Talanta 58 (2002) 1335 1341 www.elsevier.

com/locate/talanta

Advanced sample pretreatment for the monitoring of polycyclic aromatic hydrocarbons and extractable organic halogens in waste water. Flow based procedures with chromatomembrane cells
Ju rgen Simon *, Annette Kirchhoff, Oliver Gu ltzow
Institute of Chemistry /Inorganic and Analytical Chemistry, Free Uni6ersity of Berlin, Fabeckstr. 34 /36, 14195 Berlin, Germany Received 27 February 2002; received in revised form 26 March 2002; accepted 26 March 2002

Abstract Polycyclic aromatic hydrocarbons (PAH) and extractable organic halogens (EOX) pollute industrial waste waters and need to be controlled continuously by automated procedures. Their sample pretreatment requires extraction rst from complex matrices containing surfactants, humic acids, urine and electrolytes besides. When using chromatomembrane cells (CMC) for the extraction with pentane or hexane a ow based system can be established which preconcentrates the pollutants up to ratios of 100:1 at the same time. The extracted compounds become supplied to a gas chromatograph (PAH) using the split/splitless injection or to a combustion furnace (EOX), respectively. An aqueous solution later extracts the hydrogen halides from the exhaust gas for their simultaneous detection with an ion-chromatograph. The limits of lowest detection are attainable in the lower mg l 1 range by matrix adjusted calibration. The CMC is a novel device containing a bloc of biporous PTFE which enables the contact of two immiscible phases. Polar liquids ll the macropores whereas the micropores remain available for non-polar liquids or gases. 2002 Elsevier Science B.V. All rights reserved.
Keywords: Chromatomembrane cell; Extraction of PAH; Extraction of EOX; Sample pretreatment; Waste water monitoring

1. Introduction Trace analysis normally requires both, analyte separation from the matrix and an integrated preconcentration step, if possible before the selected detector becomes enabled to work suf* Corresponding author. Tel.: + 49-30-83854006; fax: + 4930-83852424. E -mail address: jsimon@chemie.fu-berlin.de (J. Simon).

ciently. In general, the techniques of ow injection analysis permit such sample pretreatment on-line, so that a fully automated setup can be established whenever the procedure of separation is agreed with the given circumstances. Solvent extraction was unquestionably the mostly used method for analyte separation in the laboratories of the past. But a lot of difculties arose in adapting the extraction procedures to the requirements of modern ow systems. The small

0039-9140/02/$ - see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 2 0 3 - 5

1336

J. Simon et al. / Talanta 58 (2002) 1335 1341

sample sizes still being necessary for the instrumentation of today call for novel devices correspondingly miniaturized. On the distinct approaches in this eld reports Fangs text book [1]. The procedures are known to solve the problem in three steps following each other but, however, they all do not include the simultaneous preconcentration of the analyte. Apart from technical facts the use of organic solvents for extractants should be reduced because of their toxicity which goes especially for halogen containing compounds such as CCl4, CHCl3 or something similar. Solid phase extraction procedures were developed [2] using columns which are packed with ion-exchange resins or surface treated beads. In this way usually occurs the separation of trace metals with chelating reagents bound as functional groups which allows remarkable preconcentration ratios. The elution of the metal ions from the surfaces is later on made possible by varying nothing but the pH in the aqueous phase. Solvents, as pentane and hexane, can be tolerated adequately from the standpoint of environmental protection. Both can be used successfully to extract polycyclic aromatic hydrocarbons (PAH) and extractable organic halogens (EOX) from waste waters so that ow based procedures could be established for their detection. A high ratio of preconcentration is the objective that is worth striving for, because only this guarantees to use the solvents sparingly and to enhance the sensitivity of the detection following. The distribution coefcients for the extraction of PAH and EOX from the aqueous phase by pentane or hexane are very favorable for approaching high preconcentration ratios. They are found for most part in a range of 10 3 and for a requested preconcentration of 100:1 (i.e. the contacting volumes have to comply with that proportion) one calculates easily from the distribution law that even then just under 10% will remain in the aqueous phase only. The excellent repeatability of ow based systems should meet the requirements for reasonable analytical procedures, but unfortunately the customary extraction coils facilitate the immiscible phases to

contact therein with a volume ratio of about 1:1 only, in special cases up to 3:1 possibly. Not any remarkable preconcentration will take place under this circumstances, and it helps very little that the analyte will be extracted almost completely. So the need on a novel material can be concluded which is for lling the extraction cells [3] of ow based systems. We proposed [3 5] in this connection the biporous PTFE which contains two types of pores, micropores and macropores. Polar liquids ll the macropores whereas the micropores remain available for non-polar liquids or gases. The capillary pressure of polar liquids prevents their penetration into micropores. The diameter of micropores is 0.1 0.5 mm, and that of macropores 250 500 mm. Each cm3 of biporous PTFE is capable to take 0.3 cm3 polar and 0.3 cm3 non-polar phase as an average. The size of the surface boundary, where the two phases are in mutual contact, was determined to be in the range of 60 cm2 cm 3 biporous PTFE [6]. Independent uxes of both phases can be realized within this material: their direction to each other is open, even the counter ow is possible. The ow ratios can vary up to 100:1, i.e. a 100-fold preconcentration is possible. The system is stable as long as the pressure of the aqueous phase is a little higher than that of the non-polar phase. Microporous PTFE membranes serve as water barriers, their positioning in the extraction cells guides the two phases to ow in directions as requested. The procedures to detect PAH [7,8] and EOX [9] as laid down of the environmental authorities require a clean up for the aqueous samples where the pollutants become extracted using separatory funnels, i.e. automation is not possible and bigger sample sizes are necessary. The separation of the extracted PAH, however, is carried out by chromatographic methods as well. The content of halogen in case of the EOX determination results from the argentometric titration as a composite only. The application of chromatomembrane cells (CMCs) proposed in this paper makes automation possible.

J. Simon et al. / Talanta 58 (2002) 1335 1341

1337

2. Experimental

2.2. Extraction of polycyclic aromatic hydrocarbons

The mode of operation is shown in Fig. 1: the aqueous sample ows continuously with 5.0 ml min 1 to the CMC. Pentane extracts the PAH (ow rate: 65 ml min 1; this equals a preconcentration of 77:1) and ows over 4 min (accumulation time) through the injector which is lled with Chromosorb G (DCMS 80/100 mesh). At an injector temperature of 50 C the extractant evaporates therein (split vent) when rinsing with a nitrogen ow rate of 14 ml min 1. The temperature of the injector becomes increased up to 250 C after and the analyte desorbs into the N2 carrier (ow rate 14 ml min 1) which is supplied then to the gas chromatograph (HP 5890 series II). The column HP-1 is lled with polydimethylsiloxane (0.53 mm 5 m 2.65 mm). The temperature of the column is 36 C over 9 min and increases after with a rate of 30 C min 1 up to a maximum of 250 C. The working temperature of the FID detector is 300 C (fuel gas H2:22 ml min 1, air:170 ml min 1). The chromatogram

2.1. The chromatomembrane cell

The extraction regimes which were applied in this paper base on the prototype of the CMC which was described several times by our group [4 6]. A block of biporous PTFE is covered with microporous membranes on the matching sides, so that the uxes of the two phases are rectangularly directed to each other. Since the ow rates can be regulated as you like, a two-dimensional chromatography became created. This goes especially for using a stopped-ow mode if multiple extraction processes actually take place. The complete extraction of the analyte is possible under such a condition which was used for the samplingprocedure of the hydrogen halides extracted from the exhaust gas of the EOX combustion furnace. The PAH and EOX extraction from the waste waters was carried out by a continuous regime which guarantees high preconcentration ratios under constant and reproducible conditions.

Fig. 1. Flow chart for PAH measurement: (1) aqueous sample; (2) extraction unit, CMC; (3) injection valve; (4) injector lled with Chromosorb G.

1338

J. Simon et al. / Talanta 58 (2002) 1335 1341

Fig. 2. Flow chart for EOX measurement: A, absorption solution; B, carrier gas (argon); C, aqueous sample; D, hexane; E, oxygen; F, aqueous waste; G, organic waste; H, gaseous waste; (1) peristaltic pump; (2) rst extraction unit; CMC, (3) injection valve with sample loop; (4) combustion unit including Pt-needle and quartz tube; (5) wash bottle with H2SO4 (optional), for gas drying; (6) second extraction unit, CMC.

for the separation of the test compounds (naphthalene, acenaphthen, uoren, anthracene, uoranthene) is registered within 16 min. The entire analysis including the simultaneous detection of the ve compounds takes about 30 min.

2.3. Extraction of extractable organic halogens

The mode of operations is shown in Fig. 2: the aqueous sample (C) ows continuously with 3.9 ml min 1 to the CMC (2). Hexane (D) extracts EOX (ow rate 0.05 ml min 1; this equals a preconcentration of 78:1) and lls a sample loop (3) with a volume of 200 ml mainly (rst extraction procedure). The EOX combustion unit (Haberkorn & Braun, Germany) is supplied with the hexane extract by the carrier gas (B) (ow rate of 25 ml min 1). It is mixed with oxygen (E) which ows with 180 ml min 1 into the combustion unit heated up to 1000 C. The exhausted gas becomes washed with sulfuric acid (5) and ows to a second CMC (6) for the hydrogen-

halide absorption into a solution of hydrazine sulfate (0.4 mmol l 1). An ion chromatograph (Metrohm 690, Switzerland) separates the HX (column PRP-X 100, type 6.1005.000). A solution of phthalic acid (3 mmol l 1) and acetone (15%) is used for their elution (pH 5.0) and separation. Compounds under study: chloroform, methylenechloride, o -chlorobenzene, pentachlorphenol, bromoform, bromobenzene, bromophenol, iodoform, iodobenzene, iodobenzoic acid.

2.4. Reagents
All the test substances were commercial products whereof stock solutions in acetone (PAH) and methanol (EOX), respectively have been prepared. They were added either to distilled water or to natural river water (Mulde River) or urine. The samples, such prepared, were analyzed directly or mixed before with electrolytes, humic acid or surfactants in case that waste water had to be simulated.

J. Simon et al. / Talanta 58 (2002) 1335 1341

1339

3. Results and discussion

3.1. The simultaneous detection of PAHs

The aqueous samples under study were prepared by adding beforehand all the ve PAHs (naphthalene, acenaphthene, uorene, anthracene, uoranthene) so that the different concentrations (5, 10, 15, 20, 25, 30, and 35 mg l 1) were attained for the respective component. The FID-signals were registered after separation within the GC. Their peak-areas (measured in nA s 1) were plotted vs. PAH concentration to obtain the numerous calibration graphs for the different matrices. The calculated gradients of the most are listed in Table 1; their unit comes to (nA s mg 1 l 1) so that LODs result in the range 1 2 mg l 1. The determination of naphthalene is less sensitive and that, however, not for reason of its extractability in the CMC but of its volatility in the injector unit of the GC. Varying pH-values do not effect noticeable differences. Electrolytes cause the PAHs to salt-out, but they do not have lasting inuence on the proposed procedure because their solubility in water decreases altogether. It should be taken into account only in case of a later addition of electrolyte to highly polluted samples. The signal sensitivity increases slightly with the content of humic acid in the aqueous samples.

Mulde water is highly stressed because the river drains a widespread industrial area, its DOCvalue amounted to 23 mg l 1. A reasonable repeatability of the procedure even results in urine samples. Therefore, it can be concluded that normal calibration graphs are adequate for data evaluation for the most part. The inuence of matrix compounds on PAH-signals should possibly be tested by standard addition in highly polluted waste water. The detection of PAH in urine matrix is not of practical importance because PAHs may be collected in human issues but urine excretes them only metabolized. The presence of surfactants in aqueous samples effects remarkable depressions of PAH signals. Surfactants act probably as solubilizer for PAHs, and their distribution coefcients vary accordingly. Moreover, an increased solvency of pentane in the aqueous phase can be observed during the extraction process inside the CMC. In such cases standard addition will be recommended for calibration.

3.2. The simultaneous detection of the halogen contents of EOX

The EOX becomes continuously extracted from the aqueous sample by hexane within the rst CMC (Fig. 2). The chosen ow rates effect a

Table 1 Calibration graphs for PAHs in different matrices PAH compound Distilled water Distilled water with humic acid (20 mg l1) R2 Gradient (nA s mg l1) 14.63 9 0.55 52.88 9 0.67 56.72 9 0.83 41.01 9 0.92 39.75 9 0.88 R2 River water (Mulde) Urine

Gradient (nA s mg l1) Naphthalene Acenaphthene Fluorene Anthracene Fluoranthene 10.01 9 0.18 44.26 9 0.38 52.83 9 0.35 43.53 9 0.33 37.47 9 0.50

Gradient (nA s mg l1) 9.71 9 0.34 55.44 9 0.55 54.24 9 0.69 38.90 9 0.36 45.09 9 0.40

R2

Gradient (nA s mg l1) 9.57 9 0.19

R2

0.993 0.998 0.999 0.999 0.996

0.994 0.999 0.999 0.998 0.999

1.000 1.000 0.999 1.000 1.000

0.99 0 0.99 6 1000 1000 1000

40.37 9 0.78 55.62 9 0.34 38.60 9 0.43 37.65 9 0.43

1340

J. Simon et al. / Talanta 58 (2002) 1335 1341

Table 2 Calibration graphs for the halogen contents in EOX polluted water Halogen Chlorine Bromine Iodine Retention time (IC) (min) 1.3 1.9 5.7 Gradient (mV s l mg1) 4.94 9 0.09 2.16 9 0.05 0.96 9 0.02 R2 0.999 0.999 0.988 Recovery (%) 98 100 83

preconcentration ratio of 78:1 and in addition, they guarantee the complete extraction as checked for the test compounds listed in chapter 2.3 (i.e. the condition for a multiple extraction process was given). The extractant collected in the sample loop was supplied to the combustion unit with argon as the carrier gas. Within the second CMC, operated in the stopped-ow-mode, an aqueous absorbing solution extracts the HX from the exhaust gas. The wash bottle with sulfuric acid removes water formed by the combustion of the sample. The volume of the absorbing solution is dened therefore because it is equal to the volume of the macropores in the PTFE-block inside the CMC. That sample is transferred to the IC for the separation of the HX and their detection. The results are listed in Table 2 in the form of the gradients for the calibration graphs. The IC is implemented with a conductivity detector and the peak areas must be evaluated. Therefore, the unit of the gradients results as (mV s mg 1 l 1), so that the following LODs can be determined: 0.3 mg l 1 chlorine, 0.5 mg l 1 bromine and 1 mg l 1 iodine, all related to the original aqueous sample. The long retention time for I effects a peak dispersion which makes evaluation more difcult. An UV-detector could be used for the more sensitive detection of iodine, but that would require another eluent which has to be free of acetone. In principle it is possible to increase the preconcentration ratio for the extraction with hexane. Certainly, its realization may be difcult using the CMC, but possible in case the CMC will be replaced with an extraction cell [3] which is lled with biporous PTFE. Such a setup allows longer contact times with the extractant, but this is not subject of this paper.

4. Conclusions The paper shows the CMC to be a simple and versatile device for extraction procedures so that sample pretreatment can be integrated into the setup for the automated monitoring of pollutants in the environment. Manual work becomes reduced and the quality of analytical results improves. Especially the separation of organical compounds from waste water is made possible by the economic use of alkanes like pentane and hexane for extractants because they facilitate an extensive preconcentration of traces at the same time. The setup proposed for the EOX determination is the rst analyzer to our knowledge that allows the halogens to differentiate with regard to their content in the sample. It may be emphasized that the procedure includes a solvent extraction step from the gas phase in addition. Such conception works successfully even in case of the absorptive sampling of pollutants from the atmosphere as recently proposed for the nitrogen dioxide monitoring [10]. Acknowledgements The authors thank Leonid N. Moskvin and Tatiana G. Nikitina, State University of St. Petersburg, Russia, for considerable discussions and productive cooperation. References
[1] Z. Fang, Flow Injection Separations and Preconcentrations, Textbook, VCH, Weinheim, 1993. [2] R.E. Majors, LC GC Int. 11 (9) (1998) 8 16. [3] J. Simon, L.N. Moskvin, Anal. Sci. 17 (Suppl.) (2001) i425 i428.

J. Simon et al. / Talanta 58 (2002) 1335 1341 [4] L.N. Moskvin, J. Simon, Talanta 41 (1994) 1765. [5] J. Simon, L.N. Moskvin, Talanta 49 (1999) 985. [6] H. Erxleben, J. Simon, L.N. Moskvin, T.G. Nikitina, J. Flow Injection Anal. 18 (2001) 39. [7] DIN 38409 Teil 13, Summarische Wirkungs- und Stoffkenngro en (Gruppe H), Bestimmung von polycyclischen aromatischen Kohlenwasserstoffen (PAK) in Trinkwasser (H 13-1 bis 3), Juni 1981.

1341

[8] US-EPA Method 610 Determination of Polyaromatic Hydrocarbons by HPLC. [9] DIN 38409 Teil 8, Summarische Wirkungs- und Stoffkenngro en (Gruppe H), Bestimmung der extrahierbaren organisch gebundenen Halogen (EOX) (H8), September 1984. [10] Y. Wei, M. Oshima, J. Simon, S. Motomizu, Talanta 57 (2002) 355.