Journal of Chromatography A, 1336 (2014) 1–9

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Determination of benzotriazoles in water samples by concurrent derivatization–dispersive liquid–liquid microextraction followed by gas chromatography–mass spectrometry J. Casado a , R. Nescatelli b , I. Rodríguez a , M. Ramil a,∗ , F. Marini b , R. Cela a a Department of Analytical Chemistry, Nutrition and Food Sciences, IIAA-Institute for Food Analysis and Research, University of Santiago de Compostela, R/Constantino Candeira SN, 15782 Santiago de Compostela, Spain b Department of Chemistry, University of Rome “La Sapienza”, P.le Aldo Moro 5, Rome 00185, Italy

a r t i c l e

i n f o

Article history: Received 7 November 2013 Received in revised form 23 January 2014 Accepted 24 January 2014 Available online 31 January 2014 Keywords: Benzotriazoles Polar emerging contaminants Water analysis Dispersive liquid–liquid microextraction Acetylation Gas chromatography–mass spectrometry

a b s t r a c t Concurrent acetylation–dispersive liquid–liquid microextraction (DLLME) combined with gas chromatography–mass spectrometry (GC–MS) has been proposed for the sensitive determination of five polar benzotriazolic compounds (1H-benzotriazole, BTri; 4 and 5-methyl-1H-benzotriazole, 4TTri and 5-TTri; 5,6-dimethyl-1H-benzotriazole, XTri; and 5-chloro-1H-benzotriazole, 5-ClBTri) in water samples. Under optimized conditions, samples (10 mL volume) were combined with 1 mL of Na2 HPO4 (8%, w/v) and mixed with the ternary acetylation–microextraction mixture, consisting of 100 ␮L of acetic anhydride, 1.5 mL of acetonitrile and 60 ␮L of toluene. Thus, analytes were simultaneously acetylated and transferred to the dispersed droplets of toluene. The proposed methodology achieved limits of quantification (LOQs) between 0.007 ng mL−1 and 0.080 ng mL−1 , enrichment factors between 93 and 172 times, good reproducibility, with relative standard deviations lower than 10%, and linearity with determination coefficients above 0.9991 for all compounds in the range between LOQs and 20 ng mL−1 . Pseudo-external calibration, with fortified ultrapure water samples submitted to the acetylation–DLLME procedure, proved to be adequate for the accurate quantification of complex aqueous matrices such as surface or wastewater, providing recoveries comprised between 86% and 112%. BTri, 4-TTri and 5-TTri were measured in environmental samples up to a concentration of 1.9 ng mL−1 for BTri in raw wastewater. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Benzotriazole derivatives are categorized as high production volume chemicals, being complexing agents widely used as anticorrosives (e.g. in engine coolants, aircraft deicers and antifreeze liquids) and for silver protection in dish washing liquids [1,2]. Toxicological studies have demonstrated that they might be hazardous to plants [3,4], mutagenic in bacteria cell systems [4] and toxic to some microorganisms [5]. Moreover, 1H-benzotriazole (BTri) has been classified as a suspected human carcinogen by the Dutch Expert Committee on Occupational Standards [4]. In the environment, benzotriazoles are considered as emerging pollutants [1,2], with sewage treatment plants (STPs) representing one of the most important discharge sources of these compounds into the aquatic media [6–8]. Thus, they have been detected in different aquatic compartments, such as surface, ground or wastewater [1,2], sludge

∗ Corresponding author. Tel.: +34 881814466; fax: +34 881814468. E-mail address: [email protected] (M. Ramil). 0021-9673/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2014.01.068

[9,10] and sediments [10]. Also, benzotriazoles appear in indoor environments [11], and even in human urine [12]. Due to their polar character (log Kow values from 1.44 to 2.25), high water solubility and low volatility, liquid chromatography (LC), usually coupled to mass spectrometry (MS), has been the preferred technique for their sensitive determination in environmental samples during last years [2]. Most water samples analysis have been carried out using triple quadrupole (QqQ) LC–MS/MS instruments, achieving methodological LOQs in the low ng L−1 range [6,13–16]; furthermore, other mass analyzers, such as LTQ FT Orbitrap MS [17], HRMS [18] and QTOF MS [19], have also demonstrated their suitability for benzotriazole determination. Limited performance of gas chromatography (GC) methods for benzotriazole compounds has been overcome by the use of ionic liquid coated columns [20], derivatization processes, such as methylation [21,22] or acetylation [23], and the use of two-dimensional gas chromatography [22,24]. But for now, no simple analytical methodologies, based on the use of a routine laboratory affordable GC-single quadrupole MS instrument, have been developed, able to reach LOQs comparable to those provided by LC–MS/MS methods.

2

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

Regarding sample preparation, solid-phase extraction (SPE), using conventional hydrophilic–lipophilic balanced polymeric materials such as OASIS HLB [6,12–14,17] or Strata X [20,22], remains as the most popular concentration technique for benzotriazoles determination in water samples. SPE is also the preferred approach to carry out multiresidue water sampling campaigns in which these emerging pollutants are often included [25–28]. Despite microextraction techniques potential advantages, such as miniaturization, low solvent consumption and high selectivity [29,30], they have just been scarcely investigated for the extraction and preconcentration of benzotriazoles. As regards solid-phase methodologies, stir-bar sorptive extraction (SBSE) has been tested for the determination of BTri in ultrapure water. Whatever the tested coating, the extraction efficiency for BTri remained below 1%, for 50 mL of ultrapure water, after 4-h sampling [31]. Benzotriazoles have also been successfully concentrated from water samples using polyethersulfone micro-sorbents [19]. However, this sample preparation method required 6 h to achieve equilibrium conditions. Slow extraction kinetics of the above methods, which are characteristic of solid-phase microextraction techniques [29], can be overcome by some liquid–liquid microextraction methodologies, such as dispersive liquid–liquid microextraction (DLLME) [32]. Following the first report by Assadi and co-workers [32] in 2006, a high number of DLLME applications have been published. Some of them, as well as the most outstanding trends in DLLME, are revised in a recent review [33]. Relevant research lines in DLLME are (1) the search for alternative extractants, displaying a lower toxicity than halogenated, high density solvents [34], and (2) the combination of DLLME with derivatization reactions [35], which permit to expand the application field of DLLME to polar and/or thermal unstable compounds, in case of GC determination, and/or improving their detectability. To the best of our knowledge, the only application of DLLME to benzotriazoles analysis considered tri-n-butylphosphate as extractant, with concentrated species determined by LC with fluorescence detection and LC–MS/MS [36]. Therefore, main aims of this work are (1) the development of a simple, easy, highly efficient, environmental friendly and low cost sample preparation proposal, based on a concurrent derivatization–DLLME extraction, and (2) the combination with a relatively inexpensive determination technique, such as GC–MS, for the sensitive and selective determination of trace levels of benzotriazolic compounds in complex aqueous matrices. The performance of the developed method is compared to that corresponding to previously published approaches. Although acetylation and DLLME have been previously combined for the effective extraction of environmental pollutants [37] and food off-flavors [38], this approach has not been explored for benzotriazol compounds.

2. Experimental 2.1. Standards, solvents and material Standards of BTri (98%), 4-methyl-1H-benzotriazole (4-TTri; 100%), 5-methyl-1H-benzotriazole (5-TTri; 98%), 5,6-dimethyl-1Hbenzotriazole (XTri; 99%) and 1H-benzotriazole-(ring-d4 ) solution (BTri-d4 ), 10 ␮g mL−1 in acetone, used as internal surrogate (IS) through derivatization and liquid microextraction steps, were purchased from Sigma–Aldrich (Milwaukee, WI, USA). Two different standards of 5-chloro-1H-benzotriazole (5-ClBTri), with nominal purities of 98% and 99% were acquired from TCI (Zwijndrecht, Belgium) and Sigma–Aldrich, respectively. Stock solutions of the above compounds and diluted mixtures, used to spike water samples employed during optimization of extraction conditions, were prepared in acetonitrile and stored at 4 ◦ C for a maximum of 2

weeks. A standard of 1-acetyl-1H-benzotriazole (97%) was also provided by Sigma–Aldrich. Methanol and acetonitrile (HPLC-grade) were from Merck (Darmstadt, Germany). Acetone, toluene, chlorobenzene, carbon tetrachloride and 1,1,1-trichloroethane (trace analysis grade) were provided by Sigma–Aldrich. Ultrapure water was obtained from a Milli-Q system (Millipore, Billerica, MA, USA). Sodium acetate, sodium chloride, acetic acid, sodium bicarbonate (NaHCO3 ), disodium hydrogen phosphate (Na2 HPO4 ) and acetic anhydride were also obtained from Sigma–Aldrich. Cellulose acetate membrane filters (0.45 ␮m pore size) were purchased from Millipore (Bedford, MA, USA). Acetylated derivatives of target compounds, used during optimization of GC–MS determination conditions, were prepared as described elsewhere for BTri [23]. In brief, 10 mL of ultrapure water, containing a 0.8% (w/v) of Na2 HPO4 , were spiked with benzotriazole standards prepared in acetonitrile. Thereafter, 150 ␮L of acetic anhydride were poured into the same vessel, 5 mL of toluene were added and vials were manually shaken for 2 min. Derivatized species were concentrated in the upper organic phase (toluene), which was recovered using a Pasteur pipette before GC–MS analysis. In the particular case of BTri, the commercially available acetylated standard was also used. Concentrations of acetylated benzotriazoles in toluene, used in this study, varied from 50 to 1500 ng mL−1 .

2.2. Samples Grab samples of treated wastewater were obtained from different STPs located in Galicia (Northwest Spain); moreover, time-proportional 24-h composite samples were received from the inlet stream of a STP serving a 100,000 inhabitants city, in the same region. River water was obtained from two pristine creeks and the river receiving the discharge of the above STP.

2.3. Optimization of sample preparation conditions Optimization of acetylation and DLLME conditions was performed with spiked (0.050–20 ng mL−1 ) aliquots of ultrapure water, adjusted at different pHs, considering also different volumes of derivatization reagent (acetic anhydride), dispersant and extractant solvents. Extractions were performed in conical bottom glass tubes (nominal volume 12 mL), provided by Afora (Barcelona, Spain), which were manually shaken during derivatization and microextraction steps. Thereafter, tubes were centrifuged and the settled drop of extractant (case of chlorinated solvents) recovered after removal of the upper aqueous phase. When using toluene as extraction solvent, the floating organic phase, together with some water, was first transferred, using a micropipette 0.2 mL tip, to a conical insert (0.3 mL volume) to improve separation between aqueous and toluene phases. Thereafter, enough volume of the toluene layer could be recovered, with a 0.1 mL gas tight glass syringe, to be handled with the autosampler of the GC–MS instrument. Sample preparation conditions were optimized using uniand multi-variate strategies based on the use of experimental factorial designs. In the latter case, the Statgraphics software (Statpoint Technologies, Warrenton, VA, USA) was used for experimental design creation and analysis. Under optimal conditions, samples (10 mL) were first mixed with 1 mL of Na2 HPO4 (8%, w/v) in the DLLME tube. Acetylation and microextraction of target compounds were simultaneously carried out by addition of a ternary mixture, consisting of 100 ␮L of acetic anhydride, 1.5 mL of acetonitrile and 60 ␮L of toluene. Reaction and centrifugation (3000 rpm) times were set at 1 and 5 min,

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

3

Table 1 Abbreviations, relevant properties, chemical structure, retention times and GC–MS recorded ions for acetylated species. log Kow

pKa

1H-Benzotriazole

BTri

1.44

1.17b 8.37a

119.1

9.76

133 (161)

4-Methyl-1H-benzotriazole

4-TTri

1.71

1.65b 8.74a

133.2

10.99

104 (175)

5-methyl-1H-benzotriazole

5-TTri

1.71

1.65b 8.74a

133.2

11.31

104 (175)

5-Chloro-1H-benzotriazole

5-ClBTri

2.13

0.48b 7.39a

153.6

11.71, 11.83

195 (197)

5,6-Dimethyl-1H-benzotriazole

XTri

2.25

1.89b 8.92a

147.2

13.16

118 (189)

1H-Benzotriazole-(ring-d4) (I.S.)

BTri-d4

1.44

1.17b 8.37a

123.2

9.74

137 (165)

a

c

Molecular weight (Da)

Retention time (min)

c

Abbreviation

b

Structure

c

Compound

Quantification (qualifier) ions, m/z values

pKa of the protonated form. pKa of the neutral form. Data referred to acetylated derivatives.

respectively. After phase separation, as described above, around 27 ␮L of toluene could be recovered for GC–MS analysis.

2.4. Evaluation of the derivatization–DLLME method performance The efficiency of the DLLME process, under optimized conditions, was evaluated using enrichment factors (EFs). They were defined as the ratio between the concentration of each compound in toluene extracts (Ce ) and that added to the water sample (Cs ) [32,33]. The concentration of BTri in the former solution was determined against a calibration curve built with a commercial standard of this acetylated compound (calibration range: 50–1500 ng mL−1 , 6 levels). The concentrations for the rest of compounds were determined against calibration curves obtained for acetylated derivatives (calibration range 50–1500 ng mL−1 , referred to the toluene solution), prepared as defined in Section 2.1, assuming that the LLE extraction with toluene yields quantitative recoveries. The absolute extraction efficiency (EE,%) of the DLLME process was calculated as follows: EE (%) = EF × (Ve /Vs ) × 100, being Ve the recovered volume of toluene and Vs the volume of sample (10 mL) [32,33]. Potential variations of extraction efficiencies among ultrapure, surface and wastewater samples were evaluated using relative recoveries (%R) defined as follows: %R = [(As − Ab )/Ar ] × 100. As is the response (analyte/IS peak areas) measured in the extract from a spiked sample, Ab is the response of the extract from a non-spiked fraction of the same sample, and Ar is the response measured in the extract from an aliquot of ultrapure water spiked at the same concentration level. The calculated %R values remained around

100%, a fact that indicates small variations in the efficiency of the acetylation–DLLME process for different matrices. 2.5. Quantification of real samples Given that the efficiency of the developed procedure was similar for model ultrapure water and environmental samples, pseudoexternal calibration was considered as quantification technique. Thus, responses measured for aliquots of different water samples (river and wastewater), corrected with the IS peak area, were compared with calibration curves obtained for spiked ultrapure water aliquots in the range of values from 0.05 ng mL−1 (0.1 ng mL−1 in case of 5-ClBTri) to 20 ng mL−1 , containing the IS at the same constant level (1 ng mL−1 ) as real water samples. Obviously, this pseudo-external calibration approach is faster than performing standard additions over each particular sample for quantification purposes. 2.6. GC–MS conditions Acetylated compounds were determined by GC–MS. The gas chromatograph was an Agilent (Wilmington, DE, USA) 7890A model, equipped with a split/splitless injector and connected to a quadrupole MS spectrometer (Agilent MSD5975C), which was furnished with an electron impact (EI) ionization source. Compounds were separated with an Agilent HP-5MS capillary column (30 m × 0.25 mm i.d., df : 0.25 ␮m) using helium (99.999%) as carrier gas, at a constant flow of 1.2 mL min−1 . The GC oven was programmed as follows: 80 ◦ C (held for 2 min), rate at 10 ◦ C min−1 to

4

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

NaHCO3

Na2HPO4

Without catalyst

Acetate buffer

Normalized peak areas (%)

120% 100% 80% 60% 40% 20% 0%

BTri

4-TTri

5-TTri

5-ClBTri

XTri

Fig. 1. Influence of the catalyst on the efficiency of the acetylation reaction. Normalized peak areas to those obtained for Na2 HPO4 , n = 3 replicates. Error bars correspond to standard deviations.

280 ◦ C (held for 6 min). Injections (2 ␮L) were done in the splitless mode (injector temperature 260 ◦ C), with the solenoid valve switching to the split mode after 1 min. EI source, quadrupole and transfer line temperatures were maintained at 230 ◦ C, 150 ◦ C and 280 ◦ C, respectively. GC–MS chromatograms were recorded in the SIM mode, selecting two different ions per compound, Table 1. Identities of acetylated benzotriazoles were confirmed using a second GC–MS system, equipped with an hybrid quadrupole time-of-flight (QTOF), 7200 model from Agilent, mass analyzer. Chromatographic conditions, EI source parameters and transfer line temperatures were set to same values as those used in the single quadrupole GC–MS system. Moreover, an equivalent capillary column was installed in the GC–QTOF–MS system. Accurate MS spectra were recorded in the m/z range from 50 to 500 units with the spectrometer operated in the 2 GHz mode (full-width halfmaximum mass resolution 5000 at m/z 131). 3. Results and discussion 3.1. Preliminary experiments Pervova et al. [23] reported, for the first time, the acetylation of BTri with the aim of improving the performance of its GC–MS determination. Thereafter, the same procedure was applied to 4TTri and 5-TTri to determine their concentrations in dish-washing detergents [39]. In both cases, acetylation was performed in aqueous media, with acetic anhydride in presence of a basic catalyst. Acetylated derivatives were further extracted, by conventional liquid–liquid extraction (LLE), with 5 mL of toluene. This derivatization strategy was extrapolated to the rest of compounds (XTri and 5-ClBTri) involved in this study, introducing some changes regarding the type of base and the volumes of acetic anhydride and toluene. Whatever the tested derivatization parameters, in agreement with previous results obtained for BTri and tolyltriazoles [23,39], all compounds rendered a single, well-defined peak corresponding to the acetylated derivative, whose identity was verified on the basis of low and high resolution MS scan spectra and, in case of BTri, by injection of a commercially available acetylated standard. However, in the case of 5-ClBTri (nominal purity above 98%), two peaks with the same MS spectra and similar intensities were observed. The extracted ion chromatogram (extraction window 50 ppm) and accurate MS spectra, acquired with the GC–QTOF–MS system, after acetylation of 5-ClBTri are provided as supplementary information, Fig. S1. Injection of nonderivatized standard solutions of this compound, obtained from

two different suppliers, led to a single peak, figure not shown. Therefore, at difference to the behavior observed for 4-TTri and 5TTri, acetylation of 5-ClBTri renders two isomeric derivatives, likely resulting from acetylation at nitrogen atoms in positions 1 and 3 of the triazolic ring. This result is in agreement with the trend of some benzotriazoles, substituted in the benzenic ring, to form tautomeric structures as it was already proposed by Benson et al. [40] using spectroscopy data. Quantification and identification ions used during this work for acetylated derivatives of target compounds, and the IS, together with their retention times are summarized in Table 1. In case of 5-ClBTri, the sum of peak areas for both acetylated forms was used as response variable.

3.2. Optimization of sample preparation 3.2.1. Derivatization conditions and DLLME setup Performance of acetylation reactions in aqueous solution can be affected by the type of catalyst and the pH of the solution. Fractions (10 mL) of a spiked (3 ng mL−1 ) ultrapure water sample were mixed with 1 mL of two different bases (NaHCO3 , pH 8; Na2 HPO4 , pH 9; both 5%, w:v). Additional experiments were also performed using ultrapure water (pH 6), without any catalyst, and samples adjusted at pH 5 with 1 mL of a sodium acetate-acetic acid (1 M) solution. Then, 0.150 mL of acetic anhydride were added and the mixture was shaken for 2 min. A binary extraction mixture, consisting of 1 mL of acetone and 0.1 mL of chlorobenzene, was used for DLLME in all cases. Fig. 1 shows the normalized signals (peak areas for each compound under tested conditions divided by those measured in presence of Na2 HPO4 and multiplied by 100) obtained for acetylated compounds under above conditions. The highest responses were achieved using Na2 HPO4 , which was selected as catalyst of the acetylation reaction in further experiments. The effects of acetic anhydride volume (50–150 ␮L, factor 1), derivatization time (2–6 min, factor 2) and Na2 HPO4 concentration (2–8%, factor 3) were simultaneously investigated using a Box–Behnken experimental factorial design, with 3 central points and a total of 15 experiments. DLLME extractions were performed under the conditions reported in the previous paragraph, with ultrapure water samples spiked again at 3 ng mL−1 . Responses for each compound in these experiments (see supplementary information, Table S1) were analyzed by the Statgraphics software to obtain the main effects and the two-factor interactions corresponding to variables involved in the design. Table 2 compiles the standardized

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

CCl4

Toluene

1,1,1-trichloroethane

5

Chlorobenzene

1,2E+06

Corrected responses

1,0E+06

8,0E+05

6,0E+05

4,0E+05

2,0E+05

0,0E+00

BTri

4-TTri

5-TTri

5-ClBTri

XTri

Fig. 2. Corrected responses (peak areas multiplied by the volume of recovered organic extract) as function of the type of extractant solvent, n = 3 replicates. Error bars correspond to standard deviations.

main effects. The sign of main effects, positive or negative, corresponds to an improvement or to a decrease in the acetylation step efficiency, respectively; whereas, the absolute values are correlated to the variation in the response of a given analyte when the associated factor moves from the low to the high level, within the domain of the design. The statistical significance boundary was established at the 95% confidence level. The Na2 HPO4 concentration (factor 3) was the most relevant variable, with a positive and statistically significant influence on the acetylation process for 4 of the 5 compounds. For XTri presented a positive influence although it did not reach the statistical significance level. Factor 2 (derivatization time) followed an opposite trend, showing a negative effect in the yield of the derivatization, being just statistically significant for BTri. Finally, the acetic anhydride volume (factor 1), despite exerting a positive influence on the process, remained non-significant. Two-factor interactions (data not shown) played negligible effects in the acetylation process. Based on above results, the concentration of Na2 HPO4 was set at the highest level (8%) and the volume of acetic anhydride fixed in the intermediate value (100 ␮L). The negative, although in most cases non-significant, effect of the derivatization time in the responses of acetylated species suggests that (1) acetylation of benzotriazoles is a fast process and that (2) derivatives might be slowly hydrolyzed to the free forms in contact with aqueous sample. Taking into account these considerations, the possibility of combining acetylation and DLLME processes in the same step, as reported in case of chlorophenol compounds [37], was further evaluated. To this end, we compared the responses obtained under above derivatization conditions, considering an acetylation time of 2 min, followed by DLLME extraction (two-step approach) and adding the acetic anhydride (100 ␮L) to the binary mixture of acetone (1 mL) and chlorobenzene (100 ␮L) (single-step procedure). In both cases, manual shaking and centrifugation (3000 rpm) times were

2 and 5 min, respectively. The obtained responses are shown in the supplementary information section (Fig. S2). No significant differences between the results provided by the two methodologies are observed for 5-ClBTri and XTri. For the other 3 benzotriazoles, responses for the single step approach represented between 90 and 95% of those attained in two steps. On the view of these results, in order to save time and to reduce sample manipulation, in further experiments analytes acetylation and concentration were simultaneously performed. 3.2.2. DLLME conditions 3.2.2.1. Extractant and dispersant solvents. Selection of a suitable extraction solvent is one of the most important issues during method development in DLLME. Three solvents with higher density than water, commonly used in DLLME [33] (chlorobenzene, carbon tetrachloride and 1,1,1-trichloroethane), and toluene, as a lighter than water alternative, were compared on the basis of their affinity for acetylated benzotriazoles. Other non-chlorinated solvents, previously employed in DLLME methods, such as 1-undecanol and linear alkanes (e.g. C16) [33] were also considered; however, they were not included in the comparative study since they overlapped the chromatographic peak of acetylated BTri. In all cases, the volume of extractant was 100 ␮L. Fig. 2 shows the responses obtained for each compound. Depicted data correspond to peak areas multiplied by the volume of the organic extract collected for each solvent: 0.092 mL for chlorobenzene, 0.075 mL for carbon tetrachloride, 0.072 mL for 1,1,1-trichloroethane and 0.082 mL for toluene. Thus, plotted responses are only affected by the affinity of each solvent by acetylated compounds and not by variations in the volume of the extract, which usually depend on the lower or higher water solubility of the extractant solvents. Carbon tetrachloride and 1,1,1-trichloroethane provided the lowest responses and the highest variability for all

Table 2 Standardized main effect values for factors involved in the Box–Behnken experimental design. Compound

Factor 1 Acetic anhydride volume (50, 100 and 150 ␮L)

Factor 2 Reaction time (2, 4 and 6 min)

Factor 3 Na2 HPO4 concentration (2, 5 and 8%)

BTri 4-TTri 5-TTri 5-ClBTri XTri

0.87 0.88 1.75 0.72 2.01

−3.05a −2.11 −1.24 −1.66 −0.81

6.28a 5.81a 3.02a 3.59a 1.42

a

Significant effects at the 95% confidence level.

6

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

Fig. 3. Influence of toluene and acetonitrile volumes on the performance of the acetylation–DLLME process, n = 3 replicates.

species, whereas practically similar responses were measured for toluene and chlorobenzene. Likely, ␲–␲ interactions established between acetylated benzotriazoles and both aromatic solvents are responsible for their higher extraction efficiencies versus chlorinated alkanes. Despite the separation of the floating toluene extract was more complex than direct collection of the settled phase of chlorobenzene, the former solvent was preferred as extractant because of its lower toxicity. As reported in the experimental section, firstly, the upper phase of the extraction tube was transferred to a narrow (i.d. 3 mm) conical insert (0.3 mL volume), where a neat, well defined interface between toluene and the aqueous phase was obtained. This procedure was preferred to the employment of narrow neck, home-made, vessels reported in previous applications of DLLME with low density extractants [33], since it minimized the risks of cross-contamination problems. The volume of the toluene layer was measured with a gas tight glass syringe (0.1 mL volume) and then, transferred to another insert to be handled with the autosampler of the GC–MS system.

The type of dispersant (methanol, acetone and acetonitrile) exerted a minor effect in the responses of derivatized compounds (data not shown); however, acetone led to a peak with a retention time close to that of 5-TTri and same nominal m/z values and methanol showed the highest variability. Therefore, acetonitrile was selected as dispersant. 3.2.2.2. Extractant and dispersant volumes. Fig. 3 compares the peak areas obtained combining two different volumes of toluene (60 and 120 ␮L) with four of acetonitrile (0.5, 1.0, 1.5 and 2 mL). Although above experiments involve a variation in the volume of recovered toluene phase, the measured peak areas are directly proportional to the achieved EFs; thus, they are correlated with the LOQs of the method. In all cases, 100 ␮L of acetic anhydride was incorporated in the extraction mixture. With the only exception of the lowest dispersant volume, higher responses were achieved using 60 ␮L of toluene than with 120 ␮L. For the former extractant volume, the increase in the responses of the analytes with the volume of

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

7

Table 3 Linearity of the concurrent acetylation–DLLME method, enrichment factors (EFs = extract concentration/sample concentration) with their standard deviations, average extraction efficiencies (EEs,%), inter-day precision and limits of quantification (LOQs) of the method. Compound

BTri 4-TTri 5-TTri 5-ClBTri XTri a b c

EFs ± SD

Linearity evaluation (Range: 0.050–20 ng mL−1 , n = 7 levelsa )

b

EEs (%)

c

Slope (SD)

Intercept (SD)

R2

0.833 (0.005) 1.523 (0.016) 2.142 (0.015) 1.114 (0.010) 2.251 (0.021)

−0.028 (0.046) −0.01 (0.13) −0.11 (0.13) −0.16 (0.09) 0.25 (0.18)

0.9997 0.9995 0.9991 0.9996 0.9993

93 ± 5 134 ± 7 134 ± 7 165 ± 9 172 ± 9

LOQs (ng mL−1 )

Reproducibility (RSDs, %) (n = 9 replicates, 3 replicates/day)

25 36 36 45 46

0.2 ng mL−1

c

2 4 6 8 7

2 ng mL−1

2 5 4 5 8

0.045 0.007 0.009 0.080 0.013

For 5-ClBTri, linearity was investigated between 0.1 and 20 ng mL−1 . Extraction efficiencies calculated taking 0.027 mL as the average volume of the toluene extract. Added concentration.

Table 4 Relative recoveries (%), with their standard deviations (%), for samples spiked at different concentrations levels using pseudo-external calibration as quantification technique, n = 4 replicates. Compound

Tap water a

BTri 4-TTri 5-TTri 5-ClBTri XTri a

0.2 ng mL−1

101 99 112 92 112

± ± ± ± ±

6 1 11 1 6

River water a

0.5 ng mL−1

103 107 106 98 97

± ± ± ± ±

14 16 15 10 2

a

0.2 ng mL−1

101 93 109 99 96

± ± ± ± ±

10 2 13 9 6

Treated wastewater a

0.5 ng mL−1

98 101 101 108 108

± ± ± ± ±

a

6 4 2 7 4

0.5 ng mL−1

97 101 103 94 99

± ± ± ± ±

3 5 5 9 10

a

Raw wastewater

2 ng mL−1

98 98 86 101 105

± ± ± ± ±

2 10 5 9 10

a

10 ng mL−1

108 109 109 106 108

± ± ± ± ±

2 4 2 3 3

a

0.5 ng mL−1

101 96 111 104 104

± ± ± ± ±

1 5 4 7 10

a

2 ng mL−1

101 98 97 100 96

± ± ± ± ±

5 2 6 8 5

a

10 ng mL−1

111 109 109 106 108

± ± ± ± ±

2 4 2 4 3

Added concentration.

acetonitrile can be explained since a more efficient dispersion of toluene droplets in the aqueous sample is achieved. At 2 mL of acetonitrile, the increased solubility of acetylated analytes in the aqueous phase led to a small reduction in the responses of target compounds. Thus, 60 ␮L and 1.5 mL were adopted as toluene and acetonitrile optimal volumes. Under these conditions, 25–30 ␮L of toluene could be recovered at the end of the phase separation process.

3.2.2.3. Other variables. The influence of the ionic strength on the efficiency of the DLLME was evaluated comparing the responses obtained without and with addition of 1 g of NaCl to water samples. No significant variations were noticed in the responses measured for acetylated compounds; thus, no salt was used in further extractions. The extraction time, after addition of the ternary acetylation–extraction mixture, was varied between 1 and 5 min, whereas centrifugation (3000 rpm) times of 5, 10 and 15 min were tested. None of these factors modified the performance of the extraction; thus, extraction and centrifugation steps were limited to 1 and 5 min, respectively. The fast kinetics of acetylation and DLLME extraction is in agreement with results obtained in previous applications of this strategy to chlorophenol compounds [37].

3.3. Performance of the method The linearity of the proposed methodology was investigated with ultrapure water aliquots fortified with increasing concentrations of target benzotriazoles at seven different levels (0.05, 0.1, 0.5, 1, 5, 10 and 20 ng mL−1 ). In the particular case of 5-ClBTri, the lowest calibration level was 0.1 ng mL−1 . The IS was maintained at 1 ng mL−1 . The corrected responses (peak area/IS peak area) for each compound were plotted against their concentrations in the water samples and fitted to a linear model. The slope, intercept and determination coefficients (R2 ) values for the obtained graphs are compiled in Table 3. R2 values varied from 0.9991 up to 0.9997, Table 3. Regarding reproducibility, nine extractions were carried out in three different days (3 extractions per day) with samples spiked at two concentration levels, 0.2 ng mL−1 and 2 ng mL−1 . The relative standard deviation (RSDs, %) of corrected responses (peak area/IS peak area) remained between 2 and 10%. Efficiency of the proposed method was first evaluated with EFs, calculated as defined in the experimental section, for a sample spiked at the 10 ng mL−1 level. Analytes were concentrated between 93 (BTri) and 172 times (XTri), Table 3. Considering 27 ␮L as the average value of the recovered toluene phase, the absolute EEs of the procedure varied between 24% and 46%, with increased yields for the less polar compounds. These EEs are similar to those attained for

Table 5 Concentrations (ng mL−1 ) of BTri and tolyltriazoles in 24-h composite raw wastewater, and masses (g day−1 ) entering an urban STP during a seven-day sampling campaign, n = 3 replicates. Day

Concentration (ng mL−1 ) ± SD

Ratio 5-/4-TTri

BTri

4-TTri

5-TTri

1 2 3 4 5 6 7

1.94 ± 0.08 1.31 ± 0.02 1.35 ± 0.02 1.43 ± 0.04 0.62 ± 0.01 0.46 ± 0.02 0.66 ± 0.02

0.47 ± 0.02 0.32 ± 0.01 0.24 ± 0.01 0.35 ± 0.01 0.32 ± 0.03 0.20 ± 0.01 0.17 ± 0.01

0.56 ± 0.04 0.37 ± 0.01 0.30 ± 0.01 0.42 ± 0.01 0.27 ± 0.01 0.21 ± 0.01 0.18 ± 0.01

1.2 1.2 1.3 1.2 0.8 1.1 1.1

Average

1.11

0.30

0.33

1.1

Water volume (m3 day−1 )

Mass (g day−1 ) BTri

4-TTri

5-TTri

58,410 58,909 62,813 61,505 58,024 66,050 70,394

113 77 85 88 36 30 46

27 19 15 22 19 13 12

33 22 19 26 16 14 13

62,301

69

19

21

8

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

other polar compounds, such as mono-chlorophenols (EEs from 29 to 38%), using chlorinated extractants [37], and much higher than the EE (around 1%) calculated for BTri using SBSE as microextraction technique [31]. The limits of quantification (LOQs) of the method were calculated as the concentration of each compound providing a response 10 times higher than the baseline noise at the retention time of each compound in procedural blanks. The attained LOQs varied between 0.007 ng mL−1 for 4-TTri and 0.08 ng mL−1 for 5-ClBTri, Table 3. Obviously, obtention of two acetylated derivatives for 5ClBTri led to higher LOQs for this compound than for the rest of benzotriazoles involved in the study. In comparison with previous microextraction applications, the above LOQs are significantly lower than those obtained by DLLME, using tri-n-butylphosphate as extractant, and LC determination (0.1–7.3 ng mL−1 ) [36] and in the same order than those reported using polyethersulfone solidphase microextraction and LC-QTOF MS (0.005–0.1 ng mL−1 ) [19], with the advantage of employing a much faster sample preparation approach. LOQs summarized in Table 3 are also equivalent to those obtained by SPE combined with LC–MS/MS [6,13–16,26–28], LC–LTQ FT Orbitrap MS [17] and GC × GC–TOF–MS [22,24] requiring a less sophisticated and much cheaper instrumentation. Potential changes in the performance of the sample preparation procedure among water samples with different complexities were investigated comparing the responses obtained for ultrapure and different water samples [37], each one spiked at different concentration levels (0.2 and 0.5 ng mL−1 for the cleaner matrices, and 0.5, 2 and 10 ng mL−1 in case of wastewater samples). Obviously, non-spiked aliquots of environmental water samples were also prepared. The relative recoveries values, calculated as described in the experimental section, varied between 86 ± 5% and 112 ± 11%, Table 4. Therefore, after IS correction, comparison of responses measured for environmental water samples with those attained for spiked aliquots of ultrapure water can be used as quantification approach, providing accurate results without requiring the use of the time-consuming standard addition methodology. To further confirm the accuracy of pseudo-external calibration as quantification technique, the slopes of calibration curves corresponding to spiked aliquots of ultrapure water were compared to those obtained for spiked aliquots of river and a 24 h composite municipal raw wastewater sample. The latter matrix presented a pH of 7.4 and chemical and biochemical oxygen demands of 288 and 150 mg O2 L−1 , which are typical values for this type of sample. Added levels were those provided at the beginning of this section. The ratios between the slopes of the plots for each analyte in environmental samples (river and raw wastewater) and those corresponding to ultrapure water aliquots varied between 97 and 110%. Again, these data point out to the fact that the performance of the derivatization–DLLME method is scarcely affected by the nature of the sample. Data for the above graphs (slopes, intercepts and R2 values) are provided as supplementary information, Table S2. Fig. 4 overlays the GC–MS chromatograms for a procedural blank, a non-spiked treated wastewater and a 2 ng mL−1 fortified ultrapure water sample. As observed, a baseline separation of isomeric tolyltriazoles was achieved. Quite often, 4-TTri and 5TTri are just partially resolved in LC-based methods [6,19], which impairs their individual quantification in environmental water samples.

3.4. Real samples analysis Table 5 reflects BTri, 4-TTri and 5-TTri levels in 24-h composite raw wastewater samples obtained, during a week, from the same STP serving a 100,000 inhabitants population. The rest of compounds remained under their LOQs; however, the two peaks

Abundance 40000

Ion 133.00

BTri

30000

20000

10000

9.20

9.40

9.60 Time

9.80

10.00

10.20

Abundance Ion 104.00 60000

5-TTri

4-TTri

50000 40000 30000 20000 10000 10.40

10.60

10.80

11.00

11.20 Time

11.40

Abundance

11.60

11.80

Ion 118.00

50000

XTri 40000 30000 20000 10000 12.00

12.40

12.80

13.20 Time

13.60

14.00

Ion 195.00

Abundance

5-ClBTri

14000 12000 10000 8000 6000 4000 2000 0

11.20 11.40 11.60 11.80 12.00 12.20 12.40 12.60 12.80 Time

Fig. 4. GC–MS chromatograms corresponding to a procedural blank (solid line), a raw wastewater sample (code 1, Table 5; dashed line), and ultrapure water fortified at 2 ng mL−1 (dotted line).

resulting from acylation of 5-ClBTri were detected in some samples. The average raw wastewater concentration of BTri (1.11 ng mL−1 ) was significantly lower than that found in German STPs influents (12 ng mL−1 ) [6,7] and other Spanish locations (7.3 ng mL−1 ) [24]. Average individual concentrations of tolyltriazoles represented around 25% of that corresponding to BTri. The ratios of their concentrations (5-TTri/4-TTri) varied from 0.8 to 1.3, with an average value of 1.1, which is in concordance with previous studies. While Weiss et al. [6] reported a 5-TTri/4-TTri ratio of 1.06, Casado et al.

J. Casado et al. / J. Chromatogr. A 1336 (2014) 1–9

9

Table 6 Concentrations (ng mL−1 ) in grab samples of river and treated wastewater, n = 3 replicates. Code

Type

Concentration (ng mL−1 ) ± SD BTri

1 2 3 4 5 6 7 8 9

River River River Sewage Sewage Sewage Sewage Sewage Sewage

0.025 0.051 0.144 0.64 0.27 0.19 0.68 0.41 0.15

Ratio 5-/4-TTri 4-TTri

± ± ± ± ± ± ± ± ±

0.003 0.003 0.005 0.01 0.01 0.01 0.02 0.01 0.01

[19] found values between 0.84 and 1.04. Taking into account the daily processed water volume (c.a. 62,000 m3 ), the global mass discharge of the above corrosion inhibitors in the plant was estimated. The average daily input of BTri was 69 g, followed by 20 g of 4- and 5-TTri. Thus, the STP receives a total of 0.11 kg day−1 of benzotriazoles, which is in a relatively low amount when compared with 9.72 kg day−1 recently reported for a STP processing a 12-times higher volume of wastewater [16]. Table 6 compiles the concentrations of BTri, 4-TTri and 5-TTri (rest of compounds remained undetected) in grab samples of river water (codes 1–3) and the outlet streams (codes 4–9) of different STPs. River water samples codes 1 and 2 were collected from relatively pristine creeks, whereas sample number 3 was taken 5-km downstream the discharge of a STP. As regards treated wastewater samples, BTri usually remained at higher levels than tolyltriazoles; however, differences between their concentrations were lower than those found for raw wastewater samples compiled in Table 5. Finally, 5-TTri/4-TTri ratios in treated wastewater again remained around the unit (Table 6), except for sample code 9. This sample corresponds to the only STP applying UV disinfection after the secondary (activated sludge) treatment tank. 4. Conclusions

n.d. 0.016 0.102 0.37 0.16 0.15 0.26 0.21 0.19

Acknowledgments

[1] [2] [3] [4]

[5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

[19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31]

This study has been supported by the Spanish Government (project CTQ2012-33080) and E.U. FEDER funds. We are also in debt with Aquagest for supplying some wastewater samples used in this study.

[32]

Appendix A. Supplementary data

[37]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma. 2014.01.068.

± ± ± ± ± ± ± ±

0.001 0.003 0.01 0.01 0.01 0.02 0.02 0.01

n.d. 0.009 0.102 0.39 0.15 0.15 0.25 0.20 0.090

± ± ± ± ± ± ± ±

0.002 0.005 0.01 0.01 0.01 0.02 0.03 0.004

– 0.6 1.0 1.1 0.9 1.0 1.0 1.0 0.5

References

[18]

A simple, rapid, and low cost methodology has been developed for the determination of several benzotriazolic derivatives in different aqueous matrices. The protocol requires a very low volume of sample and just a few microliters of extractant organic solvent for the microextraction. It enables the concurrent acetylation and microextraction processes with sample preparation requiring just 10 min, rendering a single derivative, except in case of 5-ClBTri. GC–MS, a relative accessible instrumentation, reaches LOQs comparable to those reported using more sophisticated systems such as LC–MS/MS or GC × GC–TOF–MS. In summary, the described procedure constitutes an appealing alternative to monitor the levels and the behavior of several benzotriazoles during wastewater treatments and also to investigate their fate in the aquatic environment.

5-TTri

[33] [34] [35] [36]

[38] [39] [40]

S.D. Richardson, T.A. Ternes, Anal. Chem. 83 (2011) 4614. S.D. Richardson, Anal. Chem. 84 (2012) 747. S. Castro, L.C. Davis, L.E. Erickson, Int. J. Phytorem. 5 (2003) 245. Health Council of the Netherlands: Dutch Expert Committee on Occupational Standards (DECOS), 1,2,3-Benzotriazole; publication no. 2000/14OSH, Health Council of the Netherlands, The Hague, 2000. J.S. Cornell, D.A. Pillard, M.T. Hernández, Environ. Toxicol. Chem. 19 (2000) 1465. S. Weiss, T. Reemtsma, Anal. Chem. 77 (2005) 7415. S. Weiss, J. Jakobs, T. Reemtsma, Environ. Sci. Technol. 40 (2006) 7193. T. Reemtsma, S. Weiss, J. Mueller, M. Petrovic, S. González, D. Barcelo, F. Ventura, T.P. Knepper, Environ. Sci. Technol. 40 (2006) 5451. Y.S. Liu, G.G. Ying, A. Shareef, R.S. Kookana, J. Chromatogr. A 1218 (2011) 5328. Z. Zhang, N. Ren, Y.F. Li, T. Kunisue, D. Gao, K. Kannan, Environ. Sci. Technol. 45 (2011) 3909. L. Wang, A.G. Asimakopoulos, H.B. Moon, H. Nakata, K. Kannan, Environ. Sci. Technol. 47 (2013) 4752. A.G. Asimakopoulos, A.A. Bletsou, Q. Wu, N.S. Thomaidis, K. Kannan, Anal. Chem. 85 (2013) 441. W. Giger, C. Schaffner, H.P.E. Kohler, Environ. Sci. Technol. 40 (2006) 7186. D. Voutsa, P. Hartmann, C. Schaffner, W. Giger, Environ. Sci. Pollut. Res. 13 (2006) 333. I. Carpinteiro, B. Abuin, M. Ramil, I. Rodriguez, R. Cela R, Anal. Bioanal. Chem. 402 (2012) 2471. A.G. Asimakopoulos, A. Ajibola, K. Kannan, N.S. Thomaidis, Sci. Total Environ. 452–453 (2013) 163. J.A. van Leerdam, A.C. Hogenboom, M.M.E. van der Kooi, P. de Voogt, Int. J. Mass Spectrom. 282 (2009) 99. F. Wode, C. Reilich, P. van Baar, V. Dünnbier, M. Jekel, T. Reemtsma, J. Chromatogr. A 1270 (2012) 118. J. Casado, I. Rodríguez, M. Ramil, R. Cela, J. Chromatogr. A 1299 (2013) 40. C. Dominguez, C. Reyes-Contreras, J.M. Bayona, J. Chromatogr. A 1230 (2012) 117. S.R. Corsi, S.W. Geis, J.E. Loyo-Rosales, C.P. Rice, R.J. Sheesley, G.G. Failey, D.A. Cancilla, Environ. Sci. Technol. 40 (2006) 3195. V. Matamoros, E. Jover, J.M. Bayona, Anal. Chem. 82 (2010) 699. M.G. Pervova, V.E. Kirichenko, V.I. Saloutin, J. Anal. Chem. 65 (2010) 276. E. Jover, V. Matamoros, J.M. Bayona, J. Chromatogr. A 1216 (2009) 4013. R. Loos, B.M. Gawlik, G. Locoro, E. Rimaviciute, S. Contini, G. Bidoglio, Environ. Pollut. 157 (2009) 561. R. Loos, G. Locoro, S. Comero, S. Contini, D. Schwesig, F. Werres, P. Balsaa, O. Gans, S. Weiss, L. Blaha, M. Bolchi, B.M. Gawlik, Water Res. 44 (2010) 4115. K. Nödler, T. Licha, K. Bester, M. Sauter, J. Chromatogr. A 1217 (2010) 6511. S. Huntscha, H.P. Singer, C.S. McArdell, C.E. Frank, J. Hollender, J. Chromatogr. A 1268 (2012) 74. J.M. Kokosa, A. Przyjazny, M.A. Jeannot, Solvent Microextraction. Theory and Practice, John Wiley & Sons, Inc, Hoboken, New Jersey, 2009. E.A. Souza Silva, S. Risticevic, J. Pawliszyn, Trends Anal. Chem. 43 (2013) 24. N. Gilart, N. Miralles, R.M. Marcé, F. Borrull, N. Fontenals, Anal. Chim. Acta 774 (2013) 51. M. Rezaee, Y. Assadi, M.R.M. Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, J. Chromatogr. A 1116 (2006) 1. H. Yan, H. Wang, J. Chromatogr. A 1295 (2013) 1. M.I. Leong, S.D. Huang, J. Chromatogr. A 1211 (2008) 8. M.K. Reddy Mudiam, Ch. Ratnakekhar, J. Chromatogr. A 1291 (2013) 10. M.T. Pena, X. Vecino-Bello, M.C. Casais, M.C. Mejuto, R. Cela, Anal. Bioanal. Chem. 402 (2012) 1679. N. Fattahi, S. Samadi, Y. Assadi, M.R.M. Hosseini, J. Chromatogr. A 1169 (2007) 63. I. Carpinteiro, B. Abuín, I. Rodríguez, M. Ramil, R. Cela, J. Chromatogr. A 1229 (2012) 79. W. Vetter, J. Lorenz, Environ. Sci. Pollut. Res. 20 (2013) 4435. F.R. Benson, L.W. Hartzel, W.L. Savell, J. Am. Chem. Soc. 741 (1952) 4917.