Desalination 250 (2010) 414–417

Contents lists available at ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r.com/ locate /desa l

Development of solid-phase microextraction for the determination oftrihalomethanes in drinking water from Bizerte, Tunisia☆

M. Bahri, M.R. Driss ⁎Laboratoire de Chimie Analytique et Environnement, 05/UR/12-03, Faculté des Sciences de Bizerte, 7021 Zarzouna, Tunisia

☆ Presented at the 1st Maghrebian Conference onWatHammamet, Tunisia, 7–10 December 2007.⁎ Corresponding author. Tel.: +216 72 591906; fax: +

E-mail address: MR.DRISS@fsb.rnu.tn (M.R. Driss).

0011-9164/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.desal.2009.09.067

a b s t r a c t

a r t i c l e i n f o

Available online 12 October 2009

Keywords:Trihalomethanes (THMs)Disinfection by-products (DBPs)Headspace-solid-phase microextraction(HS-SPME)Drinking water analysis

Headspace-solid-phase microextraction (HS-SPME) combined with gas chromatography-electron capturedetector (GC-ECD) has been developed and studied for the determination of trihalomethanes (THMs) intreated water samples. Experimental parameters such as the selection of thickness of the polymer coating,addition of salt, magnetic stirring, extraction temperature, and extraction time were studied. Extraction ofthe analytes was performed using HS-SPME with a 100 µm poly(dimethylsiloxane) coating followed bythermal desorption at 250 °C and GC analysis. The optimized conditions were 20 min extraction time at 35 °Cwith 25 w/v% NaCl. Analytical parameters such as linearity and limit of detection were also evaluated. Thelinear range of 1–100 µg/l was established with relative standard deviations (%RSD) within the range, 1.3–11.7%. The limits of detection (LODs) were ranged from 1.4 ng/l to 6.1 ng/l. The average THM concentrationwas 88.16 µg/l which was well within the proposed European Union directive of 100 µg/l.

er Treatment and Desalination,

216 72 590 566.

ll rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Chlorination is the most widely used disinfection method fordrinkingwater [1,2], but its disadvantage is the formation of disinfectionby-products (DBPs). Themajor DBPs are trihalomethanes (THMs). Theyare formed when free available chlorine reacts with natural organicmatter in raw water during water disinfection [3,4].The frequentlyformed THMs are chloroform (CHCl3), dichlorobromomethane (CHCl2-Br), chlorodibromomethane (CHClBr2) and bromoform (CHBr3). Theycan have adverse health effects. The US Environmental ProtectionAgency [5], the EuropeanUnion (EU) [6] has setmaximum contaminantlevels 80 or 100 µg/l, respectively, for total THM concentrations indrinking water. The World Health Organization (WHO) providesguidelines for individual THM compounds [7].

A wide number of techniques are reported in the literature for thedetermination of THMs and other volatile organic compounds in watersamples, such as liquid–liquid extraction (LLE) [8], purge-and-trap [9]and solid-phasemicroextraction (SPME).SPME developed by pawliszyn[10] is a practical solvent-free alternative for the extraction of organiccompounds. SPME integrates sampling, extraction, concentration andsample introduction into a single solvent-free step. Analytes in thesample are directly extracted and concentrated to the extraction fiber.This technique has been successfully applied to the analysis of THMs

[11–13], BTEX [14], organochlorine pesticides [15], PAHs [16], PCB [17]and volatile organic compounds [18] in water samples.

The main objective of this work was to develop, evaluate andimprove a simple, rapid and sensitive method for extraction anddetermination of THMs in drinking water by headspace-solid-phasemicroextraction (HS-SPME) combined with a capillary gas chromatog-raphy-electron capture detector (GC-ECD). Analytical methods areneeded for monitoring THMs directly in the drinking water distributionsystem and applied for an investigation program concerning the THMconcentration in Tunisia drinking water supplies.

2. Experimental

2.1. Standard solutions

A stock solution of a THM mixture (CHCl3, CHCl2Br, CHClBr2 andCHBr3) containing each compound at 2 mg/ml MeOH were purchasedfrom Supelco. Working aqueous standard solutions were prepared dailybydiluting themethanolic standardswithhighqualitywater (ultrapure)obtained using a Milli-Q water purification system (Millipore, Bedford,MA, USA) and also stored at 4 °C in the refrigerator. Final concentrationswere in the range of 1–100 µg/l for each analyte together.

2.2. HS-SPME procedure

The SPME holder and fiber assemblies for manual sampling wereprovided by Supelco. The fiber coatings assayed was poly(dimethyl-siloxane) (PDMS, 100, 30 and 7 µm).Before measurements the fiberwas conditioned according to Supelco's recommendations. The HS-

Fig. 1. Extraction efficiencies with different film thicknesses of the PDMS fiber: extractiontime of 20 min at 35 °C, desorption time of 4 min at 250 °C. THM concentration: 80 µg/l,n=5.

Fig. 2. Effect of the addition of salt and magnetic stirring on extraction: extraction timeof 20 min at 35 °C, desorption time of 4 min at 250 °C. THM concentration: 80 µg/L,n=5.

415M. Bahri, M.R. Driss / Desalination 250 (2010) 414–417

SPME extractions were performed by placing 2 mL of aqueous sampleinto 4 mL vials capped with PTFE-coated septa. The aqueous standardsolutions were freshly prepared by spiking appropriate amounts ofthe working standard solution. The samples were immersed in atemperature-controlled water bath during the sampling process. TheHS-SPME equilibrium was conducted with stirring the sample for anappropriate time period, during which analytes sorb on the stationaryphase of the fiber. After extraction, the fiber was thermally desorbedfor 4 min into the glass liner of the gas chromatograph injector at250 °C. Every day before use, the SPME fiber was conditioned for 5–15 min at 250 °C. Identification of the four analytes was deduced fromtheir retention times and quantification was performed using thepeak area measurement as well as comparison with responses of amixed THMs standard based on multi-level calibration from 1 to100 µg/l (n=6).

2.3. SPME-GC-ECD analysis

AHewlett-Packard 6890 gas chromatograph equippedwith a split/splitless injection port, an 63Ni electron capture detector (GC-ECD)and operated by HP Chemstation software was used for theexperiments to optimize HS-SPME conditions. The injector was usedin splitless mode (4 min) and held isothermally at 250 °C. The columnused for analysis was VOCOL (60 m×0.32 mm ID, 1.8 µm filmthickness, Supelco). The initial oven temperature was set at 60 °Cfor 4 min, ramped at 15 °C/min to 210 °C and held for 15 min. The ECDsystemwasmaintained at 300 °C. The Carrier gas was helium at a flowrate of 1.5 ml/min and the flow rate of themake-up gaswas 60 ml/minwith nitrogen.

2.4. Sample collection

Duplicate samples for THM analysis were collected from thedistricts' water treatment plants (WTPs) and its distributive system inBizerte area (north of Tunisia). All the samples were collected in 22-ml amber glass vials and were capped with PTFE-faced silica septa.Before sampling, a sodium sulfite (50 µl of a 1.5 g/l Na2SO3 solution)was added to bottles to eliminate any remaining residual chlorine andto stop further chlorination by-products (CBP) formation. The vialswere completely filled to avoid evaporation of volatile compounds.The samples were transported to the laboratory, transferred to arefrigerator (set at 4 °C) and analyzed within 2 days of collection.

3. Results and discussion

3.1. Development of HS-SPME procedure

The transport of analytes from the matrix into the extractionmedium begins as soon as the coated fiber has been placed in contactwith the sample. In most cases, SPME extraction is considered toterminate when the analyte concentration has reached distributionequilibrium between the sample matrix and the fiber coating. Theamount of analytes adsorbed by the fiber depends on the thickness ofthe polymer coating, salt addition, extraction temperature andextraction time. Experimental parameters were optimized and usedfor the analysis of THM in drinking water.

3.1.1. Effect of film thicknessIn this study three fibers with different film thickness, 7-µm poly

(dimethylsiloxane) (7-PDMS), 30-µm poly(dimethylsiloxane) (30-PDMS) and 100-µm poly(dimethylsiloxane) (100-PDMS) were chosento select the appropriate fiber for the analysis. New fibers were con-ditioned, following themanufacturer's recommendations.Water samples(2 ml spiked at a level of 80 µg/l of THMs)were analyzed with each fiber.The extraction time was 20 min at 35±1 °C and desorption time was4 min at 250 °C for all fibers. In order to evaluate the extraction efficiency,

thepeak areas obtained for each THMswith the differentfibers are shownin Fig. 1. Extraction efficiencies for the THMswere increased according tothe following order: 7-PDMS < 30-PDMS < 100-PDMS. The 100-PDMSfiber was found to be able to extract these compounds from aqueoussolution by HS-SPME method.

3.1.2. Effect of the addition of salt and magnetic stirringThe addition of salt can improve the extraction efficiency for

compounds. So, sodium chloride at various concentrations (from 0 to25%, w/v) was studied. The results of these experiments showed thatthe optimum responses of the compounds were obtained with theaddition of 25 w/v% NaCl (Fig. 2).

Stirring the water sample can increase extraction efficiency,because stirring can speed up the transfer of the compounds fromwater to headspace. Fig. 3 shows that stirring (at 750 rpm) affectssignificantly the response of THM.

3.1.3. Effect of extraction temperatureThe influence of temperature on the extraction yield was studied

varying the temperature between 20 and 60 °C, using 100 µm PDMSand 20 min extraction time. I t can be seen from Fig. 4 that betterresponse for most of the compound s were obtained at 35 °C.

Fig. 3. Effect of magnetic stirring on extraction: extraction time of 20 min at 35 °C,desorption time of 4 min at 250 °C. THM concentration: 80 µg/L, n=5.

Fig. 5. Effect of extraction time from 5 to 50 min at 35 °C on extraction: desorption timeof 4 min at 250 °C. THM concentration: 80 µg/L, n=5.

416 M. Bahri, M.R. Driss / Desalination 250 (2010) 414–417

3.1.4. Extraction time profilesTheHS-SPME is an equilibriumprocess of the analytes between the

vapour phase and the fiber coating, so it is important to determine thetime that analytes reach equilibrium. Analytes with high molecularweight or low Henry's constant values need longer equilibrium times[10]. In order to evaluate the extraction efficiency, the extraction timeranged from 5 to 50 min at 35 °C, using 100 µm PDMS. Acceptableequilibrium states were achieved for THM at 20 min as shown in Fig. 5.

3.2. Linear range and limits of detection

Quality parameters such as linearity and limits of detection werecalculated when the optimum conditions for the HS-SPME-GC-ECDprocedure were established. The linearity range of the HS-SPMEmethod was evaluated by plotting the calibration curves of the areaversus the concentration of each analyte. The correlation coefficients(R2) for all compounds were good, since they ranged from 0.9947 to0.9979. The relative standard deviation (RSD%) ranged from 1.3 to11.7 indicating that the method had good repeatability. The limits ofdetection (LODs), defined as the concentration of analytes in thesamplewhich causes a peakwith a signal-to-noise ratio of 3, were alsodetermined. In order to calculate them, ultrapure water containingeach analyte spiked at low concentrations (0.1 µg/l) was used. Underthese conditions, LODs of the methodwere ranged from 1.4 to 6.1 ng/l(Table 1).

Fig. 4. Effect of extraction temperature on extraction efficiency: extraction time of20 min, desorption time of 4 min at 250 °C. THM concentration: 80 µg/L, n=5.

3.3. THM concentrations in Bizerte drinking water

So far, no informationonTHMconcentrations indrinkingwater fromTunisia has been available. Thus, the proposed HS-SPME method wasdeveloped to monitor and investigate THMs at the Bizerte watertreatment plant (WTP) and its distributive system. BizerteWTP receiveswater mainly from dam (located at 2000 m from the WTP) and supplywater to nearly 180,000 people, treating 30,000 m3/day. The plantcarries out conventional treatment, consisting of prechlorination (tobreak-point), flocculation, coagulation, filtration and final chlorinationwith a lower dosage of chlorine, to guarantee a 0.9–1.5 mg/l concentra-tion in the distribution water system. THM concentrations weredetermined simultaneously at two points of the WTP and seventeenreservoirs located at variousdistances fromtheplant. Results of analysesof water samples are summarized in Table 2. CHCl3, CHCl2Br, CHClBr2and CHBr3 were detected in all water samples. The concentration ofTTHMs was ranged from 51.41 to 119.48 µg/l. The most of total THMconcentration in analyzed water samples were below the drinkingwater standards of the EU and the USEPA. Brominated THMs were thedominant and most abundant species. They represented in term ofconcentration among the TTHMs a percentage ranged from 80% (tapwater1) to 95% (tap water 15).This can be explained by the presence ofbromide ions in raw water. Many researchers have observed that thepresence of bromide increases the yield of THMs [19–21]. Also thebromide ion (Br−) influences the by-products mixture by formingbromine containing species when chlorine oxidizes the bromidehypobromous acid (HOBr), which behaves in a manner analogous tohypochlorous acid. Also, the recorded results show that the concentra-tion of brominated THM increases when the distance from the WTP

Table 1Retention time, correlation coefficients (R2) of linearity and detection limits with %RSDin parenthesis, for four trihalomethane standards (1–100 µg/l) (n=6), using the HS-SPME method and GC analysis.

Compound Retention time (min) R2 LOD (ng/l)

CHCl3 9.87 0.9968 (4.7–10.1) 1.4CHCl2Br 11.86 0.9952 (1.3–8.6) 6.1CHClBr2 13.69 0.9947 (5.3–11.7) 2.3CHBr3 15.45 0.9979 (3.8–10.5) 2.2

Table 2THM concentrations (µg/l) in Bizerte drinking water.

Distance fromWTPa (m)

CHCl3 CHCl2Br CHClBr2 CHBr3 TTHMs

WTP1b – 8.01 13.26 21.03 9.11 51.41WTP1c – 11.42 17.53 24.38 10.76 64.19Tap water 1 7000 15.60 20.27 27.59 11.22 74.68Tap water 2 20,000 13.27 19.91 29.82 22.21 85.21Tap water 3 40,000 11.68 17.24 27.71 32.47 89.10Tap water 4 44,000 14.78 22.81 33.21 27.94 98.74Tap water 5 42,000 15.01 25.40 40.83 38.24 119.48Tap water 6 43,000 14.25 21.03 31.22 30.58 97.08Tap water 7 48,000 13.18 21.37 33.32 35.51 103.38Tap water 8 45,000 13.48 22.25 35.46 28.58 99.77Tap water 9 43,000 8.63 14.42 23.09 26.52 72.66Tap water 10 45,000 8.13 16.10 27.29 33.45 84.97Tap water 11 41,000 10.54 18.02 28.48 27.59 84.63Tap water 12 46,000 10.96 18.94 32.90 40.20 103.00Tap water 13 47,000 8.92 16.90 25.33 31.36 82.51Tap water 14 58,000 6.24 15.88 29.76 39.10 90.98Tap water 15 62,000 3.63 12.81 27.24 41.05 84.73Tap water 16 59,000 5.49 14.27 28.44 40.41 88.61Tap water 17 66,000 6.02 14.56 30.33 48.93 99.84Average 10.49 18.05 29.34 30.27 88.16

a Water treatment plant.b After prechloration and before decantation.c Water at the extremity point of the WTP.

417M. Bahri, M.R. Driss / Desalination 250 (2010) 414–417

increases. This is due to the fact that the formation of THM is carried onthe distribution system.

4. Conclusions

The present results show that the combination of HS-SPMEwithGC-ECD is a powerful tool for determination and monitoring the fourtrihalomethanes commonly found in drinking water samples. Extrac-tion conditions such as thickness of the polymer coating, salt addition,extraction temperature, and time are optimized and must be keptconstant during the procedure of analysis. The 100 PDMS fiber isproposed for extracting THMs, which allows the quantitative analysis ofthis group of disinfection by-products in water samples. Equilibration

without an increase in sample temperaturewas achieved and sensitivitywas improved by addition of salt. The method has good linearity in therange of concentrations of interest with good precision between 1 and12% and it is sufficiently sensitive with limits of detection in the ng/l range. Thefirst result of THM from the survey of Bizerte drinkingwatersamples showed that the average concentration of THMs was 88.16 µg/l which was well within the proposed European Union directive of100 µg/l. It can be concluded that the HS-SPME technique has a greatpotential for the analysis of drinking water.

References

[1] X. Yang, C. Shang, Environ. Sci. Technol. 38 (19) (2004) 4995.[2] X. Yang, C. Shang, J.C. Huang, Water Res. 39 (2005) 4755.[3] T.A. Bellar, J.J. Lichtenberg, R.C. Kroner, J. Am, Water Works Assoc. 66 (1974)

703–706.[4] T.D. Lekkas, in: O. Hutzinger (Ed.), Haloforms and Related Compounds in Drinking

Water, Springer, Berlin, 2003, pp. 193–214.[5] US EPA National Primary Drinking Water Regulations, Stage 2 Disinfectants and

Disinfection By-products Rule: Proposed Rule, Fed. Reg. Nos. 49547-49681, August18, 2003.

[6] The council of theEuropeanUnionQuality ofwater intended for humanconsumption,Official Journal of the European Communities, Council Directive 98/83/EC, 1998.

[7] World Health Organization (WHO), Guidelines for Drinking Water Quality, ThirdEd., Vol. 1 Geneva, 2004.

[8] S.K. Golfinopoulos, T.D. Lekkas, A.D. Nikolaou, Chemosphere 45 (2001) 275–284.[9] N. Campillo, P. Vinas, I. Lopez-Garcia, N. Aguinaga, M. Hernandez-Cordoba,

J. Chromatogr. A 1035 (2004) 1–8.[10] J. Pawliszyn, J. Pawliszyn, Solid Phase Microextraction: Theory and Practice,

Wiley-VCH, New York, 1998.[11] C. Deok-Hee, K. Sung-Ho, O. Seong-Geun, Water Res. 37 (2003) 402–408.[12] M.A. Stack, G. Fitzgerald, S. O'Connell, K.J. James, Chemosphere 41 (2000)

1821–1826.[13] M.S.J. Pedro, D.C. José, T.T. María, J. Chromatogr. A 1139 (2007) 27–35.[14] S. Fustinoni, R. Giampiccolo, S. Pulvirenti, M. Buratti, A. Colombi, J. Chromatogr. B

723 (1999) 105–115.[15] D. Ruey-An, L. Pei-Lin, J. Chromatogr. A 918 (2001) 177–188.[16] P. Popp, C. Bauer, M. Moder, A. Paschke, J. Chromatogr. A 897 (2000) 153–159.[17] A. Derouiche, M.R. Driss, J.P. Morizur, M.H. Taphanel, J. Chromatogr. A 1138 (2007)

231–243.[18] S. Nakamura, S. Daishima, Anal. Chim. Acta 548 (2005) 79–85.[19] T. Aizawa, Y. Magara, M. Musashi, J. Wat. Sup. Res. Technol. Aqua. 38 (1989)

165–175.[20] S.K. Golfinopoulos, M.N. Kostopoulou, T.D. Lekkas, J. Environ. Sci. Hth. Part A —

Environ. Sci. Eng. 31 (1996) 67–81.[21] I. Toroz, V. Uyak, Desalination 176 (2005) 127–141.