Removal of fluoride and uranium by nanofiltration and reverse osmosis: A review

Chemosphere 117 (2014) 679–691

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Review

Removal of fluoride and uranium by nanofiltration and reverse osmosis: A review Junjie Shen a,b, Andrea Schäfer b,c,⇑ a

School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, United Kingdom Department of Water and Environmental Engineering, Nelson Mandela African Institute of Science and Technology, Arusha, Tanzania c Institute of Functional Interfaces (IFG), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 F and U occurrences and health

implications are comprehensively summarized.  Up-to-date progress on F and U removal by NF and RO are critically reviewed.  F and U removal under various conditions are illustrated with mechanistic schematics.

a r t i c l e

i n f o

Article history: Received 27 November 2013 Received in revised form 25 September 2014 Accepted 30 September 2014

Handling Editor: O. Hao Keywords: Nanofiltration Reverse osmosis Fluoride Uranium Drinking water

a b s t r a c t Inorganic contamination in drinking water, especially fluoride and uranium, has been recognized as a worldwide problem imposing a serious threat to human health. Among several treatment technologies applied for fluoride and uranium removal, nanofiltration (NF) and reverse osmosis (RO) have been studied extensively and proven to offer satisfactory results with high selectivity. In this review, a comprehensive summary and critical analysis of previous NF and RO applications on fluoride and uranium removal is presented. Fluoride retention is generally governed by size exclusion and charge interaction, while uranium retention is strongly affected by the speciation of uranium and size exclusion usually plays a predominant role for all species. Adsorption on the membrane occurs as some uranium species interact with membrane functional groups. The influence of operating conditions (pressure, crossflow velocity), water quality (concentration, solution pH), solute–solute interactions, membrane characteristics and membrane fouling on fluoride and uranium retention is critically reviewed. Ó 2014 Elsevier Ltd. All rights reserved.

Contents 1. 2. 3. 4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Worldwide occurrence of fluoride and uranium . . . . . . . . Health effects of fluoride and uranium in drinking water. Chemical characteristics of fluoride and uranium . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

680 680 681 682

⇑ Corresponding author at: Institute of Functional Interfaces (IFG), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 EggensteinLeopoldshafen, Germany. Tel.: +49 721 608 26906. E-mail address: [email protected] (A. Schäfer). http://dx.doi.org/10.1016/j.chemosphere.2014.09.090 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

680

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

5.

Fluoride and uranium removal in NF/RO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. General introduction of NF/RO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Removal of fluoride and uranium by different NF/RO membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Retention mechanism I: Solution–diffusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Retention mechanism II: Size exclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5. Retention mechanism III: Charge interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6. Retention mechanism IV: Adsorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7. Influence of operating parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.8. Influence of water chemistry: Concentration and pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.9. Influence of solute–solute interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.10. Influence of membrane fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix A. Supplementary material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Global demands for safe drinking water are increasing and inability to meet requirements is leading to increasing water conflicts (Mbonile, 2005; Sivakumar, 2011). Natural waters including groundwater, surface water (rivers and lakes) and rainwater are the main drinking water sources, while desalination of brackish and seawater is playing an increasing role. Groundwater, which constitutes 97% of global freshwater, is consumed for drinking purpose by more than 50% of the world population (Schmoll et al., 2006). In many remote and developing communities where basic water distribution systems are unavailable, groundwater serves as the most economically viable option (Ayoob and Gupta, 2006). However, groundwater often contains inorganic contaminants such as fluoride, uranium, arsenic, and boron amongst many others. Long term exposure to such contaminants causes health effects in humans. For example, excessive fluoride ingestion leads to dental and skeletal fluorosis (Fawell et al., 2006), and continuous uranium intake from drinking water has toxic effects on kidneys (Zamora et al., 1998). The occurrence of inorganic contaminants is highly geology-dependent. Therefore the practical approach to remove such ions/species/contaminants is to develop appropriate and flexible technologies for local use (Schwarzenbach et al., 2006). Conventional water treatment methods involve a combination of adsorption, coagulation, flocculation, clarification, filtration and disinfection (Binnie and Kimber, 2009). The main drawback of conventional methods is that they are generally less effective for removing trace contaminants (Schwarzenbach et al., 2006). Nanofiltration (NF) and reverse osmosis (RO) are very promising techniques compared with conventional methods, in particular for drinking water applications. NF/RO can achieve high inorganic removal as they involve a mixture of separation mechanisms including solution diffusion, size exclusion, charge repulsion and adsorption. Suitable membrane characteristics can be selected to match particular water qualities. Inorganic contaminants in groundwater are often accompanied by bacteria, viruses and micropollutants such as pesticides which are also undesirable. NF/RO can simultaneously remove those contaminants in one single process, while the removal of micropollutants depends on specific characteristics. Furthermore, NF/RO is modular in design and flexible in implementation, making it a suitable choice for remote communities. Nevertheless, the possible drawbacks of NF/RO cannot be denied, those include membrane fouling and scaling, concentrate disposal and relatively high energy consumption. Membrane fouling and scaling are inherent in the separation process but can be significantly reduced by optimizing the operation conditions while scaling depends on the likelihood of precipitate formation in a particular water. High energy consumption can be compensated by introducing renewable energy technologies

683 683 683 684 684 685 686 687 687 687 688 689 689 689 689

(Schäfer et al., 2006; Richards et al., 2008). Given the very good water quality produced by NF/RO, it is a good practice to use such water predominantly for potable purposes. Feed water quality permitting usage of concentrates for non-potable purposes provides near zero discharge opportunities. However, inorganic contaminants in water involve a wide range of chemical characteristics and the removal mechanisms by NF/RO are not well understood for all. Inorganics such as fluoride and uranium are rapidly emerging issues of water quality which are of increasing concern worldwide. Their removal mechanisms differ significantly and in consequence these two contaminants are chosen for this critical review. The occurrence of fluoride and uranium and their health implications, reported removal by NF/RO along with specific removal mechanisms are investigated in this paper. 2. Worldwide occurrence of fluoride and uranium It is evident that fluoride contamination is a worldwide issue (Table 1). Amini et al. (2008) provides a global overview of groundwaters with fluoride concentration exceeding the WHO guideline of 1.5 mg L1. The results show that areas most severely affected include East Africa, Middle East, Argentina, the United States, India, and China. The occurrence of fluoride in natural waters is closely linked to the local geology. The chemical element fluorine is abundant in the Earth’s crust (625 mg kg1) as a result of volcanic activity and fumarolic gases (Edmunds and Smedley, 2005). Fluorides are naturally released into water by the dissolution of fluoride-containing rocks and soils. The dissolution process is affected by various factors including rock chemistry, groundwater age, residence time, well depth and conditions of the pathways (Kim and Jeong, 2005). Fluoride concentration in water is strongly controlled by the solubility of minerals, especially calcium fluorite (CaF2) which has the lowest solubility of 15 mg L1 at 18 °C (Kwasnik, 1963). Therefore high fluoride concentrations are associated with minerals with low calcium contents, or high alkaline and carbonous conditions where sodium instead of calcium dominates the water composition (Amini et al., 2008). One typical example arises in the East African Rift Valley. Fluoride concentration in local soda lakes is up to 2800 mg L1 and in groundwater is as high as 330 mg L1 (Smedley et al., 2002; Fawell et al., 2006). Such concentrations are extremely high even compared to other elevated-fluoride areas in the world. In addition to natural dissolution of minerals, industrial operations, such as metallurgical industries, fertilizer plants, and semiconductor production, generate effluents with high fluoride contents (Ndiaye et al., 2005; Dolar et al., 2011). In the case of phosphate production, fluoride in the effluent can reach up to 3000 mg L1 (Ndiaye et al., 2005).

681

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691 Table 1 Global examples of elevated fluoride in drinking water. Country Tanzania

Kenya Ethiopia South Africa Turkey Argentina USA Germany China India

Fluoride concentration

Reference

1

<250 mg L in groundwater, Shinyanga Region <330 mg L1 in groundwater, Arusha Region <60 mg L1 in spring, Arusha Region <180 mg L1 in groundwater in the Rift Valley <90 mg L1 in watercourse Median 10 mg L1, max 68 mg L1 in well water of the Main Ethiopian Rift <40 mg L1 in groundwater <13.7 mg L1 in the middle and eastern parts Median 3.8 mg L1, max 182 mg L1 in well water in the southeast regions <4.3 mg L1 in water supplies in Texas <8.8 mg L1 in wells in the Muenster region <10 mg L1 in groundwater, 50% > 2 mg L1, Jilin Province <8.3 mg L1 in groundwater, Shanxi province <19 mg L1 in groundwater, North-West India <20 mg L1 in groundwater, South India

British Geological Survey (2000), Ghiglieri et al. (2012)

Gaciri and Davies (1993) Rango et al. (2012) Muller et al. (1998) Azbar and Türkman (2000) Paoloni et al. (2003) Segreto et al. (1984) Queste et al. (2001) Zhang et al. (2003), Wang et al. (2007) Agarwal et al. (1997), Fawell et al. (2006)

Table 2 Global examples of elevated uranium in drinking water, adapted mainly from Schulte-Herbrüggen (2011). Country

Uranium concentration

Reference

Kazakhstan Australia

<1920 lg L1 in well water at uranium mining sites <2000 lg L1 in groundwater, South Australia <440 lg L1 in groundwater, North Australia <700 lg L1 in private wells <845 lg L1 in wells of aboriginal people <620 lg L1 in private wells <1000 lg L1 in contaminated groundwater 52% of wells >20 lg L1 <1263 lg L1 in North Mara Gold Mine <540 lg L1 in groundwater surrounding gold mine <266 lg L1 in borehole, Central Region Median 28 lg L1, max 1920 lg L1 in well water <643 lg L1 in bedrock groundwater <80 lg L1 in stream water Median of 39 lg L1 in black shale <196 lg L1 in bedrock and near-surface groundwater <166 lg L1 in private wells with a mean of 5 lg L1 and median 1.6 lg L1

Uralbekov et al. (2011) Pirlo and Giblin (2004), Payne et al. (2001)

Canada USA Tanzania South Africa Ghana Finland

Sweden Kosovo

Uranium, the other inorganic contaminant of interest, is a naturally occurring radioactive element. It has three main isotopes: 238 U, 234U and 235U, among which 238U is the most prevalent form (WHO, 2011). Uranium is ubiquitous in the Earth’s crust with an average abundance of 2.7 mg kg1 while higher concentrations of up to 15 mg kg1 and 120 mg kg1 may occur in granite and phosphate rocks, respectively (Langmuir, 1997). Similar to fluoride, uranium in water is from both natural mineral dissolution and external human activities, such as mining, fertilizer and nuclear industries (WHO, 2011). A series of high uranium concentrations in groundwater from various locations in the world is presented in Table 2. The occurrence of uranium in European waters is in association with granite rocks and volcanic activities (Åström et al., 2009). High-uranium groundwaters in Kazakhstan, Australia and Canada are often encountered in uranium mining areas since those three countries provide about 64% of the world’s uranium production (World Nuclear Association, 2012). Leakage from nuclear power plants and military use of depleted uranium also pose a high risk of uranium contamination into natural waters (Bleise et al., 2003; Um et al., 2010). 3. Health effects of fluoride and uranium in drinking water Fluoride has been recognized to have either beneficial or detrimental effect on human health, depending on the concentration and the uptake duration (Fawell et al., 2006). Small amounts of fluoride can be beneficial to the teeth, preventing dental caries among children (Edmunds and Smedley, 2005). However, chronic ingestion of fluoride at high doses leads to a wide variety of

Zamora et al. (2009) Orloff et al. (2004), Abdelouas et al. (1998), Hakonson-Hayes et al. (2002) Almås et al. (2009) Winde and Jacobus van der Walt (2004) Rossiter et al. (2010a) Kurttio et al. (2002), Åström et al. (2009)

Åström et al. (2009) Berisha and Goessler (2013)

adverse effects. Dental fluorosis and crippling skeletal fluorosis are the first adverse effects that fluoride can have on the body, which are manifested by mottled teeth in mild cases and brittle bones and neurological complications in severe cases (Edmunds and Smedley, 2005; Fawell et al., 2006; Bhatnagar et al., 2011). Some researchers indicated that chronic exposure to fluoride is further associated with decreased birth rates, increased rates of kidney stones, impaired thyroid function, and lower intelligence in children (Wang et al., 2007; Ozsvath, 2009; Choi et al., 2012). In light of the complex effects of fluoride on human health, WHO recommends a guideline value of 1.5 mg L1 as a level below which the detrimental effect should be minimal (Fawell et al., 2006). Figure 1 depicts children and adults suffering from dental fluorosis (Fig. 1a) and crippling fluorosis (Fig. 1b and c) which is the most severe type whereby victims are facing mobility difficulties. Natural uranium is classified as both a radiotoxic and a chemotoxic agent. Dose estimation based on water consumption indicated that exposure to groundwater containing natural uranium up to 1200 lg L1 would not result in significant radiological toxicity (Hakonson-Hayes et al., 2002). On the contrary, chemical toxicity of uranium is a greater health concern than radiological toxicity. Studies showed that kidneys and bones are the principal accumulation sites of uranium and long-term ingestion of uranium may increase the risk of kidney failures (Kurttio et al., 2002; Zamora et al., 2009). In a more recent study in Germany, a significantly elevated risk of leukemia in men and kidney cancer in women was identified in areas with increased drinking water uranium content (Radespiel-Tröger and Meyer, 2013). However,

682

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

(a)

(b)

(c)

Fig. 1. (a) Dental fluorosis observed in people of all ages. (b) Crippling fluorosis in young children (note scars from leg straightening operations). (c) Fluorosis and premature aging in middle aged adults (all photos taken at Oldonyosambu Ward in Arumeru District, Tanzania in April 2013).

uranium chemical toxicity was found to be determined by its speciation, among which calcium–uranyl–carbonate complexes have the lowest toxicity (Prat et al., 2009). WHO recommends that uranium in drinking water should not exceed 15 lg L1, based on the tolerable daily intake (WHO, 2011).

Fluoride forms complexes with a number of cations due to charge attraction. Table 3 summarizes common fluoride species in water and their physicochemical properties (Kwasnik, 1963). The solubility of fluoride species varies significantly, among which calcium, magnesium and strontium fluorides are very insoluble. In addition, fluoride ion has the same charge and nearly the same ionic radius as OH. Therefore fluoride ion may replace hydroxide in mineral structures under acid conditions through ion exchange (Fawell et al., 2006). This explains why some common mineral species of low solubility contain fluoride. Uranium has several valence states: +III to +VI, of which +IV (uranous) and +VI (uranyl) are the most common in the

4. Chemical characteristics of fluoride and uranium The speciation of fluoride is pH dependent with the equilibrium coefficient of hydrofluoric acid being pKa = 3.2 (Lide, 1995) with negative fluoride ions dominating above the pKa (Fig. 2a and c).

Carbonate-type water 100

a

100

b

UO2

2+

Relative pecent (%)

+

80

80

UO2F

UO2F2 (aq) UO2F3

60

60

pKa=3.2

-

UO2CO 3 (aq) UO2(CO 3 )2

40

40

HF (aq) FNaF (aq)

20

UO2(CO 3 )3

24-

Ca2 UO2 (CO3 )3 (aq) 2-

CaUO 2(CO 3) 3

20

-

UO2(OH) 3

2-

UO2(OH) 4

0

0 0

2

4

6

8

10

12 0

2

4

pH

8

6

10

12

pH

Sulphate-type water

c

100

100

d

2+

UO 2

Relative pecent (%)

2-

80

80

UO2(SO4) 2

UO2SO4 (aq) +

UO2F

60

60

pKa=3.2

UO 2CO 3 (aq)

HF (aq) FMgF+ NaF (aq)

40

UO2F2(aq) 2-

40

UO2(CO3) 2

4-

UO2(CO3)3

Ca2UO2(CO3)3 (aq)

20

20

2-

CaUO 2(CO3 )3 UO 2OH

+ -

0

0 0

2

4

6

pH

8

10

12 0

2

4

6

8

10

UO2(OH)3

2-

UO2(OH)4

12

pH

Fig. 2. Fluoride (a, c) and uranium (b, d) speciation in two different semi-natural groundwaters modeled by Visual Minteq 3.0 (only species >1% are listed. Water composition is provided in Table S2 in Supplementary Material).

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691 Table 3 Physicochemical properties of common fluoride species (Kwasnik, 1963). Fluoride species

Molecular weight

Solubility in water (g L1)

Melting point (°C)

Boiling point (°C)

NaF KF NH4F CaF2 MgF2 SrF2 AlF3

42 58 37 78 62 126 84

43 (25 °C) 923 (18 °C) 1000 (0 °C) 0.015 (18 °C) 0.087 (18 °C) 0.117 (18 °C) 5.59 (25 °C)

993 857 – 1418 1248 1190 1260

1704 1503 – 2500 2260 2460 –

Table 4 Uranium species commonly detected in the environmental system (Langmuir, 1997). Name

Chemical formula

Uranyl ion Uranyl hydro species

UO2+ 2 + 0  UOH3+, U(OH)2+ 2 , U(OH)3, U(OH)4, U(OH)5 , 9+ , UO2OH+, (UO2)2(OH)22+, (UO2)3(OH)5+ U6(OH)15 4 UO2CO03, UO2(CO3)2 2 , UO2(CO3)3 0 2 0 USO2+ 4 , UO2SO4, UO2(SO4)2 , U(SO4)2 0 2 4 UHPO2+ 4 , U(HPO4)2, U(HPO4)3 , U(HPO4)4 , + UO2HPO04, UO2(HPO4)2 2 , UO2H2PO4, UO2(H2PO4)02, UO2(H2PO4)3 + 0  2 0 UF3+, UF2+ 2 , UF4, UF5 , UF6 , UO2F , UO2F2, UO2F3, UO2F42 UO2Cl+, UCl3+ UO2H3SiO+4

Uranyl carbonate species Uranyl sulfate species Uranyl phosphate species

Uranyl fluoride species Uranyl chloride species Uranyl silicate species

environment (see Table 4). At reducing conditions, U(IV) and its aqueous complexes predominate but the U(IV) concentrations in groundwater are usually low because of the extremely low solubility of its solids (Langmuir, 1997). At oxygenated and acidic conditions, uranium occurs predominantly in the U(VI) oxidation state and it is present as the uranyl cation (UO2+ 2 ) in aqueous systems (Langmuir, 1997; Dinh Chau et al., 2011). Uranyl cation further forms complexes with a wide range of inorganic ligands such as carbonate, fluoride, phosphate and sulfate, depending largely on the solution pH (Dinh Chau et al., 2011). Figure 2b and d show two examples of U(VI) speciation in different types of water. In the carbonate-type water with high fluoride, uranyl fluoride species dominate at acidic to neutral pH, while uranyl carbonate species become dominant at alkaline pH. In the sulfate-type water, uranyl and uranyel sulfate species dominate at acidic pH, while uranyl di- and tri-carbonate species take the predominance at neutral and alkaline pH. Unlike uranium, the speciation of fluoride in these two waters does not vary significantly, which makes the comparison of fluoride and uranium so interesting. Fluoride and uranium, two important contaminants but with extremely different characteristics, present a challenge to many treatment processes. NF and RO have been proven to effectively remove both contaminants from drinking water (Raff and Wilken, 1999; Lhassani et al., 2001; Favre-Réguillon et al., 2003, 2008; Tahaikt et al., 2007, 2008; Rossiter et al., 2010b). The unique possibility of replacing various treatment processes by a single technique makes NF/RO a very attractive and promising option for the drinking water industry.

5. Fluoride and uranium removal in NF/RO 5.1. General introduction of NF/RO In general, a membrane is a selective barrier between two phases. The permeation process takes place when a driving force is introduced to the system (Ayoob et al., 2008). The driving force

683

is generally a difference in chemical potential due to a pressure or concentration gradient across the membrane. Depending on membrane porosity and the trans-membrane gradient involved, a number of different types of membrane processes exists like RO, NF, ultrafiltration (UF) and microfiltration (MF). RO is generally accepted to be a dense membrane material with nonporous structure while UF and MF are well recognized as porous membranes. For NF, which is defined as a process between UF and RO that can separate multivalent ions from monovalent ions, it is somewhat arbitrary to claim that NF is either porous or nonporous, because the openness of NF lies in the spectrum between discrete pores (UF/MF) and dense materials (RO) (Wijmans and Baker, 1995; Bowen and Mukhtar, 1996). Some polymeric NF membranes are nonporous cross linked network structures, while inorganic NF membranes contain discrete pores with sizes in the order of 0.5– 2 nm (Mulder et al., 2005). For nonporous membranes, the transport mechanism is best described by the solution–diffusion model, wherein solutes dissolve in the dense material and then diffuse through the membrane down a concentration gradient (Wijmans and Baker, 1995). The solute transport rate and selectivity depends on interactions between solute and membrane materials. For porous membranes, solutes transfer through membrane pores following a pressure gradient. The transport rate and selectivity are primarily dominated by convective flow and sieving or size exclusion, combined with other mechanisms such as charge repulsion or sorption. As NF operates in the transition region between RO and UF, its transport mechanism is recognized to be complex and combine the characteristics of both nonporous and porous membranes (Bowen and Mukhtar, 1996). 5.2. Removal of fluoride and uranium by different NF/RO membranes Removal of fluoride by NF/RO from water has been reported to be successful (Ayoob et al., 2008). Dolar et al. (2011) published fluoride rejections of >96% with RO, >90% with tight NF and >50% with loose NF membranes. Tahaikt et al. (2008) compared fluoride rejections of three NF membranes with different configurations and found that the double pass with TR60 or NF270 membranes is comparable to the simple pass with NF90 which is similar to RO membranes. On the other hand, Lhassani et al. (2001) and Pontie et al. (2003) reported the advantage of NF over RO to be that NF rejects fluoride selectively from other halide ions. Pontie et al. (2012) further compared the performances of NF and RO membranes and indicated that NF permits to reduce partially the total salinity and remove fluoride to meet WHO guideline for feed concentrations up to 15 mg L1 with lower energy consumption than RO. Figure 3a provides a summary of fluoride removal by NF/RO reported in previous studies. The retention varies significantly with membrane types, inviting a thorough investigation on the retention mechanisms (influenced by operating conditions and water chemistry) for better understanding and predicting NF/RO performance. Uranium and its complexes are very heavy, which allows effective removal by NF/RO processes (Fig. 3b). Raff and Wilken (1999) found that NF membranes can remove 90–98% of uranium from drinking water, while uranium rejection by RO membranes is even higher (95–99%) (Gamal Khedr, 2013). Favre-Réguillon et al. (2008) indicated that NF can selectively remove uranium while allowing other trace minerals to pass through. Yurlova and Kryvoruchko (2010) found that the simultaneous use of modified montmorillonite increases uranium retention by NF to 99–99.9%. Considering the complexity of uranium speciation, the removal mechanism at different conditions is not always the same and hence uranium is an interesting contaminant to study.

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

Fluoride Retention (%)

100

a

b

100

80

80

60

60 i ii iii iv v vi vii viii

40

20

40 ix x xi xii xiii

20

Uranium Retention (%)

684

0

0 30 0LE F70 F90 270 400 C-S R3 80A A 4 60 R-1 DL HL 250 450 7 7 BW W3 N N NF NF T F FC-S TC- ESP T R S R- RB T U NT NT

30 F45 F90 C-S SR2 PA4 N T F C- ES BW N TF

Membrane Type

DK

DL

HL G10 G20

Membrane Type

Fig. 3. (a) Fluoride and (b) uranium retention by different NF/RO membranes (data adapted from i. (Richards, 2012), ii. (Tahaikt et al., 2008), iii. (Tahaikt et al., 2007), iv. (Lhassani et al., 2001), v. (the author of this paper), vi. (Hu and Dickson, 2006), vii. (Choi et al., 2001), viii. (Hong et al., 2007), ix. (Raff and Wilken, 1999), x. (Favre-Réguillon et al., 2003), xi. (Favre-Réguillon et al., 2008), xii. (Rossiter et al., 2010b), xiii. (Schulte-Herbrüggen, 2011)).

Four main NF/RO retention mechanisms will be thoroughly illustrated and discussed in this review paper. They are solution– diffusion, size exclusion, charge interaction and adsorption. The graphic abstract presented before the text demonstrates the dynamic interplay between mechanisms and affecting factors. Size exclusion and charge interaction are universal removal mechanisms for both fluoride and uranium while solution–diffusion is more important for fluoride and adsorption is more common for uranium due to their different characteristics. Water chemistry (concentration, pH, counter-ion presence), operating parameters (pressure, crossflow velocity), solute–solute interactions (and speciation), membrane characteristics and fouling are affecting certain mechanism in different ways. The schematic figure below visualizes the four mechanisms and corresponding affecting factors in more detail. It should be noted that the aim is to show principles and hence ions or molecules shown in the schematic are abstract and not to scale. For example, the structure of all uranium species is presented as a sphere for simplification. 5.3. Retention mechanism I: Solution–diffusion The solution–diffusion model is the most widely used transport model for permeation in a nonporous membrane (Wijmans and Baker, 1995). Solution–diffusion is basically a three-step chemical process (Fig. 4a): (1) solute leaves the solvent by dissolving in the membrane, (2) solute diffuses through the membrane driven by a concentration gradient, and (3) solute enters again in the solvent phase that permeates through the membrane. Step 1 and 3 are very fast relative to step 2, so diffusion through the membrane is the rate-limiting step in mass transfer. The diffusive flux (Jdiffusive, mol h1 m2) follows the general form of Fick’s first law as:

J diffusive ¼ D

dc dx

where dc/dx is the concentration gradient across the membrane (mol m4). D is the diffusion coefficient of the solute in the membrane (m2 s1). A recent defluoridation study conducted by the authors observed an ascending trend of permeate fluoride concentration with filtration volume in a stirred cell system (Fig. 5). This phenomenon can be well explained by the solution–diffusion model. In the stirred cell system, feed solution concentration increases continuously, leading to an increasing concentration gradient across the membrane. As a result, the diffusion rate of fluoride is increased, allowing more fluoride to permeate the membrane.

The diffusion rate is also influenced by the solubility limit of the solute within the polymer matrix. Selective fluoride retention from other solutes is attributed to the high solubility of fluoride in water, which makes it difficult to dissolve in the membrane. Hydration plays a key role because fluoride needs to be detached from its hydration shell to diffuse through the membrane. On the contrary, other ions such as nitrate and boron are less hydrated and are less retained by the membrane (Richards et al., 2010). 5.4. Retention mechanism II: Size exclusion Size exclusion is widely regarded as the basic mechanism of porous pressure driven membrane filtration processes (Mulder, 1996). In a size exclusion model, any solute larger than the pore size of the membrane is retained. In NF this is complicated because the sub-nanometer pore size of the membrane is not constant. Most membranes are characterized by a pore size distribution rather than a specific pore size which complicates size evaluations. Several authors have investigated the pore sizes of some commercial NF membranes and found that the approximate effective pore radius is between 0.4 and 1.5 nm with the majority in the range of 0.4–0.8 nm (Wang et al., 1995; Bowen and Mohammad, 1998; Mulder et al., 2005). This is obviously very dependent on membrane materials and structures. Equally, the generally small solutes have non-spherical and often flexible shapes. For the sake of simplification, the size exclusion model assumes that solutes are spheroids and the membrane pores are uniform cylindrical capillaries. The Stokes radius is defined as the effective size of a theoretical solid sphere that diffuses at the same speed as the target ion and is commonly used to describe the effective ion size (Bowen et al., 1997; Hussain et al., 2007). It is calculated as

rs ¼

kB T 6pgD

where rs is the Stokes radius (m), kB is the Boltzmann constant (J mol1 K1), g is viscosity (kg m1 s1), T is temperature (K). The ion with smaller Stokes radius has a higher mobility and vice versa. For dissolved ions like fluoride and uranium, the effective size needs to account for both ionic size and hydration layers. The dissolved ion is surrounded by water molecules due to the dipole nature of water and thus forms a larger entity (Tansel et al., 2006). As a result the effective radius of the hydrated ion is increased. David et al. (2001) concluded that the hydrated radius is determined by the charge density and the crystal radius of the central ion. Having

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

685

Fig. 4. Fluoride and uranium retention mechanisms in NF/RO (a) solution–diffusion; (b) size exclusion; (c) charge interactions; (d) adsorption; (e) influence of solute–solute interactions; and (f) influence of membrane fouling.

24 BW30 BW30-LE NF90 NF270 TFC-SR2 TFC-SR3

Permeate Fluoride (mg L-1)

21 18 15 12 9 6

WHO guideline 1.5 mg/L

3 0 0

10

20

30

40

50

Recovery (%) Fig. 5. Permeate fluoride concentration versus recovery by different NF/RO membranes (recovery is the ratio of permeate to feed solution volume, feed F 42 mg L1, TDS 1380 mg L1, applied pressure 1000 kPa, temperature 25 °C).

higher charge density, the smaller ions hold the water molecules more strongly. This provides implications for the higher retention of fluoride compared with other anions due to size exclusion (Richards et al., 2013). As seen from Fig. 4b, although the fluoride ion is smaller than the chloride ion, the former has a larger hydrated radius (0.35 nm) compared with the latter (0.33 nm)

(Richards, 2012). Chloride is more weakly hydrated and its hydration shell can be detached more easily, therefore it can pass through the membrane more readily (Tansel et al., 2006). Uranyl ion has a very high charge density, and tends to form a number of aqueous hydroxy species (see Table 4). The stable form of uranyl ion in aqueous solution is oxometal ion where the charge density is significantly reduced. Aaberg et al. (1983) first demonstrated the structure of the hydrated uranyl ion which has a linear O = U = O entity surrounded by five water molecules in a pentagonal geometry as the hydration shell (see simplified illustration in Fig. 4b). The distance between the uranium atom and the hydration shell is 0.24 nm. Nichols et al. (2008) further reported that uranyl ion has a second hydration shell and the relevant hydrated radius is 0.46 nm. Considering that uranyl ion is the smallest uranium species in water, it can be assumed that a majority of uranium species can be removed by NF due to size exclusion. FavreRéguillon et al. (2003) found that the membrane with larger pore size exhibited lower uranium rejection at the same uranium concentration (G20 membrane in Fig. 3b), further indicating the predominance of size exclusion in uranium removal.

5.5. Retention mechanism III: Charge interaction Charge interaction refers to the repulsion or attraction between a charged solute and a charged membrane. This mechanism reveals its importance especially when the size of the charged solute is smaller than the membrane pore size.

686

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

Solute charge varies with chemical speciation, which depends on the specific water chemistry, including pH, ionic strength, temperature, and available ligands (Morel and Hering, 1993). As mentioned in Fig. 2, fluoride exists as negative F- above pH 3.2 while the speciation of uranium is much more complicated because uranyl ion easily forms complexes with inorganic and organic ligands. 4 Uranyl carbonate UO2(CO3)2 2 and UO2(CO3)3 were shown to be the major individual components above pH 7 (Favre-Réguillon et al., 2008). Membrane charge is determined by the ionisable functional groups on the surface (Childress and Elimelech, 1996; Vankelecom et al., 2005). The isoelectric points of most NF membranes usually fall in the range of pH 3–6 so they are negatively charged at neutral and alkaline pH (Hagmeyer and Gimbel, 1999). Due to the dissociation of functional groups, charge repulsion between the polymer chains increases and forces the membrane pores to ‘open up’ (Braghetta et al., 1997). At the isoelectric point membrane functional groups are uncharged and hence the pores are believed to be smallest. However, some researchers reached contrary conclusions in that the absence of repulsion force actually contributes to the expansion of membrane pores (Childress and Elimelech, 2000). The point of contention is how charge repulsion between polymers affects the membrane pore size. More studies on the distribution of various functional groups in the membrane structure need to be carried out before a further conclusion can be drawn. Richards et al. (2010) studied the permeate flux of a number of NF/RO membranes as a function of pH. Surprisingly, the results showed that flux (permeability) is independent of pH for most membranes, indicating that the influence of membrane charge on pore size may indeed be overstated. Once the electrostatic properties of the solute and the membrane are determined, the charge interactions can be understood. Figure 4c highlights the effects of charge interactions on the behavior of fluoride and uranium in NF/RO. The first ruling principle is Donnan equilibrium as shown in Fig. 4c (i) (Schaep et al., 1998; Teixeira et al., 2005). In this mechanism charges on the membrane functional groups (represented as COO groups) prevent F co-ions from entering the membrane phase while attracting counter-ions. Meanwhile, an equivalent number of low charge counter-ions (Na+ in this case) are retained to satisfy the electroneutrality condition. This leads to the retention of the ion-pair NaF as a whole. In the case of various ion-pairs (1–1, 1–2, 2–1), the effect of charge becomes even more complex. Mulder et al. (2005) reviewed the retention of Na2SO4, NaCl and CaCl2 by a negative membrane and a positive membrane. For the negative membrane, the retention of Na2SO4 was highest and that of CaCl2 the lowest. For the positive membrane, the reversed order was observed with the highest retention for CaCl2 while Na2SO4 had the lowest retention. It was concluded that in salt mixtures the retention of single-charged co-ions decreased drastically whereas the retention of multiple-charged co-ions hardly changed (Mulder et al., 2005). Furthermore, the single-charged co-ions may even be negatively rejected, depending on the concentration ratio between single-charged and multiple-charged co-ions in the electrolyte solution (Tsuru et al., 1991; Mulder et al., 2005). The second principle is charge shielding, which refers to the shielding effect of counter-ions on membrane surface charge (Wang et al., 1997). Teixeira et al. (2005) found that calcium ions screened the ionized sites of a polyamide membrane so remarkably that zeta-potentials switched from negative to positive as the calcium concentration increased and the isoelectric point of the membrane was consequently modified. As shown in Fig. 4c (ii), the negatively charged uranyl complexes (mainly UO2(CO3)2 2 ) should be highly rejected by the membrane, but an increase in cation concentration shields the membrane charge and thus allows anion pass through. As a result, the retention of uranyl was decreased

by a factor of two (Favre-Réguillon et al., 2008). Equally, the negative effect of charge shielding on fluoride retention was also reported (Choi et al., 2001). Charge shielding is further associated with the valence of counter-ions, wherein higher counter-ion valence leads to lower retention due to more effective charge shielding (Nghiem et al., 2006). However, the shielding effect appears to be more valid for monovalent ions than divalent ions since the retention of divalent anion remains high despite the occurrence of charge shielding (Choi et al., 2001). Dielectric exclusion is a further established non-sieving mechanism in NF (Yaroshchuk, 2000, 2001; Bandini and Vezzani, 2003; Oatley et al., 2012). Dielectric exclusion occurs when an ion interacts with the bound electrical charges induced by the ion, at the interface between media of different dielectric constants. In this case the interface is between the membrane matrix and the solvent (Oatley et al., 2012). Due to the fact that the dielectric constant of the aqueous solution is significantly higher than that of the polymeric matrix, the polarization charges have the same sign as the ions in the aqueous solution. Consequently, the interaction always causes a repulsive effect on the ion (Bandini and Vezzani, 2003). The magnitude of the repulsive force can be described by the concept of ‘image force’ as illustrated in Fig. 4c (iii): the interaction between a charged ion and a polarized interface is equivalent to the interaction with a fictitious ion located at the other side of the interface (inside the membrane matrix) at the same distance from the interface as the real charge (Oatley et al., 2012). Yaroshchuk (2000) provided a comprehensive review of this exclusion mechanism and concluded that this interaction is dependent on the ion valance, the ratio of dielectric constants of the two media, and the pore geometry of the membrane. Dielectric exclusion is expected to be remarkably relevant to the high rejection of multivalent counter-ions by the uncharged membranes, for which Donnan exclusion is negligible (Bandini and Vezzani, 2003). However, the role of dielectric exclusion in the retention of fluoride and uranium species has not been sufficiently characterized, which provides fertile areas for future research. 5.6. Retention mechanism IV: Adsorption In membrane filtration processes, adsorption refers to the interactions between the solute and the membrane. Non-electrostatic adsorption mechanisms include physisorption and chemisorption. Physisorption generally refers to van der Waals type interactions while chemisorption involves formation of chemical bonds. Interactions via chemisorption are strong and expected to change the properties of the membrane (Han et al., 2012). As for fluoride, adsorption is generally not favorable due to the electrostatic repulsion between the negatively charged membrane and the anion. Besides, fluoride ions are highly hydrated in solution so hydrophobic adsorption is difficult to take place. However, it is in theory possible for fluoride to be adsorbed by the positively charged membrane, or by the positively charged groups of the negative membrane (Fig. 4d (i)). On the other hand, cations may modify the charge of the membrane and affect the adsorption of fluoride indirectly or allow fluoride to be adsorbed together with other ions as a complex. The adsorption of uranium is a combination of electrostatic interaction, chemical interaction, and hydrogen bonding. It is highly dependent on the particular species present at different conditions (Fig. 2b and Fig. 4d). For example, the positively charged species UO2(OH)+ is more adsorbed than the negatively charged species, which is mainly attributed to the charge attraction between the hydroxyl complex and the negatively charged membrane (Freger, 2003; Samuel de Lint et al., 2003). Theoretical studies indicate that UO2(OH)+ species can interact with carboxylic acid groups (Schlosser et al., 2006), while UO2(CO3)4 3 species tends to

687

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

5.7. Influence of operating parameters Operating parameters such as pressure, crossflow velocity and recovery significantly affect NF/RO performance. They may determine the dominant mechanisms by affecting the properties of the membrane and conditions at the membrane surface or boundary layer. Previous studies revealed that retention of fluoride and uranium generally increased with applied pressure (Fig. 6). Despite the contribution of diffusion, the increase in water flux surpasses the increase in convective transport of the solute, and therefore the retention increases. However, as the pressure increases, concentration polarization in the boundary layer increases and hence solute diffusion increases (Van de Lisdonk et al., 2001). Concentration polarization leads to increased osmotic pressure and thus decreased flux. This further leads to an increased risk of membrane fouling. For crossflow NF/RO, the predominant operation mode, crossflow velocity has a significant effect on concentration polarization and the ensuing permeate flux decline in addition to pressure (Bhattacharjee et al., 1999). For example, Koyuncu and Topacik (2003) observed that increased crossflow velocity resulted in higher rejection of NaCl, which was attributed to less concentration polarization under high crossflow conditions. 5.8. Influence of water chemistry: Concentration and pH The impact of concentration on fluoride retention is closely linked to diffusion since concentration gradient is the driving force for diffusive flux. High fluoride concentration at the boundary layer increases solution–diffusion and hence the permeate concentra-

Fluoride Retention (%)

100

tion. A summary of previous studies shows a descending trend of fluoride retention with increasing initial fluoride concentration (Fig. 7). Uranium, on the contrary, shows a positive correlation between concentration and retention as observed in previous studies (Prabhakar et al., 1992; Hoyer et al., 2014). This correlation is expected to be dependent on uranium speciation and membrane pore size, as uranium retention is predominantly governed by size exclusion. Solution pH is a very important parameter in NF/RO, because it affects not only solute speciation but also membrane characteristics (Childress and Elimelech, 1996; Childress and Elimelech, 2000). Therefore, pH can impact both solvent permeability and solute retention mechanisms, especially the size exclusion and charge repulsion. The influence of pH on fluoride and uranium retention is shown in Fig. 8. For fluoride, the uncharged HF is deprotonated at its acid dissociation constant (pKa = 3.2), and hence charge repulsion becomes significant. A change in fluoride species also affects hydration state and consequently hydrated radius, thus impacting retention when size exclusion is important (Richards et al., 2010). For uranium, the dominant species varies significantly with pH, but the retention is generally high and independent of speciation due to size exclusion (Hoyer et al., 2014). The strength of electrostatic interaction between ions and the membrane varies according to solution pH, with minimal interaction occurring around the isoelectric point of the membrane surface. This can be seen in Fig. 8a where fluoride retention increases with increasing membrane charge. Besides, pH may affect the openness of the membrane by controlling the dissociation of the functional groups on the membrane surface though the effect is not universal for all membranes (Braghetta et al., 1997; Childress and Elimelech, 2000). Bellona and Drewes (2005) found that an increasing pH can only increase the rejection of negatively charged organic solutes to a certain level before the effect of increased surface electronegativity is offset by the increased membrane openness. Richards et al. (2010) reported the similar phenomenon on fluoride retention but explained it using the hydration theory as they found the openness of the membrane used was not affected by pH. 5.9. Influence of solute–solute interactions In natural waters, co-exiting ions and natural organic matter (NOM) are important factors affecting fluoride and uranium retention in NF/RO. This is because, in addition to Donnan equilibrium and charge shielding effects, these solutes may interact with fluoride/uranium and consequently change the size or charge properties of the target ion. In this section, calcium ion is selected as

b

a

100

80

80

60

60 SR-1 (i) DL (i) HL (i) NF70 (ii) NF-1 (iii) NF-2 (iii) NF-20 (iii)

40 20

40 G10 (iv) DL (iv) TFC-SR2 (v) BW30 (v)

20

Uranium Retention (%)

form hydrogen bonding with amine groups (Prudden et al., 2004) (see Fig. 4d). Hydrogen bonding is a reversible process so the UO2(CO3)4 3 species can still penetrate the membrane at high applied pressure if size exclusion allows (Schulte-Herbrüggen, 2011). It can be concluded that at least three parameters play a role in the adsorption of uranium to the membrane: (1) the charge of uranium species and membrane, (2) the reactivity of uranium species towards membrane functional groups, and (3) uranium species size relative to the membrane pore size. As mentioned above, the four NF mechanisms, namely solution–diffusion, size exclusion, charge interaction and adsorption, are likely to co-exist for most water applications, but the relative contribution of each mechanism may change with operating conditions, water chemistry and, ultimately, fouling as well as the aging of membranes due to periodic cleaning. All those factors are bound to affect retention to some extent. In the following sections some of those factors will be reviewed.

0

0 0

200 400

600 800 1000 1200 1400 1600 1800 0

Applied Pressure (kPa)

200 400 600 800 1000 1200 1400 1600 1800

Applied Pressure (kPa)

Fig. 6. Effect of applied pressure on (a) Fluoride and (b) uranium retention by different NF/RO membranes (data adapted from i. (Hu and Dickson, 2006), ii. (Lhassani et al., 2001), iii. (Chakrabortty et al., 2013), iv. (Favre-Réguillon et al., 2008), v. (Schulte-Herbrüggen, 2011).

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

Fluoride Retention (%)

100

a

100

b

80

80 HL (iii) DL (iii) DK (iii) NF90 (iii) NF45 (iii) DL (iv) DK (iv) BW30 (v) ESPA4 (v) TFC-S (v) NF90 (v) M6 (vi)

60 NF270 (i) TR60 (i) NF90 (i) NF-1 (ii) NF-2 (ii) NF-20 (ii)

40 20

-- Retention required to meet WHO guideline

60 40 20

Uranium Retention (%)

688

-- Retention required to meet WHO guideline 0

0 0

5

10

15

20

25

30

35

Fluoride Concentration (mg L-1)

0

1

2

3

4

5

6

7

8

Uranium Concentration (mg L-1)

Fig. 7. Effect of concentration on (a) fluoride and (b) uranium retention by different NF/RO membranes (data adapted from i. (Tahaikt et al., 2008), ii. (Chakrabortty et al., 2013), iii. (Raff and Wilken, 1999) iv. (Favre-Réguillon et al., 2008) v. (Rossiter et al., 2010b) vi. (Hoyer et al., 2014)).

a

100

b

80

80 60 40

BW30 (i) TFC-S (i) NF90 (i) NF-1 (ii) NF-2 (ii) NF-20 (ii)

20

60

BW30 (iii) NF90 (iii) ESPA4 (iii) TFC-S (iii) HL (iv) DL (iv) DK (iv) NF90 (iv) NF45 (iv) M6 (v)

pKa=3.2

40 20

0 0

2

4

6

8

10

12

pH

14 0

2

4

6

8

10

Uranium Retention (%)

Fluoride Retention (%)

100

0 12

pH

Fig. 8. Effect of pH on (a) fluoride and (b) uranium retention by different NF/RO membranes (data adapted from i. (Richards et al., 2010), ii. (Chakrabortty et al., 2013), iii. (Rossiter et al., 2010b), iv. (Raff and Wilken, 1999) v. (Hoyer et al., 2014)).

the representative co-existing ion due to its high reactivity with both fluoride and uranium. Calcium and fluoride tend to form the insoluble CaF2 via a stoichiometric reaction in the pH range of 4–10 (Tokunaga et al., 1995). Indeed, the precipitation of CaF2 from fluoride-containing water is one of the most used conventional defluoridation methods (Ayoob et al., 2008). Fluoride concentration is the limiting factor of this interaction (Jackson et al., 2002). Once the concentration exceeds a certain level, F can even replace CO2 from 3 CaCO3 (Ksp = 3.36  109) to form the more insoluble CaF2 (Ksp = 3.45  1011) (Farrah et al., 1985; Lide, 1995). In natural groundwater, Ca2+ and F are not likely to form precipitation due to the low concentrations, but in NF/RO processes where Ca2+ and F are both concentrated at the membrane surface, such precipitation occurs and becomes a major source for membrane scaling (discussed in the next section). Uranium can also be influenced by calcium due to the formation of stable calcium–uranyl–carbonate complexes (Tsushima et al., 2002; Prat et al., 2009). Schulte-Herbrüggen (2011) studied the impact of these complexes on uranium retention in NF/RO. Ca2UO2(CO3)3 dominated uranium species at pH 8–9, and because it is a neutral species and very stable in solution, it is less likely to be affected by membrane surface charge (Fig. 4e). As a result, uranium adsorption to the membrane is much lower than without calcium. Above pH 10, calcium causes co-precipitation of uranium by CaCO3 precipitation, and leads to higher uranium retention along with a large decline of permeate flux (30–40%). In addition to calcium, NOM can form various soluble complexes with uranium (Murphy et al., 1999; Kantar, 2007). Semião et al.

(2010) investigated the influence of NOM on uranium removal by UF. The uranium–NOM complexation increases the amounts of uranium deposited to the UF membrane, but the complexes are still smaller than the UF pore size and thus cannot be retained. Considering the effective removal of NOM by NF/RO, any uranium bound to NOM is expected to be efficiently rejected (Fig. 4e). As for fluoride, it is unlikely to bond with NOM to a significant degree because it is highly hydrated and negatively charged in water (Richards et al., 2010). But cations may act as a bridge in fluoride–NOM interactions as they have a good affinity to both fluoride and NOM and given the diversity of NOM positive functional groups are likely to exist. Lund et al. (2008) investigated the binding of fluoride to a hydrophobic macromolecule using molecular dynamics simulations. The results showed that fluoride, as a strongly hydrated ion, can bind to the metal ion loaded macromolecule because specific cation–anion interactions overcome the hydrophobic repulsion. The potential fluoride–NOM interactions have great implications for improving fluoride retention in NF/RO.

5.10. Influence of membrane fouling Apart from the above-mentioned mechanisms, membrane fouling in NF/RO plays a noteworthy role on fluoride and uranium retention. Fouling can be categorized into four types: biofouling, colloidal fouling, inorganic scaling and organic fouling (Schäfer et al., 2004). In this review, the latter two will be discussed because groundwater has a high risk of inorganic scaling due to high concentrations of inorganic ions, while surface water often contains

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

NOM which is likely to affect both uranium speciation and membrane characteristics. Inorganic scaling occurs as the concentration of inorganic ions at the membrane surface exceeds the solubility limit. One of the most common scalants is Ca2+ (Schäfer et al., 2004). Ca2+ forms sparingly soluble salts such as CaCO3, CaF2 and CaSO4, and precipitates on the membrane surface to cause membrane scaling (Fig. 4f (i)). Tu et al. (2011) observed decreased retention of boron by CaSO4 fouled membrane due to cake-enhanced concentration polarization mechanism. Such fouling may further screen the negative charge of the membrane surface, and consequently reduce the retention of negative solutes. Inorganic scaling might decrease fluoride retention in the same way, but the mechanism would be more complicated as fluoride may precipitate in the filtration process. Organic fouling exhibits a variety of characteristics due to the complex interactions between functional groups of organic foulants and those of the polymeric membranes (Hong and Elimelech, 1997; Schäfer et al., 2004). It has been revealed that pore blocking is the predominant fouling mechanism at the first stage, and the latter stage is governed by cake-enhanced mechanism (Nghiem and Hawkes, 2009). Several previous studies found that organic fouling can significantly increase or decrease ion retention (Schäfer et al., 2004; Tang et al., 2006; Tu et al., 2011). The formation of a denser and more negative fouling layer increases size and charge exclusion (Fig. 4f (ii)), whereas cake-enhanced concentration polarization tends to reduce retention. Both inorganic and organic fouling can be avoided by early enough membrane cleaning. Sehn (2008) reported the anti-fouling strategies of a large scale RO plant in Finland. The plant manages to employ regular flushing and preventive cleaning to control the risk of membrane fouling. After more than three years of operation, salt and fluoride rejections remain constant.

6. Conclusions Globally, fluoride and uranium contaminations in drinking water are the results of natural and human activities, and pose a huge risk to human health. This review paper describes the application of NF/RO in the removal of fluoride and uranium from drinking water. NF/RO retention normally reaches more than 95%, although this is dependent on membrane and water characteristics. Retention is generally governed by four mechanisms including solution–diffusion, size exclusion, charge interaction, and membrane adsorption. For fluoride, both size exclusion and charge interaction contribute most to the high retention. For uranium, the retention pattern correlates closely with the speciation of uranium but size exclusion usually plays a predominant role for all species. Adsorption on the membrane occurs as some uranium species are prone to interact with membrane functional groups. The impacts of operating parameters, water quality, solute–solute interactions and membrane fouling are critically reviewed. Pressure is the driving force for convective flux and strongly affects size exclusion. Crossflow velocity affects concentration polarization and thus the solution–diffusion process. Ion concentration affects both solution–diffusion and charge repulsion. Solution pH is strongly linked to ion speciation and dissociation of membrane functional groups, therefore determining electrostatic interactions between the ion and the membrane. Solute–solute interaction may enhance the retention due to size exclusion. Membrane fouling can either reduce the retention due to cake-enhanced concentration polarization, or improve the retention because of the formation of a denser and more negative fouling layer.

689

Acknowledgements The authors thank Leverhulme Royal Society Africa Award SADWAT-Tanzania for project funding. The PhD studentship for Junjie Shen was provided by Energy Technology Partnership (ETP) Scholarship with Drinking Water Quality Regulator for Scotland (DWQR) as industrial sponsor. Godfrey Mkongo (Ngurdoto Defluoridation Research Station (NDRS), Arusha, Tanzania) is acknowledged for his local support. Prof Bryce Richards (Karlsruhe Institute of Technology, Germany) and Dr Laura Richards (University of Manchester, UK) contributed with valuable scientific comments.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.09.090. References

Aaberg, M., Ferri, D., Glaser, J., Grenthe, I., 1983. Structure of the hydrated dioxouranium(VI) ion in aqueous solution. An X-ray diffraction and proton NMR study. Inorg. Chem. 22, 3986–3989. Abdelouas, A., Lutze, W., Nuttall, E., 1998. Chemical reactions of uranium in ground water at a mill tailings site. J. Contam. Hydrol. 34, 343–361. Agarwal, V., Vaish, A.K., Vaish, P., 1997. Ground water quality: focus on fluoride and fluorosis in Rajasthan. Curr. Sci. India 73, 743–746. Almås, Å.R., Kweyunga, C., Manoko, M.L. 2009. Investigation of trace metal concentrations in soil, sediments and waters in the vicinity of gold mines in North West Tanzania, Norwegian University of Life Sciences. Amini, M., Mueller, K., Abbaspour, K.C., Rosenberg, T., Afyuni, M., Møller, K.N., Sarr, M., Johnson, C.A., 2008. Statistical modeling of global geogenic fluoride contamination in groundwaters. Environ. Sci. Technol. 42, 3662–3668. Åström, M.E., Peltola, P., Rönnback, P., Lavergren, U., Bergbäck, B., Tarvainen, T., Backman, B., Salminen, R., 2009. Uranium in surface and groundwaters in Boreal Europe. Geochem. Explor. Environ. Anal. 9, 51–62. Ayoob, S., Gupta, A.K., 2006. Fluoride in drinking water: a review on the status and stress effects. Crit. Rev. Environ. Sci. Tech. 36, 433–487. Ayoob, S., Gupta, A.K., Bhat, V.T., 2008. A conceptual overview on sustainable technologies for the defluoridation of drinking water. Crit. Rev. Environ. Sci. Tech. 38, 401–470. Azbar, N., Türkman, A., 2000. Defluoridation in drinking waters. Water Sci. Technol. 42, 403–407. Bandini, S., Vezzani, D., 2003. Nanofiltration modeling: the role of dielectric exclusion in membrane characterization. Chem. Eng. Sci. 58, 3303–3326. Bellona, C., Drewes, J.E., 2005. The role of membrane surface charge and solute physico-chemical properties in the rejection of organic acids by NF membranes. J. Membr. Sci. 249, 227–234. Berisha, F., Goessler, W., 2013. Uranium in Kosovo’s drinking water. Chemosphere 93, 2165–2170. Bhatnagar, A., Kumar, E., Sillanpää, M., 2011. Fluoride removal from water by adsorption—a review. Chem. Eng. J. 171, 811–840. Bhattacharjee, S., Kim, A.S., Elimelech, M., 1999. Concentration polarization of interacting solute particles in cross-flow membrane filtration. J. Colloid Interface Sci. 212, 81–99. Binnie, C., Kimber, M., 2009. Basic Water Treatment, fourth ed. UK, Thomas Telford Limited. Bleise, A., Danesi, P.R., Burkart, W., 2003. Properties, use and health effects of depleted uranium (DU): a general overview. J. Environ. Radioactive 64, 93–112. Bowen, W.R., Mohammad, A.W., 1998. Characterization and prediction of nanofiltration membrane performance—a general assessment. Chem. Eng. Res. Des. 76, 885–893. Bowen, W.R., Mukhtar, H., 1996. Characterisation and prediction of separation performance of nanofiltration membranes. J. Membr. Sci. 112, 263–274. Bowen, W.R., Mohammad, A.W., Hilal, N., 1997. Characterisation of nanofiltration membranes for predictive purposes—use of salts, uncharged solutes and atomic force microscopy. J. Membr. Sci. 126, 91–105. Braghetta, A., DiGiano, F.A., Ball, W.P., 1997. Nanofiltration of natural organic matter: pH and ionic strength effects. J. Environ. Eng. 123, 628–641. British Geological Survey, 2000. Groundwater quality: Tanzania. Chakrabortty, S., Roy, M., Pal, P., 2013. Removal of fluoride from contaminated groundwater by cross flow nanofiltration: transport modeling and economic evaluation. Desalination 313, 115–124. Childress, A.E., Elimelech, M., 1996. Effect of solution chemistry on the surface charge of polymeric reverse osmosis and nanofiltration membranes. J. Membr. Sci. 119, 253–268.

690

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691

Childress, A.E., Elimelech, M., 2000. Relating nanofiltration membrane performance to membrane charge (electrokinetic) characteristics. Environ. Sci. Technol. 34, 3710–3716. Choi, S., Yun, Z., Hong, S., Ahn, K., 2001. The effect of co-existing ions and surface characteristics of nanomembranes on the removal of nitrate and fluoride. Desalination 133, 53–64. Choi, A.L., Sun, G., Zhang, Y., Grandjean, P., 2012. Developmental fluoride neurotoxicity: a systematic review and meta-analysis. Environ. Health Perspect. 120, 1362–1368. David, F., Vokhmin, V., Ionova, G., 2001. Water characteristics depend on the ionic environment. Thermodynamics and modelisation of the aquo ions. J. Mol. Liq. 90, 45–62. Dinh Chau, N., Dulinski, M., Jodlowski, P., Nowak, J., Rozanski, K., Sleziak, M., Wachniew, P., 2011. Natural radioactivity in groundwater – a review. Isot. Environ. Health S. 47, 415–437. Dolar, D., Košutic´, K., Vucˇic´, B., 2011. RO/NF treatment of wastewater from fertilizer factory—removal of fluoride and phosphate. Desalination 265, 237–241. Edmunds, W.M., Smedley, P., 2005. Fluoride in natural waters. Essentials of smedical geology: impacts of the natural environment on public health. In: Selinus, O, Alloway, B, Centeno, J.A. (Eds.). Academic Press, US, pp. 301–329. Farrah, H., Slavek, J., Pickering, W., 1985. Fluoride sorption by soil components: calcium carbonate, humic acid, manganese dioxide and silica. Soil Res. 23, 429– 439. Favre-Réguillon, A., Lebuzit, G., Foos, J., Guy, A., Draye, M., Lemaire, M., 2003. Selective concentration of uranium from seawater by nanofiltration. Ind. Eng. Chem. Res. 42, 5900–5904. Favre-Réguillon, A., Lebuzit, G., Murat, D., Foos, J., Mansour, C., Draye, M., 2008. Selective removal of dissolved uranium in drinking water by nanofiltration. Water Res. 42, 1160–1166. Fawell, J., Bailey, K., Chilton, J., Dahi, E., Fewtrell, L., Magara, Y., 2006. Fluoride in drinking water. London World Health Organization. Freger, V., 2003. Nanoscale heterogeneity of polyamide membranes formed by interfacial polymerization. Langmuir 19, 4791–4797. Gaciri, S.J., Davies, T.C., 1993. The occurrence and geochemistry of fluoride in some natural waters of Kenya. J. Hydrol. 143, 395–412. Gamal Khedr, M., 2013. Radioactive contamination of groundwater, special aspects and advantages of removal by reverse osmosis and nanofiltration. Desalination 321, 47–54. Ghiglieri, G., Pittalis, D., Cerri, G., Oggiano, G., 2012. Hydrogeology and hydrogeochemistry of an alkaline volcanic area: the NE Mt. Meru slope (East African Rift – Northern Tanzania). Hydrol. Earth Syst. Sci. 16, 529–541. Hagmeyer, G., Gimbel, R., 1999. Modelling the rejection of nanofiltration membranes using zeta potential measurements. Sep. Purif. Technol. 15, 19–30. Hakonson-Hayes, A.C., Fresquez, P.R., Whicker, F.W., 2002. Assessing potential risks from exposure to natural uranium in well water. J. Environ. Radioactive 59, 29– 40. Han, J., Qiu, W., Hu, J., Gao, W., 2012. Chemisorption of estrone in nylon microfiltration membranes: adsorption mechanism and potential use for estrone removal from water. Water Res. 46, 873–881. Hong, S., Elimelech, M., 1997. Chemical and physical aspects of natural organic matter (NOM) fouling of nanofiltration membranes. J. Membr. Sci. 132, 159– 181. Hong, S.U., Malaisamy, R., Bruening, M.L., 2007. Separation of fluoride from other monovalent anions using multilayer polyelectrolyte nanofiltration membranes. Langmuir 23, 1716–1722. Hoyer, M., Zabelt, D., Steudtner, R., Brendler, V., Haseneder, R., Repke, J.-U., 2014. Influence of speciation during membrane treatment of uranium contaminated water. Sep. Purif. Technol. 132, 413–421. Hu, K., Dickson, J.M., 2006. Nanofiltration membrane performance on fluoride removal from water. J. Membr. Sci. 279, 529–538. Hussain, A.A., Abashar, M.E.E., Al-Mutaz, I.S., 2007. Influence of ion size on the prediction of nanofiltration membrane systems. Desalination 214, 150–166. Jackson, P.J., Harvey, P.W., Young, W.F., 2002. Chemistry and bioavailability aspects of fluoride in drinking water Report No: CO 5037. Henley Road, Medmenham, Marlow, Bucks, SL7 2HD, WRc-NSF Ltd. Kantar, C., 2007. Heterogeneous processes affecting metal ion transport in the presence of organic ligands: reactive transport modeling. Earth-Sci. Rev. 81, 175–198. Kim, K., Jeong, G.Y., 2005. Factors influencing natural occurrence of fluoride-rich groundwaters: a case study in the southeastern part of the Korean Peninsula. Chemosphere 58, 1399–1408. Koyuncu, I., Topacik, D., 2003. Effects of operating conditions on the salt rejection of nanofiltration membranes in reactive dye/salt mixtures. Sep. Purif. Technol. 33, 283–294. Kurttio, P., Auvinen, A., Salonen, L., Saha, H., Pekkanen, J., Makelainen, I., Vaisanen, S.B., Penttila, I.M., Komulainen, H., 2002. Renal effects of uranium in drinking water. Environ. Health Perspect. 110, 337–342. Kwasnik, W., 1963. Fluorine compounds. In: Brauer, G. (Ed.), Handbook of Preparative Inorganic Chemistry, vol. 1. Academic Press, New York, US. Langmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice Hall, New Jersey. Lhassani, A., Rumeau, M., Benjelloun, D., Pontie, M., 2001. Selective demineralization of water by nanofiltration application to the defluorination of brackish water. Water Res. 35, 3260–3264. Lide, D.R. (Ed.), 1995. CRC Handbook of Chemistry and Physics, seventysixth ed. CRC Press, US.

Lund, M., Vacha, R., Jungwirth, P., 2008. Specific ion binding to macromolecules: effects of hydrophobicity and ion pairing. Langmuir 24, 3387–3391. Mbonile, M.J., 2005. Migration and intensification of water conflicts in the Pangani Basin, Tanzania. Habitat Int. 29, 41–67. Morel, F.M.M., Hering, J., 1993. Principles and Applications of Aquatic Chemistry. Wiley-Interscience, New York. Mulder, M., 1996. Basic Principles of Membrane Technology, second ed. Kluwer Academic Publishers, The Netherlands. Mulder, M.H.V., van Voorthuizen, E.M., Peeters, J.M.M., 2005. Membrane characterization. Nanofiltration-principles and applications. In: Schäfer, A.I., Fane, A.G., Waite, T.D. (Eds.). Elsevier, Oxford, UK. Muller, W.J., Heath, R.G.M., Villet, M.H., 1998. Finding the optimum: fluoridation of potable water in South Africa. Water SA 24, 21–27. Murphy, R.J., Lenhart, J.J., Honeyman, B.D., 1999. The sorption of thorium (IV) and uranium (VI) to hematite in the presence of natural organic matter. Colloids Surf. A 157, 47–62. Ndiaye, P.I., Moulin, P., Dominguez, L., Millet, J.C., Charbit, F., 2005. Removal of fluoride from electronic industrial effluent by RO membrane separation. Desalination 173, 25–32. Nghiem, L.D., Hawkes, S., 2009. Effects of membrane fouling on the nanofiltration of trace organic contaminants. Desalination 236, 273–281. Nghiem, L.D., Schäfer, A.I., Elimelech, M., 2006. Role of electrostatic interactions in the retention of pharmaceutically active contaminants by a loose nanofiltration membrane. J. Membr. Sci. 286, 52–59. Nichols, P., Bylaska, E.J., Schenter, G.K., de Jong, W., 2008. Equatorial and apical solvent shells of the UO2+ 2 ion. J. Chem. Phys. 128, 124507–124508. Oatley, D.L., Llenas, L., Pérez, R., Williams, P.M., Martínez-Lladó, X., Rovira, M., 2012. Review of the dielectric properties of nanofiltration membranes and verification of the single oriented layer approximation. Adv. Colloid Interface 173, 1–11. Orloff, K.G., Mistry, K., Charp, P., Metcalf, S., Marino, R., Shelly, T., Melaro, E., Donohoe, A.M., Jones, R.L., 2004. Human exposure to uranium in groundwater. Environ. Res. 94, 319–326. Ozsvath, D., 2009. Fluoride and environmental health: a review. Rev. Environ. Sci. Biotechnol. 8, 59–79. Paoloni, J.D., Fiorentino, C.E., Sequeira, M.E., 2003. Fluoride contamination of aquifers in the southeast subhumid pampa, Argentina. Environ. Toxicol. 18, 317–320. Payne, T.E., Edis, R., Fenton, B.R., Waite, T.D., 2001. Comparison of laboratory uranium sorption data with ‘in situ distribution coefficients’ at the Koongarra uranium deposit, Northern Australia. J. Environ. Radioactive 57, 35–55. Pirlo, M.C., Giblin, A.M., 2004. Application of groundwater–mineral equilibrium calculationsto geochemical exploration for sediment-hosted uranium:observations from the Frome Embayment, South Australia. Geochem. Explor. Environ. Anal. 4, 113–127. Pontie, M., Buisson, H., Diawara, C.K., Essis-Tome, H., 2003. Studies of halide ions mass transfer in nanofiltration—application to selective defluorination of brackish drinking water. Desalination 157, 127–134. Pontie, M., Dach, H., Lhassani, A., Diawara, C.K., 2012. Water defluoridation using nanofiltration vs. reverse osmosis: the first world unit, Thiadiaye (Senegal). Desalin. Water Treat. 51, 164–168. Prabhakar, S., Panicker, S.T., Misra, B.M., Ramani, M.P.S., 1992. Studies on the reverse osmosis treatment of uranyl nitrate solution. Sep. Sci. Technol. 27, 349–359. Prat, O., Vercouter, T., Ansoborlo, E., Fichet, P., Perret, P., Kurttio, P.I., Salonen, L., 2009. Uranium speciation in drinking water from drilled wells in Southern Finland and its potential links to health effects. Environ. Sci. Technol. 43, 3941– 3946. Prudden, A.R., Lien, N.R., Telford, J.R., 2004. An outer-sphere ligand for uranyl carbonate. Chem. Commun., 172–173. Queste, A., Lacombe, M., Hellmeier, W., Hillermann, F., Bortulussi, B., Kaup, M., Ott, K., Mathys, W., 2001. High concentrations of fluoride and boron in drinking water wells in the Muenster region-results of a preliminary investigation. Int. J. Hyg. Environ. Health 203, 221–224. Radespiel-Tröger, M., Meyer, M., 2013. Association between drinking water uranium content and cancer risk in Bavaria, Germany. Int. Arch. Occup. Environ. Health 86, 767–776. Raff, O., Wilken, R.-D., 1999. Removal of dissolved uranium by nanofiltration. Desalination 122, 147–150. Rango, T., Kravchenko, J., Atlaw, B., McCornick, P.G., Jeuland, M., Merola, B., Vengosh, A., 2012. Groundwater quality and its health impact: an assessment of dental fluorosis in rural inhabitants of the Main Ethiopian Rift. Environ. Int. 43, 37–47. Richards, L.A., 2012. The removal of inorganic contaminants using nanofiltration and reverse osmosis. School of Engineering and Physical Sciences. Edinburgh, UK, Heriot-Watt University. PhD thesis. Richards, B.S., Capão, D.P.S., Schäfer, A.I., 2008. Renewable energy powered membrane technology. 2. The effect of energy fluctuations on performance of a photovoltaic hybrid membrane system. Environ. Sci. Technol. 42, 4563– 4569. Richards, L.A., Vuachère, M., Schäfer, A.I., 2010. Impact of pH on the removal of fluoride, nitrate and boron by nanofiltration/reverse osmosis. Desalination 261, 331–337. Richards, L.A., Richards, B.S., Corry, B., Schäfer, A.I., 2013. Experimental energy barriers to anions transporting through nanofiltration membranes. Environ. Sci. Technol. 47, 1968–1976. Rossiter, H.M.A., Graham, M.C., Schäfer, A.I., 2010a. Impact of speciation on behaviour of uranium in a solar powered membrane system for treatment of brackish groundwater. Sep. Purif. Technol. 71, 89–96.

J. Shen, A. Schäfer / Chemosphere 117 (2014) 679–691 Rossiter, H.M.A., Owusu, P.A., Awuah, E., MacDonald, A.M., Schäfer, A.I., 2010b. Chemical drinking water quality in Ghana: water costs and scope for advanced treatment. Sci. Total Environ. 408, 2378–2386. Samuel de Lint, W.B., Benes, N.E., Lyklema, J., Bouwmeester, H.J.M., van der Linde, A.J., Wessling, M., 2003. Ion adsorption parameters determined from zeta potential and titration data for a c-alumina nanofiltration membrane. Langmuir 19, 5861–5868. Schaep, J., Van der Bruggen, B., Vandecasteele, C., Wilms, D., 1998. Influence of ion size and charge in nanofiltration. Sep. Purif. Technol. 14, 155–162. Schäfer, A.I., Andritsos, N., Karabelas, A.J., Hoek, E.M.V., Schneider, R., Nyström, M., 2004. Fouling in nanofiltration. Nanofiltration: principles and applications. In: Schäfer, A.I., Fane, A.G., Waite, T.D. (Eds.). Elsevier. Schäfer, A.I., Broeckmann, A., Richards, B.S., 2006. Renewable energy powered membrane technology. 1. Development and characterization of a photovoltaic hybrid membrane system. Environ. Sci. Technol. 41, 998–1003. Schlosser, F., Krüger, S., Rösch, N., 2006. A density functional study of uranyl monocarboxylates. Inorg. Chem. 45, 1480–1490. Schmoll, O., Howard, G., Chilton, J., Chorus, I. (Eds.), 2006. Protecting Groundwater for Health. World Health Organization, UK. Schulte-Herbrüggen, H.M.A., 2011. Remote community drinking water supply – mechanisms of uranium retention and adsorption by ultrafiltration, nanofiltration and reverse osmosis. School of Engineering. Edinburgh, UK, The University of Edinburgh. PhD Thesis. Schwarzenbach, R.P., Escher, B.I., Fenner, K., Hofstetter, T.B., Johnson, C.A., von Gunten, U., Wehrli, B., 2006. The challenge of micropollutants in aquatic systems. Science 313, 1072–1077. Segreto, V.A., Collins, E.M., Camann, D., Smith, C.T., 1984. A current study of mottled enamel in Texas. J. Am. Dent. Assoc. 108, 56–59. Sehn, P., 2008. Fluoride removal with extra low energy reverse osmosis membranes: three years of large scale field experience in Finland. Desalination 223, 73–84. Semião, A.J.C., Rossiter, H.M.A., Schäfer, A.I., 2010. Impact of organic matter and speciation on the behaviour of uranium in submerged ultrafiltration. J. Membr. Sci. 348, 174–180. Sivakumar, B., 2011. Water crisis: from conflict to cooperation—an overview. Hydrol. Sci. J. 56, 531–552. Smedley, P.L., Nkotagu, H., Pelig-Ba, K., MacDonald, A.M., Tyler-Whittle, R., Whitehead, E.J., Kinniburgh, D.G., 2002. Fluoride in Groundwater from High-fluoride Areas of Ghana and Tanzania. British Geological Survey, Nottingham. Tahaikt, M., El Habbani, R., Ait Haddou, A., Achary, I., Amor, Z., Taky, M., Alami, A., Boughriba, A., Hafsi, M., Elmidaoui, A., 2007. Fluoride removal from groundwater by nanofiltration. Desalination 212, 46–53. Tahaikt, M., Ait Haddou, A., El Habbani, R., Amor, Z., Elhannouni, F., Taky, M., Kharif, M., Boughriba, A., Hafsi, M., Elmidaoui, A., 2008. Comparison of the performances of three commercial membranes in fluoride removal by nanofiltration. Continuous operations. Desalination 225, 209–219. Tang, C.Y., Kwon, Y.-N., Leckie, J.O., 2006. Characterization of humic acid fouled reverse osmosis and nanofiltration membranes by transmission electron microscopy and streaming potential measurements. Environ. Sci. Technol. 41, 942–949. Tansel, B., Sager, J., Rector, T., Garland, J., Strayer, R.F., Levine, L., Roberts, M., Hummerick, M., Bauer, J., 2006. Significance of hydrated radius and hydration shells on ionic permeability during nanofiltration in dead end and cross flow modes. Sep. Purif. Technol. 51, 40–47. Teixeira, M.R., Rosa, M.J., Nyström, M., 2005. The role of membrane charge on nanofiltration performance. J. Membr. Sci. 265, 160–166. Tokunaga, S., Haron, M.J., Wasay, S.A., Wong, K.F., Laosangthum, K., Uchiumi, A., 1995. Removal of fluoride ions from aqueous solutions by multivalent metal compounds. Int. J. Environ. Stud. 48, 17–28.

691

Tsuru, T., Urairi, M., Nakao, S.-I., Kimura, S., 1991. Negative rejection of anions in the loose reverse osmosis separation of mono- and divalent ion mixtures. Desalination 81, 219–227. Tsushima, S., Uchida, Y., Reich, T., 2002. A theoretical study on the structures of 0 0 UO2(CO3)4 3 , Ca2UO2(CO3)3, and Ba2UO2(CO3)3. Chem. Phys. Lett. 357, 73–77. Tu, K.L., Chivas, A.R., Nghiem, L.D., 2011. Effects of membrane fouling and scaling on boron rejection by nanofiltration and reverse osmosis membranes. Desalination 279, 269–277. Um, W., Icenhower, J.P., Brown, C.F., Serne, R.J., Wang, Z., Dodge, C.J., Francis, A.J., 2010. Characterization of uranium-contaminated sediments from beneath a nuclear waste storage tank from Hanford, Washington: implications for contaminant transport and fate. Geochim. Cosmochim. Acta 74, 1363–1380. Uralbekov, B.M., Smodis, B., Burkitbayev, M., 2011. Uranium in natural waters sampled within former uranium mining sites in Kazakhstan and Kyrgyzstan. J. Radioanal. Nucl. Chem. 289, 805–810. Van de Lisdonk, C.A.C., Rietman, B.M., Heijman, S.G.J., Sterk, G.R., Schippers, J.C., 2001. Prediction of supersaturation and monitoring of scaling in reverse osmosis and nanofiltration membrane systems. Desalination 138, 259–270. Vankelecom, I.F.J., Smet, K.D., Gevers, L.E.M., Jacobs, P.A., 2005. Nanofiltration membrane materials and preparation. Nanofiltration-principles and applications. In: Schäfer, A.I., Fane, A.G., Waite, T.D. (Eds.). Elsevier, Oxford, UK. Wang, X.-L., Tsuru, T., Togoh, M., Nakao, S.-I., Kimura, S., 1995. Evaluation of pore structure and electrical properties of nanofiltration membranes. J. Chem. Eng. Jpn. 28, 186–192. Wang, X.-L., Tsuru, T., Nakao, S.-I., Kimura, S., 1997. The electrostatic and sterichindrance model for the transport of charged solutes through nanofiltration membranes. J. Membr. Sci. 135, 19–32. Wang, S.X., Wang, Z.H., Cheng, X.T., Li, J., Sang, Z.P., Zhang, X.D., Han, L.L., Mao, X.Y., Wu, Z.M., Wang, Z.Q., 2007. Arsenic and fluoride exposure in drinking water: children’s IQ and growth in Shanyin county, Shanxi province, China. Environ. Health Perspect. 115, 643–647. WHO, 2011. Uranium in drinking-water. Background document for development of WHO guidelines for drinking-water quality. Geneva. Wijmans, J.G., Baker, R.W., 1995. The solution–diffusion model: a review. J. Membr. Sci. 107, 1–21. Winde, F., Jacobus van der Walt, I., 2004. The significance of groundwater–stream interactions and fluctuating stream chemistry on waterborne uranium contamination of streams—a case study from a gold mining site in South Africa. J. Hydrol. 287, 178–196. World Nuclear Association, 2012. World Uranium Mining Production. . Retrieved July 2013. Yaroshchuk, A.E., 2000. Dielectric exclusion of ions from membranes. Adv. Colloid Interface 85, 193–230. Yaroshchuk, A.E., 2001. Non-steric mechanisms of nanofiltration: superposition of Donnan and dielectric exclusion. Sep. Purif. Technol. 22–23, 143–158. Yurlova, L.Y., Kryvoruchko, A.P., 2010. Purification of uranium-containing waters by the ultra- and nanofiltration using modified montmorillonite. J. Water Chem. Technol. 32, 358–364. Zamora, M.L., Tracy, B.L., Zielinski, J.M., Meyerhof, D.P., Moss, M.A., 1998. Chronic ingestion of uranium in drinking water: a study of kidney bioeffects in humans. Toxicol. Sci. 43, 68–77. Zamora, M.L.L., Zielinski, J.M., Moodie, G.B., Falcomer, R.A.F., Hunt, W.C., Capello, K., 2009. Uranium in drinking water: renal effects of long-term ingestion by an aboriginal community. Arch. Environ. Occup. Health 64, 228–241. Zhang, B., Hong, M., Zhao, Y., Lin, X., Zhang, X., Dong, J., 2003. Distribution and risk assessment of fluoride in drinking water in the west plain region of Jilin Province, China. Environ. Geochem. Health 25, 421–431.