA values, changes in the aromatic moieties of DOM could be not inferred from the SUVA values since most of nitrogen species in MBR permeates generally exist in the form of nitrate ions due to the activities of nitrifying bacteria in bioreactors (Chon et al., 2011; Chon et al., 2012a; Chon et al., 2013c), which may interfere with accurate SUVA measurements (Potter and Wimsatt, 2005).

3.3. Removal of metals and metalloids Changes in the levels of metals and metalloids in the feed and treated wastewaters by the MBR and NF hybrid system are summarized in Table S2 (SI). The MBR process appears to be effective for the removal of boron (removal efficiency = 49 %), copper (removal efficiency = 51 %), and iron (removal

11

efficiency = 65 %) whereas some metals and metalloids were not sufficiently removed after the MBR 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

due to their very low concentrations (arsenic = 3.3 µg/L; cadmium = 0.2 µg/L; chromium = 0.2 µg/L; cobalt = 0.3 µg/L). Both the NE70 and NE90 membranes might efficiently remove most of metals and metalloids from the MBR permeate (i.e., arsenic, cobalt, chromium, copper, and iron). Nevertheless, higher removal efficiencies were achieved from the NE90 membrane as size-exclusion governs the removal of metals and metalloids by the NF membranes (Chon et al., 2012a), which was in good agreement with the removal patterns of conductivity and DOC. In the case of boron, it was poorly removed even by the NE 90 membrane (removal efficiency of NE70 < 3 %; removal efficiency of NE90 < 13 %) due to the fact that boron in the MBR permeate was in the form of non-ionized boric acid which have a small molecular size (pH of the MBR permeate = 5.4; pKa of boric acid = 9.2) (Chon et al., 2012a). These observations support the assumption that physico-chemical properties of the NF membranes and inorganic materials (i.e., MWCO and pKa) are key parameters affecting the removal of metals and metalloids by the NF membranes.

3.4. Removal of N-nitrosamines 3.4.1. Concentrations of N-nitrosamines in municipal wastewater The concentrations of the seven different N-nitrosamines in municipal wastewater and their transport through the MBR and NF hybrid system are illustrated in Fig. 3. Among the target Nnitrosamines, NDMA was found as the most abundant compound in the feed waters from two sampling campaigns (NDMA = 192 – 455 ng/L) whereas the concentrations of the other N-nitrosamines were less than 110 ng/L (NPYR = 71 – 105 ng/L, NDEA = 81 – 82 ng/L, NPIP = 16 – 19 ng/L, NMOR = 66 – 79 ng/L, NDBA = 63 – 93 ng/L, NDPHA = 95 – 106 ng/L). A great daily variation was previously found in concentrations of NDMA among seven different WWTPs in California (USA) and its concentrations ranged from 7 to 790 ng/L (median = 88 ng/L) in untreated wastewater samples (Sedlak et al., 2005). Loads of N-nitrosamines originating from human urine excretion to municipal wastewater only

12

accounted for a few ng/L (NDMA < 5 ng/L; NMOR < 5 ng/L; other N-nitrosamines < 1 ng/L) (Krauss et 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

al., 2009). Therefore, higher concentrations of N-nitrosamines above these background concentrations in municipal wastewater might be attributed to industrial and/or commercial discharges (Krauss et al., 2009). Since most of the selected N-nitrosamines (Log P of NDMA = –0.64; Log P of NPYR = 0.23; Log P of NDEA = 0.34; Log P of NPIP = 0.74; Log P of NMOR = –1.39), except NDBA (Log P = 2.31) and NDPHA (Log P = 3.16), are hydrophilic compounds and their Henry’s low constants are very low (NDMA = 1.2 × 10-6 m3·atm/mol; NPYR = 1.99 × 10-7; NDEA = 1.73 × 10-6 m3·atm/mol; NPIP = 2.81 × 10-7 m3·atm/mol; NMOR = 2.13 × 10-10 m3·atm/mol; NDBA = 9.96 × 10-6 m3·atm/mol; NDPHA = 1.38 × 10-5 m3·atm/mol), biosorption and volatilization of N-nitrosamines by activated sludge and aeration in the MBR are expected to be negligible (Table S1, SI). Based on these findings, it can be postulated that the removal of N-nitrosamines after the MBR is mainly due to biodegradation by aerobic microbial activities in the MBR. Although NF membranes was implemented to meet the Australian Guidelines for Water Recycling, the concentrations of NDMA and NMOR in the NE70 (NDMA = 22 – 66 ng/L; NMOR = 13 – 14 ng/L) and NE90 permeates (NDMA = 21 – 63 ng/L; NMOR = 11 – 12 ng/L) were higher than 10 ng/L and 1 ng/L, respectively (NRMMC et al., 2009). This implies that an additional subsequent treatment (e.g., advanced oxidation process) is required for portable uses of the reclaimed wastewater.

3.4.2. Removal mechanisms of N-nitrosamines using MBR In order to elucidate the removal mechanisms of the target N-nitrosamines (i.e., NDMA, NPYR, NDEA, NPIP, NMOR, NDBA, and NDPHA) by the MBR, their removal efficiencies were correlated to the physico-chemical (i.e., MW and hydrophobicity (LogP)) and structural properties (i.e., functional group composition) (Fig. 4). In spite of the relatively low MLSS concentration (2533 mg/L), the removal efficiencies of the N-nitrosamines by the MBR achieved in this study (NDMA: 73 – 84 %; NPYR: 58 – 61 %; NDEA: 76 – 77 %; ; NPIP: 57 – 76 %; NMOR: 45 – 57 %; NDBA: 75 – 78 %; NDPHA: 53 –

13

54 %) are comparable with their removal efficiencies (24 – 94 %) by the conventional MBR process 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

with the MLSS concentration of 5000 mg/L previously reported by Wijekoon et al. (2013). Similar removal trends of the N-nitrosamines by the MBR were observed for two sampling campaigns. However, there was no significant correlation between the removal of N-nitrosamines by the MBR and their MW and Log P since biodegradation is considered to be mainly responsible the removal of N-nitrosamines by the MBR (see section 3.4.1). Therefore, the removal efficiencies of the N-nitrosamines could be qualitatively predicted by their structural characteristics (i.e., functional group composition) closely related to biodegradation by aerobic bacteria (Wijekoon et al., 2013). The higher removal efficiencies of NDMA, NDEA, and NDBA by the MBR than NPYR, NPIP, NMOR, and NDPHA were probably due to their acyclic alkyl amine groups, which are readily subjected to the attacks by oxygenases of aerobic microorganisms (Wijekoon et al., 2013). Among the N-nitrosamines with an acyclic alkyl amine group (i.e., NDMA, NDEA, and NDBA), the highest removal efficiency was observed for NDMA as the Nnitrosamines with relatively short aliphatic chains (i.e., NDMA) are more reactive with the catabolic enzymes of aerobic bacteria (e.g., ammonia monooxygenases) compared to the N-nitrosamines with relatively long aliphatic chains (i.e., NDEA and NDBA) (Table S1, SI) (Krauss et al., 2009). The relatively low removal efficiencies of NPIP and NPYR could be attributed to the less reactivity of alicyclic alkyl amine groups with the catabolic enzymes compared to acyclic alkyl amine groups (Wijekoon et al., 2013). A possible explanation for the lowest removal efficiencies of NMOR and NDPHA by the MBR is that the morpholine and/or aromatic amine groups make NMOR and NDPHA recalcitrant to the catabolic enzymes (Wijekoon et al., 2013). These results indicate that the reactivity of amine functional groups with the catabolic enzymes can determine biodegradation rates of Nnitrosamines by aerobic microorganisms.

3.4.3. Removal mechanisms of N-nitrosamines using NF membranes

14

As shown in Fig. 5, the effects of physico-chemical properties of N-nitrosamines (i.e., MW, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

hydrophobicity (Log P), and charge characteristics) on their removal were investigated using two different kinds of NF membranes with different MWCOs (NE70 = 350 Da; NE90 = 210 Da) and surface zeta potentials (NE70 = –18 mV; NE90 = –15 mV) as the removal of trace organic contaminants using the NF membranes may be governed by size-exclusion, electrostatic repulsion, and adsorption onto membrane surfaces (Chon et al., 2012a). In this study, electrostatic repulsion does not play a major role in the removal of N-nitrosamines by the NF membranes due to the uncharged character of the Nnitrosamines at neutral pH (Asano et al., 2007; Fujioka et al., 2013b) while a relatively strong correlation was found between the removal efficiencies of the N-nitrosamines by the NF membranes and their molecular size (i.e., MW): NDMA (MW = 74.1 Da) = 8 – 59 %; NPYR (MW = 100.1 Da) = 38 – 60 %; NDEA (MW = 102.1 Da) = 43 – 75 %; NPIP (MW = 114.1 Da) = 39 – 72 %; NMOR (MW = 116.1 Da) = 51 – 74 %; NDBA (NW = 158.2 Da) = 61 – 71 %; NDPHA (MW = 198.2 Da) = 69 – 80 %). Although the removal patterns of N-nitrosamines by the NF membranes in two sampling campaigns were similar to each other, a distinctive difference was found for the removal efficiencies of NDMA. A marginal reduction was found for removal of NDMA by the NF membranes in 1st sampling campaign (removal efficiency of NE70 = 8 %; removal efficiency of NE90 = 12 %) due to the fact that MW of NDMA (74.1 Da) was much lower than MWCOs of the NF membranes (NE70 = 350 Da; NE90 = 210 Da). However, it was substantially removed by the NF membranes in 2nd sampling campaign (removal efficiency of NE70 = 57 %; removal efficiency of NE90 = 59 %). This was probably due to the formation of membrane fouling by DOM in secondary effluents which is expected to considerably contribute to the removal of N-nitrosamines with low MW, particularly NDMA (Fujioka et al., 2013b). A great decrease in the permeate flux of the NF membranes supports this observation (Chon et al., 2013c). The relatively higher removal efficiencies of NDBA and NDPHA by the NF membranes compared to the other Nnitrosamines could be attributed to their higher MW (NDBA = 158.2 Da; NDPHA = 198.2 Da) and Log P values (NDBA = 2.31; NDPHA = 3.16). Fujioka et al. (2012) reported that adsorption onto the

15

membrane surfaces may enhance the removal the N-nitrosamines by the NF membranes. The type of the 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

NF membranes also significantly influenced the removal of N-nitrosamines (Fig. 5). In spite of the fact that both the NE70 and NE90 membranes were operated under the same temperature (20 ±2 °C), retentate flux (500 mL/min), and TMP (400 kPa) conditions, the N-nitrosamines were more effectively removed by the NE90 membrane (removal efficiency = 12 – 80 %) than the NE70 membrane (removal efficiency = 8 – 72 %). Similar to the results from the removal patterns of bulk water contaminants, these findings can be explained by the differences in their MWCO (NE70 = 350 Da; NE90 = 210 Da). This suggests that size-exclusion is a key factor affecting the removal of N-nitrosamines by the NF membranes

4. Conclusions The microbial activities in the MBR and MWCO of the NF membranes may play critical roles in the removal of bulk water contaminants. Biodegradation is mainly responsible for the removal of Nnitrosamines by the MBR. Therefore, their removal efficiencies were significantly different depending on the reactivity of the amine functional groups with the catabolic enzymes (removal efficiency = 45 – 84 %). Although adsorption and formation of membrane fouling may substantially contribute to the removal of N-nitrosamines by the NF membranes, size-exclusion is found to be the dominant removal mechanism for the N-nitrosamines (removal efficiency: NE90 > NE70).

Acknowledgements This work was conducted under the framework of the Research and Development Program of the Korea Institute of Energy Research (KIER) (B5-2481) and also supported by the NRF grant funded by the Korea government (MEST) (No. 2012047029 (TOC)).

References

16

[1] Asano, T., Burton, F.L., Leverenz, R., Tchobanoglous, G., 2007. Water reuse: Issues, technologies, and 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

applications. McGraw-Hill, New York, USA. [2] Australian Capital Territory, 2004. Greywater use guidelines for residential properties in Canberra. Australian Capital Territory. Canberra, Australia. [3] Banu, J.R., Do, H.U., Athish, K., Yeom. I.T., 2001. Effect of low temperature thermochemical pretreatment on sludge reduction capacity of MBR treating secondary treated dairy wastewater. Water Qual. Res. J. Can. 46, 312–320. [4] Banu, J.R., Do, H.U., Yeom. I.T., 2009. Nutrient removal in an A2O-MBR reactor with sludge reduction. Bioresource Technol. 100, 3820–3824. [5] Carballa, M., Omil, F., Lema, J.M., 2005. Removal of cosmetic ingredient and pharmaceuticals in sewage primary treatment. Water Res. 39, 4790–4796. [6] CDPH, 2013. Draft Regulations: ground water replenishment with recycled water. CDPH, Sacramento, CA, USA. [7] CDPH, 2013. NDMA and other nitrosamines–drinking water issues. CDPH, Sacramento, CA, USA. [8] Chon, K., Cho, J., Kim, S.J., Jang, A., 2014. The role of a combined coagulation and disk filtration process as a pre-treatment to microfiltration and reverse osmosis membranes in a municipal wastewater pilot plant. Chemosphere 117, 20–26. [9] Chon, K., Cho, J., Shon, H.K., 2013a. A pilot-scale hybrid municipal wastewater reclamation system using combined coagulation and disk filtration, ultrafiltration, and reverse osmosis: Removal of nutrients and micropollutants, and characterization of membrane foulants. Bioresource Technol. 141, 109–116. [10] Chon, K., Cho, J., Shon, H.K., 2013c. Fouling characteristics of a membrane bioreactor and nanofiltration hybrid system for municipal wastewater reclamation. Bioresource Technol. 130, 239– 247.

17

[11] Chon, K., Cho, J., Shon, H.K., 2013d. Advanced characterization of algogenic organic matter, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

bacterial organic matter, humic acids and fulvic acids. Water Sci. Technol. 67, 2228–2235. [12] Chon, K., Kim, S.J., Moon, J., Cho, J., 2012b. Combined coagulation-disk filtration process as a pretreatment of ultrafiltration and reverse osmosis membrane for wastewater reclamation: An autopsy study of a pilot plant. Water Res. 46, 1803–1806. [13] Chon, K., Lee, Y., Jacqueline. T., von Gunten, U., 2013b. Quantification and characterization of dissolved organic nitrogen in wastewater effluents by electrodialysis treatment followed by sizeexclusion chromatography with nitrogen detection. Water. Res. 47, 5381–5391. [14] Chon, K., Sarp, S., Lee, S., Lee, J.H., Lopez-Ramirez, J.A., Cho, J., 2011. Evaluation of a membrane bioreactor and nanofiltration for municipal wastewater reclamation: Trace contaminants control and fouling mitigation. Desalination 272, 128–134. [15] Chon, K., Shon, H.K., Cho, J., 2012a. Membrane bioreactor and nanofiltration hybrid system for reclamation of municipal wastewater: Removal of nutrients, organic matter, micropollutants. Bioresource Technol. 233, 181–188. [16] Clara, M., Strenn, B., Gans, O., Martinez, E., Kreuzinger, N., Kroiss, H., 2005. Removal of selected pharmaceuticals, fragrances and endocrine disrupting compounds in a membrane bioreactor and conventional wastewater treatment plants, Water Res. 39, 4797–4807. [17] Do, H.U., Banu, J.R., Son, D.H., Yeom, I.T., 2012. Influence of ferrous sulphate on thermochemical sludge disintegration and on performance of wastewater treatment in an anoxic/oxic membrane bioreactor coupled with sludge disintegration step. Biochem. Eng. J. 66, 20–26. [18] Dolar, D., Gros, M., Rodriguez-Mozaz, S., Moreno, J., Comas, J., Rodriquez-Roda, I., Barcelo, D., 2012. Removal of emerging contaminants from municipal wastewater with an integrated membrane system, MBR-RO. J. Hazard. Mater. 239, 64–69.

18

[19] Fujioka, T., Khan, S.J., McDonald, J.A., Henderson, R.K., Poussade, Y., Drewes, J.E., Nghiem, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

L.D., 2013b. Effects of membrane fouling on N-nitrosamine rejection by nanofiltration and reverse osmosis membranes. J. Membrane Sci.427, 311–319. [20] Fujioka, T., Khan, S.J., McDonald, J.A., Roux, A., Poussade, Y., Drewes, J.E., Nghiem, L.D., 2013c. N-nitrosamine rejection by nanofiltration and reverse osmosis membranes: the importance of membrane characteristics. Desalination 316, 67–75. [21] Fujioka, T., Nghiem, L.D., Khan, S.J., McDonald, J.A., Poussade, Y., Drewes, J.E., 2012. Effects of feed solution characteristics on the rejection of N-nitrosamines by reverse osmosis membranes. J. Membrane Sci.409–410, 66–74. [22] Fujioka, T., Stuart J.K., James, McDonald, J.A., Roux, A., Poussade, Y., Drewes, J.E., Nghiem, L.D., 2013a. N-nitrosamine rejection by reverse osmosis membranes: A full-scale study. Water Res. 47, 6141–6148. [23] Krauss, M., Longrée, P., Dorusch, F., Ort, C., Hollender, J., 2009. Occurrence and removal of N– nitrosamines in wastewater treatment plants, Water Res. 43, 4381–4391. [24] Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zuggg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999– 2000: a national reconnaissance. Environ. Sci. Technol. 36, 1202–1211. [25] NRMMC, EPHC, AHMC, 2008. Australian guidelines for water recycling: Managing health and environmental risks (Phase 2): Augmentation of drinking water supplies. Environment Protection and Heritage Council, National Health and Medical Research Council, Natural Resource Management Ministerial Council, Canberra, Australia. [26] Mitch, W.A., Sedlak, D.L., 2002. Factors controlling nitrosamine formation during wastewater chlorination. Water Sci. Technol.: Water Supply 2, 191–198. [27] Munch, J.W., Bassett, M.W., 2004. EPA Method 521: Determination of nitrosamines in drinking water by solid phase extraction and capillary column gas chromatography with large volume

19

injection and chemical ionization tandem mass spectrometry/National Exposure Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Laboratory, Office of Research and Development, EPA, Cincinnati, OH, USA. [28] Potter, B.B., Wimsatt, J.C., 2005. Determination of total organic carbon and specific UV absorbance at 254 nm in source water and drinking water. EPA/600/R-05/055. United States Environmental Protection Agency, Cincinnati, OH, USA. [29] Raj, S.E., Banu, J.R.,, Kaliappan. S., Yeom, I.T., Kumar, A., 2013. Effects of side-stream, low temperature phosphorus recovery on the performance of anaerobic/anoxic/oxic systems integrated with sludge pretreatment. Bioresource Technol. 140, 376–384. [30] Sacher, F., Schmidt, C.K., Lee, C.C., von Gunten, U., 2008. Strategies for minimizing nitrosamine formation during disinfection. AWWA Research Foundation, Denver, CO, USA. [31] Sedlak, D.L., Deeb, R.A., Hawley, E.L., Mitch, W.A., Durbin, T.D., Mowbray, S., Carr, S., 2005. Sources and fate of nitrosodimethylamine and its precursors in municipal wastewater treatment plants. Water Environ. Res. 77, 32–39. [32] Ternes, T.A., Joss, A., 2006. Human pharmaceuticals, hormones and fragrances. The challenge of micropollutants in urban water management. IWA Publishing, London, UK. [33] U.S. EPA, 1993. Integrated risk information system. U.S. EPA, Washington, DC, USA. [34] Wijekoon, K.C., Fujioka, T., McDonald, J.A., Khan, S.J., Hai, F.I., Price, W.E., Nghiem, L.D., 2013. Removal of N-nitrosamines by an aerobic membrane bioreactor. Bioresource Technol. 141, 41–45. [35] Xing, C.H., Tardieu, E., Qian, Y., Wen, X.H., 2000. Ultrafiltration membrane bioreactor for urban wastewater reclamation. J. Membrane Sci. 177, 73–82.

20

Table 1. Physico-chemical characteristics of the MF and NF membranes. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Type

Code

Manufacturer

Materials

Pore size (µm)

MWCO (Da)

Zeta potential (mV)

Roughness (nm)

Contact angle (°)

MF

Cleanfil -S30V

Kolon Industries

PVDF

0.1

N.A.

N.A.

N.A.

N.A.

350 a

–18 (at pH 5.4) b

11.3 b

41.2c

210 a

–15 (at pH 5.6) b

38.1 b

55.1c

NE70 Toray Chemical

NF

PA TFC

N.A.

NE90

N.A.: not available; PVDF: polyvinylidene fluoride; PA TFC: polyamide thin-film composite; a

Adapted from Chon et al., 2012a.

b

Adapted from Chon et al., 2013c.

c

Roughness presents the root mean-square-average (Rq) of the roughness profile ordinates.

21

Table 2. Water characteristics of the feed and treated wastewaters (n=3). 1 2 3 4 5 pH 6 7 8 Conductivity (µS/cm) 9 10 CODMn (mg/L) 11 12 13 DOC (mg/L) 14 15 16 UVA254 (/cm) 17 18 SUVA (L/mgC·m) 19 20 N.D.: not detected. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Feed

MBR

NE70

NE90

7.7 (±0.1)

5.4 (±0.2)

5.5 (±0.1)

5.5 (±0.2)

320 (±10)

286 (±39)

190 (±3)

74 (±3)

9.7 (±0.3)

5.8 (±0.2)

N.D.

N.D.

14.59 (±0.25)

6.01 (±0.11)

0.70 (±0.13) a

0.40 (±0.09)

0.162 (±0.001)

0.125 (±0.001)

0.024 (±0.001)

0.020 (±0.002)

1.11 (±0.02)

2.08 (±0.03)

5.91 (±0.60)

5.13 (±0.82)

22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Stirrer with a crossed-blade impeller

2 way automatic water valve

Automatic water level control valve

Back washing Micro pump

Flow meter

Magnetic gear pump

Temperature controller

Effluents from primary sedimentation Air diffuser

Submerged MBR

Magnetic gear pump

Reservoir for NE70 permeates Magnetic stirrer

NE90

MF Recycling pump

Flat-sheet NF membranes NE70

Magnetic gear pump

Reservoir for NE90 permeates Magnetic stirrer

Magnetic stirrer Reservoir for MBR permeates

Fig. 1. System description of the MBR and NF hybrid system for municipal wastewater reclamation.

23

(a)

UVA response (mV)

40000

920 830

Feed MBR NE70 NE90

670 30000

20000

35,200

10000

0 1e+1

1e+2

1e+3

1e+4

1e+5

MW (Da)

(b) 42000

Fluorescence response (mV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

35000

Feed MBR NE70 NE90

570 860 520 930

28000 21000

36,660 29,510

210 150 50 120

14000 7000 0 1e+1

1e+2

1e+3

1e+4

1e+5

MW (Da)

Fig. 2. Changes in MW distribution of DOM in the raw and treated wastewaters by the MBR and NF hybrid system: (a) aromatic substances and (b) protein-like substances.

24

(a) 500 Feed MBR NE70 NE90

Concentration (ng/L)

400 300 100 80 60 40 20 0 NDMA

NPYR

NDEA

NPIP

NMOR

NDBA NDPHA

(b) 250

Concentration (ng/L)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Feed MBR NE70 NE90

200

150

100

50

0 NDMA

NPYR

NDEA

NPIP

NMOR

NDBA NDPHA

Fig. 3. Concentrations of N-nitrosamines in the feed and treated wastewaters: (a) 1st sampling campaign and (b) 2nd sampling campaign (NDMA: N-nitrosodimethylamine; NPYR: Nnitrosopyrrolidin;

NDEA:

N-nitrosodiethylamine;

NPIP:

N-nitrosopiperidine;

nitrosomorpholine; NDBA: N-nitrosodibutylamine; NDPHA: N-nitrosodiphenylamine).

25

NMOR:

N-

(b)

(a)

100 NDMA NPIP (74.1) NDEA (112.1) (102.1) NMOR (116.1)

90 80

NDBA (158.2)

NDPHA (198.2)

NPYR (100.1)

70

Removal efficiency (%)

Removal efficiency (%)

100

60 50 40

NDMA (-0.64)

90

NDEA (0.34)

NDBA (2.31)

NMOR (-1.39)

80

NDPHA (3.16)

NPYR NPIP (0.23) (0.74)

70 60 50 40

60

80

100

120

140

160

180

200

220

-2

-1

0

MW (Da)

1

2

3

4

Log P

(d)

(c) 100

100 NDMA NPIP (74.1) NDEA (112.1) (102.1)

90

NDBA (158.2)

80

Removal efficiency (%)

Removal efficiency (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

NDPHA (198.2) NPYR (100.1)

70

NMOR (116.1)

60

50

NDMA (-0.64)

90 NMOR (-1.39)

80

NDEA (0.34) NPIP (0.74)

NDBA (2.31)

NDPHA (3.16)

NPYR (0.23)

70 60 50 40

60

80

100

120

140

160

180

200

220

-2

MW (Da)

-1

0

1

2

3

4

Log P

Fig. 4. Removal efficiencies (%) of N-nitrosamines by the MBR as a function of MW and Log P: (a)–(b) 1st sampling campaign and (c)–(d) 2nd sampling campaign (NDMA: N-nitrosodimethylamine; NPYR: N-nitrosopyrrolidin; NDEA: N-nitrosodiethylamine; NPIP: N-nitrosopiperidine; NMOR: Nnitrosomorpholine; NDBA: N-nitrosodibutylamine; NDPHA: N-nitrosodiphenylamine).

26

(a)

NDBA (158.2)

NDMA (74.1)

40

20

80

NPYR (0.23)

60

NDBA (2.31)

NDPHA (3.16)

NDMA (-0.64)

40

20

0

0 60

80

100

120

140

160

180

200

-2

220

-1

0

NE70 100

NDEA NPIP (102.1) (112.1) NPYR NMOR (100.1) (116.1)

90 80

(d)

NE90

2

3

4

NDBA (158.2)

NDPHA (198.2)

NPYR (0.23) NDEA (0.34) NDMA (-0.64) NPIP NMOR (0.74) (-1.39)

90

NDMA (74.1)

60 50 40

NE90

NE70 100

Removal efficiency (%)

(c)

1

Log P

MW (Da)

70

NE90

NDEA NPIP (0.34) (0.74)

NMOR (-1.39)

NDPHA (198.2)

Removal efficiency (%)

Removal efficiency (%)

60

NE70 100

NPIP NMOR (112.1) (116.1) NDEA (102.1) NPYR (100.1)

80

(b)

NE90

NE70 100

Removal efficiency (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

80 70

NDBA (2.31)

NDPHA (3.16)

60 50 40

30

30 60

80

100

120

140

160

180

200

220

-2

MW (Da)

-1

0

1

2

3

4

Log P

Fig. 5. Removal efficiencies (%) of N-nitrosamines by the NF membranes as a function of MW and Log P: (a)–(b) 1st sampling campaign and (c)–(d) 2nd sampling campaign (NDMA: Nnitrosodimethylamine; NPYR: N-nitrosopyrrolidin; NDEA: N-nitrosodiethylamine; NPIP: Nnitrosopiperidine; NMOR: N-nitrosomorpholine; NDBA: N-nitrosodibutylamine; NDPHA: Nnitrosodiphenylamine).

27

Highlights 

MWCO of NF membranes is a key factor affecting removal of bulk water contaminants.



Reactivity of amine groups with catabolic enzymes determines removal of N-nitrosamines.



Removal of N-nitrosamines by NF membranes can be mainly governed by size-exclusion.



Adsorption onto membrane surfaces significantly contributes to removal of N-nitrosamines.



Formation of membrane fouling may enhance removal of N-nitrosamines with low MW.