- Email: [email protected]
Application of probe sonication extraction for the determination of linear alkylbenzene sulfonates from sewage sludge. Comparison with other extraction methods
Microchemical Journal 90 (2008) 164–170
Contents lists available at ScienceDirect
Microchemical Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / m i c r o c
Application of probe sonication extraction for the determination of linear alkylbenzene sulfonates from sewage sludge. Comparison with other extraction methods M. Villar, R. Fernández-Torres, M. Callejón ⁎, P. Villar, J.C. Jiménez Department of Analytical Chemistry, Faculty of Chemistry, University of Seville., c/ Profesor García González s/n, 41012, Seville, Spain
A R T I C L E
I N F O
Article history: Received 7 April 2008 Received in revised form 14 May 2008 Accepted 14 May 2008 Available online 21 May 2008 Keywords: Linear alkylbenzene sulfonates Sewage sludge Extraction Liquid chromatography
A B S T R A C T A new method based on probe sonication extraction (USP) prior to high performance liquid chromatography (HPLC) has been developed for the determination of linear alkylbenzene sulfonates (LAS) from sewage sludge. The optimized method was designed to be cost effective compared to existing extraction methods (ultrasonic assisted extraction, Soxhlet or pressurized liquid extraction) which may require large quantities of organic solvents, or costly instrumentation or equipment. The main factors affecting the extraction efficiency (extractant volume, ultrasounds power and extraction time) were optimized using compost sludge. The detection limit of total LAS in the sludge was 10 mg kg− 1. The extraction of C10–C13 homologues is carried out using an extraction time of 7 min with 10 mL of methanol. Liquid chromatography with fluorescence (FL) detector is used for determination of LAS homologues. A mobile phase acetonitrile-water containing 0.1 M NaClO4 (65:35) and isocratic elution was used. Compounds were eluted over 6 min at a flow rate of 1 mL/min. Polar interferences are eluted between 0 and 2 min and no purification of the samples is required prior to the final determination by high performance liquid chromatography (HPLC). The recoveries of LAS in spiked sewage sludge were between 84.0% and 97.0%, which reflect the efficiency of the method for extraction of these analytes from sewage sludge. Concentration levels found were between 11,858 mg kg− 1 for digested sludge and 2379 mg kg− 1 for compost sludge. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Linear alkylbenzene sulfonates (LAS) are the most important synthetic anionic surface-active agents widely used as the principal constituents of commercial detergents. The European consumption of LAS in detergents applications was about 350 kt in 2005. This represents more than 80% of the total European consumption of LAS, which was estimated to be about 430 kt in the year 2005 [1]. They are used as complex mixtures of C10–C13 homologues and of positional isomers where the benzenesulfonate is located at various alkyl carbon positions from the second to the center. After application, LAS are usually discharged through the sewage infrastructure to municipal wastewater treatment plants which are subjected to physical and biological treatment. In wastewater treatment plants, 10% or more of the LAS present in wastewater are eliminated by adsorption/ precipitation processes [2], together with suspended solids during the primary treatment, while the rest are biodegraded in the aerobic stage of the treatment process. Nevertheless, as a consequence of the
⁎ Corresponding author. Tel.: +34 954557169; fax: +34 954557168. E-mail address: [email protected] (M. Callejón). 0026-265X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.microc.2008.05.005
low biodegradability of LAS under anaerobic conditions, the sludge that is stabilized by anaerobic processes shows concentrations of LAS as high as 3000–30,000 mg kg− 1 of dry solid [3–5]. During recent decades a variety of procedures have been used to extract LAS in solid samples. Soxhlet and ultrasounds methods using methanol as extraction solvent are mainly employed [4–7]. Soxhlet and ultrasounds extraction methods are relatively costly, time consuming, labor intensive and require the used of large volume of organic solvents. Some attempts performed in order to reduce both the volume of organic solvents used and the time needed for the complete extraction have been based on supercritical fluid extraction [8] and pressurized liquid extraction with methanol [9,10], all of which require complex and expensive instrumentation. High performance liquid chromatography (HPLC) employing UV or fluorescence detection (FL) is the most frequently used method for the selective determination of LAS [11–15]. The objective of this study was to develop and validate a routine method for analysis of LAS in sewage sludge based on probe sonication extraction (USP) and HPLC-FL determination. The main factors affecting the extraction efficiency were studied and optimized, paying special attention in reducing the volume of organic solvent as well as the time needed for extraction. The method did not require clean-up
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
165
2. Materials and methods
extraction in ultrasound probe extraction. Centrifugation system was a Sigma k3-15. The microwave extraction system was a Microwave Ethos 900 apparatus (Milestone, Sorisole, Italy) with a programmable power and irradiation time. The apparatus is equipped with a carousel that is able to hold six extraction vessels.
2.1. Chemicals and reagents
2.3. Sewage sludge collection and preparation
HPLC grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany), just as analytical grade sodium perchlorate. Petroquímica Española (Cádiz, Spain) supplied the commercial LAS mixture (Petrelab P-550) with the following homologue distribution: C-10 (12.3%), C-11 (32.1%), C-12 (30.8%), C-13 (23.4%). Standard solutions of LAS were prepared in ultrapure water. Ultra-high-quality water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA).
Sewage sludge samples (digested and compost sludge) were collected from Seville (Spain) wastewater treatment plant. Digested sludge is obtained when primary sludge undergoes a process of digestion (aerobic or anaerobic) for its stabilization. 1000 mL of digested sludge were collected in different points to obtain a representative sample. Compost sludge is obtained by exhibition to the sun, with a process of natural fermentation that takes place helped by ventilation. 500 mL of compost sludge were collected in different points of the storage system and different depths to obtain a representative sample. Before analysis, sludge was dried in an oven at 40 °C until humidity b10% (digested sludge between 5 and 7 days and compost sludge between 2 and 3 days) and ground in an agate mortar. The sludge was sieved in order to obtain a fraction b1 mm. This fraction was stored at 4 °C until it was analyzed.
or preconcentration steps and no special equipment is needed. The proposed method has been satisfactorily applied to sewage sludge samples from Seville (Spain) wastewater treatment plant.
2.2. Instrumentation All measurements were made with an Agilent (Palo Alto, CA, USA) 1100 series liquid chromatograph equipped with a fluorescence detector (FL), an injector with a loop 20 μL, a quaternary pump, a vacuum degasser and a thermostated column compartment. The analytical cartridge column was a Zorbax (Agilent) XDB-C8, 150 mm × 4.6 mm ID, 5 μm particle size. A C-18 guard column was installed to protect it from contamination. An ultrasound bath Ultrasons-P (Selecta, Spain) was used in ultrasounds bath extraction. Sonoplus ultrasonic homogeneiser (Bandelin, Germany) fitted with a HF generator 2200 and a microtip of 2 mm diameter was used for LAS
2.4. Preparation of spiked samples Two portions of 20 g of compost sludge were spiked with commercial LAS at two different concentrations (2000, 4000 mg kg− 1 of total LAS). One sample was not spiked in order to evaluate the original presence of LAS in sludge. Spiked samples were prepared adding 40
Fig. 1. Optimization of probe sonication extraction (USP).
166
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
and 80 mL of standard solution of 1000 mg L− 1 of total LAS to 20 g of sludge. Samples were allowed to interact with the natural organic matter during one day under stirring and nitrogen to prevent aerobic degradation.
Table 2 Limits of detection (LOD), limits of quantification (LOQ), repeatability (%RSD) and recoveries (%R)
Method 1
2.5. Sample treatment Probe sonication extraction was performed on 0.5 g of dried sewage sludge. Sewage sludge was dried in an oven at 40 °C and ground in an agate mortar. The sludge was sieved in order to obtain a fraction b1 mm prior to the analysis. The main factors affecting the extraction efficiency were optimized using compost sludge from a Seville (Spain) wastewater treatment plant. Methanol was used as extraction solvent and the extracts were filtered through glass wool and analyzed by HPLC.
Method 2
Method 3
3. Results and discussion 3.1. Chromatographic conditions LAS were determined as previously described by Villar et al. [16], using as mobile phase acetonitrile-water containing 0.1 M NaClO4 (65:35), isocratic elution and the flow rate was 1 mL min− 1. The column was thermostated (25 °C) and fluorescence detector (λex =225 nm, λem = 295 nm) was used for the determination of LAS. Retention time for LAS C10, LAS C-11, LAS C-12 and LAS C-13 were 2.38, 2.82, 3.43 and 4.32 min, respectively. 3.2. Optimization of probe sonication extraction (USP) The optimization of the extraction procedure was performed with compost sludge from a Seville (Spain) wastewater treatment plant. The different factors that could influence in the process of extraction (extractant volume, extraction time and ultrasounds power) were studied and optimized. Three samples of 0.5 g of dried compost sludge were subjected to USP extraction with methanol. The first parameter optimized was the extractant volume. It was tested between 2 and 20 mL of methanol (extraction time fixed at 3 min and power fixed at 20% of the maximum power (2200 W) of the ultrasounds probe). Maximum extraction was achieved for the four LAS when 10 mL of methanol were used (Fig. 1). Secondly, the irradiation power was tested between 10 and 50% of the maximum power (extractant volume fixed at 10 mL and extraction time at 3 min). All LAS showed maximum extraction when 50% of power was used (Fig. 1). Finally, the extraction time was tested between 1 and 15 min (extractant volume 10 mL and ultrasonic power 50%). All LAS showed maximum extraction when 7 min of extraction was used (Fig. 1). Optimal conditions for extraction of linear alkylbenzene sulfonates of sewage sludge using probe sonication extraction were: 10 mL of methanol, 50% of irradiation power and 7 min of extraction time.
Method 4
Method Method Method Method
1: 2: 3: 4:
LAS C-10
LAS C-11
LAS C-12
LAS C-13
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 5.42
2.65 6.62 3.50
4.58 15.22 4.21
2.05 6.85 5.31
94 85
93 83
90 83
84 86
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 3.48
2.65 6.62 4.46
4.58 15.22 3.72
2.05 6.85 5.40
94 88
93 89
93 82
85 87
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 3.96
2.65 6.62 4.58
4.58 15.22 4.38
2.05 6.85 4.51
95 88
96 91
97 84
93 89
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.33 1.11 2.10
1.06 3.53 4.70
1.83 6.09 4.02
0.82 2.74 5.42
94 90
102 95
100 85
98 89
Ultrasonic and centrifugation. Stirred and centrifugation. Ultrasonic probe extraction. Microwave-assisted extraction.
research on the linearity, matrix effect, precision, detection and quantification limits and recovery studies. Calibration plots were obtained by using standard solutions of the four homologues of LAS in water. Each level was analyzed by triplicate. Regression coefficients (r), slope standard deviations (RSDb), linearity (%) and linear dynamic ranges are shown in Table 1. Good linearity was observed and the regression coefficients were between 0.9966 for LAS C-11 and 0.9999 for LAS C-10 (Table 1). Detection limits (LOD) for each analyte were calculated as the mass of analyte which gives a signal 3σ above the mean blank signal (where σ is the standard deviation of the blank signal). The LODs obtained were between 0.02 mg L− 1 for LAS C-10 and 0.09 mg L− 1 for LAS C-12. The quantification limits (LOQ), expressed as the mass of analyte which gives a signal 10σ above the mean blank signal, ranged between 0.06 mg L− 1 for LAS C-10 and 0.30 mg L− 1 for LAS C-12. Taking into account the amounts of sample extracted and the volume of extractant used, the detection limits of LAS in sludge samples were between 0.82 mg kg− 1 for LAS C-10 and 4.58 mg kg− 1 for LAS C-12.
3.3. Analytical data and validation of the method A validation process was carried out to confirm the suitability of the method for the intended use. In order to do so, we performed a
Table 1 Regression coefficients (r), slope standard deviations (RSDb), linearity (%) and linear dynamic ranges LAS
r
RSDb (%)
% linearity
Linear dynamic ranges (mg L− 1)
LAS C-10 LAS C-11 LAS C-12 LAS C-13
0.9999 0.9966 0.9981 0.9989
0.07 0.38 0.29 0.24
99.81 99.88 99.12 99.34
0.06–86.10 0.21–321.00 0.42–215.60 0.22–163.80
Fig. 2. Chromatogram of compost sewage sludge (2500 mg kg− 1 approximately).
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
of the slopes was carried out according to González et al. [19]. We calculated R as:
Table 3 LAS (mg kg− 1) in digested and compost sludge (based on three replicates) LAS LAS LAS LAS LAS
C-10 C-11 C-12 C-13
167
Digested
Compost
907 ± 8 3142 ± 12 4207 ± 15 3602 ± 5
89 ± 3 527 ± 5 887 ± 8 876 ± 8
R¼
bSAM b
Where: bSAM is the slope obtained by standard addition b is the slope obtained by external calibration
Results are shown in Table 2 and as can be seen low concentrations of LAS can be detected in environmental sludge samples. To evaluate the repeatability and the intermediate precision of the proposed method, 0.5 g of dried compost sludge of approximately 100 mg kg− 1 of LAS C-10, 560 mg kg− 1 of LAS C-11, 940 mg kg− 1 of LAS C-12 and 900 mg kg− 1 of LAS C-13 were subjected to the entire analytical procedure and measured in one single day and two days per week during three weeks, respectively. All these values are similar to those obtained by other authors [17,18]. This fact confirms the reproducibility of the proposed method. Analytical data for each homologue are summarized in Table 2. Since sewage sludge is a complex matrix, it is very important to study the influence of it on the proposed method. A calibration was realized by standard addition and the slopes were compared with those obtained by an external calibrate. A statistical comparison of standard deviation
The absence of matrix effect corresponds to bSAM = b, or, in terms of recovery, R = 1. This must be checked for statistical significance: t¼
jR−1j uðRÞ
With the uncertainty given by: uðRÞ ¼
2 1=2 u ðbSAM Þ b2SAM u2 ðbÞ þ 2 4 b b
According to the LGC/VAM protocol [20], if the degrees of freedom associated with the uncertainty of consensus recovery are known, t is compared with the two-tailed tabulated value, ttab (ν,P) for the appropriate number of degrees of freedom at P% confidence. If t ≤ttab,
Fig. 3. Optimization of ultrasounds bath plus centrifugation extraction.
168
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
the consensus recovery is not significantly different from 1. Alternatively, instead of ttab, coverage factor k may be used for the comparison. Typical values are k = 2 or k = 3 for 95% or 99% confidence, respectively, so: – if jR−1j uðRÞ Vk, the recovery is not significantly different from 1; and, – if
jR−1j uðRÞ Nk,
the recovery is significantly different from 1 and the
analytical result must be corrected by R. For the four LAS studied the value of t was inferior to 2. Therefore, it could be concluded that the method is not affected by any matrix effect. As no standard reference material was available, spiked sludge obtained as previously described in Section 2.4 was used in order to study the accuracy of the method. The samples were analyzed, by triplicate, using the proposed methods. Recovery values were between 84% and 97% (Table 2). These values are in accordance with those reported previously by García et al. and by Prats et al. [11,21] who used Soxhlet extraction for 8 h (recoveries between 85 and 99%), and by Küchler et al. [22] who used Soxhlet extraction for 24 h (recoveries N86%) and by Sarrazin et al. [23] who carried out an ultrasonic extraction (recoveries N85%). Recoveries obtained by Villar et al. [16] using MAE were between 85% and 102%.
to the procedure described in Section 2.3. Fig. 2 shows a typical chromatogram of compost sludge (2400 mg kg− 1 approximately) and as it is shown LAS C-10 is the first compound eluted at a retention time of 2.38 min followed by LAS C-11, LAS C-12 and finally LAS C-13 with retention times of 2.82 min, 3.43 min and 4.31 min, respectively. Since polar interferences are eluted between 0 and 2 min, they do not interfere in the analysis. Concentration levels of LAS were between 11,858 mg kg− 1 for digested sludge and 2379 mg kg− 1 for compost sludge as can be seen in Table 3. 4. Optimization and comparison with other extraction methods The proposed method has been compared with the method developed by Villar et al. [16] and with other two extraction methods based on previously described methodologies (ultrasounds bath [24] and stirred [25]) but modified, optimized and validated in our laboratory. The main factors affecting the extraction procedures (volume of extractant, time of extraction and stirring time) have been modified and a new step of centrifugation have been also included in the process. 4.1. Optimization of ultrasounds bath plus centrifugation extraction
3.4. Application of the proposed method to sewage sludge from Seville wastewater treatment plant The optimized and validated method was applied to the determination of homologues of LAS in digested and compost sludge from a Seville wastewater treatment plant. Samples were prepared according
Extractant volume, ultrasounds bath time, centrifugation velocity and centrifugation time were studied and optimized. The four LAS achieved maximum extraction when 30 min of ultrasound bath, 10 mL of methanol, 20 min of centrifugation time and 6000 rpm were used (Fig. 3).
Fig. 4. Optimization of magnetic stirring plus centrifugation extraction.
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
169
4.2. Optimization of magnetic stirred plus centrifugation extraction
Table 4 Optimal extraction conditions for the methods
The different factors that can be influenced in the process of extraction (extractant volume, magnetic stirring time, centrifugation velocity and centrifugation time) were optimized. Optimal values were: 5 min of stirring time, 10 mL of extractant, 10 min of centrifugation time and 6000 rpm of velocity of centrifugation (Fig. 4).
Method/conditions
Ultrasonic and centrifugation
Stirred and centrifugation
Ultrasonic probe extraction
Microwaveassisted extraction
Extraction time (min) Extractant volume (mL)
50 10
15 10
7 10
10 5
4.3. Microwave-assisted extraction Microwave-assisted extraction optimization has been previously published [16]. Extractant volume, power of microwave irradiation and irradiation time were chosen to be 5 mL, 250 W and 10 min respectively (Fig. 5). 4.4. Comparison of the methods Results of repeatability, intermediate precision, detection and quantification limits and recovery studies are show in Table 2 for each method. Similar values of repeatability and intermediate precision have been obtained for the tested methods. However, both detection and quantification limits, present lower values when microwaveassisted extraction was used, the range of detection limit was between 0.33 for LAS C-10 and 1.83 mg kg− 1 for LAS C-12. Recoveries obtained in spiked sludge at two concentration levels are given in Table 2. It can be observed that in the four methods studied recoveries values are N82%. As we can see in Table 4, ultrasounds bath, magnetic stirred and USP extraction needed 10 mL of methanol while MAE only needed 5 mL. Probe sonication extraction is carried out in only 7 min versus
10, 15 and 50 min needed in MAE, magnetic stirred extraction and ultrasounds bath extraction, respectively. Despite the fact that MAE needed less quantity of solvent, this itself does not represent a great advantage with respect to the other methods because the volume of solvent required in the other cases is only 10 mL which, comparatively, does not represent a great waste of solvent. What it really means an advantage compared to the other methods is that probe sonication extraction shortens significantly the time of analysis with respect to magnetic stirred extraction and ultrasounds bath extraction and although this time of treatment seems to be similar between MAE and USP the former needs a long cooling time that lengthens too much time of analysis. Besides, the extraction procedure using USP is cost effective, simple and rapid compared to MAE. 5. Conclusions A simple, rapid and practical method has been developed for determination of LAS homologues in sewage sludge. No special equipment is required and extractions are not dependent on the length of the alkyl chain of LAS and no purification of the samples is required prior to the
Fig. 5. Optimization of microwave-assisted extraction (MAE).
170
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
final determination by HPLC. USP is an attractive, affordable and effective alternative to existing extraction methods. Ultrasounds bath, magnetic stirred and USP extraction needed 10 mL of methanol while MAE only needed 5 mL, but the main advantage of the proposed method in comparison to these other methods is the low extraction time, USP extraction needed only 7 min for the extraction while 50 min,15 min and 10 min are required in ultrasounds bath, magnetic stirring and microwave-assisted extraction respectively (Table 4). Probe sonication extraction (USP) has proved to constitute reliable, reproducible and efficient method for the extraction of LAS in sewage sludge. Acknowledgements The authors would like to thank the Empresa Municipal de Abastecimiento y Saneamiento de Aguas de Sevilla (EMASESA) for supplying sludge samples, Petroquímica Española (PETRESA) and Programa de Becas de Formación de Profesorado Universitario (FPU, Ministerio de Educación y Ciencia, Spain). References [1] HERA (2007). Human and Environmental Risk Assessment of Linear Alkylbenzene Sulphonate. [2] W. Giger, A.C. Alder, P.H. Brunner, A. Marcomini, H. Siegrist, Tenside. Surfact. Det. 26 (1989) 95. [3] E. Matthijs, H. de Henau, Tenside. Surfact. Det. 24 (1987) 193.
[4] D. Prats, F. Ruiz, B. Vázquez, D. Zarzo, J.L. Berna, A. Moreno, Environ. Toxicol. Chem. 12 (1993) 1599. [5] E. Sanz, D. Prats, M. Rodríguez, A. Camacho, Waste Manage. 26 (2006) 1237. [6] K. Bester, N. Theoblad, H.Fr. Schröder, Chemosphere 45 (2001) 817. [7] S.O. Petersen, K. Henriksen, G.K. Mortensen, P.H. Krogh, K.K. Brandt, J. Sorensen, T. Madsen, J. Petersen, C. Gron, Soil Tillage Res. 72 (2003) 139. [8] A. Kreisselmeier, H.W. Dürbeck, J. Cromatogr. A 775 (1997) 187. [9] S. Álvarez-Muñoz, M. Sáez, P.A. Lara-Martín, A. Gómez-Parra, E. González-Mazo, J. Chromatogr. A 1052 (2004) 33. [10] S. González, M. Petrovic, D. Barceló, J. Chromatogr. A 1052 (2004) 111. [11] M.T. García, E. Campos, J. Sánchez-Leal, I. Ribosa, Water Res. 40 (2006) 2958. [12] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1030 (2004) 109. [13] L. Carlsen, M.B. Metzon, J. Kjelsmark, Sci. Total Environ. 290 (2002) 225. [14] P.A. Lara-Martín, M. Petrovic, A. Gómez-Parra, D. Barceló, E. González-Mazo, Environ. Pollution 144 (2006) 483. [15] M. Nimer, O. Ballesteros, A. Navalón, G. Crovetto, C. Verge, I. López, J.L. Berna, J.L. Vílchez, Anal. Bioanal. Chem. 387 (2007) 2175. [16] M. Villar, M. Callejón, J.C. Jiménez, E. Alonso, A. Guiráum, Anal. Chim. Acta 599 (2007) 92. [17] M.T. García, E. Campos, I. Ribosa, A. Latorre, J. Sánchez-Leal, Chemosphere 60 (2005) 1636. [18] S. Morales-Muñoz, J.L. Luque-García, M.D. Luque de Castro, J. Chromatogr. A 1026 (2004) 41. [19] A.G. González, M.A. Herrador, Trends Anal. Chem. 26 (2007) 227. [20] V.J. Barwick, L.R. Ellison (2000) VAM Project 3.2.1. Development and Harmonisation of Measurement Uncertainty Principles. Part d: Protocol for Uncertainty Evaluation from Validation Data, report No: LGC/VAM/1998/088 [21] D. Prats, F. Ruiz, B. Vázquez, M. Rodriguez-Pastor, Water Res. 31 (1997) 1925. [22] T. Küchler, W. Schnaak, Chemosphere 35 (1997) 153. [23] L. Sarrazin, C. Diana, E. Wafo, P. Rebouillon, Inter. J. Environ. Studies 60 (2003) 229. [24] M. Petrovic, D. Barceló, Anal. Chem. 72 (2000) 4560. [25] C.F. Tabor, L.B. Barber, Environ. Sci. Technol. 30 (1996) 161.
Contents lists available at ScienceDirect
Microchemical Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / m i c r o c
Application of probe sonication extraction for the determination of linear alkylbenzene sulfonates from sewage sludge. Comparison with other extraction methods M. Villar, R. Fernández-Torres, M. Callejón ⁎, P. Villar, J.C. Jiménez Department of Analytical Chemistry, Faculty of Chemistry, University of Seville., c/ Profesor García González s/n, 41012, Seville, Spain
A R T I C L E
I N F O
Article history: Received 7 April 2008 Received in revised form 14 May 2008 Accepted 14 May 2008 Available online 21 May 2008 Keywords: Linear alkylbenzene sulfonates Sewage sludge Extraction Liquid chromatography
A B S T R A C T A new method based on probe sonication extraction (USP) prior to high performance liquid chromatography (HPLC) has been developed for the determination of linear alkylbenzene sulfonates (LAS) from sewage sludge. The optimized method was designed to be cost effective compared to existing extraction methods (ultrasonic assisted extraction, Soxhlet or pressurized liquid extraction) which may require large quantities of organic solvents, or costly instrumentation or equipment. The main factors affecting the extraction efficiency (extractant volume, ultrasounds power and extraction time) were optimized using compost sludge. The detection limit of total LAS in the sludge was 10 mg kg− 1. The extraction of C10–C13 homologues is carried out using an extraction time of 7 min with 10 mL of methanol. Liquid chromatography with fluorescence (FL) detector is used for determination of LAS homologues. A mobile phase acetonitrile-water containing 0.1 M NaClO4 (65:35) and isocratic elution was used. Compounds were eluted over 6 min at a flow rate of 1 mL/min. Polar interferences are eluted between 0 and 2 min and no purification of the samples is required prior to the final determination by high performance liquid chromatography (HPLC). The recoveries of LAS in spiked sewage sludge were between 84.0% and 97.0%, which reflect the efficiency of the method for extraction of these analytes from sewage sludge. Concentration levels found were between 11,858 mg kg− 1 for digested sludge and 2379 mg kg− 1 for compost sludge. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Linear alkylbenzene sulfonates (LAS) are the most important synthetic anionic surface-active agents widely used as the principal constituents of commercial detergents. The European consumption of LAS in detergents applications was about 350 kt in 2005. This represents more than 80% of the total European consumption of LAS, which was estimated to be about 430 kt in the year 2005 [1]. They are used as complex mixtures of C10–C13 homologues and of positional isomers where the benzenesulfonate is located at various alkyl carbon positions from the second to the center. After application, LAS are usually discharged through the sewage infrastructure to municipal wastewater treatment plants which are subjected to physical and biological treatment. In wastewater treatment plants, 10% or more of the LAS present in wastewater are eliminated by adsorption/ precipitation processes [2], together with suspended solids during the primary treatment, while the rest are biodegraded in the aerobic stage of the treatment process. Nevertheless, as a consequence of the
⁎ Corresponding author. Tel.: +34 954557169; fax: +34 954557168. E-mail address: [email protected] (M. Callejón). 0026-265X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.microc.2008.05.005
low biodegradability of LAS under anaerobic conditions, the sludge that is stabilized by anaerobic processes shows concentrations of LAS as high as 3000–30,000 mg kg− 1 of dry solid [3–5]. During recent decades a variety of procedures have been used to extract LAS in solid samples. Soxhlet and ultrasounds methods using methanol as extraction solvent are mainly employed [4–7]. Soxhlet and ultrasounds extraction methods are relatively costly, time consuming, labor intensive and require the used of large volume of organic solvents. Some attempts performed in order to reduce both the volume of organic solvents used and the time needed for the complete extraction have been based on supercritical fluid extraction [8] and pressurized liquid extraction with methanol [9,10], all of which require complex and expensive instrumentation. High performance liquid chromatography (HPLC) employing UV or fluorescence detection (FL) is the most frequently used method for the selective determination of LAS [11–15]. The objective of this study was to develop and validate a routine method for analysis of LAS in sewage sludge based on probe sonication extraction (USP) and HPLC-FL determination. The main factors affecting the extraction efficiency were studied and optimized, paying special attention in reducing the volume of organic solvent as well as the time needed for extraction. The method did not require clean-up
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
165
2. Materials and methods
extraction in ultrasound probe extraction. Centrifugation system was a Sigma k3-15. The microwave extraction system was a Microwave Ethos 900 apparatus (Milestone, Sorisole, Italy) with a programmable power and irradiation time. The apparatus is equipped with a carousel that is able to hold six extraction vessels.
2.1. Chemicals and reagents
2.3. Sewage sludge collection and preparation
HPLC grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany), just as analytical grade sodium perchlorate. Petroquímica Española (Cádiz, Spain) supplied the commercial LAS mixture (Petrelab P-550) with the following homologue distribution: C-10 (12.3%), C-11 (32.1%), C-12 (30.8%), C-13 (23.4%). Standard solutions of LAS were prepared in ultrapure water. Ultra-high-quality water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA).
Sewage sludge samples (digested and compost sludge) were collected from Seville (Spain) wastewater treatment plant. Digested sludge is obtained when primary sludge undergoes a process of digestion (aerobic or anaerobic) for its stabilization. 1000 mL of digested sludge were collected in different points to obtain a representative sample. Compost sludge is obtained by exhibition to the sun, with a process of natural fermentation that takes place helped by ventilation. 500 mL of compost sludge were collected in different points of the storage system and different depths to obtain a representative sample. Before analysis, sludge was dried in an oven at 40 °C until humidity b10% (digested sludge between 5 and 7 days and compost sludge between 2 and 3 days) and ground in an agate mortar. The sludge was sieved in order to obtain a fraction b1 mm. This fraction was stored at 4 °C until it was analyzed.
or preconcentration steps and no special equipment is needed. The proposed method has been satisfactorily applied to sewage sludge samples from Seville (Spain) wastewater treatment plant.
2.2. Instrumentation All measurements were made with an Agilent (Palo Alto, CA, USA) 1100 series liquid chromatograph equipped with a fluorescence detector (FL), an injector with a loop 20 μL, a quaternary pump, a vacuum degasser and a thermostated column compartment. The analytical cartridge column was a Zorbax (Agilent) XDB-C8, 150 mm × 4.6 mm ID, 5 μm particle size. A C-18 guard column was installed to protect it from contamination. An ultrasound bath Ultrasons-P (Selecta, Spain) was used in ultrasounds bath extraction. Sonoplus ultrasonic homogeneiser (Bandelin, Germany) fitted with a HF generator 2200 and a microtip of 2 mm diameter was used for LAS
2.4. Preparation of spiked samples Two portions of 20 g of compost sludge were spiked with commercial LAS at two different concentrations (2000, 4000 mg kg− 1 of total LAS). One sample was not spiked in order to evaluate the original presence of LAS in sludge. Spiked samples were prepared adding 40
Fig. 1. Optimization of probe sonication extraction (USP).
166
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
and 80 mL of standard solution of 1000 mg L− 1 of total LAS to 20 g of sludge. Samples were allowed to interact with the natural organic matter during one day under stirring and nitrogen to prevent aerobic degradation.
Table 2 Limits of detection (LOD), limits of quantification (LOQ), repeatability (%RSD) and recoveries (%R)
Method 1
2.5. Sample treatment Probe sonication extraction was performed on 0.5 g of dried sewage sludge. Sewage sludge was dried in an oven at 40 °C and ground in an agate mortar. The sludge was sieved in order to obtain a fraction b1 mm prior to the analysis. The main factors affecting the extraction efficiency were optimized using compost sludge from a Seville (Spain) wastewater treatment plant. Methanol was used as extraction solvent and the extracts were filtered through glass wool and analyzed by HPLC.
Method 2
Method 3
3. Results and discussion 3.1. Chromatographic conditions LAS were determined as previously described by Villar et al. [16], using as mobile phase acetonitrile-water containing 0.1 M NaClO4 (65:35), isocratic elution and the flow rate was 1 mL min− 1. The column was thermostated (25 °C) and fluorescence detector (λex =225 nm, λem = 295 nm) was used for the determination of LAS. Retention time for LAS C10, LAS C-11, LAS C-12 and LAS C-13 were 2.38, 2.82, 3.43 and 4.32 min, respectively. 3.2. Optimization of probe sonication extraction (USP) The optimization of the extraction procedure was performed with compost sludge from a Seville (Spain) wastewater treatment plant. The different factors that could influence in the process of extraction (extractant volume, extraction time and ultrasounds power) were studied and optimized. Three samples of 0.5 g of dried compost sludge were subjected to USP extraction with methanol. The first parameter optimized was the extractant volume. It was tested between 2 and 20 mL of methanol (extraction time fixed at 3 min and power fixed at 20% of the maximum power (2200 W) of the ultrasounds probe). Maximum extraction was achieved for the four LAS when 10 mL of methanol were used (Fig. 1). Secondly, the irradiation power was tested between 10 and 50% of the maximum power (extractant volume fixed at 10 mL and extraction time at 3 min). All LAS showed maximum extraction when 50% of power was used (Fig. 1). Finally, the extraction time was tested between 1 and 15 min (extractant volume 10 mL and ultrasonic power 50%). All LAS showed maximum extraction when 7 min of extraction was used (Fig. 1). Optimal conditions for extraction of linear alkylbenzene sulfonates of sewage sludge using probe sonication extraction were: 10 mL of methanol, 50% of irradiation power and 7 min of extraction time.
Method 4
Method Method Method Method
1: 2: 3: 4:
LAS C-10
LAS C-11
LAS C-12
LAS C-13
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 5.42
2.65 6.62 3.50
4.58 15.22 4.21
2.05 6.85 5.31
94 85
93 83
90 83
84 86
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 3.48
2.65 6.62 4.46
4.58 15.22 3.72
2.05 6.85 5.40
94 88
93 89
93 82
85 87
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.82 2.78 3.96
2.65 6.62 4.58
4.58 15.22 4.38
2.05 6.85 4.51
95 88
96 91
97 84
93 89
LOD (mg kg− 1) LOQ (mg kg− 1) % RSD (n = 6) % Recovery 2000 mg kg− 1 4000 mg kg− 1
0.33 1.11 2.10
1.06 3.53 4.70
1.83 6.09 4.02
0.82 2.74 5.42
94 90
102 95
100 85
98 89
Ultrasonic and centrifugation. Stirred and centrifugation. Ultrasonic probe extraction. Microwave-assisted extraction.
research on the linearity, matrix effect, precision, detection and quantification limits and recovery studies. Calibration plots were obtained by using standard solutions of the four homologues of LAS in water. Each level was analyzed by triplicate. Regression coefficients (r), slope standard deviations (RSDb), linearity (%) and linear dynamic ranges are shown in Table 1. Good linearity was observed and the regression coefficients were between 0.9966 for LAS C-11 and 0.9999 for LAS C-10 (Table 1). Detection limits (LOD) for each analyte were calculated as the mass of analyte which gives a signal 3σ above the mean blank signal (where σ is the standard deviation of the blank signal). The LODs obtained were between 0.02 mg L− 1 for LAS C-10 and 0.09 mg L− 1 for LAS C-12. The quantification limits (LOQ), expressed as the mass of analyte which gives a signal 10σ above the mean blank signal, ranged between 0.06 mg L− 1 for LAS C-10 and 0.30 mg L− 1 for LAS C-12. Taking into account the amounts of sample extracted and the volume of extractant used, the detection limits of LAS in sludge samples were between 0.82 mg kg− 1 for LAS C-10 and 4.58 mg kg− 1 for LAS C-12.
3.3. Analytical data and validation of the method A validation process was carried out to confirm the suitability of the method for the intended use. In order to do so, we performed a
Table 1 Regression coefficients (r), slope standard deviations (RSDb), linearity (%) and linear dynamic ranges LAS
r
RSDb (%)
% linearity
Linear dynamic ranges (mg L− 1)
LAS C-10 LAS C-11 LAS C-12 LAS C-13
0.9999 0.9966 0.9981 0.9989
0.07 0.38 0.29 0.24
99.81 99.88 99.12 99.34
0.06–86.10 0.21–321.00 0.42–215.60 0.22–163.80
Fig. 2. Chromatogram of compost sewage sludge (2500 mg kg− 1 approximately).
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
of the slopes was carried out according to González et al. [19]. We calculated R as:
Table 3 LAS (mg kg− 1) in digested and compost sludge (based on three replicates) LAS LAS LAS LAS LAS
C-10 C-11 C-12 C-13
167
Digested
Compost
907 ± 8 3142 ± 12 4207 ± 15 3602 ± 5
89 ± 3 527 ± 5 887 ± 8 876 ± 8
R¼
bSAM b
Where: bSAM is the slope obtained by standard addition b is the slope obtained by external calibration
Results are shown in Table 2 and as can be seen low concentrations of LAS can be detected in environmental sludge samples. To evaluate the repeatability and the intermediate precision of the proposed method, 0.5 g of dried compost sludge of approximately 100 mg kg− 1 of LAS C-10, 560 mg kg− 1 of LAS C-11, 940 mg kg− 1 of LAS C-12 and 900 mg kg− 1 of LAS C-13 were subjected to the entire analytical procedure and measured in one single day and two days per week during three weeks, respectively. All these values are similar to those obtained by other authors [17,18]. This fact confirms the reproducibility of the proposed method. Analytical data for each homologue are summarized in Table 2. Since sewage sludge is a complex matrix, it is very important to study the influence of it on the proposed method. A calibration was realized by standard addition and the slopes were compared with those obtained by an external calibrate. A statistical comparison of standard deviation
The absence of matrix effect corresponds to bSAM = b, or, in terms of recovery, R = 1. This must be checked for statistical significance: t¼
jR−1j uðRÞ
With the uncertainty given by: uðRÞ ¼
2 1=2 u ðbSAM Þ b2SAM u2 ðbÞ þ 2 4 b b
According to the LGC/VAM protocol [20], if the degrees of freedom associated with the uncertainty of consensus recovery are known, t is compared with the two-tailed tabulated value, ttab (ν,P) for the appropriate number of degrees of freedom at P% confidence. If t ≤ttab,
Fig. 3. Optimization of ultrasounds bath plus centrifugation extraction.
168
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
the consensus recovery is not significantly different from 1. Alternatively, instead of ttab, coverage factor k may be used for the comparison. Typical values are k = 2 or k = 3 for 95% or 99% confidence, respectively, so: – if jR−1j uðRÞ Vk, the recovery is not significantly different from 1; and, – if
jR−1j uðRÞ Nk,
the recovery is significantly different from 1 and the
analytical result must be corrected by R. For the four LAS studied the value of t was inferior to 2. Therefore, it could be concluded that the method is not affected by any matrix effect. As no standard reference material was available, spiked sludge obtained as previously described in Section 2.4 was used in order to study the accuracy of the method. The samples were analyzed, by triplicate, using the proposed methods. Recovery values were between 84% and 97% (Table 2). These values are in accordance with those reported previously by García et al. and by Prats et al. [11,21] who used Soxhlet extraction for 8 h (recoveries between 85 and 99%), and by Küchler et al. [22] who used Soxhlet extraction for 24 h (recoveries N86%) and by Sarrazin et al. [23] who carried out an ultrasonic extraction (recoveries N85%). Recoveries obtained by Villar et al. [16] using MAE were between 85% and 102%.
to the procedure described in Section 2.3. Fig. 2 shows a typical chromatogram of compost sludge (2400 mg kg− 1 approximately) and as it is shown LAS C-10 is the first compound eluted at a retention time of 2.38 min followed by LAS C-11, LAS C-12 and finally LAS C-13 with retention times of 2.82 min, 3.43 min and 4.31 min, respectively. Since polar interferences are eluted between 0 and 2 min, they do not interfere in the analysis. Concentration levels of LAS were between 11,858 mg kg− 1 for digested sludge and 2379 mg kg− 1 for compost sludge as can be seen in Table 3. 4. Optimization and comparison with other extraction methods The proposed method has been compared with the method developed by Villar et al. [16] and with other two extraction methods based on previously described methodologies (ultrasounds bath [24] and stirred [25]) but modified, optimized and validated in our laboratory. The main factors affecting the extraction procedures (volume of extractant, time of extraction and stirring time) have been modified and a new step of centrifugation have been also included in the process. 4.1. Optimization of ultrasounds bath plus centrifugation extraction
3.4. Application of the proposed method to sewage sludge from Seville wastewater treatment plant The optimized and validated method was applied to the determination of homologues of LAS in digested and compost sludge from a Seville wastewater treatment plant. Samples were prepared according
Extractant volume, ultrasounds bath time, centrifugation velocity and centrifugation time were studied and optimized. The four LAS achieved maximum extraction when 30 min of ultrasound bath, 10 mL of methanol, 20 min of centrifugation time and 6000 rpm were used (Fig. 3).
Fig. 4. Optimization of magnetic stirring plus centrifugation extraction.
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
169
4.2. Optimization of magnetic stirred plus centrifugation extraction
Table 4 Optimal extraction conditions for the methods
The different factors that can be influenced in the process of extraction (extractant volume, magnetic stirring time, centrifugation velocity and centrifugation time) were optimized. Optimal values were: 5 min of stirring time, 10 mL of extractant, 10 min of centrifugation time and 6000 rpm of velocity of centrifugation (Fig. 4).
Method/conditions
Ultrasonic and centrifugation
Stirred and centrifugation
Ultrasonic probe extraction
Microwaveassisted extraction
Extraction time (min) Extractant volume (mL)
50 10
15 10
7 10
10 5
4.3. Microwave-assisted extraction Microwave-assisted extraction optimization has been previously published [16]. Extractant volume, power of microwave irradiation and irradiation time were chosen to be 5 mL, 250 W and 10 min respectively (Fig. 5). 4.4. Comparison of the methods Results of repeatability, intermediate precision, detection and quantification limits and recovery studies are show in Table 2 for each method. Similar values of repeatability and intermediate precision have been obtained for the tested methods. However, both detection and quantification limits, present lower values when microwaveassisted extraction was used, the range of detection limit was between 0.33 for LAS C-10 and 1.83 mg kg− 1 for LAS C-12. Recoveries obtained in spiked sludge at two concentration levels are given in Table 2. It can be observed that in the four methods studied recoveries values are N82%. As we can see in Table 4, ultrasounds bath, magnetic stirred and USP extraction needed 10 mL of methanol while MAE only needed 5 mL. Probe sonication extraction is carried out in only 7 min versus
10, 15 and 50 min needed in MAE, magnetic stirred extraction and ultrasounds bath extraction, respectively. Despite the fact that MAE needed less quantity of solvent, this itself does not represent a great advantage with respect to the other methods because the volume of solvent required in the other cases is only 10 mL which, comparatively, does not represent a great waste of solvent. What it really means an advantage compared to the other methods is that probe sonication extraction shortens significantly the time of analysis with respect to magnetic stirred extraction and ultrasounds bath extraction and although this time of treatment seems to be similar between MAE and USP the former needs a long cooling time that lengthens too much time of analysis. Besides, the extraction procedure using USP is cost effective, simple and rapid compared to MAE. 5. Conclusions A simple, rapid and practical method has been developed for determination of LAS homologues in sewage sludge. No special equipment is required and extractions are not dependent on the length of the alkyl chain of LAS and no purification of the samples is required prior to the
Fig. 5. Optimization of microwave-assisted extraction (MAE).
170
M. Villar et al. / Microchemical Journal 90 (2008) 164–170
final determination by HPLC. USP is an attractive, affordable and effective alternative to existing extraction methods. Ultrasounds bath, magnetic stirred and USP extraction needed 10 mL of methanol while MAE only needed 5 mL, but the main advantage of the proposed method in comparison to these other methods is the low extraction time, USP extraction needed only 7 min for the extraction while 50 min,15 min and 10 min are required in ultrasounds bath, magnetic stirring and microwave-assisted extraction respectively (Table 4). Probe sonication extraction (USP) has proved to constitute reliable, reproducible and efficient method for the extraction of LAS in sewage sludge. Acknowledgements The authors would like to thank the Empresa Municipal de Abastecimiento y Saneamiento de Aguas de Sevilla (EMASESA) for supplying sludge samples, Petroquímica Española (PETRESA) and Programa de Becas de Formación de Profesorado Universitario (FPU, Ministerio de Educación y Ciencia, Spain). References [1] HERA (2007). Human and Environmental Risk Assessment of Linear Alkylbenzene Sulphonate. [2] W. Giger, A.C. Alder, P.H. Brunner, A. Marcomini, H. Siegrist, Tenside. Surfact. Det. 26 (1989) 95. [3] E. Matthijs, H. de Henau, Tenside. Surfact. Det. 24 (1987) 193.
[4] D. Prats, F. Ruiz, B. Vázquez, D. Zarzo, J.L. Berna, A. Moreno, Environ. Toxicol. Chem. 12 (1993) 1599. [5] E. Sanz, D. Prats, M. Rodríguez, A. Camacho, Waste Manage. 26 (2006) 1237. [6] K. Bester, N. Theoblad, H.Fr. Schröder, Chemosphere 45 (2001) 817. [7] S.O. Petersen, K. Henriksen, G.K. Mortensen, P.H. Krogh, K.K. Brandt, J. Sorensen, T. Madsen, J. Petersen, C. Gron, Soil Tillage Res. 72 (2003) 139. [8] A. Kreisselmeier, H.W. Dürbeck, J. Cromatogr. A 775 (1997) 187. [9] S. Álvarez-Muñoz, M. Sáez, P.A. Lara-Martín, A. Gómez-Parra, E. González-Mazo, J. Chromatogr. A 1052 (2004) 33. [10] S. González, M. Petrovic, D. Barceló, J. Chromatogr. A 1052 (2004) 111. [11] M.T. García, E. Campos, J. Sánchez-Leal, I. Ribosa, Water Res. 40 (2006) 2958. [12] F.J. Ruiz, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A 1030 (2004) 109. [13] L. Carlsen, M.B. Metzon, J. Kjelsmark, Sci. Total Environ. 290 (2002) 225. [14] P.A. Lara-Martín, M. Petrovic, A. Gómez-Parra, D. Barceló, E. González-Mazo, Environ. Pollution 144 (2006) 483. [15] M. Nimer, O. Ballesteros, A. Navalón, G. Crovetto, C. Verge, I. López, J.L. Berna, J.L. Vílchez, Anal. Bioanal. Chem. 387 (2007) 2175. [16] M. Villar, M. Callejón, J.C. Jiménez, E. Alonso, A. Guiráum, Anal. Chim. Acta 599 (2007) 92. [17] M.T. García, E. Campos, I. Ribosa, A. Latorre, J. Sánchez-Leal, Chemosphere 60 (2005) 1636. [18] S. Morales-Muñoz, J.L. Luque-García, M.D. Luque de Castro, J. Chromatogr. A 1026 (2004) 41. [19] A.G. González, M.A. Herrador, Trends Anal. Chem. 26 (2007) 227. [20] V.J. Barwick, L.R. Ellison (2000) VAM Project 3.2.1. Development and Harmonisation of Measurement Uncertainty Principles. Part d: Protocol for Uncertainty Evaluation from Validation Data, report No: LGC/VAM/1998/088 [21] D. Prats, F. Ruiz, B. Vázquez, M. Rodriguez-Pastor, Water Res. 31 (1997) 1925. [22] T. Küchler, W. Schnaak, Chemosphere 35 (1997) 153. [23] L. Sarrazin, C. Diana, E. Wafo, P. Rebouillon, Inter. J. Environ. Studies 60 (2003) 229. [24] M. Petrovic, D. Barceló, Anal. Chem. 72 (2000) 4560. [25] C.F. Tabor, L.B. Barber, Environ. Sci. Technol. 30 (1996) 161.