Ion-pair in-tube solid phase microextraction for the simultaneous determination of phthalates and their degradation products in atmospheric particulate matter

Journal of Chromatography A, 1520 (2017) 35–47

Contents lists available at ScienceDirect

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

Ion-pair in-tube solid phase microextraction for the simultaneous determination of phthalates and their degradation products in atmospheric particulate matter夽 M. Fernández-Amado, M.C. Prieto-Blanco ∗ , P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez Universidade da Coru˜ na, Grupo QANAP, Instituto Universitario de Medio Ambiente (IUMA), Centro de Investigacións Científicas Avanzadas (CICA), na, Spain Departamento de Química, Facultade de Ciencias, Zapateira, 15071, A Coru˜

a r t i c l e

i n f o

Article history: Received 5 July 2017 Received in revised form 23 August 2017 Accepted 3 September 2017 Available online 6 September 2017 Keywords: Phthalates Degradation products Atmospheric particulate matter Ion-pair On-line in-tube solid phase microextraction-HPLC-DAD

a b s t r a c t An in-tube solid phase microextraction, coupled with high-performance liquid chromatography with diode array detection (IT-SPME-HPLC-DAD) method, has been developed for the simultaneous determination of 13 diesters (from dimethyl to dioctylphthalate plus diisobutyl, benzylbutyl, di-2-ethylhexyl, diisononyl and diisodecylphthalate) and 2 monoesters of phthalic acid (mono-butyl and mono-(2ethylhexyl) phthalate) in particulate matter (PM10 ). Triethylamine at pH = 3 was used as an ion-pair reagent with a double function, of regulating the chromatographic retention of the monoesters and the most hydrophilic diesters on a monolithic silica column, and of improving their extraction on a porous polymer with divinylbenzene-4-vinylpyridine capillary. The chromatographic separation was achieved in 13 min. A previous ultrasound-assisted extraction from PM10 filters was also optimized using methanol as solvent. The method detection limits were 0.09–0.52 ng m−3 , the inter-day precision at concentration of 20 ng mL−1 was between 4.2% and 12.7% (n = 15), and the average recovery was 87.3%. The average absolute IT-SPME recovery was 26.2% and the linear range reached up to 109 ng m−3 for most analytes. The method was applied to PM10 samples from different environments collected in Galicia (Spain). DiBP was the major phthalate, followed by its isomer DnBP in urban sites and by DEP in the suburban area. In all samples, DEHP quantified correlates with the isomers of dibutylphthalate. Total PAE concentration was between 14.5 and 245.5 ng m−3 . To the best of our knowledge, this is the first time that a method allows the simultaneous determination of 13 phthalates and their degradation products in particulate matter. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Phthalic acid esters (phthalates or PAEs) are a group of compounds used mainly as plasticizer substances. Due to their high production volume, ubiquity and endocrine disruptor properties, phthalates are target compounds to monitor in the environment, food and some consumer products [1,2]. For these reasons, some of them are included in regulations. The widely used di-2-ethylhexylphthalate (DEHP) is considered a priority hazardous substance in the field of European water policy and

夽 Selected paper from the 19th International Symposium on Advances in Extraction Technologies (ExTech 2017), 27–30 June, 2017, Santiago de Compostela, Spain. ∗ Corresponding author at: Departamento de Química, Universidade da Coruna, ˜ ˜ Spain. Campus da Zapateira s/n, 15071 A Coruna, E-mail address: [email protected] (M.C. Prieto-Blanco). http://dx.doi.org/10.1016/j.chroma.2017.09.010 0021-9673/© 2017 Elsevier B.V. All rights reserved.

Environmental Quality Standards (EQS), which have established an annual average of 1.3 ␮g L−1 for this compound in surface waters [3]. The US Environmental Protection Agency (EPA) has also set a maximum admissible concentration (MAC) for this phthalate in water systems (6 ␮g L−1 ). Along with DEHP, other phthalates such as di-n-butylphthalate (DnBP), diisobutylphthalate (DiBP), benzylbutylphthalate (BBP) and di-n-pentylphthalate (diamylphthalate, DAP) are considered toxic for reproduction, whereas di-n-octylphthalate (DnOP), diisononylphthalate (DiNP) and diisodecylphthalate (DiDP), despite not being considered toxic, are allowed under certain use restrictions [4]. The long-chain alkylphthalates DiNP and DiDP, and in a lower proportion (because of its dangers and the consequent regulation) DEHP are the most used phthalates, accounting for around 70–80% of the phthalates used in Europe over the past years [1,2,5]. They are employed for the lamination of polyvinyl chloride (PVC), the most com-

36

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

monly used plastic in the world. The pollution of phthalates is markedly associated with anthropogenic sources [6]. As phthalates are not chemically bonded to polymers, they can migrate, leach or evaporate, ending up in a lot of products and in the environment, including the atmosphere [2,7]. Their presence in ubiquitous polymeric materials is therefore the major source of PAEs in the atmosphere [8]. One of the main degradation pathways for phthalates is the hydrolysis to phthalic acid, which may occur via monoester phthalates [6]. The formation of monoester phthalates as a result of biodegradation has been widely reported, especially for DEHP, which leads to mono(2-ethylhexyl) phthalate (MEHP). This metabolite is used as a biomarker for the presence of diester in humans [9]. Mono-n-butyl phthalate (MBP) is a monoester from another widely extended phthalate, DnBP. MEHP, which can also be formed by burning, is a known cause of asthma and its toxicity may be higher than that of its precursor, DEHP [9,10]. In addition to the interest in their determination, due to their potential health risk, the detection of the presence of these degradation products in particulate matter (PM) could allow a better understanding of the reactions involving phthalates in the atmosphere. Hydrolysis of particle-adsorbed DEHP to MEHP has been observed under laboratory conditions when a monolayer was absorbed on highlydispersed powders of certain oxides [11], thus these hydrolysis reactions can be produced onto particulate matter. The presence of monoesters in PM could also be due to adsorption of the already formed monoester onto the particles. Adsorption of MEHP onto dust has been suspected before [10]. Despite this interest and potential usefulness, to the authors’ best knowledge, MBP and MEHP have been never determined in PM. In-tube solid phase microextraction (IT-SPME) is a miniaturized sample preparation technique with wide-range applications over the past years [12]. Despite its increasing use, the atmospheric matrices are not among the most analyzed using this technique. Some recent works have analyzed rainwater with IT-SPME [13–15], but its application to particulate matter has been very limited [16,17]. IT-SPME has been applied to the analysis of phthalates [18–20], including even MEHP [9,21], but never for their determination in atmospheric particulate matter samples. On the other hand, ion-pair has been used in IT-SPME [22], but this is a relatively unexplored approach that nevertheless has a great potential. In view of the methods employed in the literature for the analysis of phthalates in PM which use sample preparation of off-line multiple stages and expensive instrumentation (GC–MS or LC–MS), the purpose of this work is to develop a reliable and fast method for the determination of 15 phthalates in particulate matter. An inexpensive and commonly used instrumentation (HPLC-DAD, US) was chosen, in an attempt to accomplish a reduction of sample treatment steps and volume of organic solvents. On-line IT-SPME, with an ion-pair approach, was used to achieve these goals. The high number of analytes with different hydrophobicity includes the most commonly used phthalates (DiNP and DiDP) and the degradation products MBP and MEHP, which can be very useful in order to understand the fate of phthalates in the atmosphere. The method was applied to several PM10 samples from different environments (suburban, urban and urban-traffic) to explore the possible differences in concentrations and relative distribution of the analyzed phthalates according to the type of area.

2. Experimental 2.1. Apparatus The chromatographic system (Waters Corp, Milford, MA, USA) consisted of a 2695 Alliance module and a photodiode array detec-

tor, DAD (Waters 996), with the software Empower 2.0. A six-port injection valve (Rheodyne Model 7725i) was coupled for the ITSPME procedure. 2.2. Reagents and materials EPA 606-M Phthalate Esters Mix containing dimethyl(DMP), diethyl- (DEP), di-n-butyl- (DnBP), butylbenzyl- (BBP), di-2-ethylhexyl- (DEHP) and di-n-octyl- (DnOP) phthalates (200 ␮g mL−1 each in methanol) was obtained from Supelco (Bellefonte, PA, USA). Di-n-hexyl- (DHP), di-n-heptyl- (DHepP), diisodecyl- (DiDP), mono-n-butyl- (MBP) and mono(2-ethylhexyl)(MEHP) phthalates were purchased from Sigma-Aldrich (St. Louis, MO, USA). Diisononylphthalate (DiNP) and solutions in acetone of di-n-propyl- (DPP) (100 ␮g mL−1 ), diisobutyl- (DiBP) (2500 ␮g mL−1 ) and di-n-amyl- (DAP) (5000 ␮g mL−1 ) phthalates were obtained fromChemService (West Chester, PA, USA). Working solutions in acetonitrile were prepared from the commercial products. All solutions were stored at −18 ◦ C in amber glass vials (Waters, Milford, MA, USA, and Supelco, Steinheim, Germany). All solvents were of Gradient HPLC grade. Acetonitrile (ACN) (J.T. Baker, Phillipsburg, NJ, USA) and LC–MS grade methanol (MeOH) (Panreac, Barcelona, Spain) were filtered through 0.45 ␮m PTFE membranes (Teknokroma, Barcelona, Spain). Milli-Q quality water, filtered through 0.20 ␮m nylon membranes (Millipore, Darmstadt, Germany), was used for the mobile phase preparation and UltraResi-Analyzed Water for Environmental Inorganic and Organic Analysis (J.T. Baker, Phillipsburg, USA) was used for the preparation of blanks and standards. Synthesis grade triethylamine (TEA) was obtained from Merck (Darmstadt, Germany). Solutions of 0.05% (for conditioning and as modifiers of the sample) and 0.1% (for mobile phase) were prepared every 2 days. Their pH was adjusted to approximately 3, with orthophosphoric acid 1 M, prepared from the commercial solution 85% (Panreac, Barcelona, Spain). 2.3. Sampling, extraction and filtration of samples Particulate matter (PM10 ) samples were collected at 5 sampling sites of different types (suburban, urban, traffic) in Galicia (Northwestern Spain). The collection was accomplished on quartz-fiber filters (␾ = 150 mm, MK360 Munktell, Falun, Sweden) for 24-h periods, using Digitel DHA-80 high-volume samplers (Digitel Electronik AG, Hegnau, Switzerland). In order to remove organic compounds (among which there are the phthalates), the filters were pre-baked at 400 ◦ C overnight before use and they were stored in baked aluminum foil. All filters were weighed before and after sampling, in both cases after 48 h of conditioning at constant temperature (20 ± 1 ◦ C) and relative humidity conditions (50 ± 5%), according to the EN 12341 European Norm [23]. Samples were stored at that temperature and humidity conditions until extraction. Two circles (␾ = 1.8 and 2 cm) of the filter were extracted with 10 mL of methanol by sonication in ultrasound bath over a period ® of 20 min. Extracts were filtered through Discmic -13HP syringe filters (PTFE, 0.50 mm, 13 mm, Advantec MFS, Dublin, CA, USA). 10 mL of Milli-Q water had been previously passed through the filter to avoid phthalate leaking to the extract. After the extract, TEA 0.05% at pH ≈ 3, which accounted for 60% of the entire volume, was passed through the filter. Extracts were stored at −18 ◦ C until analysis, accomplished within 2 days. The protocol to minimize blanks consisted of cleaning glassware material in alkaline detergent for 24 h, and then rinsing it sequentially with abundant tap water, Milli-Q water and acetonitrile. Before the first use of new glass vials, they were rinsed sequentially with hexane, tetrahydrofuran, methanol and Milli-Q water. After being cleaned, all the glassware was carefully stored, and just

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47 Table 1 Optimized gradient for the separation of the 15 phthalates. Time (min)

Flow (mL min−1 )

%ACN

% (0.1% TEA, pH = 3 in water)

0 3.0 7.5 10.0 13.0 19

0.3 0.55 1.00 1.00 1.00 0.3

40 55 55 100 100 35

60 45 45 0 0 65

before its use it was rinsed with LC–MS grade methanol and dried in an oven at 130 ◦ C. 2.4. In-tube SPME procedure For the IT-SPME-HPLC-DAD procedure, the loop of the six-port injection valve was replaced by a 70 cm-long GC PLOT column ® Rt -S-BOND (divinylbenzene with 4-vinylpyridine, 0.32 mm i.d., 10 ␮m) purchased from Restek (Bellefonte, PA, USA). Capillary connections were facilitated by the use of a 2.5 cm sleeve of 1/16 in. polyether ether ketone (PEEK) tubing at each end of the capillary. In “load” position, the capillary was conditioned by passing 0.5 mL of TEA 0.05% in water (pH ≈ 3). Subsequently, 5 mL of PM methanolic extract with 60% (v/v) TEA 0.05% in water (pH ≈ 3) were passed through the capillary, followed by 60 ␮L of Milli-Q water in order to displace the remaining sample in the capillary. The run is started by programming a sample volume equal to zero for the autosampler. At the time the gradient starts, the valve is rotated to the “inject” position, and the mobile phase passes through the capillary, desorbing analytes. To avoid memory effects, the all-glass sample syringe (Ruthe, Portugal) was rinsed with acetonitrile between runs, while the valve and capillary were flushed with 2 mL of acetonitrile before conditioning of the next sample. 2.5. Chromatographic conditions The analytical column was an Onyx Monolithic C18 100 mm × 3 mm (Phenomenex, Torrance, CA, USA). Ion-pair chromatography with gradient elution mode was used, with a mixture of acetonitrile and TEA 0.1% in water, with the pH adjusted to approximately 3 with phosphoric acid (Table 1). The column temperature was set to 30 ± 5 ◦ C. The 15 phthalates were separated within 13 min. DAD data were recorded between 210 and 400 nm. Integration was performed at 225 nm until 9.5 min (till DnBP) and then at 223.8 nm (from DAP). 3. Results and discussion 3.1. Optimization of the chromatographic separation The optimization was carried out using the initial conditions consisting of a monolithic column and an acetonitrile/water gradient elution. The gradient starts with 45% ACN at a flow rate of 0.3 mL min−1 , reaches 71.5% ACN at 0.55 mL min−1 at minute 5.4 and finishes with 100% acetonitrile at 1.0 mL min−1 at minute 9. Under these conditions, the main difficulties, apart from the simultaneous separation of mono- and diesters, were the separation of two isomers (DnBP and DiBP) and of the phthalates containing a long alkyl chain (DiNP and DiDP). The retention of two monoester degradation products (MBP and MEHP) should be controlled since the protonated and ionic forms could be present due to their carboxylic acid group. The protonated forms (pH < 3.1) were selected by adjusting pH = 3 in the aqueous phase of the mobile phase [9]. Furthermore, an ion-pair reagent such as triethylamine (TEA) was

37

added in order to reinforce a reproducible retention and to reduce peak tails. In the literature, TEA has been applied to the separation of acidic compounds such as nucleic acids [24] or chloroacetic acids [25] at neutral pH, and to the separation of decarboxy-␤-cyanins in the range of pH 2.3–6.7 [26]. TEA may act as a stabilizer for the monoesters and the residual silanol groups of the column. In both cases, an increase of the retention times of monoesters should be expected. Using 0.1% TEA and pH = 3, an increase of the retention of two monoesters (more significant for MEHP) was observed. Besides, under these conditions, the pair MEHP-DEP (partially separated without TEA addition) was resolved and the elution order of DMP and MBP was changed. Phosphoric acid was selected for pH adjustment because as a result of UV detection, the baseline drift was lower than the one obtained with formic acid. TEA concentration was optimized to 0.1%, as higher concentrations do not improve the peak shape. It should be noted that by achieving a good resolution, the retention times of monoesters (MBP at 3.5 min and MEHP at 6.2 min) are comparable to those obtained with UHPLC (around 5 min and 6.5 min respectively) [27]. The other main difficulty was the separation of isomers DnBP and DiBP, increased by the retention of BBP, similar to theirs. In the literature, their separation must be addressed even when the detection is performed using MS since the spectra, and therefore quantifier and qualifier ions, are equal for both compounds [28]. An initial percentage of 60% aqueous phase was necessary for the separation of the first compounds (DMP and MBP), followed by an isocratic step containing 45% to accomplish the partial separation of BBP, DiBP and DnBP (in this elution order). This step is isocratic with respect to the mobile phase composition, but a flow gradient (between 0.55–1.0 mL min−1 ) was required to reduce the retention times and consequently the analysis time. The duration of the isocratic step was optimized between 3 and 7.5 min as a compromise between the resolution of three compounds and the analysis time (Fig. 1). For the phthalates with the longer alkyl chain, from C5 to C10 (including DEHP), an increase of acetonitrile percentage up to 100% is able to resolve all the peaks in just 13 min. As shown in Fig. 1, the peaks of DiNP and DiDP are more broadened than the other diesters due to the fact they are a mixture of isomers. 3.2. Optimization of the extraction in particulate matter A simple ultrasound-assisted extraction with methanol was optimized, which included an end step of filtration since the extracts should be filtered before passing through the IT-SPME capillary. Methanol (MeOH) has been successfully used to extract some phthalates from particulate matter [29]. Due to its moderate toxicity and its miscibility with solvents used for IT-SPME, it was considered a good choice as an extraction solvent. For each experiment, two circles (␾ = 1.8 and 2 cm) of the PM filter from a real sample spiked with 20 ␮g mL−1 of the analytes were extracted. Only 5.7 cm2 of the sample (3.7% of the total area) were needed, a sufficient amount of filter remaining for analyzing other compounds. Volume and time were simultaneously determined by comparing two different procedures: 10 mL of MeOH and 20 min of extraction time versus 5 mL of MeOH and 10 min of extraction time (Fig. 2a). Although similar recoveries were obtained for several target analytes, suggesting that 10 min may be enough in terms of extraction time, a poorer precision was achieved with the shorter extraction time. In addition, a few compounds such as DiBP presented too high recoveries. These results indicate that the extraction time is not enough to obtain reproducible and accurate extraction and 20 min was therefore selected. Moreover, the degradation product MBP presented a very poor extraction with only 10 min and 5 mL (26%), which is doubled with the alternative procedure. The extraction of both monoesters is not complete but it is reproducible and the final results should be corrected with

38

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

Fig. 1. Chromatogram of a standard solution (20 ng mL−1 in TEA 0.05% pH = 3/methanol, 60:40) under optimized conditions.

Fig. 2. Effect on US-assisted PM extraction of (a) volume and time and (b) addition of TEA 0.25%.

recoveries for these compounds. Regarding volume, despite the fact that with 5 mL it does not seem that saturation of the solvent is achieved, there is a limit volume where the sample aliquot could be immersed, which may also negatively affect precision. In addition, as different types of samples are to be analyzed, higher levels of phthalates may be present. Therefore, to ensure reproducibility and avoid possible saturation, 10 mL was selected. In order to improve the extraction of monoesters, other parameters, such as the addition of TEA or an acid medium, were tested. Acidified MeOH (pH = 3 with phosphoric acid to protonate MBP and MEHP) did not lead to significant changes with respect to non-acidified MeOH. On the other hand, the extraction with a mixture MeOH/TEA 40:60, compatible with the IT-SPME step [18], was tested (Fig. 2b). The addition of TEA in the US extraction improved the extraction of the most hydrophilic compounds (DMP, DEP and especially MBP), but the extraction of the most hydrophobic compounds (DEHP, DnOP, DiNP and DiDP) decreased, with recoveries

below 80% (only 45% for DiDP). The extraction without TEA was selected, as a better compromise for all the analytes is achieved (with all recoveries over 60%). Under the optimized conditions, no matrix effects were observed. The milder conditions of extraction compared to other techniques, such as pressurized liquid extraction (PLE), do not favor that high concentrations of interferents pass to the extract. The PM extracts were filtered through hydrophilic PTFE syringe filters which had a better performance than glass and quartz fiber filters for low volumes of sample. When aqueous PM extracts were filtered, and then methanol (40% of total volume) was passed through the filter, the recoveries were in the range from 90 to 105%. In this case, the methanolic extracts were filtered by hydrophilic PTFE filters and next, the aqueous solution was passed through. Total recoveries of the two processes (extraction and filtration) were calculated. 10 mL of Milli-Q water previously passed through

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

39

Fig. 3. Effect on IT-SPME of the addition of TEA as a modifier.

Fig. 4. Chromatogram of a standard solution (20 ng mL−1 in TEA 0.05% pH = 3/methanol 60:40) showing the effect of conditioning the capillary with TEA 0.1%, pH = 3. Black line: with conditioning; red line: without conditioning. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the filter to the sample proved to be enough to avoid contamination from the filter.

3.3. Optimization of the IT-SPME procedure The initial conditions were those optimized in a previous work, which focused on a simpler case of IT-SPME for phthalates in rainwater samples. Thus, a PLOT capillary of intermediate polarity and 5 mL of sample solved in water/methanol 60:40 were selected. Methanol is the solvent used for phthalate extraction from the particulate matter and water is added to improve the extraction of most hydrophilic compounds. Despite the fact that the aqueous phase would imply a dilution of the organic extract, the 60:40 proportion of water/methanol has been proven to be the best compromise solution for hydrophilic and hydrophobic phthalates. An effective and simultaneous extraction of monoesters and diesters is one of the proposed objectives. Thus, the pH of aqueous phase was adjusted to 3, in order to achieve the protonated forms of MBP and MEHP which should show a behavior more similar to that of diesters than that of ionic forms. Phosphoric acid, which had the best performance for acidifying the mobile phase, was selected for this purpose (Fig. 3). Ion-pair has been applied to IT-SPME for the analysis of lauralkonium chloride in water samples [22], but to the authors’ knowledge, ion-pair interactions have not been used again in IT-SPME. In this work, the suitability of TEA to create ion-pair interactions with the analytes in the extraction was also explored. The use of TEA (pH = 3 adjusted with H3 PO3 ) as a modifier produced an improvement in the extraction of the most hydrophilic compounds, especially from DMP to DPP. On the other hand, the extraction of the most hydrophobic phthalates is not practically influenced by TEA (Fig. 3). However, the TEA effect is observed only if a previous condition-

ing of the capillary with the TEA solution is performed before the phthalates pass through. As shown in Fig. 4, a clear decrease of peak areas occurs for the most hydrophilic compounds (from DMP to DAP, with dramatic losses of 86–94% for DMP, DEP, DPP and MBP) if there is no TEA conditioning. This fact can be explained considering that TEA competes with the most hydrophilic analytes for the stationary phase (divinylbenzene with 4- vinylpyridine) of the capillary. A certain adsorption of TEA on stationary phase during the conditioning step could minimize this competition and allow a more effective extraction. It is expected that the monoesters have an ion-pair-like interaction with TEA that favors the extraction. However, other mechanisms have a contribution in this regard, since the TEA effect is greater for MBP than MEHP, and certain diesters without ionizable groups, from DMP to DnBP, are also influenced. The presence of TEA in both liquid and solid phases could decrease the difference of polarity between phases, increasing the affinity of the hydrophilic compounds for the solid phase, and therefore favoring their migration to the solid phase. DMP and DEP are also more concentrated in the capillary with the addition of TEA (Fig. 3). Another parameter influencing the extraction is the TEA concentration. Concentrations of TEA of 0.05% and 0.1% were tested. By increasing TEA concentration up to 0.1%, no improvement of analytical response was obtained (Fig. 3). Thus, the methanolic extracts were diluted with 60% of the total volume of 0.05% TEA solution at pH = 3 before the IT-SPME procedure.

3.4. Analytical performance characteristics The proposed method was validated by analyzing its linearity, limits of detection and quantitation, precision and recoveries. The obtained figures are presented in Tables 2 and 3. Linear ranges

40

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

Table 2 Analytical parameters of the proposed method. Linear ranges and limits are in ng m−3 .

DMP MBP DEP DPP MEHP BBP DiBP DnBP DAP DHP DHepP DEHP DnOP DiNP DiDP *

Slope (b)b ± sb

Intercept (a)a ± sa

R2

Linear range

MDL

LOQ

RL

%RSD* (n = 15)

1233 ± 46 3830 ± 67 2605 ± 86 3115 ± 30 2309 ± 43 2641 ± 46 3272 ± 32 2915 ± 36 2738 ± 22 2721 ± 31 2357 ± 21 1873 ± 12 1431 ± 9 841 ± 10 691 ± 13

2816 ± 1225 −5611 ± 2443 9123 ± 3337 −173 ± 1112 −4840 ± 2286 −2385 ± 2427 2954 ± 1814 −759 ± 1887 −1811 ± 1143 −2289 ± 1651 −1212 ± 1087 225 ± 636 640 ± 459 1152 ± 555 1425 ± 755

0.9945 0.9982 0.9946 0.9994 0.9975 0.9979 0.9994 0.9989 0.9996 0.9991 0.9995 0.9997 0.9997 0.9992 0.9977

LOQ-60 LOQ-116 LOQ-73 LOQ-73 LOQ-137 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-109 LOQ-148

0.52 0.22 0.33 0.14 0.20 0.14 0.28 0.18 0.14 0.11 0.14 0.18 0.09 0.51 0.40

1.65 0.71 1.05 0.44 0.65 0.43 0.89 0.56 0.43 0.33 0.43 0.56 0.27 1.62 1.27

1.65 21.9 1.05 1.17 26.9 0.43 0.89 3.06 0.43 0.33 5.39 3.00 0.27 1.62 1.27

11.4 14.6 8.0 4.2 10.2 5.8 7.0 5.3 5.4 5.0 4.6 5.1 6.3 10.9 12.7

20 ng mL−1 .

Table 3 Analytical recoveries (%R) ± SD (n = 6) for the whole procedure and absolute recoveries for the IT-SPME step (%R abs.) ± SD (n = 3).

DMP MBP DEP DPP MEHP BBP DiBP DnBP DAP DHP DHepP DEHP DnOP DiNP DiDP a

%R whole procedure

IT %R abs. 50 ng

IT %R abs. 100 ng

IT %R abs. Average

76 ± 5a 62 ± 8 94 ± 10 92 ± 6 79 ± 10 95 ± 4 93 ± 5 95 ± 10 93 ± 2 89 ± 2 92 ± 4 89 ± 5 98 ± 4 90 ± 6 73 ± 9

8±1 16 ± 1 23.5 ± 0.3 32 ± 1 19 ± 2 28 ± 2 42 ± 2 37 ± 1 34 ± 2 32 ± 2 32 ± 2 30 ± 2 28 ± 2 22 ± 2 18 ± 2

5±1 33 ± 4 23 ± 1 29 ± 2 21 ± 3 29 ± 2 36 ± 2 36 ± 2 33 ± 2 32 ± 2 31 ± 2 29 ± 2 26 ± 2 20 ± 3 16 ± 2

7±3 25 ± 11 23.3 ± 0.2 30 ± 2 20 ± 1 28 ± 1 39 ± 4 36 ± 1 33 ± 1 31.8 ± 0.3 31 ± 1 29 ± 1 27 ± 1 21 ± 1 17 ± 1

n = 5.

reach up to 109 ng m−3 for most analytes, with R2 > 0.995 for all the target compounds. Because of the different extraction behaviors, a compromise solution was chosen for both the US and capillary extractions. For this reason, the most hydrophilic (DMP, DEP, DPP) and the most hydrophobic (DiNP, DiDP) phthalates present higher limits and worse precision (and a shorter linear range for the most hydrophilic) than the intermediate compounds (Table 2). In addition, MBP, MEHP, DiNP and DiDP have the broadest peaks affecting their precision. The analytical recoveries for the whole procedure were obtained using the PM10 filter spiked with 0.5 ␮g of each of the analytes (20 ng mL−1 for the IT-SPME) and the values are provided in Table 3. Recoveries of 89–98% were obtained for almost all target analytes, except for DMP, MBP, MEHP and DiDP due to the compromise conditions selected for the extraction, as previously explained. For the 4 analytes with recoveries below 80%, the obtained values (including limits and ranges in Table 2) were corrected taking recoveries into account. Due to their ubiquity and trace levels, blank problems in phthalates analysis have to be addressed. In a previous work [18], the effect of some potential sources was analyzed, and protocols were established to control and minimize blanks for an IT-SPME method. Those protocols (see Section 2.3 for description) were followed in the proposed method and blanks corresponding to IT-SPME and determination steps were controlled daily. Whole-procedure blanks (including also US extraction and extract filtration) were analyzed too and remained within the levels of the IT-SPME step blanks, not increasing the background signal. DEHP and DiBP

were always found in the blanks, with levels of 0.33 ± 0.10 and 0.79 ± 1.28 ␮g mL−1 , respectively. In the case of DiBP, the standard deviation was clearly higher than the average value obtained for the blanks, making it assimilable to zero. The other compounds usually found in the blanks were DnBP (0.34 ± 0.31 ␮g mL−1 ) and DiNP (0.10 ± 0.05 ␮g mL−1 ). All of them were below their respective limits of quantitation (LOQ). The method detection limits (MDL) were calculated with 99% confidence, according to 40 CFR, Part 136, Appendix B [30]. The limits of quantitation (LOQ) were calculated as 10 times the standard deviation obtained in the 7 replicates used for the MDL calculation. All limits (Table 2) were experimentally tested. MDL and LOQ were in the ranges 0.09–0.52 and 0.27–1.65 ng m−3 , respectively. Once again, the highest values were obtained for the most hydrophilic and the most hydrophobic compounds for the reasons previously exposed. Blank subtraction is not recommended for phthalates because sometimes they present an irregular pattern [31]. They were considered by means of reporting levels (RL) (Table 2), calculated as 10 times the blank average concentration [32]. When a concentration is between LOQ and RL, only an estimation of the value is given. RL were below 3 ng m−3 for most analytes (Table 2). A few compounds that rarely appear in the blanks presented high peak areas when they do, increasing reporting limits. This effect was particularly pronounced for the two monoesters, with RL higher than 20 ng m−3 . As these compounds are not usually present in blanks (in only 8–12% of the cases) in most cases the LOQ value would be sufficient as a limit, although below RL it would only provide an estimation of the concentration. In the latter situation, the reliability of the result could not be assured with the same certainty as if the value was above RL. Absolute recoveries of the IT-SPME (efficiency of the nonexhaustive extraction) were calculated by comparing the peak areas corresponding to 5 mL of 10 and 20 ng mL−1 (50 and 100 ng of each analyte, respectively) processed by IT-SPME to those obtained by direct injection of 10 and 20 ␮L of 5 ␮g mL−1 (50 and 100 ng, respectively). The obtained values (Table 3) were similar to those reported in the literature for IT-SPME (10%–30%). DMP value was slightly lower and the DnBP and DiBP was approximately 40%.

3.5. Application to real PM10 samples The optimized method was applied to 12 PM10 samples from 5 sampling sites with different typology, all of them from Galicia (NW Spain) (Fig. 5). Samples annotated SU are from a suburban area, U from an urban area, and T from three different urban traffic stations of three different cities (AC, SC and C). The obtained values are shown in Fig. 6 and Table S1. High correlation coefficients

Table 4 Characteristics of methods for determination of phthalates in particulate matter reported in the literature and levels found. DCHP: dicyclohexyl phthalate; DDP: didecylphthalate; DiOP: diisoctyl phthalate; DNP: dinonylphthalate; n.e.: not specified. Steps

Organic solvent vol.

Deter. technique

LOD (ng m−3 )

Levels (ng m−3 )

Sampling period and site, number of samples

Type of sample

Reference

DMP, DEP, diallylphthalate, DPrP, DnBP, DiBP, BBP, butyloctyl phthalate, DHepP, DEHP, DnOP, DiOP, DiNP, DiDP BBP, DnBP, DiBP, DEHP, butyl isobutyl phthalate

1. Soxhlet extraction (40 h) 2. SPE purification 3. Concentration

>70 mL

GC–MS

0.0004–0.0013

Wang et al. [33]

n.e.

GC–MS

0.26 for DEHP

TSP (day and night) (High and medium volume). Urban

Li et al. [43]

DMP, DEP, diallyl phthalate, butyloctyl phthalate, BBP, DPP, DnBP, DiBP, DHepP, DEHP, DnOP, DiOP, DiNP, DiDP

1. 2. 3. 4. 5. 6.

US Extraction (30 min) Filtration Concentration SPE purification Concentration Redissolution

>100 mL

HPLC-UV

0.01–0.06 ng


Apr, Jul and Oct 2005 and Jan 2006, Nanjing (Jiangsu, China), n = 170 Jan-Feb 2009, Xi’an (China) and Nov 2006-Feb 2007 and Jan 2008, New Delhi (India). n = 28 (n = 15 for 24 h) Apr 2009-Apr 2010, 10 sites in Nanjing (China), n = 450

12h-TSP (High volume). Urban and suburban

1. US Extraction 2. Concentration 3. Derivatization (3 h, for other analytes)

0.05–17.9 (urban) and 0.01–7.8 (suburban). DnBP is the major 3.3–164 (Xi’an, 24 h) and 12–689 (New Delhi, 24 h). DEHP is the major

TSP. Different types of area

Wang et al. [8]

DMP, DEP, DnBP, BBP, DEHP, DnOP

1. 2. 3. 4.

Soxhlet extraction (16 h) SPE purification Concentration Derivatization (1 h, for other analytes) 5. Volume adjustment

190 mL

GC–MS

5–34.4 ng


29 Feb–10 Mar 2004, North Sea, n=6

TSP (High volume)

Xie et al. [47]

DMP, DEP, DnBP, DiBP, BBP, DEHP

1. 2. 3. 4.

Soxhlet extraction (24 h) SPE purification Concentration Volume adjustment

>85 mL

GC–MS

0.0002–0.024

0.001-0.735. DEHP is the major

Summer 2004, The Arctic, n = 6

TSP (High volume)

Xie et al. [49]

DMP, DEP, DnBP, BBP, DEHP, DnOP

1. Volatilization-condensation extraction (12 h) 2. Concentration 3. Redissolution

100 mL

GC-ECD

0.025–0.075


May 2002–Apr 2003, Paris (France), n = 20

TSP (High volume). Urban

Teil et al. [48]

DMP, DEP, DnBP, DiBP, BBP, DEHP, DnOP

1. Soxhlet extraction (16 h) 2. Evaporation 3. Storage for water residues elimination (overnight) 4. SPE purification 5. Evaporation 6. Derivatization (for other analytes)

210 mL

GC–MS

0.001–0.013


GKSS Resarch Centre (Germany) and North Sea, n = 16

TSP (High volume). Suburban and ocean

Xie et al. [28]

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

Phthalates analyzed

41

42

Table 4 (Continued) Steps

Organic solvent vol.

Deter. technique

LOD (ng m−3 )

Levels (ng m−3 )

Sampling period and site, number of samples

Type of sample

Reference

DEP, DnBP, DEHP

1. Microwave-assisted microsolid phase extraction (15 min) 2. Rinsing and drying 3. Desorption with US (25 min) 4. Filtration 5. Evaporation by N2 stream 6. Redissolution

∼10 mL

HPLC-UV

2–5.7

0.57–5.56 (urban), 4.5–11.9 (business center), 19.2–68.8 (industrial). DEHP is the major

Dongguan (China), n=3

24h-TSP (High volume). Urban, business center and industrial sites

Jiao et al. [45]

DMP, DEP, DnBP, DiBP, BBP, DEHP

1. US extraction (3 × 10 min) 2. Concentration 3. Derivatization (3 h, for other analytes)

15 mL

GC–MS

n.e.

0.3–1748. DEHP is the major.

Jan, Jun and Jul 2003, 14 cities (China), n = 56

24h-PM2.5 (mini-volume)

Wang et al. [34]

DMP, DEP, DnBP, DEHP

1. Spiking with surrogate standard (overnight) 2. PLE (3 × 8 min) 3. Concentration by N2 stream evaporator 4. Redissolution

n.e.

LC–MS/MS

0.002–0.010


˜ (Spain), La Coruna n=8

24h-PM2.5 (High volume). Industrial, urban and suburban

Salgueiro-González et al. [29]

DMP, DnBP, DiBP, DEHP, DnOP

1. 2. 3. 4.

180 mL

GC–MS

n.e.

0.02–191.82. DEHP is the major

Jul 2006 and Jan 2007, Taizhou (China), n = 26

24h-PM2.5 (High volume). Urban

Gu et al. [44]

DMP, DEP, BBP, DnBP, DEHP, DnOP, bis(2-ethylhexyl)adipate

1. Thermal desorption

–

GC–MS

0.023–0.822 ng/sample


24h-PM2.5 (High volume). Urban

Wang et al. [36]

DnBP, BBP, DEHP, DHepP, DNP, DDP

1. Thermal desorption

–

GC–MS

n.e.

0.2–2.2 (outdoor) and 0.2–55 (indoor). DEHP is the major

PM2.5 and PM2.5–15 (84 h outdoor and 168 h indoor). Offices

Weschler [46]

DMP, DEP, BBP, DnBP, DEHP, DnOP

1. 2. 3. 4.

20 mL

GC–MS

0.002–0.016


Winter and summer 2012/2013, Guangzhou (China), n = 24 Autumn-Winter 1981–1982 (Wichita, USA); Winter-Spring 1982 (Lubbock, USA), n = 36 Jan, Apr and Jul 2010, 7 sites in Tianjin (China), n = 150

24h-PM2.5 and PM10 (medium volume). Urban

Kong et al. [39]

US Extraction (3 × 15 min) Filtration Concentration Redissolution

US Extraction (2 × 15 min) Filtration Concentration by N2 stream Volume adjustement

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

Phthalates analyzed

GC–MS

0.05 (LOQ)


Oct 2011-Aug 2012, Shanghai (China), n = 77

24h-PM2.5 and PM10 (High volume sampler). Suburban residential area

Ma et al. [38]

US Extraction (3 × 30 min) SPE Purification Evaporation Redissolution

n.e.

GC–MS

n.e.


Jan-Feb 2007, Thessaloniki (Greece), n = 20

24h-PM10 (Low volume). Urban (traffic and industrial)

Salapasidou et al. [42]

1. 2. 3. 4.

US Extraction (3 × 10 min) Filtration Concentration Derivatization (3 h, for other analytes)

n.e.

GC–MS

n.e.


Jan, Feb and May 2007, Chennai (India), n = 49

PM10 (day and night) (High volume)

Fu et al. [40]

DMP, DEP, DiBP, DnBP, BBP, DEHP

1. 2. 3. 4. 5.

US Extraction (3 × 20 min) Volume reduction Fractionation in column Evaporation Redissolution

n.e.

GC–MS

n.e.

0.01-31. DEHP is the major in most sampling points (DEP, DiBP and DnBP in the other)

27 Jul–11 Aug 2013, Southern and Eastern Mediterranean Sea, n = 40

PM10 (High volume sampler)

Romagnoli et al. [41]

DMP, DEP, DPP, DiBP, DnBP, DAP, DHP, DHepP, DEHP, DnOP, DiNP, DiDP, BBP, MBP, MEHP

1. US Extraction (20 min) 2. Filtration 3. IT-SPME

10 mL

HPLC-DAD

0.09–0.51


5 sites in Galicia (NW Spain), n = 12

24h-PM10 (High volume). Urban, suburban and traffic sites

This work

1. Spiking with deuterated standards and equilibration (3 h) 2. Extraction assisted by US and shaking in an orbital shaker (3 × (30 + 20) min) 3. Centrifugation (5 min) 4. Concentration by N2 stream

DMP, DEP, DnBP, BBP, DEHP, DnOP

1. 2. 3. 4.

DMP, DEP, DnBP, DiBP, DEHP

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

12 mL

DMP, DEP, DiBP, DnBP, BBP, DHP, DCHP, DEHP, DnOP

DCHP: dicyclohexyl phthalate; DDP: didecylphthalate; DiOP: diisoctyl phthalate; DNP: dinonylphthalate; n.e.: not specified.

43

44

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

Fig. 5. Chromatogram of a PM sample (suburban area1, black line) and an IT-SPME solvent blank of IT-SPME step (red line) under optimized conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. PAEs determined in PM10 from several sampling sites (Galicia, Spain). (A) Group I, (T-SC1 is not included). (B) Group II.

were obtained for the sampling sites, except for the city with TC1 and T-C2. In addition, one of samples (U1) collected from the urban area (near the seaport) also showed a poor correlation with

other samples. Out of the 15 target phthalates, only 4 were found above MDL for these samples: DEP, DiBP, DnBP and DEHP. DiBP (3.2-202.9 ng m−3 ) is the major phthalate found in almost all sam-

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

ples (Fig. 6A, group I), except for U1, T-C1 and T-C2. For these three samples (Fig. 6B, group II), the isomer DnBP is the major phthalate found. On the contrary, DnBP concentrations are below the reporting level for 4 out of 5 samples from the suburban area. For the samples which showed good correlations (nine out of the twelve samples under study, group I), the potential correlation between phthalates was examined. DEHP and DnBP show excellent positive correlations (r2 = 0.95), as well as DEHP and DiBP (r2 = 0.85). This fact could suggest a common origin. A lower correlation between DnBP and DiBP (r2 = 0.78) was observed. Fu et al. pointed out that DnBP and DiBP underwent different atmospheric processes and had different lifetime [35]. Thus, the differences between the two samples (U1 and U2), from the seaport could also be due to a different ageing degree of dibutylphthalate isomers in aerosol. As in other reported studies [8,28,33,36], dibutylphthalate isomers are the dominant phthalates in terms of concentration levels and frequency of detection. DEP was quantified in all the 5 suburban samples and in 3 out of 5 traffic samples. For the two urban samples, it is below MDL. DEP seems to be an important compound in the suburban area, while it is negligible in the city (AC) near to this suburban area. It has no significant correlation with other compounds. Lebedev et al. [37] reported an increase of DEP in environmental samples over the past years. In this sampling site, our QANAP research group quantified DEP in rainwater from short rainfall events at ng mL−1 concentration levels [18] and in PM2.5 at ng m−3 [29]. Therefore, its presence in PM10 was expected. The other detected phthalate is DEHP, with generally low levels and below RL for 6 samples. A significant detail is that DEHP was found in relatively low concentrations. According to the literature, DEHP is usually the major phthalate (Table 4), but in the analyzed samples it has lower levels than DiBP and DnBP. In addition, DEHP is quantified at lower concentrations in suburban than in urban areas. However, the total concentration of phthalates, except for T-SC1, did not show significant differences between urban and suburban area. In the literature, higher values are mentioned for urban than for suburban areas, but these areas belong to larger cities than those studied in this work [33]. The most polluted sample (T-SC1) contains around ten times the total concentration of the other samples, which have a similar profile (r2 from 0.67 to 0.99). Unlike the other samples, T-SC1 is taken in July, with a maximum daily temperature of 25 ◦ C. Its higher PAE concentration may be due to higher emission from sources [34], resulting from higher temperature. However, no photochemical reactions were observed, which otherwise could result in a decrease of the phthalate concentration [6,33]. In T-SC1, the total concentration of PAEs of 245.5 ng m−3 , accounts for 1% of PM10 . In the analyzed samples, PAEs (14.5–245.5 ng m−3 ) contribute in average 0.3% to the total PM10 , ranging from 0.1% in T-AC, T-C and U to 0.6% in T-SC. Finally, the U1, T-C1 and T-C2 samples showed excellent, but negative correlations between T-C2 with U1 and T-C1. The maximum temperature in T-C2 was also around 25 ◦ C. For this reduced group of samples, the two isomers had a positive correlation (Pearson coefficient 0.88) and DnBP correlated negatively with DEHP (Pearson coefficient −0.80). In this group, the phthalates may have been emitted from a different source other than that of the other group of nine samples. Other effects, such as the abovementioned secondary aerosol, a different distribution and atmospheric compartment (air, particulate matter) could also contribute to different profiles. The analysis of phthalates in PM (total suspended particulate matter TSP, PM10 and PM2.5 ) was carried out in several sites of different typologies (urban, suburban, industrial) in Asia (China and India), America and Europe (Table 4). In this work, the concentration levels in PM10 are lower than the values reported in

45

China [38,39] and India [40], and similar (except for TC-SC1) to those obtained by the Mediterranean Sea [41] and in the city of Thessaloniki [42]. However, there are differences in the dominant phthalate (DiBP and in some cases DnBP in this study, and DEHP in those reported by Romagnoli et al. [41] and Salapasidou et al. [42]). 3.6. Comparison with the literature methods for particulate matter The literature methods in terms of number of analyzed PAEs, sample treatment, chromatographic separation and limits of detection were compared (Table 4). Most of the reported literature methods have analyzed only a few phthalates (5–7), and none of them has analyzed the degradation products MBP and MEHP. Their determination could be very interesting when it comes to understanding the degradation processes of phthalates in the atmosphere and/or the air/particulate/rainwater partition of these degradation products. Most hydrophobic phthalates DiNP and DiDP have only been determined in a few studies [8,33]. In addition, using several methods no distinction was performed between DnBP and DiBP. This distinction may be significant, since DiBP is sometimes the main isomer ([40,41,43], this work), while in some areas DnBP may have a higher presence ([8,33,44], this work). The behavior of these compounds in the formation of secondary aerosol is another interesting task which needs adequate analytical methodologies. In the proposed method, owing to the fact that IT-SPME and separation are a single step, the separation time is also relevant. Thus, a more rapid chromatographic separation was obtained compared with those reported, which employed between 16 and 26 min [29,45] or even a minimum of 30 min [39] for the separation of a lower number of compounds. One of the main advantage of the presented method is related to the sample preparation. Only a few methods that use thermal desorption [36,46] have a simpler sample preparation. Despite the fact that US-assisted extraction is a widely extended technique for the phthalate extraction from the filter, after this extraction (or longer Soxhlet extraction) most methods require several steps of purification and concentration (Table 4). Using the proposed method, after a simple and fast filtration, only on-line IT-SPME is needed. In addition, very good recoveries were obtained with only one extraction, while most methods using US extraction require two or three extractions [34,38–42,44]. As a result, this method saves both time and organic solvents, avoiding also some tedious steps such as concentration. Methanol is less dangerous than dichloromethane, which is commonly used (alone or mixed with other solvents) for the extraction. Only 10 mL per sample are consumed with the method proposed herein, while volumes over 70 mL (up to 210 mL) are needed for some of the reported methods [8,28,33,47–49]. Concerning detection limits, it should be noted that most authors do not consider the well-known background contamination during the procedure, and calculate the limit of detection as 3 times the signal-to-noise ratio. These instrumental limits (LOD) do not reflect the real limits of the entire method such as MDL, which are higher, but more representative. In addition, most methods use a more expensive technique, mass spectrometry, for phthalate detection. The proposed method improves more than 10 times the limits presented by [45] with HPLC-UV (2–5.7 ng m−3 ) and even the value for DEHP is a little lower than that reported by [43] with GC–MS determination. 4. Conclusion In this work a reliable, rapid and inexpensive method was developed for the determination of 15 phthalates, including the degradation products MBP and MEHP, in atmospheric particu-

46

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47

late matter. A simple US-assisted extraction with only 10 mL of methanol, followed by fast syringe filtration and on-line IT-SPME coupled to HPLC-DAD is sufficient to perform the analysis. The separation of the 15 analytes was carried out in only 13 min. Ion-pair interactions were used to improve the peak shape and reproducibility of monoesters. Ion-pair was also applied to the in-tube solid phase microextraction, an approach hardly explored in the literature that in this case has improved significantly the extraction efficiency for the most hydrophilic compounds. The proposed method avoids the long and tedious steps widely extended in other methods for determination of phthalates in PM, such as purification and concentration, saving time and organic solvents consumption. In addition, only one US extraction is needed, in contrast to the usually reported 2–3 extractions (and with more toxic solvents). The presented method has been validated, with acceptable inter-day reproducibility, recoveries and method detection limits between 0.09 and 0.52 ng m−3 . Blank issues were taking into account by calculating reporting levels, below which only an estimation of the concentration is provided to assess the reported results. The proposed method was applied to 12 PM10 samples collected from different environments (urban, suburban and urban-traffic) in Galicia (NW Spain). Only DEP, DnBP, DiBP and DEHP were over LOQ, and PAEs ranged from 14.5 to 245.5 ng m−3 . DiBP was the major phthalate found in 75% of the samples, followed by its isomer DnBP (major in the other 25%) in urban and urban traffic sites, and by DEP in the suburban area. Low concentrations (<8 ng m−3 ) were found for the widely used DEHP, except for the most contaminated sample (18.2 ng m−3 ). Despite the fact that the number of samples is not high, two different profiles of samples were detected, as well as correlations between DEHP and dibutylphthalate isomers and the potential effect of temperature on the phthalate concentration. Both monoesters were not detected in the analyzed samples, as the amount of samples was limited and these two analytes may be present in PM samples from other places and/or periods. Acknowledgements The authors acknowledge support from Xunta de Galicia (10MSD164019PR, GRC2013-047) and Ministerio de Ciencia e Innovación (CGL2010-18145, CTM2013-48194-C3-2-R). M. F. A. is grateful to the FPU Program for her grant (AP2012-5486). The authors would like to thank P. Esperón (PTA2013-8375-I) for her support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2017.09. 010. References [1] P. Ventrice, D. Ventrice, E. Russo, G. De Sarro, Phthalates European regulation, chemistry, pharmacokinetic and related toxicity, Environ. Toxicol. Pharmacol. 36 (2013) 88–96. [2] U. Heudorf, V. Mersch-Sundermann, J. Angerer, Phthalates: toxicology and exposure, Int. J. Hyg. Environ. Health 210 (2007) 623–634. [3] Directive 2013/39/EU of the European Parliament and of the Council of 12 August 2013 amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy. [4] Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 Concerning the Registration, Evaluation, authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals agency, amending Directive 1999/45/EC and repealing Council Regulation (EEC) No 793/93 and Commission Regulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/105/EC and 2000/21/EC.

[5] ECPI, European Council for Plasticisers and Intermediates (ECPI) estimations based on IHS published data about the use of phthalate plasticizers in Europe in 2005–2014. http://www.plasticisers.org/plasticisers/orthophthalates/, 2015 (Accessed 04.07.17). [6] D.W. Gao, Z.D. Wen, Phthalate esters in the environment: a critical review of their occurrence, biodegradation, and removal during wastewater treatment processes, Sci. Total Environ. 541 (2016) 986–1001. [7] N. Salgueiro-González, M.J. López de Alda, S. Muniategui-Lorenzo, D. Prada-Rodríguez, D. Barceló, Analysis and occurrence of endocrine-disrupting chemicals in airborne particles, TrAC – Trends Anal. Chem. 66 (2015) 45–52. [8] W. Wang, Y. Zhang, S. Wang, C.Q. Fan, H. Xu, Distributions of phthalic esters carried by total suspended particulates in Nanjing, China, Environ. Monit. Assess. 184 (2012) 6789–6798. ˜ ˜ Y. Moliner-Martínez, R. [9] N. Jornet-Martínez, M. Munoz-Ortu no, Herráez-Hernández, P. Campíns-Falcó, On-line in-tube solid phase microextraction-capillary liquid chromatography method for monitoring degradation products of di-(2-ethylhexyl) phthalate in waters, J. Chromatogr. A 1347 (2014) 157–160. [10] K. Nakamiya, H. Takagi, T. Nakayama, H. Ito, H. Tsuruga, J.S. Edmonds, M. Morita, Microbial production and vaporization of mono-(2-ethylhexyl) phthalate from di-(2-ethylhexyl) phthalate by microorganisms inside houses, Arch. Environ. Occup. Health 60 (2005) 321–325. [11] W. Behnke, F. Nolting, C. Zetzsch, The atmospheric fate of di(2-ethylhexyl-)phthalate, adsorbed on various metal oxide model aerosols and on coal fly ash, J. Aerosol. Sci. 18 (1987) 849–852. [12] M. Fernández-Amado, M.C. Prieto-Blanco, P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez, Strengths and weaknesses of in-tube solid-phase microextraction: a scoping review, Anal. Chim. Acta 906 (2016) 41–57. [13] M. Fernández-Amado, M.C. Prieto-Blanco, P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez, A novel and cost-effective method for the determination of fifteen polycyclic aromatic hydrocarbons in low volume rainwater samples, Talanta 155 (2016) 175–184. [14] H. Asiabi, Y. Yamini, M. Shamsayei, Determination of ultra-trace amounts of chlorophenols in rain, tap and river water by an electrochemically controlled in-tube solid phase microextraction method, RSC Adv. 6 (2016) 94564–94573. [15] Y. Bu, J. Feng, M. Sun, C. Zhou, C. Luo, Facile and efficient poly(ethylene terephthalate) fibers-in-tube for online solid-phase microextraction towards polycyclic aromatic hydrocarbons, Anal. Bional. Chem. 408 (2016) 4871–4882. [16] M.C. Prieto-Blanco, P. López-Mahía, P. Campíns-Falcó, On-line analysis of carbonyl compounds with derivatization in aqueous extracts of atmospheric particulate PM10 by in-tube solid-phase microextraction coupled to capillary liquid chromatography, J. Chromatogr. A 1218 (2011) 4834–4839. [17] M.C. Prieto-Blanco, Y. Moliner-Martínez, P. López-Mahía, P. Campíns-Falcó, Determination of carbonyl compounds in particulate matter PM2.5 by in-tube solid-phase microextraction coupled to capillary liquid chromatography/mass spectrometry, Talanta 115 (2013) 876–880. [18] M. Fernández-Amado, M.C. Prieto-Blanco, P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez, A comparative study of extractant and chromatographic phases for the rapid and sensitive determination of six phthalates in rainwater samples, Chemosphere 175 (2017) 52–65. ˜ ˜ A. Argente-García, Y. Moliner-Martínez, J. Verdú-Andrés, R. [19] M. Munoz-Ortu no, Herráez-Hernández, M.T. Picher, P. Campíns-Falcó, A cost-effective method for estimating di(2-ethylhexyl)phthalate in coastal sediments, J. Chromatogr. A 1324 (2014) 57–62. ˜ ˜ Y. Moliner-Martínez, S. Cogollos-Costa, R. [20] M. Munoz-Ortu no, Herráez-Hernández, P. Campíns-Falcó, A miniaturized method for estimating di(2-ethylhexyl) phthalate in bivalves as bioindicators, J. Chromatogr. A 1260 (2012) 169–173. [21] N. Jornet-Martínez, C. Antón-Soriano, P. Campíns-Falcó, Estimation of the presence of unmetabolized dialkyl phthalates in untreated human urine by an on-line miniaturized reliable method, Sci. Total Environ. 532 (2015) 239–244. [22] M.C. Prieto-Blanco, Y. Moliner-Martínez, P. López-Mahía, P. Campíns-Falcó, Ion-pair in-tube solid-phase microextraction and capillary liquid chromatography using a titania-based column: application to the specific lauralkonium chloride determination in water, J. Chromatogr. A 1248 (2012) 55–59. [23] EN 12341:2014 – Ambient air – Standard gravimetric measurement method for the determination of the PM10 or PM2.5 mass concentration of suspended particulate matter. [24] E.D. Close, A.O. Nwokeoji, D. Milton, K. Cook, D.M. Hindocha, E.C. Hook, H. Wood, M.J. Dickman, Nucleic acid separations using superficially porous silica particles, J. Chromatogr. A 1440 (2016) 135–144. [25] M.C. Prieto-Blanco, M.F. Alpendurada, P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez, S. Machado, C. Gonc¸alves, Improving methodological aspects of the analysis of five regulated haloacetic acids in water samples by solid-phase extraction, ion-pair liquid chromatography and electrospray tandem mass spectrometry, Talanta 94 (2012) 90–98. [26] S. Wybraniec, Effect of tetraalkylammonium salts on retention of betacyanins and decarboxylatedbetacyanins in ion-pair reversed-phase high-performance liquid chromatography, J. Chromatogr. A 1127 (2006) 70–75. [27] L. Herrero, S. Calvarro, M.A. Fernández, J.E. Quintanilla-López, M.J. González, B. Gómara, Feasibility of ultra-high performance liquid and gas chromatography coupled to mass spectrometry for accurate determination of primaryand secondary phthalate metabolites in urine samples, Anal. Chim. Acta 853 (2015) 625–636.

M. Fernández-Amado et al. / J. Chromatogr. A 1520 (2017) 35–47 [28] Z. Xie, J. Selzer, R. Ebinghaus, A. Caba, W. Ruck, Development and validation of a method for the determination of trace alkylphenols and phthalates in the atmosphere, Anal. Chim. Acta 565 (2006) 198–207. [29] N. Salgueiro-González, M.L. De Alda, S. Muniategui-Lorenzo, D. Prada-Rodríguez, D. Barceló, Determination of 13 estrogenic endocrine disrupting compounds in atmospheric particulate matter by pressurised liquid extraction and liquid chromatography-tandem mass spectrometry, Anal. Bional. Chem. 405 (2013) 8913–8923. [30] J.A. Glaser, D.L. Foerst, G.D. McKee, S.A. Quave, W.L. Budde, 40 CFR part 136, Appendix B, Environ. Sci. Technol. 15 (1981) 1426–1435. [31] J. Munch, J. Eichelberger, T. Behymer, W. Budde, Method 525.2, US EPA, Washington D.C., USA, 1995. [32] K.K. Barnes, D.W. Kolpin, E.T. Furlong, S.D. Zaugg, M.T. Meyer, L.B. Barber, A national reconnaissance of pharmaceuticals and other organic wastewater contaminants in the United States – I) Groundwater, Sci. Total Environ. 402 (2008) 192–200. [33] P. Wang, S.L. Wang, C.Q. Fan, Atmospheric distribution of particulate- and gas-phase phthalic esters (PAEs) in a Metropolitan City Nanjing, East China, Chemosphere 72 (2008) 1567–1572. [34] G. Wang, K. Kawamura, S. Lee, K. Ho, J. Cao, Molecular, seasonal and spatial distributions of organic aerosols from fourteen Chinese cities, Environ. Sci. Technol. 40 (2006) 4619–4625. [35] P.Q. Fu, K. Kawamura, K. Okuzawa, S.G. Aggarwal, G. Wang, Y. Kanaya, Z. Wang, Organic molecular compositions and temporal variations ofsummertime mountain aerosols over Mt. Tai, North China Plain, J. Geophys. Res 113 (2008) 19107. [36] J. Wang, S.S.H. Ho, S. Ma, J. Cao, W. Dai, S. Liu, Z. Shen, R. Huang, G. Wang, Y. Han, Characterization of PM2.5 in Guangzhou, China: uses of organic markers for supporting source apportionment, Sci. Total Environ. 550 (2016) 961–971. [37] A.T. Lebedev, O.V. Polyakova, D.M. Mazur, M.A. Bol’shov, I.F. Seregina, Estimation of contamination of atmosphere of Moscow in winter, J. Anal. Chem. 67 (2012) 1039–1049. [38] J. Ma, L.L. Chen, Y. Guo, Q. Wu, M. Yang, M.H. Wu, K. Kannan, Phthalate diesters in Airborne PM2.5 and PM10 in a suburban area of Shanghai: seasonal distribution and risk assessment, Sci. Total Environ. 497-498 (2014) 467–474.

47

[39] S. Kong, Y. Ji, L. Liu, L. Chen, X. Zhao, J. Wang, Z. Bai, Z. Sun, Spatial and temporal variation of phthalic acid esters (PAEs) in atmospheric PM10 and PM2.5 and the influence of ambient temperature in Tianjin, China, Atmos. Environ. 74 (2013) 199–208. [40] P.Q. Fu, K. Kawamura, C.M. Pavuluri, T. Swaminathan, J. Chen, Molecular characterization of urban organic aerosol in tropical India: contributions of primary emissions and secondary photooxidation, Atmos. Chem. Phys. 10 (2010) 2663–2689. [41] P. Romagnoli, C. Balducci, M. Perilli, E. Perreca, A. Cecinato, Particulate PAHs and n-alkanes in the air over Southern and Eastern Mediterranean Sea, Chemosphere 159 (2016) 516–525. [42] M. Salapasidou, C. Samara, D. Voutsa, Endocrine disrupting compounds in the atmosphere of the urban area of Thessaloniki, Greece, Atmos. Environ. 45 (2011) 3720–3729. [43] J. Li, G. Wang, S.G. Aggarwal, Y. Huang, Y. Ren, B. Zhou, K. Singh, P.K. Gupta, J. Cao, R. Zhang, Comparison of abundances compositions and sources of elements, inorganic ions and organic compounds in atmospheric aerosols from Xi’an and New Delhi two megacities in China and India, Sci. Total Environ. 476–477 (2014) 485–495. [44] Z. Gu, J. Feng, W. Han, M. Wu, J. Fu, G. Sheng, Characteristics of organic matter in PM2.5 from an e-waste dismantling area in Taizhou, China, Chemosphere 80 (2010) 800–806. [45] Z. Jiao, Z. Guo, S. Zhang, H. Chen, H. Xie, S. Zeng, Novel extraction for endocrine disruptors in atmospheric particulate matter, Anal. Lett. 48 (2015) 1355–1366. [46] C.J. Weschler, Indoor-outdoor relationships for nonpolar organic constituents of aerosol particles, Environ. Sci. Technol. 18 (1984) 648–652. [47] Z. Xie, R. Ebinghaus, C. Temme, A. Caba, W. Ruck, Atmospheric concentrations and air-sea exchanges of phthalates in the North Sea (German Bight), Atmos. Environ. 39 (2005) 3209–3219. [48] M.J. Teil, M. Blanchard, M. Chevreuil, Atmospheric fate of phthalate esters in an urban area (Paris–France), Sci. Total Environ. 354 (2006) 212–223. [49] Z. Xie, R. Ebinghaus, C. Temme, R. Lohmann, A. Caba, W. Ruck, Occurrence and air-sea exchange of phthalates in the arctic, Environ. Sci. Technol. 41 (2007) 4555–4560.