Determination of Surfactants and Some of Their Metabolites in

---

Environ. Sci. Technol. 2002, 36, 4156-4161

Determination of Surfactants and Some of Their Metabolites in Untreated and Anaerobically Digested Sewage Sludge by Subcritical Water Extraction Followed by Liquid Chromatography-Mass Spectrometry FEDERICA BRUNO, ROBERTA CURINI, ANTONIO DI CORCIA,* IGOR FOCHI, MANUELA NAZZARI, AND ROBERTO SAMPERI Dipartimento di Chimica, Universita` “La Sapienza”, Piazza Aldo Moro 5, 00185 Roma, Italy

Enormous amounts of sewage sludge are worldwide generated and released into the environment. Analysis of the most common and/or toxic chemicals in sludge should be mandatory before deciding its destination. Surfactants and some of their breakdown products are invariably the most common organic contaminants in domestic sewage sludge. For determining these compounds, we have developed a method based on extraction with subcritical water followed by liquid chromatography-mass spectrometry. On extracting surfactants and their metabolites from 50 mg of sludge, the efficiency of the water extraction device was evaluated in terms of pH of the extractant, temperature, and time of the static extraction. The best extraction conditions were obtained by using carbonate buffer (pH 9.4) at 200 °C as extractant, 10 min of static extraction at the pressure of 100 bar followed by 17 min of dynamic extraction. Analyte collection was performed by inserting a solid-phase extraction cartridge downstream the extraction cell. Compared to 16-h Soxhlet extraction with methanol, this procedure was remarkably more efficient in extracting anionic surfactants and acidic metabolites of nonylphenol ethoxylates (NPECs). A short survey was conducted to estimate concentration changes of target compounds after 14-d sludge anaerobic digestion. Results showed that 5474% of both neutral and weakly acidic ethoxylate species were removed after residence of the sludge in the digester. On the contrary, little, if any, removal of anionic surfactants was observed after the digestion treatment. As expected, the level of nonylphenol increased under anaerobic conditions.

Introduction Sewage treatment plants (STPs) are the barriers of the environment, where undesired substances dissolved in sewages are biotransformed and/or sorbed to sludge. Dis* Corresponding author phone: +39-06-49913752; fax: +39-06490631; e-mail: antonio.dicorcia@uniroma1. 4156

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

posal of sewage sludge poses serious problems to municipalities. In Japan alone, it has been estimated that about 50 million m3 of sewage sludge are annually generated (1). The Directive COM 91/271 (2) issued in Europe states that towns with more than 2000 inhabitants will have to provide STPs by 2005. In 1999, it was predicted this Directive would have almost doubled the sewage sludge production in Spain in 1 year (3). Most of these enormous quantities of sewage sludge are disposed of by landfilling or burning. Minor fractions of sewage sludge variable from country to country are used as effective fertilizer in agricultural soils. Sewage sludge contains a large number of organic contaminants of anthropogenic origin whose nature and concentration may vary depending on season, source of wastewater, and type of sewage treatment. Whatever the sludge destination is, persistent chemicals contained in it may contaminate aquifers via leaching through soil or enter the food chain. Therefore, analysis of the most common and/or toxic substances in sludge should be done before deciding the destiny of sewage sludge. Among the known organic contaminants, surfactants are the most abundant species in STP sludge receiving domestic sewage (4). Sewage sludge is considered one of the most complex matrices to be analyzed. In the past, analytical procedures were developed for measuring only one (5-8) or two classes of surfactants similar in nature (9) or some of their most toxic breakdown products, mainly those deriving from biotransformation of nonylphenol ethoxylates (NPEs) (10-14). Only recently, efforts have been addressed to simultaneously determine a large number of surfactants and some of their metabolites (15). Extraction procedures adopted in many standardized analytical methodologies for determining contaminants in solid matrices are time-consuming and make use of large volumes of toxic, expensive, and inflammable solvents. Hawthorne et al. (16, 17) reported that the “environmentally friendly” water efficiently extracted both polar and nonpolar species from solid matrices by selecting an appropriate extraction temperature. This finding was explained considering that the polarity of water steadily decreases as the temperature is increased, so making water more and more capable of competing with nonpolar organics for adsorption on a solid matrix. In addition, water should be abler than conventional solvents to extract polar analytes, when their sorption to solid matrices is mainly controlled by specific interactions. In our lab, water alone (18) or containing a phosphate buffer (19, 20) has been effectively used as extractant for analyzing several classes of herbicides and their degradation products in soil. The purpose of this work has been that of evaluating the feasibility of using hot water as extractant and liquid chromatography (LC)/mass spectrometry (MS) with an electrospray (ES) ion source for simultaneously analyzing several classes of surfactants, i.e., linear alkylbenzenesulfonates (LAS), secondary alkanesulfonates (SAS), alkylsulfates (AS), alkylethoxysulfates (AES), NPE and alcohol ethoxylates (AE), major LAS coproducts, i.e., dialkyltetralinsulfonates (DATS), octylphenol (OP) and breakdown products of NPE, i.e., nonylphenol (NP) and nonylphenol ethoxy carboxylates (NPEC), in sewage sludge.

Experimental Section Reagents and Chemicals. Individual pure C-10 through C-18 ethoxylates with an even number of carbon atoms and six ethoxy units, denominated as A10E6-A18E6, and A12E with 2, 3, 4, 6, and 8 ethoxy units, branched NP and 4-tert-OP were 10.1021/es020002e CCC: $22.00

 2002 American Chemical Society Published on Web 09/05/2002

all supplied by Fluka, Buchs, Switzerland. Marlon A and Marlophen 810 are two commercial surfactant mixtures supplied by Chemische Werke Hu ¨ ls AG, Marl, Germany. Marlon A is a calibrated mixture of C10-C14 LAS. Marlophen 810 is an uncalibrated mixture of NPE alkyl chain isomers and oligomers with an average of 11 and a range of 2-20 ethoxy units. “Sirene-AlCl3”, a mixture of C10-C14 LAS containing about 10% DATS, pure C12-LAS, and C6-DATS (unknown purity) were kindly supplied by L. Cavalli (Enichem Augusta, Milan, Italy). C6-DATS was purified following a previously described procedure (21). AEs used were a commercial blend designated as Lialet 125 and kindly supplied by Henkel. This mixture contains an undetermined content of AE homologues having alkyl chain lengths from C-12 to C-15 and oligomers with an average of 10 ethoxy units. Three individual pure SAS having respectively alkyl chains with 12, 14, and 18 carbon atoms were kindly supplied by J. A. Field, while an uncalibrated mixture of C13-C18 SAS (Hostapur 60) was from Hoechst. Three individual pure AS species with alkyl chains containing respectively 12, 14, and 16 carbon atoms, an uncalibrated C12-C18 AS mixture, and a commercial mixture of AES surfactants having alkyl chain lengths from C-12 to C-15 and an average of 3 ethoxy units were kindly supplied by Enichem Augusta. Pure A10E6, C-8 LAS, and octylphenol with a linear alkyl side chain (LOP) were purchased by Fluka and were used as internal standards. A calibrated solution of NPE1C was kindly supplied by A. Marcomini. Stock standard solutions containing individual species or mixtures of them were prepared by dissolving them in methanol. Working standard solutions were prepared by further diluting with methanol. For LC-MS analysis, methanol “Plus” of gradient grade was obtained from Carlo Erba, Milan, Italy. All other solvents and chemicals were of analytical grade (Carlo Erba), and they were used as supplied. Extraction cartridges filled with 1 g Carbograph 4 and drilled cylindrical Teflon pistons and Luer tips for analyte elution in the back-flushing mode (22) were supplied by LARA, Rome, Italy. Carbograph 4 is an example of the graphitized carbon black sorbent family with a surface area of about 200 m2/g. It is commercially referred to also as “Carboprep” (Restek, Bellefonte, PA). The SPE cartridge was fitted into a sidearm filtration flask, and liquids were forced to pass through the cartridge by vacuum (water pump). Before processing sludge extracts, the cartridge was washed with 20 mL of HCl-acidified water (pH 2), followed by 5 mL of water. Sludge Samples. A composite (24-h) sludge sample that was anaerobically digested for 14 days at 33 °C and a corresponding 24-h composite untreated sludge sample were collected in September 2000 from the Roma Nord activated sludge STP located in the area of Rome, Italy. This plant receives mainly domestic wastewaters. Four months before, a grab sample of untreated sewage sludge was taken at the same plant and used for developing the analytical method. All samples were air-dried at room temperature for 48 h and then ground to a fine powder in a coffee grinder. After grinding, sludge samples were sieved to a particle size of less than 2 mm. Residual water in sludge particles was determined by heating a small sludge sample at 100 °C for 5 h. By weighing the sludge sample before and after thermal treatment, a 12% of residual water was measured. When unused, sludge samples were stored at -18 °C. Extraction Apparatus. Compared to the apparatus used in other works (18, 19), whose design is shown in a previous paper (18), the extraction apparatus was modified by inserting a needle valve (static/dynamic valve) between the extraction cell and the Carbograph 4 cartridge. In addition, a fusedsilica capillary restrictor (50 µm i.d.) was used downstream of the needle valve to maintain a pressure of 100 bar inside

the extraction cell during dynamic extraction. To avoid plugging of the extraction device by fine sludge particles, the lower frit of the extraction cell was renewed after each extraction. Extractant Preparation. A carbonate buffer (pH 9.4) was used as extractant. It was prepared by dissolving 45 and 10 mmol of respectively NaHCO3 and Na2CO3 in 1 L of distilled and He-deaerated water. Extraction Procedure. The extraction cell was half-filled with quartz sand (prewashed by 8-h Soxhlet extraction with methanol). Fifty milligrams of sludge was then placed into the extraction cell. Any void space was filled with quartz sand. After introducing the extraction cell into the GC oven, the following procedure was followed: 1. The oven was heated to 200 °C. After the oven reached the selected temperature, 5 min were allowed for equilibration. 2. Opening the needle valve, the extraction cell was filled with 3 mL of extractant. 3. With the pump still working, the needle valve was closed and the pump was switched off after the extractant pressure in the extraction cell reached 100 bar. Under this condition, 10 min were allowed for static extraction. 4. After opening the valve, the pump was again switched on allowing 17 mL of the extractant flushing the extraction cell at a flow rate of 1 mL/min and at the pressure of 100 bar. 5. The extraction procedure was terminated after a total of 20 mL of the extractant flowed through the Carbograph 4 cartridge. The Carbograph 4 cartridge was then disconnected and fitted into a sidearm filtering flask. Inorganic salts were removed from the cartridge by passing 100 mL of distilled water at a flow rate of ca. 80 mL/min. Water remaining in the Carbograph 4 cartridge was partially removed by drawing room air through the cartridge for 1 min. The water content was further decreased by slowly passing 1 mL of methanol. Again, the trap was air-dried for 1 min. By exploiting a singular feature of the Carbograph 4 material (23), neutral and weakly acidic analytes were isolated from strongly acidic ones by differential elution, following partially a procedure reported elsewhere (23). Briefly, nonacidic and weakly acidic compounds were eluted by passing through the sorbent bed 2 mL of methanol followed by 12 mL of CH2Cl2/CH3OH (80:20, v/v) acidified with 50 mmol/L of formic acid (fraction A). This fraction contained AEs, NPEs, NPEC, OP, and NP. Then, a Teflon piston was inserted into the cartridge until it touched the upper frit, the cartridge was reversed, and anionic species were back eluted by 12 mL of CH2Cl2/CH3OH (80:20, v/v) solution basified with 5 mmol/L NaOH (fraction B). This fraction contained components of LAS, DATS, AS, AES, and SAS. No significant amount of the analytes in fraction A was found in fraction B and vice versa. The flow rate at which both eluent systems passed through the cartridge was about 4 mL/min obtained by suitably regulating the vacuum in the apparatus. The two above fractions were separately collected in 1.4 cm i.d. glass vials with conical bottoms. Before solvent removal, fraction B was in part neutralized by adding 90 µL of HCl, 0.5 mol/L. Solvents were in part or totally removed at 40 °C in a water bath under a gentle flow of nitrogen. In particular, fraction A was concentrated down to about 100 µL to minimize evaporative losses of OP and NP, while fraction B was dried. Before LC-MS analysis, 10 µL of a methanol/ water (70:30, v/v) solution containing respectively 20 and 200 ng/µL of A10E6 (IS) and LOP (IS) was added to the concentrated fraction A. The residue of the fraction B was reconstituted with 500 µL of a methanol/water (60:40, v/v) solution containing 5 ng/µL C-8 LAS (IS). LC-MS Analysis. Liquid chromatography was carried out with a Varian Model 9010. The analytes were chromatographed on an “Alltima” 25 cm × 4.6 mm i.d. column filled VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4157

TABLE 1. Some ES/MS Instrumental Conditions and Ions Selected for Identifying and Quantifying Target Compounds compound

molecular ions, m/z

acquisition mode

full-scan, m/z

A12+RE2+βa,b

275+14R+44βc 292+14R+44βd 297+14R+44β,e 326+44β,d 331+44βe 277+44β 205 219 297+14R 295+14R 265+14R 277+14R 309+14R+44β

PI

270_610 in 3s

NI

200_500 in 2s

NPE2+β NPE1+βC OP NP C10+R-LAS C6+R-DATS A12+RS C14+R-SAS A12+RE1+βS

a R ) 0 to 3-5 (additional number of alkanoyl groups). protonated, ammoniated, and sodiated adduct ions.

265_440 in 2s

b

9

89, 133 89, 133 219 132, 133 119, 133 80, 183 209 96 80 96

β ) 0 to 2-5 (additional number of ethoxy units). c -em/z refers respectively to

with 5-µm C-18 reversed-phase packing (Alltech, Sedriano, Italy). For fractionating neutral, weakly acidic and strongly acidic compounds, the mobile phase consisted invariably of methanol (solvent A) and water (solvent B), both containing 0.2 mmol/L ammonium acetate. The flow rate of the mobile phase was 1 mL/min and 50 µL of the column effluent was diverted to the ES source. A “Platform” benchtop mass spectrometer (Micromass, Manchester, U.K.) consisting of a pneumatically assisted ES interface and a single quadrupole was used for detecting and quantifying target compounds in the LC column effluent. A voltage of 4 kV was applied to the capillary when the ES/MS system was operated in the positive ionization (PI) mode, while the voltage was decreased to 3 kV when operated in the negative ionization (NI) mode. The source temperature was maintained at 70 °C. Neutral (AEs and NPEs) and weakly acidic analytes (OP, NP, NPECs) present in the fraction A were separately analyzed by injecting two times 20-µL aliquots of the extract A. One chromatographic run served to detect neutral analytes in the PI mode, while the second run was used for detecting weakly acidic analytes in the NI mode. In both cases, the solvent A was linearly increased from 75% to 100% after 30 min. Strong acidic surfactants (LAS, DATS, AES, AS, SAS) in fraction B were chromatographed by linearly increasing the solvent A percentage from 60 to 100 over 40 min and detected in the NI mode. For determining anionic surfactants other than C11-C13 LAS, 50 µL of the fraction B was injected into the LC column. Quantitation of C11-C13 LAS was performed by reinjecting only 10 µL of this fraction B, under the same instrumental conditions. The latter chromatographic run was necessary for accurate LAS measurements because, under the former condition, the amounts of C11-C13 LAS injected exceeded the upper limit of the linear dynamic range of the ES/MS detector. In all cases, MS data acquisition was performed by setting the sample cone voltage to a low value (20 V). Under this condition, only (quasi-)molecular ions of all of the analytes considered were detected by the MS system, this greatly simplifying the quantification procedure. Confirmatory ions (fragment ions) for unambiguous identification of target compounds were obtained by occasionally reanalyzing extracts, after suitably increasing the skimmer cone voltage. Ions considered for identification and quantitation of target compounds together with some specific instrumental conditions are listed in Table 1. Quantitation. It has to be said that quantitation of those analytes for which no standard was available was rather laborious and could suffer from some inaccuracy. Quantitation of AyEx andNPEx Species. As shown elsewhere (24), the ES/MS detector gives a poor response for any compound bearing only one ethoxy unit. Therefore, the presumably low concentrations of the various AyE1 (y ) 124158

confirmatory ions, m/z

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

16) homologues and NPE1 in sewage sludge could not be quantified by this method. Quantification of AE2-AE6 with an even number of carbon atoms in the hydrophobic chain was accomplished as previously reported (24). Quantification of any AE with an odd number of carbon atoms was made by averaging molar response factors of the two homologues preceding and following it. No individual standard of NPEs were available for this study. To quantify these species in sludge, we assumed that, like AEs, their molar response factors were influenced by the ethoxy chain length and independent of the nature of the hydrophobic moiety (24, 25). Quantitation of OP, NP, and NPExC. With the ES/MS system operated in the NI mode, molar response factors of NP and NPE1C relative to LOP (IS) were calculated following a procedure analogous to that reported above. LOP has a linear alkyl chain and was used as IS since it did not interfere with the analysis of native OP in sludge. The latter compound has a multibranched alkyl chain, thus eluting well before LOP. Quantitation of NPEC oligomers higher than NPE1C was performed after standardization of a mixture of NPEC mixture obtained by oxidizing a NPE mixture (Marlophen 810) with the Jones reagent, under conditions reported elsewhere (26). After one 100-fold dilution of the reaction mixture with water, separation of NPE2C-NPE5C oligomers from each other and from unreacted NPEs was achieved by injecting a known aliquot of the above solution into a LC apparatus equipped with a C-8 column (Supelco, Bellefonte, PA) and a fluorimetric (FL) detector, following conditions reported elsewhere (26, 27). Under the same experimental conditions, a known volume of the NPE1C standard solution was injected into the column. Assuming the fluorescence molar quantum efficiency for each of the above NPECs is unaffected by the ethoxy chain length, quantification of NPE2C-NPE5C oligomers in the water-diluted reaction mixture was performed by comparing their peak areas with that of NPE1C. After standardization, the NPEC mixture was reinjected into the LC-ES-MS apparatus, and molar response factors of NPE2C-NPE5C relative to LOP were calculated. We observed that ionization efficiencies of NPE2C-NPE5C oligomers were not appreciably influenced by the ethoxy chain length and gave molar responses about 0.8 times lower than that of NPE1C. Quantitation of Anionic Surfactants. Calibration curves of C12-LAS, C6-DATS, C14-SAS, C12-AS, and C8-LAS (IS) were constructed by flow injection analysis. The electrosprayed solution was composed of a water/methanol (80:20, v/v) solution containing 0.2 mmol/L ammonium acetate. These curves showed that the response of the ES/MS detector was linearly related to injected amounts of all the above compounds up to 2 µg. Moreover, molar response factors for all the above compounds did not differ significantly from each other. Thus, we reasonably assumed that all the anionic

surfactants here considered for which standards were not available gave the same molar response factors as that of C8-LAS.

TABLE 2. Concentrations of Selected Analytes Measured in Sewage Sludge by Increasing the Extraction Temperature concentrationa (RSD, %), mg/kg (dry weight)

Results and Discussion Recovery Studies. Assessing the extraction efficiency for organics in environmental solid matrices by spike recovery studies may not be a reliable means of representing the extraction behavior of native analytes (28). Therefore, recovery studies performed by us were addressed to ascertain if analyte losses could occur owing to their thermal decomposition or saturation of the sorbent material used to collect analytes coming out from the extraction cell. These experiments were performed by filling the extraction cell with quartz sand spiked with amounts of target compounds comparable to those ones that may be present in sewage sludge and analyzing. Analyses were performed as reported in the Experimental Section. Triplicate experiments showed that analyte recoveries were better than 87% with relative standard deviations not larger than 8%. When analyzing complex matrices by LC-ES-MS, partial signal suppression of a target compound can occur if it is coeluted with huge amounts of another compound (29). In this work, the extent of the matrix effect was evaluated by injecting analytes from pure solutions, undigested and digested sewage sludge extracts. Addition of the analytes to sludge extracts were made with the criterion of approximately doubling their original concentrations. Results of these experiments showed that an about 30% signal weakening occurred only for NPE3C and C-14 SAS. This effect could result from coelution of the former compound with a large amount of an unknown compound and partial coelution of the latter compound with C-11 LAS. On the basis of these observations, the apparent concentrations of the two above analytes in sludge samples were increased by a factor 1.3. Choice of the Extractant. When developing a method for analyzing a large number of species having a broad range of chemical characteristics in a solid matrix, the first challenge is the selection of a suitable extractant. In our previous works (18-20), water alone or phosphate buffered water (0.5 mol/ L, pH 7.5) were employed to extract pesticides from soil. Extraction experiments of target compounds from STP sludge were initially performed by using three different extractants, i.e., the two ones mentioned above and carbonate buffered water (0.055 mol/L, pH 9.4). The rationale behind the inclusion of the latter solution in the list of potential extractants was that basified water should enhance solubility of very weakly acidic compounds, such as OP and NP. As anticipated, triplicate comparative measurements carried out as reported in the Experimental Section showed that carbonate buffered water removed remarkably larger amounts of weakly acidic analytes from sewage sludge, especially of OP and NP. Therefore, this extractant was used for subsequent experiments. Effect of the Temperature on Extraction Efficiency. As the temperature increases, polarity, surface tension, and viscosity of water decrease (15, 16). Consequently, water becomes more and more efficient in extracting lipophilic organics from solid matrices. On the other hand, a risk inherent to the use of water at elevated temperatures is that it could decompose those compounds that are thermolabile and/or prone to hydrolytic attack. Therefore, we evaluated the temperature effect on analyte recovery in STP sludge by performing extractions at three different temperatures. Other extraction parameters are reported in the Experimental Section. At each temperature, three extractions were carried out and results are reported in Table 2. Raising the temperature of the extractant from 150 to 200 °C had the effect of remarkably improving the extraction yield especially of those analytes having the largest hydrophobic moieties. By

compound

150b

200b

230b

A12E2 A12E6 A15E2 A16E2 NPE2 NPE6 NPE1C NPE4C NP C10-LAS C14-LAS A12S A18S C14-SAS C17-SAS A12E1S A15E1S

7.3 (7) 0.54 (12) 22 (10) 1.9 (12) 29 (9) 0.77 (12) 14 (10) 7.0 (12) 101 (9) 76 (8) 20 (11) 15 (9) 1.1 (13) 1.8 (9) 4.1 (11) 3.8 (13) 3.1 (12)

13 (4) 0.71 (11) 60 (7) 6.7 (8) 62 (8) 1.1 (9) 24 (6) 9.7 (9) 191 (6) 98 (6) 30 (9) 22 (7) 1.9 (9) 3.0 (11) 6.6 (8) 6.0 (9) 5.5 (8)

15 (9) 0.60 (11) 71 (8) 8.3 (9) 54 (10) 0.73 (11) 19 (7) 7.9 (10) 228 (10) 95 (8) 32 (9) 19 (8) 1.9 (11) 3.2 (10) 7.2 (8) 5.5 (10) 5.1 (10)

a

Mean values from triplicate measurements.

b

Temperature, °C.

TABLE 3. Concentrations of Selected Analytes Measured in Sewage Sludge by Using Dynamic Extraction and Static/ Dynamic Extraction concentrationa (RSD, %), mg/kg (dry weight) compound

0b

5b

10b

20b

A12E2 A12E6 A15E2 A16E2 NPE2 NPE6 NPE1C NPE4C NP C10-LAS C14-LAS A12S A18S C14-SAS C17-SAS A12E1S A15E1S

6.8 (12) 0.58 (13) 33 (13) 3.5 (14) 32 (12) 0.78 (13) 18 (10) 7.1 (12) 162 (11) 85 (8) 19 (12) 17 (10) 1.0 (11) 2.1 (12) 4.1 (10) 5.0 (10) 3.8 (13)

10 (7) 0.68 (9) 53 (9) 5.8 (9) 52 (8) 0.89 (10) 21 (8) 8.3 (8) 183 (7) 93 (8) 25 (9) 20 (8) 1.6 (10) 2.8 (11) 5.9 (10) 5.4 (9) 4.7 (11)

13 (4) 0.71 (11) 60 (7) 6.7 (8) 62 (8) 1.1 (9) 24 (6) 9.7 (9) 191 (6) 98 (6) 30 (9) 22 (7) 1.9 (9) 3.0 (11) 6.6 (8) 6.0 (9) 5.5 (8)

14 (6) 0.70 (8) 63 (8) 6.5 (7) 61 (9) 1.1 (8) 22 (7) 9.0 (6) 195 (7) 98 (7) 32 (8) 23 (7) 1.8 (8) 3.2 (10) 6.7 (9) 5.9 (8) 5.6 (10)

a Mean values from triplicate measurements. duration, min.

b

Static extraction

further raising the extraction temperature at 230 °C, the amounts of anionic surfactants removed from sludge no more significantly increased. Vice versa, the amounts extracted of both NPEs and NPECs significantly decreased. This decrease was accompanied by an increase of the NP amount. The behavior of AE species was controversial, since the increment of 30 °C in the extraction temperature produced extracts richer of oligomers with short ethoxy chains but poorer of long-chain oligomers. The suspect that these strange behaviors were actually produced by partial decomposition of ethoxylated species occurring at 230 °C was confirmed by extraction experiments with quartz sand spiked with the species mentioned above. Effects of the Extraction Mode on Extraction Efficiency. Extraction with hot liquids can be performed in both static and dynamic modes. In the past, we developed methods for analyzing pesticides in soil based on dynamic extraction with hot water (18-20). Cramers and co-workers (30) employed VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4159

TABLE 4. Concentrations of Selected Analytes in Sewage Sludge Measured by This Method Compared to Those by Two Other Extraction Methods concentrationc (RSD, %), mg/kg (dry weight) compound

method aa

A12E2 A12E6 A15E2 A16E2 NPE2 NPE6 NPE1C NPE2C NPE3C NPE4C NP C10-LAS C14-LAS A12S A18S C14-SAS C17-SAS A12E1S A15E1S

12 (5) 0.57 (10) 53 (7) 6.3 (8) 60 (7) 1.0 (8) 7.7 (9) 75 (8) 25 (9) 14 (10) 185 (6) 60 (9) 20 (11) 16 (7) 1.5 (8) 2.1 (10) 5.1 (7) 4.5 (8) 5.0 (9)

method bb

11 (7) 95 (6) 33 (7) 19 (8)

this method 13 (4) 0.71 (11) 60 (7) 6.7 (8) 62 (8) 1.1 (9) 24 (6) 115 (5) 38 (6) 21 (7) 191 (6) 98 (6) 30 (9) 22 (7) 1.9 (9) 3.0 (11) 6.6 (8) 6.0 (9) 5.6 (8)

a 16-h Soxhlet extraction with methanol. b Dynamic extraction at 75 °C with 40 mL of a water/ethanol (70:30, v/v) solution (13). c Mean values from triplicate measurements.

successfully a combination of static and dynamic extraction with hot hexane as extractant for analyzing monomers and

oligomers in a polymeric sample. In this work, we evaluated if the static/dynamic extraction mode was abler than dynamic extraction alone to remove targeted compounds from STP sludge. In the combined extraction mode, the effect of the duration of the static extraction period on the amounts extracted of the analytes was also considered. Under both extraction modes, 20 mL of carbonate buffer heated at 200 °C was used as extractant. When the extraction system was operated exclusively in the dynamic mode, the same extraction pressure as that used in the static extraction mode, i.e., 100 bar, was selected. In the mixed-mode extraction, 3 mL of extractant was introduced in the extraction cell, and, after a time variable from 5 to 20 min, the cell was flushed with the remaining 17 mL of extractant at the same flow rate adopted for dynamic extraction, that is 1 mL/min. In any case, three measurements were performed and results are shown in Table 3. The introduction of a static extraction step had the effect of remarkably enhancing the extraction yield of those compounds having relatively low affinity for water. This finding indicates that removal of substances having a dominating hydrophobic moiety from a solid matrix by water is a process requiring a certain equilibration time. In this work, 10 min of static extraction sufficed to establish an equilibrium state. Method Comparison. In term of extraction efficiency, we compared our extraction procedure with a conventional one, that is 16-h Soxhlet extraction with methanol as extractant (procedure A), and a recently reported procedure that was specifically designed to analyze NPECs in sewage sludge (procedure B) (14). Briefly, procedure B involves dynamic extraction at 75 °C for 20 min with 40 mL of 30% ethanol in

TABLE 5. Concentrations of Target Compounds Measured in Sewage Sludge before and after Anaerobic Digestion Treatment (14 Days) concentration,a mg/kg dry weight

concentration,a mg/kg dry weight

compound

before treatment

after treatment

removal, %

compound

before treatment

after treatment

removal, %

A12E2 A12E3 A12E4 A12E5 A12E6 A13E2 A13E3 A13E4 A13E5 A13E6 A14E2 A14E3 A14E4 A14E5 A14E6 A15E2 A15E3 A15E4 A15E5 A15E6 A16E2 A16E3 A16E4 A16E5 A16E6 NPE2 NPE3 NPE4 NPE5 NPE6 NPE1C NPE2C NPE3C NPE4C

13 4.3 2.2 1.1 1.0 57 6.7 1.5 0.63 1.0 69 6.9 4.3 1.8 1.9 106 8.1 3.1 1.9 1.8 8.7 5.3 2.0 0.71 0.83 100 15 8.4 3.5 3.2 6.8 69 11 3.0

7.1 2.8 1.4 0.47 0.53 29 2.9 1.0 0.29 0.56 26 3.0 1.9 0.63 0.69 57 1.9 0.90 0.39 0.36 1.2 1.5 0.53 0.19 0.24 40 8.4 3.5 1.8 1.8 1.9 18 4.2 1.1

45 35 36 57 47 49 57 33 54 44 62 57 56 65 64 46 77 71 79 80 86 72 74 73 71 60 44 58 49 40 97 74 62 63

OP NP C10-LAS C11-LAS C12-LAS C13-LAS C14-LAS C6-DATS C7-DATS C8-DATS C9-DATS C10-DATS A12S A13S A14S A15S A16S A18S C14-SAS C15-SAS C16-SAS C17-SAS A12E1S A12E2S A12E3S A13E1S A13E2S A13E3S A14E1S A14E2S A14E3S A15E1S A15E2S A15E3S

14 242 150 950 1770 1730 66 4.8 28 73 91 9.6 40 3.6 10 4.1 3.6 4.7 6.4 13 21 23 22 12 7.3 6.4 1.5 0.78 12 4.3 1.7 4.3 2.0 1.0

17 308 142 880 1690 1630 59 4.5 25 67 84 8.7 28 2.8 8.1 3.3 3.3 4.5 6.0 12 20 20 21 11 6.4 6.2 1.3 0.74 11 3.4 1.5 4.1 1.8 0.86

5 7 5 6 11 6 11 8 8 9 30 22 19 20 8 6 6 8 5 13 5 8 12 3 13 5 8 21 12 5 10 14

a

Mean values from quadruplicate measurements. In all cases, RSD ranged between 6 and 12%.

4160

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

water. To analyze sludge extracts obtained by procedures A and B, aliquots of the two extractants were first diluted with water so that the final solvent percentage was not larger than 5%. Analytes in the two solutions were then extracted by Carbograph 4 cartridges, and finally the rest of the procedure reported in the Experimental Section was followed. The performance of the extraction cartridge was not affected by the presence of the small percentage of either methanol or ethanol in water, as ascertained by reanalyzing effluents from the cartridges with fresh cartridges. With every method, extractions were made in triplicate and results are shown in Table 4. Compared to Soxhlet extraction, the extraction efficiency of this method for neutral species showed to be of comparable value. Conversely, the extraction procedure developed by us was remarkably more efficient in removing ionogenic compounds from sludge. Probably, solvation effects by water weaken specific interactions between organic anions and some unknown sludge components. An alternative explanation could be that water at 200 °C is abler than methanol to solubilize that fraction of anionic surfactants present in sludge as calcium salts. Compared to procedure B, this method extracted larger amounts of NPECs. Most likely, addition of a basic agent coupled to an extraction temperature much higher than that used in procedure B makes water more effective to solubilize weakly acidic species, such as NPECs. Effects of Anaerobic Digestion of Sewage Sludge on the Content of Target Compounds. Anaerobic digesters are neither designed nor intended for the biodegradation of surfactants. Their function is the reduction of the biomass, stabilization of sludge, and elimination of pathogens. Nevertheless, Giger et al. (10) showed that anaerobically treated sludge contained NP amounts 4-5 times larger than those in aerobically stabilized sludge. This result was accounted for by anaerobic biotransformation of short-chain NPEs, specifically NPE1 and NPE2. Moreover, some authors (31, 32) estimated, on the basis of mass balance calculations, that a 15-20% LAS biotransformation occurred after 20-30 days of sludge anaerobic digestion. We shortly investigated the effect of a relatively short residence time (14 d) in an anaerobic digester on concentration levels of target compounds. Table 5 lists concentrations of target compounds before and after anaerobic treatment. Since a single pre- and postanaerobic digestion sample were analyzed, the discussion of our results reported below should not be regarded as a general conclusion of the anaerobic digestion process. The LAS total concentration found by us in anaerobically stabilized sludge was 4.7 g/kg that is within the 3.8-7.5 g/kg concentration range measured in six Swiss treated STP sludges (9). We observed a 7% LAS removal after 14-d residence time in the digester. It is difficult to establish whether this slight decrease was casual or due to some anaerobic biotransformation. After LAS, AE surfactants are probably the most widely used synthetic surfactants for domestic purposes. Nevertheless, relatively small amounts, i.e., 312 mg/kg as a total, were found in untreated sludge. This concentration is about 16 times lower than that of LAS. Considering that the AE consumption is only four times lower than that of LAS, our data suggest that the former species are even more readily biodegraded in activated sludge STP than the latter ones. The original amounts of neutral ethoxylate surfactants in sludge were more than halved after the anaerobic digestion treatment. This finding is consistent with that of a 28-day batch anaerobic digester die-away test (33) showing that about 85% of A16E7 underwent ultimate biotransformation.

Analogously to AEs, the sludge anaerobic digestion process removed 57% of NPEs. Even more pronounced was the NPEC removal, amounting to 74%. An approximate mass balance calculation suggested that the removed fractions of both NPEs and NPECs were not exclusively converted to NP. Finally, concentration changes of anionic surfactants, other than LAS, seem to indicate that the anaerobic digestion treatment provoked scarce, if any, removal of them from sludge.

Literature Cited (1) Inoue, S.; Sawayama, S.; Ogi, T.; Yokoyama, S. Biomass Bioenergy 1996, 10, 37. (2) Commission of the European Communities. Council Directive concerning Urban Wastewater Treatment; Official Journal of the European Communities L135740-52. (3) Eljarrat, E.; Caixach, J.; Rivera, J. Environ. Sci. Technol. 1998, 33, 2493. (4) Klo¨pffer, W. Chemosphere 1996, 33, 1067. (5) McEvoy, J.; Giger, W. Environ. Sci. Technol. 1986, 20, 376. (6) Matthjis, E.; DeHenau, H. Tenside, Surfactants, Deterg. 1987, 4, 193. (7) Threy, M.; Gledhill, W. E.; Orth, R. G. Anal. Chem. 1990, 62, 2581. (8) Breen, D. G. P. A.; Horner, J. M.; Bartle, K. D.; Clifford, A. A.; Waters, J.; Lawrence, J. G. Water Res. 1996, 30, 476. (9) Field, J. A.; Miller, D. J.; Field, T. M.; Hawthorne, S. B.; Giger, W. Anal. Chem. 1992, 64, 3161. (10) Giger, W.; Brunner, P. H.; Schaffner, C. Science 1984, 225, 623. (11) Ahel, M.; Giger, W. Anal. Chem. 1986, 57, 1577. (12) Lee, H. B.; Peart, T. E. Anal. Chem. 1995, 67, 1976. (13) Lin, J. G.; Arunkumar, R.; Liu, C. H. J. Chromatogr. 1999, 840, 71. (14) Field, J. A.; Reed, R. L. Environ. Sci. Technol. 1999, 33, 2782. (15) Petrovic, M.; Barcelo´, D.; Anal. Chem. 2000, 72, 4560. (16) Hawthorne, S. B.; Yang, Y.; Miller, D. J. Anal. Chem. 1994, 66, 2912. (17) Yang, Y.; Hawthorne, S. B.; Miller, D. J. Environ. Sci. Technol. 1997, 31, 430. (18) Crescenzi, C.; D’Ascenzo, G.; Di Corcia, A.; Nazzari, M.; Marchese, S.; Samperi, R. Anal. Chem. 1999, 71, 2157. (19) Di Corcia, A.; Barra-Caracciolo, A.; Crescenzi, C.; Giuliano, G.; Murtas, S.; Samperi, R. Environ. Sci. Technol. 1999, 33, 3271. (20) Crescenzi, C.; Di Corcia, A.; Nazzari, M.; Samperi, R. Anal. Chem 2000, 72, 3050. (21) Crescenzi, C.; Di Corcia, A.; Marchiori E.; Marcomini, A.; Samperi, R. Water Res. 1996, 30, 722. (22) Di Corcia, A.; Bellioni, A.; Madbouly, M. D.; Marchese, S. J. Chromatogr. 1996, 733, 383. (23) Andreolini, F.; Borra, C.; Caccamo, F.; Di Corcia, A.; Samperi, R. Anal. Chem. 1987, 59, 1720. (24) Crescenzi, C.; Di Corcia, A.; Marcomini, A.; Samperi, R. Anal. Chem 1995, 67, 1797. (25) Di Corcia, A.; Crescenzi, C.; Marcomini, A.; Samperi, R. Environ. Sci. Technol. 1998, 32, 711. (26) Di Corcia, A.; Marcomini, A.; Samperi, R. Environ. Sci. Technol. 1994, 28, 850. (27) Crescenzi, C.; Di Corcia, A.; Passariello, G.; Samperi, R.; Turnes Carou, M. I. J. Chromatogr. 1996, 733, 41. (28) Burford, M. D.; Hawthorne, S. B.; Miller, D. J. Anal. Chem. 1993, 65, 1497. (29) Di Corcia, A.; Crescenzi, C.; Lagana`, A.; Sebastiani, E. J. Agric. Food Chem. 1996, 44, 1930. (30) Lou, X.; Janssen, H.-G.; Cramers, C. A. Anal. Chem. 1997, 69, 1596. (31) Berna, J. L.; Ferrer, J.; Moreno, A.; Prats, D.; Ruiz Bebia, F. Tenside Surf. Det. 1989, 26, 101. (32) Giger, W.; Brunner, P. H.; Ahel, M.; McEvoy, J.; Marcomini, A.; Schaffner, C. Gas Wasser Abwasser 1987, 67, 111. (33) Steber, J.; Wierich, P. Surf. Congr. 1984, 1, 176.

Received for review January 3, 2002. Revised manuscript received May 13, 2002. Accepted July 12, 2002. ES020002E

VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4161