Extraction of Nonylphenol and Nonylphenol ... - Springer Link

149 downloads 0 Views 234KB Size Report
Extraction of Nonylphenol and Nonylphenol. Ethoxylates from River Sediments: Comparison of Different Extraction. Techniques. V. Croce1 / L. Patrolecco2 / S.
Extraction of Nonylphenol and Nonylphenol Ethoxylates from River Sediments: Comparison of Different Extraction Techniques 2003, 58, 145–149

V. Croce1 / L. Patrolecco2 / S. Polesello3* / S. Valsecchi3 1 2 3

Department of Chemical, Physical and Mathematical Sciences, University of Insubria, Via Valleggio 11, Como, Italy C.N.R. Istituto di Ricerca sulle Acque, Via Reno 1, Roma, Italy C.N.R. Istituto di Ricerca sulle Acque, Via Mornera 25, Brugherio (MI), Italy; E-Mail: [email protected]

Key Words Column liquid chromatography Extraction methods River sediments Nonylphenol and nonylphenol ethoxylates

Summary Five different extraction techniques (Soxhlet, automated Randall, accelerated solvent extraction, microwave-assisted solvent extraction and extraction with a surfactant solution) have been evaluated for the determination of nonylphenol (NP) and nonylphenol ethoxylates (NPEO) in river sediments. All the techniques were applied to the same three samples collected from northern Italian rivers. The analyses were performed with two RP columns, with different stationary reversed phases—a classical C18 phase and a hexyl–phenyl phase. The recoveries and reproducibility of the different extraction techniques were comparable and all the methods gave reliable results. The variance of the results was dominated by the variance in sample homogeneity, sample preparation, and chromatographic analysis. A choice between the methods can be made on the basis of the cost and safety of each technique. Preliminary results obtained from use of a water-based extraction method with a surfactant solution (Tween-80), and its application to analysis of sediment and of worm tissue, are also presented.

Introduction Nonylphenol ethoxylates (NPEO) have been used for more than 40 years as detergents, emulsifiers, wetting agents, and dispersing agents. NPEO-containing products are used in many sectors,

Presented at: Chemical Analysis and Risk Assessment of Emerging Contaminants, Barcelona, Spain, November 28–30, 2002

Original DOI: 10.1365/s10337-003-0032-8 0009-5893/03/08 $03.00/0

including textile processing, pulp and paper processing, paints, resins, and protective coatings. Mainly because of their low cost NPEO are still being used in substantial amounts in institutional and industrial detergency [1]. NPEO can be biodegraded in the environment by stepwise loss of ethoxy groups to form lower ethoxylated congeners, carboxylated products, and nonylphenols (NP). The environmental dynamics of these compounds depend

largely on their physicochemical properties. Estimates based on logKows (nonylphenol, 4.48; nonylphenol-1-ethoxylate, 4.17; nonylphenol-2-ethoxylate, 4.20) and laboratory experiments show that these contaminants are hydrophobic and can accumulate in soils and sediments, but information on environmental concentration levels, persistence, and bioconcentration in organisms are sparse [1]. Because NP have been reported to cause estrogenic responses in a variety of aquatic organisms [2] studies have been performed to determine the environmental occurrence and fate of NP and its parent compounds [1–3]. Field studies have shown that the concentration of NP is usually low in treated effluents, because it degrades and is adsorbed by sludge particles. Adsorbed NP tends to be deposited and to accumulate in river sediments. Extraction of NP and parent compounds from sediments has been achieved by conventional means such as Soxhlet extraction with non-polar [4] or polar solvents [5, 6], or by ultrasonication in static [7, 8] or flowthrough mode [9]. Growing interest in the simultaneous determination of NPEO and their degradation product, NP, has prompted exploration of the feasibility of methods which use less solvent and are less time-consuming, for example dynamic supercritical-fluid extraction (SFE) with methanol-modified carbon dioxide [10, 11], pressurized liquid extraction with methanol in an SFE apparatus [12], or accelerated solvent extraction (ASE) with non-polar [13] or polar solvent mixtures [14]. Focused microwave energy and a

Chromatographia 2003, 58, August (No. 3/4)  2003 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH

145

polar solvent mixture (dichloromethane– methanol, 2:1) have been used to extract 4-tert-octylphenol from tissue samples [15], but application of microwaveassisted solvent extraction (MASE) to sediments has not been reported. The inclusion of 4-nonylphenol and octylphenol in the EC Priority List in the field of water policy [16] makes it urgent to review and discuss the different approaches to the extraction and determination of alkylphenol in sediments. Knowledge of the performance of the different extraction methods is the basis of setting realistic quality standards for different aquatic compartments. A few years ago we started a study comparing different extraction techniques for the simultaneous determination of NP and NPEO in river sediments starting with traditional methods, such as Soxhlet and Randall, and including more innovative techniques, such as ASE and MASE, to find a simple extraction method for monitoring purposes. Results from the comparison of ASE and Soxhlet have already been published [17]. The aim of this work was to collect and critically discuss all the data collected in our laboratory by use of different extraction techniques. Preliminary results on a water-based extraction method with a surfactant solution, and its application to sediment and to worm tissue analysis, are also presented.

Experimental Reagents Pesticide-grade acetone, pesticide-grade hexane, and HPLC gradient-grade methanol (Riedel–de Hae¨n, Seelze, Germany) were used without further purification. Technical grade 4-nonylphenol (4-NP; from Aldrich Chemie, Steinheim, Germany), 4-nonylphenol ethoxylate with 1 and 2 ethoxy units (NPEO(1,2); from ChemService, West Chester, PA, USA), and 4-NPEO with 9 and 10 ethoxy units (NPEO (9,10); ChemService) were used as standards. Technical t-octylphenol ethoxylate with an average number of 9.5 ethoxy units (Triton X-100, Fluka Chemie, Buchs, Switzerland) was used as internal standard. Polyoxyethylene(20) sorbitan monooleate (Tween 80) and Florisil were purchased from Aldrich. Neutral aluminium oxide for chromatography was purchased from

146

Fluka. Water for chromatography was purified (18 MW cm–1 quality) by means of a Milli-Q system (Millipore, Bedford, USA).

Sample Preparation Sediments with different organic carbon and alkylphenol content were collected with a grab from the Po river (Northern Italy), upstream (sample denoted ‘‘Po1’’) and downstream (sample ‘‘Po2’’) of the confluence of its most polluted tributary, the Lambro river, and from the Lambro river itself (sample ‘‘Lambro’’). The day after the collection the sediments were homogenised, freeze-dried, and stored in tightly closed brown glass bottles in a desiccator at room temperature until extraction. Laboratory-reared oligochaetes (Lumbriculus variegatus) were exposed to Po river sediment. At the end of the period of exposure the sediment was sieved and the worms were kept in clean water for gut purging. The worms were then frozen at )80 C until extraction.

Sample Extraction Soxhlet and Automated Randall Extraction

Freeze-dried sediment (approx. 5 g) was extracted in a Soxhlet apparatus with 400 mL methanol for 10 h (9 cycles h)1). An automated extractor (SER 148, Velp Scientifica, Usmate, Italy), based on Randall technique, was also used. The sample (5 g), in a glass fibre thimble, was immersed for 2 h in 120 mL hot methanol (instrument plate temperature 260 C) and then washed with solvent under reflux for 3 h. The frozen oligochaete samples (approx. 1 g f.w.) were mixed with anhydrous sodium sulphate and extracted with hexane–acetone (1:1 v/v) in the Soxhlet or Randall apparatus. The other conditions were the same as reported for sediment samples.

Accelerated Solvent Extraction (ASE)

ASE extraction (ASE 200, Dionex, Sunnyvale, CA, USA) was performed by loading an 11-mL stainless steel extraction cell with the freeze-dried sediment (approx. 6 g). Two 5-min cycles of extraction with methanol (approx.

15 mL) were performed at 100 C and 100 atm, in static mode (flush 60%, purge 90 s).

Microwave-Assisted Solvent Extraction (MASE)

Microwave-assisted solvent extractions were performed using an Ethos SEL system (Milestone, Sorisole, Italy). Freezedried sediment (5 g) was weighed into the PTFE liner of the extraction vessel and 40 mL methanol were added. The extractions were performed with magnetic stirring at 120 C for 20 min. After cooling of the samples the supernatant was decanted and combined with a 5-mL methanol rinse of the extracted sample. The combined solution was centrifuged at 6000 rpm for 20 min and the collected supernatant was treated as described in the ‘‘Clean-up’’ section (below).

Extraction with a Surfactant (Tween 80)

Freeze-dried sediments (5 g) were added to 100 mL of 10 g L)1 Tween 80 aqueous solution. The suspension was then mixed thoroughly at room temperature by magnetic stirring (approx. 300 rpm) for 3 h and then centrifuged at 10 000 rpm for 15 min. After centrifugation the supernatant was subjected to a preconcentration and clean-up step by solid-phase extraction (SPE) on C18 cartridges (Strata C18-E cartridges, 50 lm, 500 mg/6 mL; Phenomenex, Torrance, CA, USA). The cartridges were conditioned with methanol, acetone, and Milli-Q water (5 mL of each) at a flow rate of 3 mL min)1. After loading of the sample (at approximately 5 mL min)1) and subsequent washing with 10 mL Milli-Q water, the cartridges were dried under vacuum for 45 min. Analyte elution was achieved with 3 · 10 mL acetone. The extracts obtained in this way were then concentrated to an approximate volume of 0.5 mL under a gentle stream of nitrogen and reconstituted in methanol–water, 60:40 (v/v), to a final volume of 1 mL for HPLC analysis. Frozen oligochaetes (approx. 1 g) were mixed with anhydrous sodium sulphate and added to 20 mL of 10 g L)1 Tween 80 aqueous solution. The extraction was performed in the same manner as for the sediment samples.

Chromatographia 2003, 58, August (No. 3/4)

Original

Clean-up of Solvent Extracts Methanol extracts were concentrated to 1–2 mL by rotary evaporation under vacuum and transferred on to a column (1.5 cm i.d. · 4.5 cm length) containing neutral alumina deactivated with 15% water; the column had previously been washed with methanol. Hexane–acetone extracts were evaporated just to dryness, re-dissolved in 1–2 mL of methanol, and purified on an activated Florisil column (1.5 cm i.d., 4.5 cm length) which had previously been washed with methanol. Both alumina and Florisil columns were then eluted with 15 mL 10% acetic acid in methanol. The purified extracts were concentrated to 0.5 mL under a gentle stream of nitrogen using a TurboVap-II workstation (Zymark, Hopkinton, MA, USA) and filtered through a 0.45-lm PTFE filter before injection.

Figure 1. Chromatograms obtained from analysis of: (a) a river sediment (sample Po2) on column 1, after extraction in a Soxhlet apparatus and (b) the same sample extracted with Tween-80 surfactant and analysed on column 2.

Analysis Analytical separations were performed with a HPLC system comprising a PU1580 gradient pump, with 10-lL sample loop, and an FP-920 fluorescence detector set at kexc. 230 nm and kem. 302 nm (Jasco, Tokyo, Japan). Two columns were used for the analysis. Column 1 was 250 mm · 4.6 mm i.d. containing 5 lm LiChrospher 60 RP-select B (Merck, Darmstadt, Germany), at 25 C. The mobile phase, at a constant flow rate of 1 mL min)1, began with a 5 min isocratic step at 30% water and 70% methanol, followed by a 20 min linear gradient to 20% water and 80% methanol then another 5-min isocratic step at the final mobile phase ratio. This enabled separation of NP and total NPEO with TritonX as internal standard (Figure 1a). Column 2 was 150 mm · 4.6 mm i.d. containing 5 lm Luna Phenyl–Hexyl (Phenomenex, Torrance, CA, USA) at 25 C. The mobile phase, at a constant flow rate of 1 mL min)1, began with a 5 min isocratic step at 40% water and 60% methanol, followed by a 25 min linear gradient to 20% water and 80% methanol then another 10 min isocratic step at the final mobile phase ratio. This enabled separation of NP and the first oligomers of NPEO, nonylphenol(1)ethoxylate (NPEO-1) and nonylphenol(2)ethoxylate (NPEO-2) (Figure 1b). Original

Results and Discussion Soxhlet extraction is the standard method for analysis of sediment but it takes long time and needs high-quantity solvent. For this reason we tested the feasibility of alternative methods of extraction for simultaneous determination of NP and its ethoxylates from river sediments and biota. The methods tested were automated Randall extraction, accelerated solvent extraction (ASE), and microwave-assisted solvent extraction (MASE) using the same solvent—methanol. Optimisation and validation of ASE have been described elsewhere [17]. Together with these instrumental extraction methods we are testing the feasibility of using an aqueous surfactant solution (polyoxyethylene(20)sorbitan monooleate, trade name Tween-80) as extracting medium. The surfactant has been used for the determination of chemicals in living cells in metabolic pathway studies [18] and we are applying it in an extraction method for the determination of lipophilic molecules such as NP and NPEO in sediment and biota (Patrolecco et al., in preparation). A useful property of this surfactant is that it is transparent to fluorescence detection. Except the Tween 80 extracts, which underwent a specific concentration step Chromatographia 2003, 58, August (No. 3/4)

by SPE, other extracts were subjected to a similar clean up procedure. All the extracts were analysed by RP HPLC with direct fluorescence detection at kexc. 230 nm and kem. 302 nm. RP HPLC with fluorimetric detection is considered a reliable method for routine analysis of alkylphenols and their ethoxylates in environmental samples, because it enables separation and quantification of the different homologues and oligomers by alkyl chain length [6]. Because C18 columns cannot resolve the ethoxylate oligomers, nonylphenol ethoxylate was determined as the total sum of NPEO oligomers (NPEOtot) (Figure 1a). To obtain baseline resolution of the first ethoxylate oligomers (NPEO1, NPEO2), which have a significant estrogenic effect, we used an alternative separation column (Luna Phenyl–Hexyl) with a shorter and less lipophilic chain and greater selectivity for aromatic compounds (Figure 1b). The two columns were shown to be complementary—the former enabled determination of total NPEO content and the latter enabled quantification of single low-molecular-weight ethoximers with more efficiency and longer column lifetime than a classical normal-phase column. The comparability of the different extraction methods was evaluated by

147

Table I. Nonylphenol (NP) and nonylphenol ethoxylate (NPEO) concentrations obtained by use of different extraction techniques. Results are expressed as mean ± SD; n = number of replicates. Extraction technique Lambro sample Soxhlet Automated Randall ASE Tween-80 Po1 sample Soxhlet Automated Randall MASE ASE Tween-80 Po2 sample Soxhlet Automated Randall MASE ASE Tween-80

Column

NP (lg g)1)

NPEO1 (lg g)1)

NPEO2 (lg g)1)

NPEOtot (lg g)1)

n

Ref.

1 2 2 1 2

4.7 ± 0.3 2.1 2.6 2.9 ± 0.6 2.2 ± 0.4

– 1.1 1.5 – 1.2 ± 0.3

– 0.6 0.5 – 0.6 ± 0.2

4.1 ± 0.1 – – 5.7 ± 0.7 –

3 2 2 5 6

[17] This work This work [17] This work

1 2 2 1 1 2

0.4 ± 0.3 0.3 0.3 0.34 ± 0.04 0.3 ± 0.1 0.3

– 0.5 1.0 – – 0.5

– 0.2 0.2 – – 0.2

2.9 ± 0.8 – – 1.4 ± 0.1 1.5 ± 0.5 –

6 2 2 3 5 2

[17] This This This [17] This

1 2 2 1 1 2 2

2.8 2.6 2.4 2.2 2.9 2.6 2.4

– 4.0 3.4 – – 3.0 3.6 ± 0.2

– 1.2 1.4 – – 1.3 0.92 ± 0.04

11.2 ± 2.5 – – 11.4 ± 0.9 10.4 ± 0.6 – –

4 2 2 3 5 2 5

[17] This This This [17] This This

± 0.8 ± 0.4 ± 0.6 ± 0.3

analysing the same three samples collected in the Po river basin (Northern Italy), in the most industrialised and urbanised region of Italy. These samples had been freeze-dried and stored in a desiccator for the last three years and can be considered a secondary reference material for our laboratory. Table I shows that results from all the extraction methods are in good agreement except those from Soxhlet extraction and C18 column (Column 1) analysis of the Lambro river sediment. This is because this sample has the most difficult matrix and Soxhlet extraction suffers from interferences. The interference problem has been overcome by using a more selective column, for example Column 2, based on the phenyl–hexyl chain. Analyses with the C18 column usually gave slightly higher NP concentrations than those performed with the phenyl– hexyl column, because of partial overlapping of the NP peak with that of NPEO (Figure 1a). Nevertheless, the complementary use of the two columns furnished more complete information about nonylphenol and nonylphenolrelated compound distribution in the environment. Because no certified reference materials (CRM) for NP and NPEO were available commercially, the accuracy of some of the extraction techniques was verified by performing recovery experiments on sediment samples spiked at the 10 lg g)1 level. The mean recoveries obtained by use of ASE (85 ± 22%

(n ¼ 13) and 87 ± 13 (n ¼ 15) for 4-NP and 4-NPEO, respectively), Soxhlet extraction (79 ± 24% and 82 ± 12% for 4-NP and 4-NPEO, respectively) [17], and the Tween-80 method (85–94%) (Patrolecco et al., in preparation) are comparable with those reported by Shang et al. (65–93%) [4] and Petrovic et al. (89–94%) [14] by use of ASE. No recovery studies have been performed for the Randall and MASE methods; the results listed in Table I show, however, that these two methods are as accurate as the other techniques. Limits of detection (LOD) of the analytical determination, calculated as three times the standard deviation of analyte standard solutions (n ¼ 10) at concentrations near the detection limits, were 1 mg L)1 for column 1 and 0.5 mg L)1 for column 2; for a sample mass of 5 g and a final volume of 0.5 mL these correspond to method LOD of 0.1 and 0.05 lg g)1, respectively. Taking into account the total analytical variance and defining acceptable precision as better than 30% [19], an effective quantitation limit of 0.1–0.3 lg g)1 for MASE, ASE, and Tween extraction, and 0.5 lg g)1 for Soxhlet can be estimated. The precision of the extraction methods was evaluated from the standard deviations of a series of extractions (n ¼ 3–5) performed on different days (Table I). RSD for all the methods usually ranged from 10 to 30%, with the worst values (33–78%) for the lowest concentration range (0.3–0.4 lg g)1),

148

Chromatographia 2003, 58, August (No. 3/4)

work work work work work work work work work

near the detection limit. RSD values are an estimate of the total analytical variance, including sample homogeneity, extraction, and chromatographic determination, the latter having coefficients of variation of 2% at the highest calibration level (20 mg L)1) and 30% at the lowest (0.5 mg L)1). Our results were rather worse than RSD (4–5%) reported by Ding and Fann [12] for determination of 4-NP in sediments by pressurised liquid extraction; in the latter work, however, samples were spiked with a much higher concentration (20 lg g–1) of 4-NP and a more sensitive and precise analytical method (GC–MS) was used. Gas chromatographic analysis cannot be used for simultaneous determination of nonylphenol and its ethoxylates because the latter are not sufficiently volatile. LC coupled with mass spectrometric detection might be the technique of choice for this kind of analysis because the detection limit (0.5– 2 ng g)1) is very low and reproducibility is good, with RSD ranging from 3.6 to 13.6%, after ASE extraction of sediment samples spiked at the 0.1 lg g)1 level [14]. Preliminary results from the determination of NP, NPEO1, and NPEO2 in worm tissue with three different techniques (Soxhlet, Randall and Tween-80 extraction) are shown in Table II. The agreement among the three methods is rather surprising but it can be explained by the very good homogeneity of the samples: the worms analysed had, indeed, Original

Table II. Determination of nonylphenol (NP) and nonylphenol-1 and 2-ethoxylates (NPEO1 and NPEO2) in oligochaetes after extraction by use of different techniques; n = number of replicates. Extraction technique

Column

NP (lg g)1)

NPEO1 (lg g)1)

NPEO2 (lg g)1)

n

Ref.

Soxhlet Automated Randall Tween-80

2 2 2

0.9 0.9 1.0

3.8 4.0 4.1

0.5 0.3 0.2

1 1 2

This work This work This work

Table III. Comparison of solvent volumes and extraction times among different extraction techniques. Extraction technique

Organic solvent volume (mL)

Extraction and cooling time (min)

Soxhlet Automated Randall MASE ASE Tween-80

400 120 40 30 0

600 300 40 30 180

been exposed to carefully homogenised sediment under controlled laboratory conditions. These preliminary data confirm that reduction of the total variance can be achieved by careful homogenisation of environmental samples.

Conclusions The recovery and the reproducibility of the different extraction methods are comparable and all methods give reliable results. The variance of the results is dominated by the variance in sample homogeneity, sample preparation, and chromatographic analysis. The choice between the methods can be made on the basis of the cost and safety of each extraction technique. Differences between solvent and time consumption among the methods are listed in Table III. The small amount of solvent consumed, the reduced extraction time, and a real improvement in operator safety are the most important advantages of innovative extraction methods such as ASE and MASE. One significant drawback of MASE compared with ASE is the need for sample centrifugation and filtration, which can have critical effects on

Original

analytical accuracy. This drawback can be overcome by using proper accessories produced for automatic sample handling. An interesting environmentally-friendly approach is the use of aqueous surfactant solution (Tween-80) for extraction of lipophilic compounds from sediment and biological matrices. This method needs SPE cartridges, instead of solvent, but the main drawback is that the final solution contains part of the surfactant used for extraction and cannot be injected into a gas chromatograph without further purification.

Acknowledgements Authors thank Velp Scientifica, Dionex Italia and FKV for providing the instrumental facilities for method comparison. The support of A. Pagnoni and G. Passoni in analytical work is gratefully acknowledged.

References 1. Environmental risk assessment of nonylphenol straight chain and branched chain carried out in the framework of Council

Chromatographia 2003, 58, August (No. 3/4)

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

regulation (ECC) 793/93 on the evaluation and control of the risk of existing substances, Bruxelles, 2001 Servos M (1999) Water Qual Res J Can 34:123–177 Bennie DT (1999) Water Qual Res J Can 34:79–122 Shang DY, Ikonomou MG, Macdonald RW (1999) J Chromatogr A 849:467–482 Marcomini A, Pojana G, Sfriso A, Quiroga Alonso JM (2000) Environ Toxicol Chem19:2000–2007 De Voogt P, De Beer K, Van Der Wielen F (1997) Trends Anal Chem 16:584–595 Blackburn MA, Kirby SJ, Waldock MJ (1999) Mar Pollut Bull 38:109–118 Petrovic M, Barcelo D (2000) Anal Chem 72:4560–4567 Ferguson PL, Iden CR, Brownawell BJ (2000) Anal Chem 72:4322–4330 Kreisselmeier A, Du¨rbeck HW (1997) J Chromatogr A 775:187–196 Daniels WM, House WA, Rae JE, Parker A (2000) Sci Total Environ 253:81–92 Ding W, Fann JCH (2000) J Chromatogr A 866:79–85 Shang DY, Macdonald RW, Ikonomou MG (1999) Environ Sci Technol 33:1366– 1372 Petrovic M, Lacorte S, Viana P, Barcelo D (2002) J Chromatogr A 959:15–23 Pedersen SN, Lindholst C (1999) J Chromatogr A 864:17–24 Decision no. 2455/2001/EC of the European Parliament and of the Council of 20/ 1/2001 (2001) Off J Eur Commun L331:1–5 Valsecchi S, Polesello S, Cavalli S (2001) J Chromatogr A 925:297–301 Nandakumar R, Gounot AM, Mattiasson B (2000) J Biotechnol 83:211–217 Eurachem (1998) Eurachem guide The fitness for purpose of analytical methods. Laboratory of the Government Chemist, Teddington UK Received: Dec 20, 2002 Revised manuscript received: Mar 18, 2003 Accepted: Apr 28, 2003

149