Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) by RO membranes

Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) by RO membranes

Journal of Membrane Science 245 (2004) 71–78 Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) b...

148KB Sizes 0 Downloads 38 Views

Journal of Membrane Science 245 (2004) 71–78

Rejection of neutral endocrine disrupting compounds (EDCs) and pharmaceutical active compounds (PhACs) by RO membranes Katsuki Kimuraa,∗ , Shiho Toshimaa , Gary Amyb , Yoshimasa Watanabea b

a Department of Urban and Environmental Engineering, Hokkaido University, Sapporo 060-8628, Japan Department of Civil, Environmental, and Architectural Engineering, University of Colorado at Boulder, ECOT 441, CO 80303, USA

Received 16 April 2004; received in revised form 27 July 2004; accepted 27 July 2004 Available online 30 September 2004

Abstract As high quality drinking water becomes scarcer, unintentional indirect potable water reuse, where wastewater effluent is used as a part of a downstream drinking water source, has become a great concern throughout the world. In this case, a variety of organic micro-pollutants contained in wastewater effluent could create problems. High pressure-driven membranes such as nanofiltration (NF) or reverse osmosis (RO) might be a powerful option to deal with such micro-pollutants, however, a lack of information on their performance is apparent. This study examined the ability of RO membranes to retain neutral (uncharged) endocrine disrupting compounds (EDCs) and pharmaceutically active compounds (PhACs). A total of 11 compounds were chosen so that a certain range of molecular weights and octanol–water distribution coefficients (Kow ) could be studied. With respect to membranes, two different materials (polyamide and cellulose acetate) were examined. Generally, the polyamide membrane exhibited a better performance in terms of the rejection of the selected compounds but the retention was not complete (57–91%). It was found that salt rejection or molecular weight cut-off (MWCO) that are often used to characterize membrane rejection properties did not provide quantitative information in terms of EDCs/PhACs rejection by NF/RO membranes. Molecular weight of the tested compounds could generally indicate the tendency of rejection for the polyamide membranes (size exclusion dominated the retention by the polyamide membrane) while polarity was better able to describe the retention trend of the tested compounds by the cellulose acetate membrane. The results obtained in this study imply that each membrane polymer material for NF/RO membranes, including ones that will be newly developed in the future, would exhibit different trends in terms of rejection of organic micro-pollutants, which is determined by physico-chemical properties of the compounds. © 2004 Elsevier B.V. All rights reserved. Keywords: Reverse osmosis; Endocrine disrupting compounds; Pharmaceutically active compounds; Molecular weight; Polarity

1. Introduction Among various types of organic micro-pollutants with low molecular weight, compounds that are categorized as endocrine disrupting compounds (EDCs) and pharmaceutically active compounds (PhACs) have been receiving a considerable attention recently. With the rapid development of analytical techniques, it has been reported that many aquatic environments are polluted with low concentrations of EDCs [1–4] and PhACs [2,5,6]. Sewage treatment plant (STP) ef∗

Corresponding author. Tel.: +81 11 706 6267; fax: +81 11 706 6267. E-mail address: [email protected] (K. Kimura).

0376-7388/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2004.07.018

fluents are often considered to be a major source of this type of pollution. As good sources of drinking water are becoming scarcer, unintentional indirect potable water reuse, in which wastewater effluent is contained in drinking water sources to some extent, is being carried out in many places. Pollution of drinking water sources with organic micro-pollutants is of great concern in such situations. Their concentrations in the raw water would be influenced by the percentage of treated wastewater. Actually, with respect to several PhACs, tap water in Germany was reported to contain lower ng/L concentrations of PhACs [7,8]. Efficiency of conventional drinking water production technology as a barrier for EDCs/PhACs at low concentrations is not clear


K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78

at present although Ternes et al. [9] described the performance of real drinking water facilities in terms of rejection of a limited number of PhACs (i.e., bezafibrate, clofibric acid, carbamazepine, diclofenac, and primidone). Although health effects of the consumption of EDCs/PhACs at low concentration levels are not fully elucidated yet, drinking water should be relatively free of such compounds. High pressure-driven membranes such as nanofiltration (NF) and reverse osmosis (RO) would be powerful options to remove such organic micro-pollutants. Near complete retention of low molecular weight organic compounds, particular pesticides, by NF/RO membranes has been reported by several researchers [10–15]. However, it was recently demonstrated that EDCs with high molecular weight such as 17␤-estradiol (MW: 279 g/mol) were still detected in RO permeate, although at very low concentrations [16]. A more fundamental understanding of EDCs/PhACs rejection by NF/RO membranes is currently required. In our previous study, limited numbers of EDCs/PhACs were examined in terms of their rejection by polyamide NF/RO membranes [17] and it was found that negatively charged compounds would be significantly rejected by NF/RO membranes due to electrostatic repulsion between the compounds and membranes. The high rejection (>90%) associated with negative charge was observed even when compounds with a small molecular weight (e.g., 110) and a rather loose membrane (i.e., NF) were examined. Therefore, charged compounds will not be considered in this paper. When it comes to neutral (uncharged) compounds, rejection will depend on physical/chemical properties of compounds. Relationships between physical/chemical properties of compounds and their rejection by NF/RO membranes have been discussed by several researchers [18,19]. However, in these other studies, neutral EDCs/PhACs which have a certain range of physico-chemical properties were not examined and therefore there is still a lack of information on the efficiency of NF/RO membranes for reduction of such compounds. Based on the context described above, rejection of neutral EDCs/PhACs by RO membranes was the focus of this study. As data on EDCs/PhACs rejection by membrane processes are currently still rare, data presented in this paper would be important to predict membrane performance. Also, consideration of important physical/chemical properties of target compounds that influence rejection would provide a better understanding of the retention mechanisms by RO membranes.

2. Materials and methods 2.1. Compounds Fig. 1 and Table 1 shows the structures and the physical/chemical properties of the tested compounds, respectively. As mentioned earlier, all of the tested compounds

Fig. 1. Structural formulas of the compounds examined in this study.

were chosen not to exhibit a charge under the tested conditions. These compounds were chosen so that certain ranges of molecular weight and octanol–water distribution coefficient (Kow ) could be examined. In addition, the Henry’s law constant of each compound was taken into consideration as volatilization during experiments makes interpretation of data difficult. Based on our preliminary tests, Henry’s law constant values of the order of 10−8 atm m3 /mol would not cause such problems. Compounds with rather large molecular weights such as >300 are expected to be efficiently removed by RO membranes and therefore were excluded in this study. Table 1 also summarizes the values of dipole moment calculated for the tested compounds. Calculation was carried out with a commercial software (Hyperchem [21]). 17␤-Estradiol and bisphenol A were examined as they are representative EDCs. 4-Phenylphenol is used as an antioxidant and is a potential EDC. NAC-standard (carbaryl) is used as a pesticide, many of which are considered as EDCs. Carbamazepine, isopropylantipyrine, phenacetine, primidone, and sulfamethoxazole are PhACs. Carbamazepine and primidone are antiepileptics, isopropylantipyrine and phenacetine are analgesics, and sulfamethoxazole is antibiotic. Caffeine might be categorized as PhAC (stimulant), but is often considered as an indicator of pollution of aquatic environment with wastewater effluent. It is used in this study as it fills a gap in the range of physical/chemical properties that was intended to be examined in this study. 2-Naphthol was used as

K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78


Table 1 Tested compounds in this study [20,21] Compound


Molecular weight

Water solubility (mg/L)

log Kow

Henry’s law constant (atm m3 /mol)


Dipole moment (debye)

2-Naphthol 4-Phenylphenol Phenacetine Caffeine NAC standard Primidone Bisphenol A Isopropylantipyrine Carbamazepine Sulfamethoxazole 17␤-Estradiol

Surrogate EDC PhAC Surrogate EDC PhAC EDC PhAC PhAC PhAC EDC

144 170 179 194 201 218 228 231 236 253 272

7.55E + 02 56.2 7.66E + 02 2.16E + 04 1.11E + 02 5.00E + 02 1.20E + 02 3.00E + 06 17.7 6.10E + 02 3.6

2.70 3.20 1.58 −0.07 2.36 0.91 3.32 1.94 2.45 0.89 4.01

2.74E−08 5.23E−08 2.13E−10 1.90E−19 3.27E−09 1.94E−10 1.00E−11 1.84E−09 1.08E−10 6.42E−13 3.64E−11

9.50 9.55 N/A N/A N/A N/A N/A N/A N/A N/A N/A

0.873 1.124 1.675 3.862 1.989 2.696 0.709 3.851 3.286 6.318 0.798

well for the same reason. Stock solutions were prepared in methanol at a concentration of 1 mg/L. 2.2. Membranes Two RO membranes made of different materials were used in this study. One is made of polyamide (XLE, Dow-FilmTec) and the other is made of cellulose acetate (SC-3100, Toray). Table 2 summarizes properties of the membranes employed. In this study, salt rejection and molecular weight cut-off (MWCO) of each tested membranes were estimated under the same conditions as the rejection tests (described later). For evaluation of salt rejection, 1000 ppm of NaCl was used. MWCO was estimated by filtering solutions with a certain molecular weight range of poly-ethylene glycol (PEG) (Wako Pure Chemical, Osaka, Japan). Historically, cellulose acetate was the first material used for manufacturing RO membranes but is now becoming less common due to its low water production capability. Instead, polyamide has now become the preferred material. In this study, however, cellulose acetate was also examined since properties of this material are significantly different from those of polyamide, which might cause differences in rejection of EDCs/PhACs. 2.3. Filtration tests In this study, all experiments were carried out in a crossflow membrane unit with a flat-sheet membrane cell (C70-F, Nitto Denko, Osaka, Japan). Effective membrane area in the cell was 32 cm2 . Fig. 2 shows the schematic diagram of the filtration test unit used in this study. In all of the experiments, the applied pressure was fixed at 0.5 MPa and resulted in an averaged permeate flux of 1.05 and 0.35 ml/min for XLE and

SC-3100, respectively. All experiments were performed in a recycle mode, with both permeate and retentate recycled back into the feed reservoir. All parts of the experimental apparatus used in this study were made of stainless steel in order to prevent undesirable adsorption of tested compounds onto it. Control experiments were carried out to determine whether adsorption of compounds onto the apparatus occurred with respect to the first and second most hydrophobic compounds among the tested compounds (i.e., 17␤-estradiol and bisphenol A). Milli-Q water containing 100 ␮g/L of 17␤-estradiol or bisphenol A was circulated for 24 h without inserting a membrane coupon and the decrease of the concentration in the reservoir (indication of adsorption on the experimental apparatus) was monitored. With respect to bisphenol A, a decrease of the concentration was not recognized, implying that adsorption onto the apparatus could be neglected for most of the tested compounds. In contrast, for 17␤-estradiol, about a 5% decrease in the concentration was observed in the control test. However, the decrease could be minimized to <1% by replacing the feed water after the “first” control test. Thus, with respect to 17␤-estradiol, this conditioning of the experimental apparatus was carried out prior to filtration tests. Pressures and flow were adjusted by manual valves. The retentate flow, which is not passed through the membrane, is recycled to the feed water reservoir. The permeate flow was also recycled to the feed tank except for small sample aliquots. By running in a recycle mode, it was possible to account for compound depletion by adsorption onto/into membrane. A new mem-

Table 2 Characteristics of the membrane used in this study Name




Salt rejection (%)a

XLE SC-3100

Dow-FilmTec Toray

Polyamide Cellulose acetate

<200 200–300

90 94


Determined in this study.

Fig. 2. Schematic diagram of the filtration unit.


K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78

brane was used for each experiment. Prior to each experiment, membranes were rinsed with Milli-Q water and compacted by filtering Milli-Q water overnight. In all of the experiments, the feed solution contained only one component (i.e., single solute experiment). At the beginning of an experiment, an aliquot of the stock solution was spiked into Milli-Q water in the manner described elsewhere [22] so that the concentration in the feed water was approximately 100 ␮g/L. All of the experiments were carried out with control of tempereture (20 ◦ C). Prior to the filtration test, the pH of the solution was adjusted to 7.0 ± 0.1 by adding 0.1 M NaOH. Recovery was also fixed at 3.5 and 1.5% with respect to XLE and SC-3100, respectively. Filtration of the feed water containing a target compound was carried out for 24 h in order not to cause overestimation of rejection (details shown and discussed later). 2.4. Analysis With respect to compounds that have high sensitivity to fluorescence (i.e., 2-naphthol, 4-phenylphenol, and NAC standard), concentrations were determined by using a fluorometer (Shimadzu RF-5300PC, Kyoto, Japan) with 1 cm cuvette. The wavelengths employed (emission/excitation, nm) were 356/326, 332/277, and 333/291 with respect to 2naphthol, 4-phenylphenol, and NAC standard, respectively. 17␤-estradiol, bisphenol-A, and sulfamethoxazole were analyzed with HPLC with a fluorescence detector according to the methods of Yoon et al. [23]. PhACs excluding isopropylantipyrine were analyzed by GC/MS coupled with solid phase extraction (SPE) according to Reddersen and Heberer [24]. Concentrations of caffeine and isopropylantipyrine were determined by LC/MS coupled with SPE. The LC system used in this study was equipped with 30 mm × 4.6 mm XDB-C18 column (Agilent) and the mobile phase was 80/20, water/methanol (isocratic conditions). Mass spectra were acquired in positive ion electrospray (ESI(+)) mode. The drying gas was operated at a flow rate of 13.0 ml/min at 350 ◦ C. The nebulizer pressure was 60 psi, the capillary was set at 4000 V, and the fragmentor was set at 100 V.

crease. Regarding this decrease of concentration in the feed water, it can be only attributed to adsorption onto/in the membranes as adsorption onto the experimental apparatus was shown to be negligible due to the use of stainless steel, as described in the experimental section. Thus, conditioning of the membrane was essential for a reasonable evaluation of the rejection properties of the membranes. In this study, all of the determined rejections were based on 24 h of filtration to assume quasi-saturation of the membranes. 3.2. Overview of rejection properties of the tested membranes Table 3 summarizes rejection of the tested compounds by both membranes. The percent rejection of a compound was calculated by:   Cp × 100 Rejection = 1 − Cf where Cp and Cf are the concentrations of a tested compound detected at the termination of each filtration test (24 h) in the permeate and the feed water, respectively. XLE exhibited a rejection of between 57 and 91% while rejection determined with SC-3100 considerably fluctuated and sometimes exhibited no rejection (2-naphthol and NAC standard). Salt rejection is a characteristic that is often used

3. Results and discussion 3.1. Time-dependence of compound concentration during filtration tests In our previous study [22] using surrogate compounds, it was shown that short-term filtration with NF/RO membranes would lead to overestimation of rejection. A similar trend was observed in this study using EDCs and PhACs as well. As a representative example, Fig. 3 shows the time course change in 4-phenylphenol concentration during filtration tests. As shown in Fig. 3, with both membranes tested, feed water concentration decreased while concentration in the permeate increased. As a result, the retention efficiency of 4-phenylphenol by both RO membranes continued to de-

Fig. 3. Change in concentration of 4-phenylphenol during the filtration tests.

K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78 Table 3 Rejection (%) of the tested compounds by XLE and SC-3100 aa


2-Naphthol 4-Phenylphenol Phenacetine Caffeine NAC standard Primidone Bisphenol A Isopropylantipyrine Carbamazepine Sulfamethoxazole 17␤-Estradiol

57 61 74 70 79 87 83 78 91 70 83

SC-3100 0 11 10 44 0 85 18 69 85 82 29

by NF and RO membrane manufacturers to describe membrane rejection properties. Given that the salt rejection of SC-3100 is higher than XLE (Table 2), it was surprising that XLE apparently exhibited better rejection than SC-3100 for the EDCs/PhACs examined in this study. Thus, salt rejection, often used to describe the rejection efficiency of RO membranes, was found to be inappropriate to indicate potential rejection of organic micro-pollutants. Another index that is often used to characterize membrane rejection is MWCO. Based on the MWCO evaluated in this study, XLE should exhibit better rejection compared with SC-3100, which is in accordance with the results shown in Table 3. Thus, when dealing with rejection of organic micropollutants by RO membranes, the MWCO should be considered instead of salt rejection that is usually used to express the rejection properties of RO membranes. MWCO, however, was not so descriptive that rejection of micro-pollutants could be predicted. According to the convention that MWCO is the molecular weight of a standard compound (in this case PEG) that can be rejected by 90%, compounds with molecular weight of >200 should be rejected more than 90%, at least by the XLE given that the MWCO of XLE was evaluated to be <200. As seen from Table 3, however, this is not the case. Namely, only rejections of carbamazepine exceeded 90% while rejections of NAC standard, primidone, bisphenol A, isopropylantipyrine, sulfamethoxazole, and 17␤-estradiol


could not reach 90% in spite of the fact that their molecular weights are larger than 200. Thus, MWCO cannot be relied on for the accurate prediction of rejection properties of membranes, especially in terms of organic micro-pollutant rejection. This can be easily understood by the fact that the physical/chemical properties of a standard compound (e.g., PEG) used for determination of MWCO is often totally different from those of target compounds. Based on the argument described above, it can be concluded that MWCO should be used only for semi-quantitative purpose and cannot be used for the accurate prediction of organic micro-pollutant rejection by RO membranes. MWCO is, however, still more useful than salt rejection when it comes to organic micro-pollutant rejection by RO membranes. As mentioned above, XLE generally exhibited better rejection performance than SC-3100. However, one exception can be found in the rejection of sulfamethoxazole. In addition, rejections of primidone, isopropylantipyrine, carbamazepine by SC-3100 were comparable to those by XLE while there was a significant difference between the two membranes in terms of the rejections of the other compounds. These results imply that compounds that can be easily rejected by RO membranes might be different depending on properties of both the compounds and membranes. This point will be further discussed later. 3.3. Relationship between rejection and physical/chemical properties Physico-chemical properties of a compound are considered to influence its rejection by RO membranes. Many researchers have discussed this topic. Van der Bruggen et al. [13] reported that a high dipole moment of a compound would lead to decrease in rejection by membrane. Kiso et al. [14] reported that hydrophobicity represented by Kow would influence rejection. Ozaki and Li [18] reported that molecular size would be the dominant factor in rejection by RO. In this section, the rejection data obtained in this study will be analyzed based on these physico-chemical properties of the tested compounds.

Fig. 4. Relationship between rejection by XLE (A) and SC-3100 (B) vs. molecular weight of the tested compounds.


K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78

Fig. 5. Relationship between rejection by XLE (A) and SC-3100 (B) vs. log Kow of the tested compounds.

Fig. 4 shows the relationship between rejections determined in this study and molecular weight of the compounds tested. With respect to XLE (Fig. 4A), the trend was rather clear. The higher the molecular weight was, the higher the rejection. The most easily accessible parameter that indicates the size of a molecule, molecular weight, remains useful for the description of retention by XLE. Similar results were demonstrated by Van der Bruggen et al. [19] and Ozaki and Li [18] who examined a wide range of organic molecules not including EDCs/PhACs. In contrast, considerable scatter in rejection is observed with respect to SC-3100 when analyzed with molecular weight (Fig. 4B). It is considered that sieving is the dominant factor in rejection by XLE while other factor(s) needs to be taken into consideration with respect to SC-3100. Fig. 5 shows the relationship between the rejection and log Kow of the tested compounds. With respect to XLE (Fig. 5A), no significant correlation is recognized between the rejection and log Kow . As seen from Fig. 5B, Kow could not explain the trend of the rejection by SC-3100, either. To better explain rejection by SC-3100, polarity needs to be considered. Rejection observed in this study was analyzed based on the calculated dipole moment (Table 1) as well. Rejection as

a function of the calculated dipole moment is presented in Fig. 6. Dipole moment could not explain the rejection tendency with respect to XLE (Fig. 6A). This is not too surprising in that dipole moment is somewhat inversely related to Kow . Compared to molecular weight and log Kow , however, a better correlation with rejection is obtained for SC-3100 (Fig. 6B). Apparently, the higher the dipole moment was, the higher the rejection by SC-3100 was. Van der Bruggen et al. [13] showed that molecules with dipole moment of >3 debye consistently showed a lower rejection than molecules of approximately the same size but with a lower dipole moment. When looking at the pairs of caffeine (MW: 194) and NAC standard (MW: 204), isopropylantipyrine (MW: 231) and bisphenol A (MW: 228), and sulfamethoxazole (MW: 253) and 17␤-estradiol (MW: 272), that was the case with the XLE membrane in this study: high dipole moment brought about poor rejection. However, the effect of polarity on the rejection was somewhat minor. When rejection by SC-3100 is considered with respect to the same pairs of compounds, the influence of polarity was much more significant but appeared to act in the opposite way: high dipole moment caused better rejection. This difference may be explained by the difference in membrane material. Van der Bruggen et al. [13] derived

Fig. 6. Relationship between rejection by XLE (A) and SC-3100 (B) vs. dipole moment of the tested compounds.

K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78


membranes would be affected, especially when the material is cellulose acetate. It will be quite interesting to examine whether this effect can be recognized with other membrane materials.

4. Summary

Fig. 7. Relationship between the calculated dipole moment and log Kow of the tested compounds.

their conclusion by using membranes made of polyamide and sulfonated poly-ethersulfone while we observed a significant polarity effect with a cellulose acetate membrane. 3.4. Characteristics of compounds that bring about high rejection by cellulose acetate membrane As described above, XLE generally exhibited a better rejection for the tested compounds compared with SC-3100. However, with respect to sulfamethoxazole, SC-3100 showed a better rejection over XLE. In addition, rejections of primidone, isopropylantipyrine, carbamazepine by SC-3100 were comparable to those of XLE although there was a significant difference in rejection of the other compounds between the two tested membranes. It is considered that primidone, isopropylantipyrine, carbamazepine, and especially sulfamethoxazole would have some common characteristics that facilitate a better rejection by SC-3100 (cellulose acetate membrane). The relationship between dipole moment and Kow of the compounds might provide insight into this point. Fig. 7 shows the relationship between dipole moment and log Kow of the compounds tested in this study. In Fig. 7, the compounds whose rejections by SC-3100 were high (i.e., sulfamethoxazole, primidone, isopropylantipyrine, and carbamazepine) are represented by hollow triangles instead of squares. Kow , an indicator of hydrophobicity, and dipole moment, an indicator of polarity, cannot be easily separated as water is a polar solvent. In general, hydrophilic compounds (represented by low log Kow ) are likely to have higher values of dipole moment. This trend can be recognized in Fig. 7. As seen from Fig. 7, however, compounds that were somewhat surprisingly rejected by the SC-3100 deviated from the general trend that was assumed (the line in Fig. 7). Depending on arrangement of atomic and functional groups, a molecule could exhibit a much higher (or lower) dipole moment than would be expected from its Kow value. In that case, rejection of such compounds by RO

In this study, rejection of neutral (uncharged) EDCs/ PhACs by RO membranes was investigated based on the bench-scale crossflow experiments. Tested compounds were chosen to cover a certain range of molecular weight and Kow , and two membranes made of different materials were used. Results obtained in this study can be summarized as follows. Salt rejection often used to express rejection properties of RO membranes does not necessarily represent rejection of EDCs/PhACs. Rather, MWCO would be more useful for evaluating the rejection of EDCs/PhACs by RO membranes although use of that parameter is not common for RO membranes. MWCO, however, cannot be used for precise prediction of rejection of EDCs/PhACs by RO membranes since properties of standard compounds used for MWCO determination and those of target EDCs/PhACs are considerably different. Comparing the tested two membranes, the polyamide membrane (XLE) generally exhibited better rejections than the cellulose acetate membrane (SC-3100). However, even XLE did not exhibit complete rejection for the tested compounds. Rejection of the compounds by XLE was in the range of 57–91%, which could be reasonably correlated to molecular weight. With respect to SC-3100, rejection was highly dependent on polarity. SC-3100 appeared to be good in rejection of polar compounds. Based on these results, it was hypothesized that the dominant rejection mechanism for RO membranes would be different depending on membrane material and the physico-chemical properties of target compounds. Better understanding of this aspect could lead to development/modification of new membrane materials for better rejection of EDCs/PhACs.

Acknowledgements This study was partially supported by the Ministry of Education, Science, Sports and Culture of Japan, Grantin-Aid for Young Scientists (B), 10292054, 2003. The authors thank Dr. Hiroto Tachikawa of Department of Molecular Chemistry, Hokkaido University, for his instruction in calculation of the dipole moments. The authors also thank Mr. Makoto Takeda of Department of Urban and Environmental Engineering, Hokkaido University, for his important contributions in the analysis using LC/MS. The authors acknowledge the donation of the XLE membrane by the Dow Chemical Company, and the SC-3100 membrane by Toray Industries Inc.


K. Kimura et al. / Journal of Membrane Science 245 (2004) 71–78

References [1] C. Baronti, R. Curini, G. D’ascenzo, A. Di Corcia, A. Gentili, R. Samperi, Monitoring natural and synthetic estrogens at activated sludge sewage treatment plants and in a receiving river water, Environ. Sci. Technol. 34 (2000) 5059–5066. [2] D.W. Kolpin, E.T. Furlong, M.T. Meyer, E.M. Thurman, S.D. Zuggg, L.B. Barber, H.T. Buxton, Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999–2000: a national reconnaissance, Environ. Sci. Technol. 36 (2002) 1202–1211. [3] J.K. Fawell, D. Sheahan, H.A. James, M. Hurst, S. Scott, Oestrogens and oestrogenic activity in raw and treated water in severn trent water, Water Res. 35 (2001) 1240–1244. [4] P. Roefer, S. Snyder, R.E. Zegers, D.J. Rexing, J.L. Fronk, Endocrine-disrupting chemicals in a surface water, J. AWWA 92 (8) (2000) 52–58. [5] Th.A. Ternes, Occurrence of drugs in German sewage treatment and rivers, Water Res. 32 (1998) 3245–3260. [6] Th. Heberer, Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data, Toxicol. Lett. (2002) 5–17. [7] T.h.A. Ternes, Pharmacetuticals and Personal Care Products in the Environment: Scientific and Regulatory Issue, in: Ch.G. Daughton, T.L. Jones-Lepp (Eds.), ACS Symposium Series 791, American Chemical Society, Washington, DC, 2001. ISBNO-8412-3739-5. [8] Th. Heberer, Tracking persistent pharmaceutical residues from municipal sewage to drinking water, J. Hydrol. 266 (2002) 175–189. [9] Th.A. Ternes, M. Meisenheimer, D. Mcdowell, F. Sacher, H.-J. Brauch, B. Haist-Gulde, G. Preuss, U. Wilme, N. Zulei-Seibert, Removal of pharmaceuticals during drinking water treatment, Environ. Sci. Technol. 36 (2002) 3855–3863. [10] Y. Kiso, Y. Sugiura, T. Kitao, K. Nishimura, Effect of hydrophobicity and molecular size on rejection of aromatic pesticides with nanofiltration membranes, J. Membr. Sci. 192 (2001) 1–10. [11] J.A.M.H. Hofman, E.F. Beerendonk, H.C. Folmer, J.C. Kruithof, Removal of pesticides and other micropollutants with cellulose-acetate, polyamide and ultra-low pressure reverse osmosis membranes, Desalination 113 (1997) 209–214.

[12] C. Ventresque, G. Turner, G. Bablon, Nanofiltration: from prototype to full scale, J. AWWA 89 (10) (1997) 65–76. [13] B. Van der Bruggen, J. Schaep, W. Maes, D. Wilms, C. Vandecasteele, Nanofiltration as a treatment method for the removal of pesticides from ground waters, Desalination 117 (1998) 139–147. [14] Y. Kiso, Y. Nishimura, T. Kitao, K. Nishimura, Rejection properties of non-phenylic pesticides with nanofiltration membranes, J. Membr. Sci. 171 (2000) 229–237. [15] I.C. Escobar, S. Hong, A.A. Randall, Removal of assimilable organic carbon and biodegradable dissolved organic carbon by reverse osmosis and nanofiltration membranes, J. Membr. Sci. 175 (2000) 1–17. [16] A.T. Salveson, D.A. Requa, R.D. Whitley, G. Tchobanoglous, Potable versus reclaimed water quality, regulatory issues, emerging concerns, in: Proceedings of Annual Conference Water Environment Federation, WEFTEC, Anaheim, CA, 2000. [17] K. Kimura, G. Amy, J.E. Drewes, Th. Heberer, T.-U. Kim, Y. Watanabbe, Rejection of organic micropollutants (disinfection by-products, endocrine disrupting compounds, and pharmaceutically active compounds) by NF/RO membranes, J. Membr. Sci. 227 (2003) 113–121. [18] H. Ozaki, H. Li, Rejection of organic compounds by ultra-low pressure reverse osmosis membrane 36 (2002) 123–130. [19] B. Van der Bruggen, D. Schaep, C. Wilms, C. Vandecasteele, Influence of molecular size, polarity and charge on the retention of organic molecules by nanofiltration, J. Membr. Sci. 156 (1999) 29–41. [20] [21] Hyperchem, Release 6 for Windows, Molecular Modeling System, Hypercube Inc., Gainesville, FL. [22] K. Kimura, G. Amy, J. Drewes, Y. Watanabe, Adsorption of hydrophobic compounds onto NF/RO membranes: an artifact on overestimation of rejection, J. Membr. Sci. 221 (2003) 89–101. [23] Y. Yoon, P. Westerhoff, S.A. Snyder, in: Proceeding of AWWA Annual Conference, CD-ROM, HPLC-fluorescence detection and adsorption of endocrine disrupting compounds on powdered activated carbon (2003). [24] K. Reddersen, T. Heberer, Multiple methods for the detection of pharmaceutical residues in various waters applying solid phase extraction (SPE) and gas chromatography with mass spectrometric (GC–MS) detection, J. Sep. Sci. 26 (2003) 1–8.