Boron removal in seawater desalination by reverse osmosis membranes - The impacts of operating conditions

Contents xv

Radiochronological Methods as Tools to Study Environmental Pollution 1015H.N. Erten

Are Certain Invertebrate Species Sensitive Bioindicatorsof the Air Pollution? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023Liliana Vasiliu-Oromulu, Viorica Honciuc, Sanda Maican,Cristina Munteanu, Minodora Stanescu, Cristina Fiera, Mihaela Ion,and Dorina Purice

Determination of Heavy Metal Pollution in Some HoneySamples from Yozgat Province, Turkey . . . . . . . . . . . . . . . . . . 1037Ahmet Aksoy, Zeliha Leblebici, and Yavuz Bagci

Removal of Direct Orange-46 from Aqueous Solutions UsingMN-Diatomite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1045Selay Aksoy, Mesut Tekbas, Güleda Engin, and Nihal Bektas

Environmental Problems from the Open Dump in GümüshaneProvince and Investigation of Biological Recyclingfor the Organic Solid Wastes . . . . . . . . . . . . . . . . . . . . . . . . 1055S. Serkan Nas and Adem Bayram

Adsorption Behavior of Radionuclides, 137Cs and 140Ba, ontoSolid Humic Acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065O. Çelebi and H.N. Erten

The Fate of Chlortetracycline During the Anaerobic Digestionof Manure from Medicated Calves . . . . . . . . . . . . . . . . . . . . . 1087Osman A. Arikan

Adsorption of Methylene Blue from Aqueous Solution onto Bentonite . 1097J. Krstic, Z. Mojovic, A. Abu Rabi, D. Loncarevic, N. Vukelic,and D. Jovanovic

The Evaluation of the Pb(II) Removal Efficiency of DuckweedLemna Minor (L.) from Aquatic Mediums at Different Conditions . . . 1107Yagmur Uysal and Fadime Taner

Biodegradation of a Tannery and Chemical Plant ProducingAsetilsalisilikat Wastewater Mixture . . . . . . . . . . . . . . . . . . . . 1117E.U. Cokgor, O. Karahan, and D. Orhon

Boron Removal in Seawater Desalination by Reverse OsmosisMembranes – the Impacts of Operating Conditions . . . . . . . . . . . 1127H. Köseoglu, N. Kabay, M. Yüksel, S. Sarp, Ö. Arar, and M. Kitis

Respirometric Evaluation of Strong Wastewater ActivatedSludge Treatment for a Complex Chemical Industry . . . . . . . . . . . 1139E. Ubay Cokgor, G. Insel, E. Aydın, S. Ozdemir, and D. Orhon

Boron Removal in Seawater Desalinationby Reverse Osmosis Membranes – the Impactsof Operating Conditions

H. Köseoglu, N. Kabay, M. Yüksel, S. Sarp, Ö. Arar, and M. Kitis

Abstract Production of drinking water through seawater desalination using reverseosmosis (RO) membranes is becoming increasingly attractive especially in coastalareas with limited freshwater sources. However, one challenge in such conventionaldesalination RO plants is the difficulty of meeting boron standards in product waters.Therefore, most of the current desalination plants employ additional treatmentsteps including pH adjustment of feedwater, dilution of RO permeate with othersources, ion exchange post-treatment of RO permeate, and/or double-pass stagingfor permeate. All these further treatment options increase the cost of desalination.Although membrane manufacturers have been developing modified RO membraneswith enhanced boron removal capacities such membranes still should be improvedfrom operational flux and pressure perspectives. The main objective of this work wasto determine the impacts of operational conditions (membrane pressure, cross-flowvelocity and flux) and water chemistry on boron rejections using two commer-cial RO membranes specified for enhanced boron removal (TorayTM UTC-80-ABand FilmtecTM SW30HR). A lab-scale cross-flow flat-sheet configuration test unit(SEPA CF II, Osmonics) was used for all RO experiments. Seawater samples werecollected from the Mediterranean Sea, Alanya-Kızılot shores, south Turkey. For allexperiments, mass balance closures were between 91 and 107%, suggesting rela-tively low loss of boron on membrane surfaces during 14 h of operation. Boronrejections were relatively constant (a maximum change of ±3%) during the 14 hof operation period for all experiments, suggesting that steady state dynamic mem-brane conditions were immediately achieved within couple hours. Boron rejectionsobtained with Toray and FilmTec membranes at pH of original seawater (8.2) andat other various operating conditions ranged between 85 and 92%, resulting in per-meate boron concentrations of about 0.2–0.9 mg/L. On the other hand, for bothmembranes, much higher boron removals were achieved at a pH of 10.5 (>98%),resulting in permeate boron concentrations less than 0.1 mg/L. The charged boron

M. Kitis (B)Department of Environmental Engineering, Suleyman Demirel University, 32260 Isparta, Turkeye-mail: [email protected]

1127H. Gökçekus et al. (eds.), Survival and Sustainability, Environmental Earth Sciences,DOI 10.1007/978-3-540-95991-5_106, C© Springer-Verlag Berlin Heidelberg 2011

1128 H. Köseoglu et al.

species are expected to be dominant at pH values >9.24 (pKa of boric acid) com-pared to the neutral boric acid. Therefore, as expected, both membranes exhibitedhigher boron rejections at a pH of 10.5. Salt rejections (as measured by conductiv-ity) were generally 97–99% at both pH values. Boron rejections were independentof feed water boron concentrations up to 6.6 mg/L. For each membrane type, per-meate fluxes at constant pressure were generally lower at pH of 10.5. The rangesof permeate fluxes measured in all experimental conditions were 11–15, 13–17 and19–21 L/m2-h for 600, 700 and 800 psi (41, 48 and 55 bar) pressures, respectively,after an operation period of 14 h. For all experimental conditions, permeate fluxesgradually decreased during the 14 h operation although a leveling off was observedafter 12 h. At constant membrane pressure of 800 psi and pH of 8.2, feed flowratethus the cross-flow velocity (0.9 and 0.5 m/s) did not exert any significant impact onboron rejection.

Keywords Boron · Desalination · Membrane · Reverse osmosis · Seawater · SWRO

1 Introduction

Due to increasing demand for water, both potable and for irrigation, coupled with adecrease in suitable water sources suppliers have to turn to alternatives. Seawaterdesalination or treatment of high saline, eventually contaminated surface watershave become standard [1, 2]. The cost of seawater desalination by reverse osmo-sis (RO) membranes has significantly decreased. Both greater competition andimproved technology have contributed to this reduction in desalted water prices [3].Producing potable water by RO desalination has been performing successfully fornearly 25 years [4].

Seawater (4–5 mg/L), municipal wastewaters (0.5–2 mg/L) and groundwaters (upto 8 mg/L in Cyprus, Italy and Greece) in some areas contain high concentrationsof boron. European Union (EU) is classifying boron as a pollutant in potable waterregulations [5]. Laboratory studies showed impediments on male reproductive sys-tem caused by boron [6]. Also, sensitive crops, including most citrus species, havea boron tolerance of only 0.40–0.75 mg/L, while vegetables are more boron tolerantwith maximum thresholds of 1–4 mg/L [7]. As a result, World Health Organization(WHO) set 0.5 mg/L for maximum boron concentration in potable waters [8].

Boron exists as non-ionic boric acid form at the natural pH level of seawaterwhich is about 8.2. From the level of pH 9.5, concentrations of borate and otherionic forms of boron become dominant compared to non-ionic forms, which is verycritical by the point of boron rejection in RO membranes. While the rejection ofboric acid in conventional RO membranes is only about 60–70%, it is 90–98% forborate ion [9]. Thus, boron rejection at RO desalination plants operating at natu-ral pH level of seawater is relatively low. Non-ionic boron species, such as boricacid, permeate through the RO membranes more than the charged, ionic species. Inrecent years, membranes with boron rejections of 91–93% even at natural seawater

Boron Removal in Seawater Desalination by Reverse Osmosis Membranes 1129

pH are taking place on the market [10]. Although membrane manufacturers havebeen developing modified RO membranes with enhanced boron removal capacitiesmost of the current desalination plants employ additional treatment steps includ-ing pH adjustment of feedwater, dilution of RO permeate with other sources, ionexchange post-treatment of RO permeate, and/or double-pass staging for permeate.In addition, a number of process configurations have been proposed to achieve lowboron concentrations in seawater RO desalination plants [3, 10]. All these furthertreatment options increase the cost of desalination. Although membrane manufac-turers have been developing modified RO membranes with enhanced boron removalcapacities such membranes still should be improved from operational flux and pres-sure perspectives. The main objective of this work was to determine the impacts ofoperational conditions (membrane pressure, cross-flow velocity and flux) and waterchemistry on boron rejections using two commercial RO membranes specified forenhanced boron removal (TorayTM UTC-80-AB and FilmtecTM SW30HR).

2 Methodology

2.1 Test Unit

A lab-scale cross-flow flat-sheet configuration test unit (SEPA CF II, Osmonics)was used for all membrane separation experiments (Fig. 1), which simulates theflow dynamics of larger, commercially available spiral-wound membrane elements.

Fig. 1 The schematic of the membrane test system

1130 H. Köseoglu et al.

The membrane test unit accommodates any 19 cm × 14 cm (7.5 inch × 5.5 inch)flat-sheet membrane for a full 140 cm2 (22 inch2) of effective membrane area.Maximum operating pressure of the unit is 69 bar (1,000 psi). Permeate flowrateis approximately proportional to driving pressure.

2.2 Membranes

Two commercial desalination RO membranes, TorayTM (UTC-80-AB) andFilmtecTM (SW30HR), were tested for boron rejections. Membrane sheets wereobtained from the manufacturers and used as-received.

2.3 Experiments

Seawater samples were collected from the Mediterranean Sea, Alanya-Kızılotshores (south Turkey). All seawater samples were filtered (1 μm cartridge filters)and kept at dark in the laboratory prior to use for membrane tests. Membrane oper-ating pressures tested were 600, 700 and 800 psi. The tested feed water pH valueswere 8.2 (seawater pH) and 10.5, which simulated the high pH level necessary forthe formation of charged, ionic boron species. The tested membrane feed flowrateswere 3.8 and 2 L/min with corresponding cross-flow velocities of about 0.9 and 0.5m/s, respectively. A new membrane was used for each test after conditioning themembrane for about 5–6 h of feeding with distilled and deionized water (DDW)at the conditions of the experiment to be conducted. Two operation modes wereemployed in the experimental matrix: batch concentration and constant feed tankvolume. The duration of each membrane test was about 14 h. Samples from feedtank and permeate were taken each hour for boron, total dissolved solids (TDS), con-ductivity, pH, and temperature measurements. To maintain constant pH in the feedtank depending on the experimental matrix, pH was monitored and adjusted withvarious concentrations of HCl and/or NaOH solutions, if necessary. Flowrates ofconcentrate and permeate streams, membrane unit and pump outlet pressures werealso recorded each hour. In an effort to calculate overall mass balances on boron,the volumes remaining in feed and permeate tanks and tubings after the experi-ments, sample volumes and the associated boron concentrations in these volumeswere measured.

2.4 Analytical Measurements

Spectrophotometric curcumine method was employed for boron analysis. In thismethod, curcumine solution forms an orange-red complex compound with borateions and the absorbance of this compound is measured at λmax of 543 nm.

Boron Removal in Seawater Desalination by Reverse Osmosis Membranes 1131

Conductivity and TDS were measured using WTW Inolab Cond. Level 1 conduc-tivity meter. pH was measured using a bench-scale Schott Handylab 1 pH meter.All chemicals used were reagent grade. Distilled and deionized water was used forstock solution preparations and dilutions.

3 Results and Discussions

Table 1 presents the mass balance calculations performed on boron for each experi-ment. Percent mass balance closure represents the ratio of the boron mass obtainedfrom all points (i.e., feed tank, permeate collection tank, tubings and samples)at the end of the experiment to the boron mass initially applied to the system.For all experiments, mass balance closures were between 91 and 107%, suggest-ing relatively low loss of boron on membrane surfaces during 14 h of operation.It should be noted that many independent parameters including volumes of vari-ous tanks/samples and boron concentrations are measured and take place in massbalance calculations. Therefore, experimental errors in volume and concentrationmeasurements are expected, which may result in mass balance closures lower orhigher than 100%. Boron measurements were conducted in duplicates. Percent coef-ficients of variations between duplicate measurements were generally less than 12%,averaging about 5%. Thus, an error degree of approximately ±5% is inherent inmass balance calculations.

Figure 2 shows the impact of solution pH on boron rejection in seawater exper-iments. For both Toray and FilmTec membranes, much higher boron rejectionswere obtained at pH of 10.5 (>98%) than those at original seawater pH (about86–90%). Permeate boron concentrations less than 0.1 mg/L were easily achievedat pH 10.5 by both membranes. The acid dissociation constant (pKa) of boric acidis 9.24. Therefore, the charged boron species are expected to be dominant at pHvalues >9.24 compared to the neutral boric acid. At a pH of 8.2, about 80% ofboron is in neutral form. For many compounds, it is known that charged speciesare rejected to a greater extent by many RO membranes through electrostatic repul-sion. Therefore, as expected, both membranes exhibited higher rejections at a pHof 10.5.

Figure 3 shows the impact of operation mode on boron rejections. In the batchconcentration mode, feed tank boron concentration increases with operation period.After 3 h of operation to reach steady state rejections, operation mode did not haveany impact on boron rejections, which were consistently larger than 98%. This resultindicates that boron rejections are independent of feed water boron concentrations

Table 1 Mass balance calculations on boron in seawater experiments

Exp. No. 1 2 3 4 5 6 7 8 9

% Mass balance closure 91.0 94.6 99.2 93.8 107.7 100.7 97.6 98.6 100.2

1132 H. Köseoglu et al.

75

80

85

90

95

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Bo

ron

Rej

ecti

on

(%

)

Filmtec SW pH=8.2

Filmtec pH=10.5

Toray SW pH=8.2

Toray pH=10.5

Fig. 2. Impact of solution pH on boron rejection by Toray and FilmTec membranes (batch con-centration mode, initial feed boron concentration: 5.1 mg/L, pressure: 700 psi, feed flowrate: 3.8L/min, feed temp: 22–24◦C)

90

92

94

96

98

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

B-R

ejec

tio

n (

%)

Batch-Concentration

Constant Feed Tank Volume

Fig. 3 Impact of operation mode on boron rejection (Toray membrane, initial feed boron con-centration: 5.1 mg/L, pressure: 700 psi, pH: 10.5, feed flowrate: 3.8 L/min, feed temp: 23–24◦C)

Boron Removal in Seawater Desalination by Reverse Osmosis Membranes 1133

resulting in similar permeate boron concentrations. However, it should also be notedthat the operation period was only 14 h and that concentration factors in batch modeoperations were only about 1.1–1.3 making feed water boron concentrations in therange of 5.6–6.6 mg/L. Higher operation periods and higher feed boron concentra-tions may lead to higher boron passage through membranes due to concentrationpolarization thus enhanced diffusion effects.

Boron rejections were relatively constant (a maximum change of ±3%) duringthe 14 h operation period for all experiments, suggesting that steady state dynamicmembrane conditions were immediately achieved in seawater. However, it shouldbe noted that before each test a new membrane was conditioned for about 5–6 hof feeding with DDW only at the conditions of the experiment to be conducted.It was found in the preliminary single solute (boron) experiments in DDW thatabout 4–5 h of operation was necessary to reach steady state boron rejections. Thistime period had also varied depending on membrane type or experimental condi-tion. Although boron rejections were relatively constant in seawater experimentsduring the 14 h operation period, permeate fluxes gradually decreased in all exper-iments. This observation suggests that membranes immediately reach steady stateconditions for boron rejection but not for permeate production rate in the first 14 hof use.

Figure 4 shows the impact of pH and membrane type on permeate fluxes. Foreach membrane type, permeate fluxes at constant pressure were generally lower at

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Per

mea

te F

lux

(L/m

2 -h

)

Filmtec SW pH = 8.2

Filmtec pH = 10.5

Toray SW pH = 8.2

Toray pH = 10.5

Fig. 4 Impact of pH and membrane type on permeate flux (batch concentration mode, initial feedboron conc: 5.1 mg/L, pressure: 700 psi, feed flowrate: 3.8 L/min, feed temp: 22–24◦C)

1134 H. Köseoglu et al.

pH of 10.5. This impact was more pronounced for FilmTec membrane, i.e., about15–40% reduction in fluxes was observed at pH of 10.5 compared to pH of 8.2. Thisphenomenon can be partially explained by membrane fouling and enhanced scaleformation by Mg and Ca compounds at higher pH values. Scale formation on themembrane surfaces caused by super saturation is the disastrous effect of the con-centration polarization which results in considerable flux decreases. In addition, theuse of NaOH for elevating the pH to 10.5 further increased the solute concentra-tion in the feed water, resulting in an increase in osmotic pressure and concentrationpolarization effect, which both gradually decrease permeate flux.

Figure 5 shows the impact of pressure on permeate flux. As expected, as themembrane pressures were increased from 600 to 800 psi permeate fluxes alsoincreased for both membrane types. Similar fluxes were obtained for Toray andFilmTec membranes at constant pressure. However, for all experimental conditions,permeate fluxes gradually decreased during the 14 h operation period although aleveling off was observed after 12 h at all tests. This result suggests that longer oper-ation period than 14 h may be required to reach steady state permeate productionrate. As discussed previously, fouling and concentration polarization appear to beimportant factors in flux reductions even at the very beginning of experiments withnew membranes. Typical average permeate flux in commercial seawater RO desali-nation plants with spiral wound membranes ranges between 14 and 16 L/m2-h. Theranges of permeate fluxes measured in all different experimental conditions in thiswork were 11–15, 13–17 and 19–21 L/m2-h for 600, 700 and 800 psi (41, 48 and55 bar) pressures, respectively, after an operation period of 14 h (new membraneswere used for each experiment). The range of flux values obtained in this work issomewhat higher those in commercial RO systems. This can be explained by twofacts: 1) flat-sheet test units as used in this study exert lower degree of pressure losscompared to spiral wound commercial systems, and 2) prolonged operation peri-ods after 14 h may further reduce the fluxes, based on the gradual-decrease trendobserved in this study.

The impact of pressure on boron rejections is shown in Fig. 6. Boron rejec-tions varied between 85 and 92% for all pressures at a pH of 8.2. As a generaltrend, Toray membranes exhibited somewhat higher rejections as the pressure wasincreased from 600 to 800 psi although there was some scatter in the data. Similarboron rejections were found for Filmtec membrane at pressures 700 and 800 psi;again the data was somewhat scattered. Since the differences in boron reductions arerelatively small the inherent experimental errors in permeate boron measurementsshould also be considered.

Figure 7 shows the impact of pressure on salt rejections. After steady stateconditions were achieved, similar salt rejections were found (97–99%) for bothmembranes independent of pressure. This result indicates that the rejection of saltsin seawater may not correlate well with boron rejection at constant conditions.Figure 8 shows the impact of feed flowrate on boron rejection. At constant mem-brane pressure of 800 psi and pH of 8.2, feed flowrate thus the cross-flow velocitydid not exert any significant impact on boron rejection.

Boron Removal in Seawater Desalination by Reverse Osmosis Membranes 1135

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Per

mea

te F

lux

(L/m

2 -h

)

Toray 600 psiToray 700 psiToray 800 psiFilmTec 700 psiFilmTec 800 psi

Fig. 5 Impact of pressure on permeate flux (batch concentration mode, initial feed boron conc:5.1 mg/L, pH: 8.2, feed flowrate: 3.8 L/min, feed temp: 22–24◦C)

80

82

84

86

88

90

92

94

96

98

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Bo

ron

Rej

ecti

on

(%

)

Toray 600 psiToray 700 psiToray 800 psiFilmTec 700 psiFilmTec 800 psi

Fig. 6 Impact of pressure on boron rejection (batch concentration mode, initial feed boron conc:5.1 mg/L, pH: 8.2, feed flowrate: 3.8 L/min, feed temp: 22–24◦C)

1136 H. Köseoglu et al.

80

82

84

86

88

90

92

94

96

98

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Co

nd

uct

ivit

y R

ejec

tio

n (

%)

Toray 600 psiToray 700 psiToray 800 psiFilmTec 700 psiFilmTec 800 psi

Fig. 7 Impact of pressure on salt (as conductivity) rejection (batch concentration mode, initialfeed boron conc: 5.1 mg/L, pH: 8.2, feed flow rate: 3.8 L/min, feed temp: 22–24◦C)

80

82

84

86

88

90

92

94

96

98

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Operation Time (h)

Bo

ron

rej

ecti

on

(%

)

2 L/min

3.8 L/min

Fig. 8 Impact of feed flowrate (cross-flow velocity) on boron rejection (Toray membrane, batchconcentration mode, initial feed boron conc: 5.1 mg/L, pressure: 800 psi, pH: 8.2, feed temp:22–24◦C)

Boron Removal in Seawater Desalination by Reverse Osmosis Membranes 1137

4 Conclusions

Mass balance closures were between 91 and 107%, suggesting relatively low lossof boron on membrane surfaces during 14 h of operation. Although membranesreached steady state conditions for boron rejection within about 4–5 h of operationpermeate fluxes gradually decreased in the 14 h of operation. However, a generalleveling off was observed for fluxes after 12 h. For both Toray and FilmTec mem-branes, much higher boron rejections were obtained at pH of 10.5 (>98%) thanthose at original seawater pH of 8.2 (about 85–92%). Permeate boron concentrationsless than 0.1 mg/L were easily achieved at pH 10.5 by both membranes. Similarsalt rejections were found (97–99%) for both membranes independent of pressureand pH. Operation modes did not have any impact on boron rejections, indicatingthat boron rejections were independent of feed water boron concentrations up to6.6 mg/L. For each membrane type, permeate fluxes at constant pressure were gen-erally lower at pH of 10.5, which may be partially explained by membrane foulingand enhanced scale formation by Mg and Ca compounds from concentration polar-ization effect at higher pH values. As the membrane pressures were increased from600 to 800 psi permeate fluxes also increased for both membrane types as expected.At constant membrane pressure of 800 psi and pH of 8.2, feed flowrate thus thecross-flow velocity did not exert any significant impact on boron rejection.

Acknowledgement The financial support of the Middle East Desalination Research Center(MEDRC) (Project Number: 04-AS-004) for this work is well acknowledged. The authors thankToray Industries, Inc. and FilmTec Corp. for providing the membranes.

References

1. Redondo J (2001) Brackish, sea and waste water desalination as sustainable water sources.Technical-economical and ecological answers based on membrane technology. Desalination138:29–40

2. Redondo J, Busch M, De Witte JP (2003) Boron removal from seawater using FILMTECTMhigh rejection SWRO membranes. Desalination 156:229–238

3. Wilf M, Bartels C (2005) Optimization of seawater RO systems design. Desalination 173:1–124. Busch M, Mickols WE (2004) Reducing energy consumption in seawater desalination.

Desalination 165:299–3125. Polat H, Vengosh A, Pankratov I, Polat M (2004) A new methodology for removal of boron

from water by coal and fly ash. Desalination 164:173–1886. Mastromatteo E, Sullivan F (1994) Summary: International symposium on the health effects

of boron and its compounds. Environ Health Perspect 102:139–1417. Bick A, Oron G (2005) Post treatment design of seawater reverse osmosis plants: Boron

removal technology selection for potable water production and environmental control.Desalination 178:233–246

8. Taniguchi M, Fusaoka Y, Nishikawa T, Kurihara M (2004) Boron removal in RO seawaterdesalination. Desalination 167:419–426

9. Nadav N (1999) Boron removal from seawater reverse osmosis permeate utilizing selectiveion-exchange resin. Desalination 124:131–135

10. Glueckstern P, Priel M (2003) Optimization of boron removal in old and new SWRO systems.Desalination 156:219–228