Anal. Chem. 1998, 70, 2831-2838
Chlorinated Diphenoquinones: A New Class of Dioxin Isomeric Compounds Discovered in Fly Ashes, Slags, and Pyrolysis Oil Samples by Using HPLC/ELCD and HRGC/MS Frank Otto, Gu 1 nter Leupold, and Harun Parlar*
Institute of Chemical Technical Analysis and Chemical Food Technology, Technical University of Munich, D-85350 Freising-Weihenstephan, Germany Reiner Rosemann,† Mu 1 fit Bahadir,† and Henning Hopf‡
Institute of Ecological Chemistry and Waste Analysis and Institute of Organic Chemistry, Technical University of Braunschweig, Hagenring 30, D-38106 Braunschweig, Germany
Chlorinated 4,4′-diphenoquinones, previously unknown byproducts of thermal processes in the presence of chlorine, have been discovered in a variety of samples such as oil products from a waste pyrolysis plant, a fly ash from a municipal waste incinerator, and a sample of Kieselrot slag. The pyrolysis plant products also contained tetramethylated and unsubstituted 4,4′-diphenoquinone. All compounds may be generated by oxidative coupling of phenols at elevated temperaturessa reaction that is considered to compete with the formation of the structurally isomeric dioxins. Once built-up, 4,4′-diphenoquinones form a redox pair with the corresponding 4,4′dihydroxybiphenyls, with variable equilibrium positions depending on the prevailing conditions. For analysis, a cleanup method which involves liquid/liquid partition and solid-phase extraction was developed. The method is suitable for the determination of all substances of this class. Quantitative analysis was carried out by HPLC using an electrochemical detector for the fly ash and slag samples. Pyrolysis oil samples were quantitatively analyzed by GC/MS after derivatization with trifluoroacetic acid anhydride. Substance identification in all samples was performed by GC/MS. The samples were found to be contaminated with chlorinated diphenoquinones in the lower and middle ppb range. Pyrolysis oil products yielded up to 1 ppm methylated diphenoquinones. The amount of diphenoquinones in the samples is compared with their yield of chlorinated dioxins and furans as well as polychlorinated biphenyls (PCBs). Kieselrot especially was contaminated, with 25 ppm of polychlorinated dibenzop-dioxins/-furans (PCDD/F) and 94 ppm of hexachlorobenzene, while the pyrolysis oil samples contained up to 360 ppb of PCDD/F and up to 25 ppm of PCB. The formation of dioxins from chlorinated phenols at elevated temperatures has been well investigated. Mechanistically, one * To whom correspondence should be addressed. Tel.: (++)49 8161/713283. Fax: (++)49 8161/71-4418. † Institute of Ecological Chemistry and Waste Analysis. ‡ Institute of Organic Chemistry. S0003-2700(97)01390-5 CCC: $15.00 Published on Web 06/17/1998
© 1998 American Chemical Society
assumes a condensation of phenolic molecules under elimination of two molecules of HCl.1 However, under oxidative conditions, phenols can also react by coupling in the para positions, which leads to double quinoid systems. The resulting compounds, the diphenoquinones,2 are structurally isomeric with the dioxins. While there are numerous publications about the formation of dioxins under special conditions, such as combustion in waste incinerators,3 it has never been considered that the same conditions can generate diphenoquinones, nor have these substances been indentified in environmental samples. Laboratory preparation of diphenoquinones can be carried out by oxidation of phenols with chromium(VI) salts. Prerequisites for the reaction are unsubstituted meta positions in the phenolic substrates. Substituents in these positions prevent the formation of the quinoid double bond.4 4,4′-Diphenoquinones form a redox pair with the 4,4′-dihydroxybiphenyls. The redox potential for the unsubstituted compounds is +0.95 V and can be shifted by electronic effects from substituents in the positions marked with R in Figure 1.5 The equilibrium position in environmental samples depends on the prevailing conditions. Principally, conditions that facilitate the phenol coupling also have enough oxidative potential to form diphenoquinones from structurally suitable dihydroxybiphenyls. The same applies to the enzymatic formation of 4,4′-dihydroxybiphenyl from phenol and its further oxidation to 4,4′-diphenoquinone under the influence of human myeloperoxidase, which has been observed indirectly by Eastmond et al.6 After their formation, the quinones themselves can act as oxidants. Without being stabilized, e.g. by alkyl substituentssas observed in the case of the tetramethylated compoundsa slow reduction to the dihy(1) Buser, H.-R. J. Chromatogr. 1975, 114, 95-108. (2) Waters, W. A. J. Chem. Soc. B, 1971, 2026-2029. (3) Tong, H. Y.; Shore, D. L.; Karasek, F. W. Anal. Chem. 1984, 56, 24422447. (4) van Auwers, K.; Wittig, G. Chem. Ber. 1924, 57, 1270-1275. (5) Musso, H.; Figge, K.; Becker, D.-J. Chem. Ber. 1961, 94, 1107-1116. (6) Eastmond, D. A.; Smith, M. T.; Ruzo, L. O.; Ross, D. Mol. Pharmacol. 1986, 30, 674-679.
Analytical Chemistry, Vol. 70, No. 14, July 15, 1998 2831
Figure 1. Formation of diphenoquinones from phenolic compounds under oxidative conditions and redox equilibrium (R1 ) R2 ) H, CH3, OCH3, or Cl).
droxybiphenyls takes place in aqueous solution. However, as a result of the comparatively low redox potential, the reformation of the quinoid form is now far easier than its first generation from the phenolic materials. Toxicological data are available only for 3,3′,5,5′-tetrachloro4,4′-diphenoquinone (TCDQ). In a study published by McKinney et al.7 in 1987, TCDQ was ranked as a “weak binder” between the highly toxic “Seveso dioxin” 2,3,7,8-TCDD and the less toxic OCDD because of its affinity for the nuclear T4 receptor. The main reason for the acute toxicity of TCDQ may be its structural similarity to the Seveso dioxin together with its planar configuration. Investigations on possible mutagenic effects of diphenoquinones using 3,3′,5,5′-tetramethyl- (TMDQ), -tetramethoxy(TMxDQ), and -tetrachloro-4,4′-diphenoquinone (TCDQ), as well as the unsubstituted 4,4′-diphenoquinone (DQ), are presently being carried out in cooperation with the GSF, Munich, Germany. The results will be published at a later date. This paper describes a new analytical cleanup method for different diphenoquinones in three different matrixes as well as the determination of the compounds by HPLC using an electrochemical detector (ELCD) and HRGC after derivatization with TFAA. In the case of the fly ash and slag samples, identification of the substances was carried out by comparison of the hydrodynamic voltammograms of samples and standards and, additionally, by GC/MS measurements. HPLC/ELCD was used for the quantification of the compounds. Diphenoquinones in waste pyrolysis products were identified and quantitatively determined by GC/MS analysis after derivatization with TFAA. Generally, the aim of this investigation was to compare the levels of these novel pollutants with those of already known ones, such as chlorinated dioxins, furans, and polychlorinated biphenyls (PCBs). EXPERIMENTAL SECTION (A) Chlorinated Diphenoquinones. Materials. All solvents and chemicals for the cleanup were of analytical grade from Merck KGaA, Darmstadt, Germany. RP-phenyl cartridges (500 mg) for the solid-phase extraction were obtained from Varian GmbH, Darmstadt, Germany. TFAA for the derivatization of the analytes was purchased from CS-Chromatographie Service, Langerwehe, Germany. All analytical standards, both diphenoquinone and dihydroxybiphenyl derivatives, were supplied by Euronorm Standard Institute, Technical University of Munich, Freising-Weihenstephan, Germany. Cleanup. The cleanup and measurement conditions were developed using dihydroxybiphenyl (DHB) and three different (7) McKinney, J.; Fannin, R.; Chae, K.; Rickenbacher, U.; Peddersen, L. J. Med. Chem. 1987, 30, 79-86.
2832 Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
tetrasubstituted compounds with chlorine (TCDHB), methyl (TMDHB), and methoxy (TMxDHB) groups in those positions marked with R in Figure 1. A liquid/liquid separation was chosen for the first cleanup step after the initial extraction. Under the given conditions, the analytes separated into two groups with different properties: a polar group, consisting of DHB and TMxDHB that could be extracted in two steps from hexane using water as a solvent, and a nonpolar group, consisting of TMDHB and TCDHB, which could be extracted in two steps from water using hexane. Because of the hydroxy functions in the molecules, both groups are quantitatively extractable with 0.1 N NaOH from hexane. Solid-phase extraction using RP-phenyl cartridges (500 mg) was employed as a second cleanup step. After the elution from the cartridges, the solvent was evaporated to dryness by rotary evaporation. HPLC samples were directly dissolved in the desired amount of the eluting agent to be ready for injection. GC samples were derivatized with TFAA prior to injection. To validate the method, a statistical evaluation under exclusion of matrix effects was performed. For this, 10 Soxhlet extractor thimbles were each spiked with 1 g of a standard solution containing 200 ppb of each analyte in acetonitrile. The dried thimbles were extracted for 36 h with acetone. The resulting extracts were prepared and measured by HPLC as described below. The derivatization reaction for the GC samples was separately evaluated. Toward this end, six identical 1 g solutions containing 200 ppb of each analyte in acetonitrile were evaporated to dryness, derivatized with TFAA, and automatically injected into a GC/ECD system as described below. For the results of the evaluations, see Tables 1 and 2 in the Results and Discussion section. HPLC Conditions. All analyses were performed with a Gynkotek model 480 pump equipped with an Antec Decade electrochemical detector (HPLC/ELCD). Gynkosoft software was used for recording and data analysis. The injection volume was 20 µL for every analysis, the detector oven, containing injection port, pulse dampener, column, and measuring cell, was thermostated to 35 °C. The detector cell was equipped with a glassy carbon working electrode and an Ag/AgCl reference electrode filled with a saturated LiCl solution in MeOH/H2O (50:50). The optimum working electrode potential for the measurement of the dihydroxybiphenyls was +0.7 V. A Chromasil 100 C18 5-µm, 250mm × 4.6-mm-i.d. column was used for the chromatograpic separation of the substances. The composition of the eluting agent is determined by the polarity of the analytes. The more polar substances DHB and TMxDHB were eluted with MeOH/H2O/ THF (55:40:5) adjusted to pH 4 with citrate buffer. TCDHB and TMDHB were eluted with MeOH/H2O (80:20) adjusted to pH 4 with citrate buffer. Quantitative analysis with HPLC was carried out by an external calibration method. Under the chosen conditions, the detector showed a reproducible linear response in a concentration range between 2 and 500 ppb. The daily validity check of the calibration was performed by injecting standard solutions of 10, 100, and 200 ppb prior to analysis. All native samples were analyzed twice, together with a blank. HRGC Conditions. HRGC/ECD was used to evaluate the TFAA derivatization reaction. All analyses were carried out on a Chrompack CP 9002 gas chromatograph equipped with a
Chrompack autosampler and controlled by a Chrompack Maestro data system. The injection volume was 1 µL, splitless. The injection port temperature was 230 °C, and the detector operated at 280 °C. A DB5 30-m × 0.25-mm column with a layer thickness of 0.25 µm, and hydrogen as carrier gas with a prepressure of 0.8 bar, were used to separate the substances. The column temperature program was initial 100 °C, hold for 1 min, ramp to 160 °C at 20 °C/min, ramp to 250 °C at 4 °C/min, hold for 5 min. HRGC/HRMS measurement was performed on a HewlettPackard 5890 Series II gas chromatograph, fitted with an SE54 30-m × 0.25-mm-i.d. fused silica column with a layer thickness of 0.25 µm and connected through a heated transfer line, maintained at 280 °C, to a Finnigan 8200 double-focusing mass spectrometer controlled by a MASPEC data system (Version 2.11) for Windows 95 (MSS, Manchester, UK). All samples were injected manually; the injection volume was 1 µL, splitless. The injection port temperature was held at 260 °C, and a helium carrier gas flow rate of 1.5 mL/min was maintained (measured at 60 °C, split 1:10). The column temperature program was initial 50 °C, hold for 1 min, ramp to 150 °C at 70 °C/min, hold for 5 min, ramp to 260 °C at 2 °C/min, hold for 30 min. EI mass spectra of GC effluents were scanned from m/z 520 to 33 at a rate of 1 s/decade, with an interscan delay of 0.2 s, at an ion source temperature of 250 °C, an electron beam current of 1 mA, and an ion acceleration voltage of 3 kV. The electron beam energy was 70 eV. In the SIM mode, FC43 was used as mass calibrant. Cleanup for Waste Pyrolysis Oils and Sludge. Cleanup for Nonpolar Dihydroxybiphenyls (Standard Compounds, TMDHB and TCDHB). One gram of the sample was dissolved in 30 mL of acetone, and the solution was filtered through a folded filter. After the filter was washed with an additional 20 mL of acetone, 50 mL of 0.1 M NaOH was added, and the organic solvent was removed by rotary evaporation. The remaining solution was washed with 50 mL of hexane and acidified to pH 3 with 1 M HCl. Subsequently, the solution was washed twice with 50 mL of hexane, the combined hexane phases were washed twice with 0.1 M NaOH, and the combined aqueous phases were prepared for solidphase extraction by acidifying to pH 3 with 1 M HCl. Cleanup for Polar Dihydroxybiphenyls (Standard Compounds, DHB and TMxDHB). One gram of the sample was dissolved in 30 mL of toluene, and the solution was filtered through a folded filter. The filter was washed with an additional 20 mL of toluene, and the dihydroxybiphenyls were extracted from the organic phase by partitioning two times with 50 mL of NaOH. The collected aqueous phase was acidified to pH 3 with 1 M HCl and, prior to the application of the solid-phase extraction, cleaned by partitioning with 50 mL of hexane. (B) Chlorinated Dibenzo-p-dioxins/-furans and Biphenyls. PCDD/F. Twenty grams of Kieselrot was spiked with internal 13C-labeled standard and extracted for 16 h with toluene. Cleanup followed the method of Hagenmeyer et al.,8 i.e., successive steps of liquid chromatography on alumina and silica. In addition to this, pyrolysis oil samples were pretreated by dissolving 2.5 g of the oil in hexane, adding silica (44% H2SO4), and refluxing for half an hour (H2SO4 treatment). Furthermore, it was necessary to repeat single steps within the cleanup (8) Hagenmeyer, H.; Brunner, H.; Haag, R.; Kunzendorf, H.-J.; Kraft, M.; Tichaczek, K.; Weberruss, U. Stand der Dioxin-Analytik. In VDI-Berichte 634; VDI-Verlag: Du ¨ sseldorf, 1987; pp 61-89.
procedure to separate interfering substances from the analytes. After addition of 1,2,3,4-TCDD standard, GC/MS determination was performed on DB-5 and CP-Sil 88 capillary columns using the isotopic dilution method. For determination of PCDD/F in Kieselrot samples, a rapid method has been developed:9 ultrasonic extraction with methanol and hexane for 15 min each was performed. After sedimentation, an aliquot was taken and analyzed by GC/MS on a DB-5 capillary collumn. PCB. One gram of pyrolysis oil was dissolved in hexane. H2SO4 treatment was used for cleanup. After the solvent was evaporated, quantitative analysis was performed by external calibration using GC/MS. Reference standards were PCB 28, 52, 101, 138, 153, and 180. HRGC/MS Parameters. HRGC/MS measurement was performed on a Hewlett-Packard 5890 Series II gas chromatograph, equipped with a temperature-programable KAS 2 injection system from Gerstel, and a Hewlett-Packard 5970 B MSD. Helium was used as carrier gas, with a flow rate of 0.7 mL/min. For analysis, different capillary columns were employed under different conditions. The determination of the sum of PCDD/F isomers as well as the determination of PCB and chlorinated benzenes (CBs) was performed on a DB5 30-m × 0.25-mm-i.d. column with a layer thickness of 0.25 µm from J&W. The temperature programs for PCDD/F were as follow: (1) column, 80 °C, 1 min, 30 °C/min to 200 °C, 10 °C/min to 300 °C, 35 min; (2) KAS 2, 80 °C, 12 °C/s to 300 °C, 2 min, 12 °C/s to 350 °C, 2 min. The injection volume was 2-3 µL. The temperature programs for the determination of PCB and CBs were as follow: (1) column, DB-5 (PCB, CB), 60 °C, 4 min, 10 °C/min to 300 °C, 15 min; (2) KAS 2, 60 °C, 12 °C/s to 300 °C, 2 min, 12 °C/s to 350 °C, 2 min. In this case, the injection volume was 1 µL. To determine the 2,3,7,8-congeners of PCDD/F, a CP-Sil 88 50-m × 0.25-mm-i.d. column with a layer thickness of 0.2 µm from Chrompack was used. After injection of 2-3 µL of the sample, the following temperature programs were started: (1) column, 80 °C, 1 min, 30 °C/min to 180 °C, 10 °C/min to 240 °C, 60 min; (2) KAS 2, 80 °C, 12 °C/s to 300 °C, 2 min, 12 °C/s to 350 °C, 2 min. RESULTS AND DISCUSSION The method developed is based on two assumptions. First, it was assumed that the formation of diphenoquinones in native samples does not depend on the presence of chlorine as a substituent. Hence, it appeared important to develop a method suitable also for the determination of other compounds of this class of substances. The second assumption was that the products of many different thermal and oxidative processes are in consideration as matrixes that may contain the analytes. As a result, the cleanup procedure should be applicable to a variety of matrixes with little or no modifications. The most important property of this class of substances that was taken into consideration for the development of the analytical method was their redox behavior. Experience has shown that Soxhlet extraction with acetone as well as liquid/liquid partitioning (9) Bahadir, M.; Lorenz, W.; Schmidt, C. Fresenius Environ. Bull. 1992, 364369.
Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
2833
with hexane and water shifted the equilibrium completely to the dihydroxybiphenyls side. Similar observations were made during the instrumental analysis of the substances. Analysis of diphenoquinone and dihydroxybiphenyl standards by gas chromatography as well as by HPLC resulted in peaks with identical retention times. Therefore, the position of the redox equilibrium in the samples is no longer ascertainable after the analytical cleanup. Cleanup. For the solid-phase extraction, an RP-phenyl phase was tried and tested. The phase retained all four analytes by nonpolar and by π-electron interactions. Previous tests with an RP-18 phase showed the passage of both substances of the polar group during the application of a 100-mL sample volume. The solid-phase extraction in this method has mainly been employed for a change of solvent but also provides a further cleanup effect. After solid-phase extraction, the further workup depends on the instrumental method of analysis. HPLC samples can simply be dissolved in the desired amount of the HPLC eluting agent for injection. Prior to gas chromatography, derivatization with TFAA is necessary for the following reasons. First, the chromatographic behavior of the substances has significantly improved after derivatization. The derivatives have shorter retention times and show no tailing on the commonly used DB-5 column. Second, derivatization leads to better solubility of the analytes in hexane, which is especially important for the more polar substances. Finally, the derivatization reaction yields products with six fluoro atoms per molecule, thereby allowing the use of an electron capture detector as an alternative to the mass spectrometer for the determination of all four analytes. HPLC/ELCD Analysis. An electrochemical detector (ELCD) was chosen to facilitate selective and sensitive analysis of the substances. Additionally, electrochemical detectors offer the possibility to identify the analytes by comparing the hydrodynamic voltammograms from standards and samples. The use of hydrodynamic voltammograms making use of an electrochemical detector has been described in the literature.10 In the analytical method development, it is indispensable for the determination of the optimum measuring potential (Figures 3 and 4). The main disadvantage of the electrochemical detection method is its limitation to isocratic analysis. HRGC/ECD and HRGC/MS Analysis. In addition to the rapid HPLC method, a gas chromatographic method was applied. The preparation of the GC samples is more time-consuming because of the derivatization with TFAA. Nevertheless, gas chromatographic support was important for the following reasons. First, GC/MS analysis was developed to have an additional method of substance confirmation. Second, the resolution in gas chromatography is considerably higher (Figure 5). This was necessary, should there be the need to analyze samples containing traces of monophenolic substancessexperience has shown that, despite the selectivity of the detector, the isocratic HPLC method was not suitable for such samples. Phenols of polarity similar to that of the analytes are not eliminated by the cleanup and, hence, interfere with the analytes. With GC, the resolution is high enough to avoid coelutions. The GC method is also suitable for quantitative determination of the derivatives. In this case, 4,4′-dibromooctafluorobiphenyl (10) Wang, J. Electroanalytical Techniques in Clinical Chemistry and Laboratory Medicine; VCH Publishers: New York, 1989.
2834 Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
Figure 2. Workup scheme for the analysis of diphenoquinones and dihydroxybiphenyls.
Figure 3. Hydrodynamic voltammograms of the nonpolar standards TMDHB and TCDHB. The response plateau for both substances is reached at a measuring potential of +0.7 V, while the detector shows no response for 2,6-dichlorophenol (DCPh).
has been used as internal standard. All standard compounds and the internal standard showed linear response in the SIM mode of the GC/MS in a range between 10 and 400 ppb. Statistical Evaluation. The statistical evaluation was carried out as described in the Experimental Section. The cleanup and the derivatization reaction were evaluated separately in order to ascertain the different deviations arising from the use of GC or HPLC. Table 1 shows the data for the cleanup. After cleanup, three of the four substances showed very good recovery data. However, for the tetramethoxy compound, the recovery was found to be very low and of poor reproducibility. TMxDHB seems to disappear partly during extraction and also during the liquid/liquid partition. The chemical reaction which is responsible for the loss of the substance, possibly on oxidation of the electron-rich derivative, is yet unknown. As a result,
Figure 4. Hydrodynamic voltagrams of the polar standards TMxDHB and DHB. The optimum measuring potential for DHB is +0.6 V, although at this potential 2,6-dimethoxyphenol is also detectable. TMxDHB can already be detected at +0.4 V.
Figure 6. HPLC/ELCD chromatogram of the Kieselrot slag after analytical cleanup.
Figure 7. Substance identification by the comparison of hydrodynamic voltagrams.
Figure 5. HRGC/ECD chromatogram of a 300 ppb standard solution of each of the dihydroxybiphenyls after derivatization with TFAA and the internal standard 4,4′-dibromooctafluorodihydroxybiphenyl. Table 1. Statistical Data for the Evaluation of the Cleanup without Matrix Effects (Results after Running the Method 10 Times recovery (%)
av SD
DHB
TMxDHB
TMDHB
TCDHB
92 3.97
∼35
89 2.78
93 3.25
Table 2. Average Area and Standard Deviation of Six Similarly Treated Samples of a Standard Containing 200 ppb of Each of the Analytes, Measured after Derivatization in an HRGC/ECD System Equipped with an Autosampler area (n ) 6) DHB av SD SD (%)
369 039 20 651 5.6
TMxDHB
TMDHB
TCDHB
259 354 48 996 18.9
362 605 14 563 4.0
397 130 26 042 6.6
quantitative statements on the presence of the substance cannot be given. Table 2 shows the statistical evaluation of the derivatization reaction. Also during the reaction, a partial decomposition of TMxDHB was observed. The substance shows the lowest response but the highest standard deviation of all four compounds. Fly Ash and Kieselrot Samples. Workup. Two fly ash samples collected from the electrical precipitator of two different German incineration plants in 1993 and one sample of Kieselrot were analyzed for the nonpolar substances TMDHB and TCDHB.
Twenty-gram portions of each sample were used for the analysis. After the implementation of the cleanup, each sample residue was dissolved in 1 g of solvent, and the subsequent quantitative analysis was carried out by HPLC. Soxhlet extraction with acetone as a solvent was initially designated as the first extraction step of the method. However, after analysis of fly ash samples from a municipal waste incinerator, the recovery of the spiked analytes was found to be very low. The reason was that the highly active fly ash surface kept hold on more than 95% of the spiked analytes even after 36 h of Soxhlet extraction. With the help of acid pulping as described in the literature11 the recovery rate increased with good reproducibility up to 32% for TCDHB and 25% for TMDHB. This method was considered to be acceptable for the quantitative analysis of both fly ash and slag samples. Each sample was suspended in 5 M HCl for 30 min under constant stirring and subsequently filtered through a folded filter. The filter cake was washed with distilled water to neutral pH and dried before the subsequent 36-h Soxhlet extraction. Results of Kieselrot Samples. The HPLC chromatogram of the Kieselrot sample (Figure 6) shows a peak fitting the retention time of the TCDHB standard. Hydrodynamic voltammograms of the sample and of the standard were recorded during the same run to get a better indication of the identity of the substance (Figure 7). The diagrams were qualitatively in good agreement with each other. Of particular note, the response plateau was reached in both cases at a working electrode potential of +0.7 V, while at 0.55 V both peaks were hardly detectable. The quantitative difference of the graphs is a result of the different concentrations of TCDHB in sample and standard solution. The standard contained 200 ppb, while the concentration of the substance in the sample was found to be 149 ppb. Regarding the concentration factor of the cleanup and the recovery rate of 32%, the slag was (11) Ontario Ministry of Environment and Energy, Method E3141B, 1993.
Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
2835
Table 3. Dihydroxybiphenyls with Different Degrees of Chlorine Substitution Detected in the Kieselrot Sample no. of chlorine atoms
monitored m/z retention time (min)
1
2
3
3
4
5
6
412 + 315 nda
446 + 349 25.07
480 + 482 29.58
480 + 482 32.42
514 + 516 34.13
550 + 552 32.41
550 + 552 36.06
no. of chlorine atoms
monitored m/z retention time (min) a
5
6
6
6
6
6
6
550 50.06
584 26.36
584 26.58
584 28.05
584 29.22
584 34.48
584 37.53
Not detected.
Table 4. Quantitative Results of Chloroorganica Found in Kieselrot Samples substance
ppb
substance
ppb
hexachlorobenzene pentachlorobenzene pentachlorobenzonitrile pentachloroaniline hexachlorobenzofuran
94 200 20 000 9 000 1 000 15 000
(trichloroethyl)pentachlorobenzene hexachlorobenzothiophene hexachloronaphthalene nonachlorobiphenyl decachlorobipheyl
5 000 12 000 4 000 6 000 12 000
TeCDF PeCDF HxCDF
1 015 635 1 890
TeCDD PeCDD HxCDD
58 1 075 387
Furans HpCDF OCDF
9 520 8 560
Dioxins HpCDD OcCDD
525 1 350
contaminated with 23 ppb of TCDHB. The presence of this substance was additionally confirmed by HRGC/MS. The GC/MS/SIM technique was used to check the Kieselrot sample for chlorinated dihydroxybiphenyls. The compounds were not available as analytical standards, except for TCDHB and the 3,3′-dichlorinated compound. Therefore, it was decided to calculate the SIM data of the TFAA derivatives of the substances. These selected m/z values were monitored during the whole chromatogram run time. The SIM data were obtained for the monochlorinated compounds choosing the mole peak (M) and the expected main fragment (M - 97), which typically originates from TFAA derivatives by loss of COCF3. For the other derivatives with up to six chlorine atoms, the first two masses in the chloro cluster were calculated. Table 3 shows a list of the detected substances. A peak was counted as one of the possible isomers when the corresponding m/z values showed roughly the correct intensity relation at the same retention time. With respect to the workup procedure and the molecular mass of the substances, there is good certainty that the detected peaks are authentic TFAA derivatives of chlorinated dihydroxybiphenyls. The compounds containing two and four chlorine substituents are identified by retention time as 3,3′-dichlorodihydroxybiphenyl and 3,3′,5,5′tetrachlorodihydroxybiphenyl. All compounds containing less than five chlorine atoms are possible reduction products of diphenoquinones, unless the position ortho to the central biphenyl bond is substituted, as in the case of higher chlorinated compounds. Monochlorinated dihydroxybiphenyls were not found. A possible reason is the polarity of the compounds. The cleanup starts with an acid pulping step, with subsequent rinsing with 2836 Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
water. During this step, polar analytes in particular can be rinsed out. During the investigation on the distribution of Kieselrot in the region of Braunschweig, Germany, the slag was found to be contaminated with 21.60 ppm of PCDF and 3.40 ppm of PCDD,9 the PCDD/F with higher level of chlorination dominating. The highest concentrations were found for HpCDF (9.50 ppm) and OcCDF (8.50 ppm). These concentrations resulted in a calculated amount of 290 µg/kg TE. Moreover, hints for the existence of other chlorinated compounds were found during the GC/MS screening. Substances such as, e.g., pentachlorobenzene, hexachlorobenzofuran, hexachloronaphthalene, and decachlorobiphenyl could also be identified at the ppm level (Table 4). The Kieselrot slags from the copper smelting works in Marsberg, Germany, were dumped in a landfill during World War II. Later, Kieselrot was used to construct sport fields and tennis courts because of its stone characteristics and red color. Up to 800 000 tons of Kieselrot were distributed all over Germany.9 The generally high contamination of Kieselrot with chlorinated organics is a consequence of the presence of chlorides during the copper smelting process, which is performed in a temperature range between 550 and 600 °C.12 Notably, the average contamination with the extremely toxic 2,3,7,8-tetrachlorodibenzodioxin was found to be in the lower ppb range.13 It can be assumed that the structurally isomeric compounds 2,3,7,8-TCDD and 3,3′,5,5′-TCDQ (12) Krause, G. H. M.; Delschen, T.; Fu ¨ rst, P.; Hein, D. UWSF-Z. Umweltchem. O ¨ kotox. 1993, 5, 194-203. (13) Schuller, E.; Heinz, H.; Stoffers, H. UWSF-Z. Umweltchem. O ¨ kotox. 1995, 7 (1), 9-14.
Table 5. Appearance of Different Substituted Dihydroxybiphenyls in Products of the Waste Pyrolysis Plant concn (ppb)
light oil 1 light oil 2 light oil 3 heavy oil 1 heavy oil 2 decanter a
Figure 8. Chromatogram cutting of fly ash sample 1. Despite a coelution, TCDHB can be measured without interference at a working electrode potential of +0.7 V.
were formed during this process in comparable concentrations. Results of Fly Ash Samples. From both investigated fly ashes, sample 2 contained no TCDHB. Sample 1 showed a peak at 8.13 min, corresponding to the retention time of TCDHB standard. The hydrodynamic voltammogram showed increasing response in the working electrode potential between range +0.55 and +0.7 V. This is in agreement with the response of the standard. However, at potential values above +0.7 V, the peak area increased rapidly, and the retention time of the peak shifted slightly to 8.09 min (measured at +0.8 V, Figure 8). This variation of the retention time was observed repeatedly. These data indicate the presence of TCDHB and a slightly shifted, coeluting substance which is detectable only at higher working electrode potentials, above +0.7 V. The presence of TCDHB in the sample was also confirmed by GC/MS analysis. The quantitative result was calculated from the HPLC/ELCD data. Fly ash sample 1 was contaminated with 0.8 ppb of tetrachlorodiphenoquinone. Both fly ash samples were obtained from different German incineration plants and have to be distinguished in one important point. Sample 1 was collected from an electrostatic precipitator which operated in the temperature range between 250 and 400 °C. Precipitators such as these were found to be the major source of PCDD/PCDF emissions from waste incinerators in the past.14 In view of the dioxin problem, the electrostatic precipitator was replaced after the withdrawal of the sample by a model operating at lower temperatures. Sample 2 was obtained from a plant that was already equipped with an electrostatic precipitator operating at temperatures below 250 °C at the time of sample collection. Due to this, the formation of chlorinated organic substances is generally minimized. With the decrease of dioxin levels, the concentration of TCDQ in the fly ash obviously is reduced below the detection limit. Pyrolysis Oil Samples. Cleanup. Waste pyrolysis15 is developed as an alternative to waste incineration. During this (14) Leichsenring, S.; Lenoir, D.; Kettrup, A.; Mu ¨ tzenich, G. UWSF-Z. Umweltchem. O ¨ kotox. 1996, 8 (4), 197-206. (15) Scholz, R. Thermische Verfahren zur Abfallbehandlung; Prozessfu ¨ hrung, Bausteine und Bewertung. In Thermische Abfallentsorgung; VDI-Verlag, Du ¨ sseldorf, 1995; pp 1-47.
TMDHB
TCDHB
DHB
40 25 230 71 39 1021
nda 8 13 nda 60 352
243 nda 7 nda 193 not treated
Not detected.
process, the waste is heated under exclusion of oxygen up to 700 °C in a rotary kiln. The resulting crude gas is cooled to produce a heavy oil, a light oil, and a BTX oil fraction as well as cleaned gas. As a byproduct in the oil decanters, a sludge is precipitated. Currently, all these fractionssexcept the gassare fed back into the pyrolysis system. For the future there are considerations to recycle oils and gases as basic input materials for the chemical industry. It is generally known that the waste pyrolysis process works in a temperature range where the formation of dioxins and furanes is favored.16 Additionally, the literature suggests several secondary reactions that can lead to the formation of a wide variety of organic structures under these conditions.17 Analyses of light and heavy oils and of decanter sludge from a waste pyrolysis plant15 were performed in order to decide whether the samples would also contain diphenoquinones/dihydroxybiphenyls. From each sample, 1 g was used for the analysis. The basic cleanup was applied with changes as described in the Experimental Section. GC/MS samples were eventually taken up in 0.3 g of hexane containing 75 ppb of the internal standard. Recovery data were obtained by spiking 1-g portions of a chosen sample with 0.1 g of a solution containing 10 ppm of each analyte in acetonitrile. After the evaporation of the solvent, the samples were analyzed as described above. The following recovery data were found (n ) 3): DHB, 85 ((4)%; TMDHB, 20 ((7)%; TCDHB, 32 ((8)%. TMxDHB showed no recovery after the workup. Results. The results of the analysis are summarized in Table 5. The compounds were found in each waste pyrolysis product at lower ppb levels. The analyses show comparable concentrations of the substances in light and heavy oils, while they seem to be enriched in the sludge of the decanters, which shows the highest contents with 350 and 1020 ppb of TCDHB and TMDHB, respectively. Because the compounds are not used in any industrial application, their formation during the pyrolytic process appears to be plausible. This correlates with investigations on the occurrence of PCDD/F and PCB in the pyrolysis oils. The samples contained 50-362 ppb of PCDF, primarily TeCDF and PeCDF. These substances did not result from the input material, which contained lower amounts of PCDD/F (primarily OcCDD, 0.07-64 ppb), but were formed during the pyrolytic process. Furthermore, a linear relationship between the amounts of PCB (16) Buser, H. R.; Bosshardt, H. P. Chemosphere 1978, 1, 109-119. (17) Ballschmiter, K. Nachr. Chem. Technol. Lab. 1991, 39, 988-1000.
Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
2837
(sum of indicator congeners, 6000-25 000 ppb) and PCDF was observed. The PCBs were also formed during the pyrolytic process. CONCLUSIONS Diphenoquinones/dihydroxybiphenyls seem to be formed whenever thermal oxidative processes occur. Regarding this fact, the mechanism of their formation is an interesting aspect. Direct oxidation of the phenols is one possible way for the formation of the redox couple. Furthermore, organic de novo synthesis has to be taken into consideration as a source for these compounds. Numerous organic structures can be generated as unintended byproducts in the presence of carbon, oxygen, and halogen atoms during thermal processes. A corresponding scheme of formation reactions was published in 1991.17 In this context, the discovery of the chlorinated diphenoquinones and dihydroxybiphenyls significantly enlarges the scheme. However, it should be pointed out that the reaction leading to their formation is not limited to chloro-substituted compounds. An important aspect concerns the toxicological potential of the novel products. Although our investigations have just begun, the structure of the compounds advises caution. The more or less free rotation of the two phenyl rings in the dihydroxybiphenyl form allows a coplanar conformation of the molecule. And it is this property which causes a pronounced increase in toxicity of similarily substituted PCB congeners. In the oxidized form, the coplanar conformation is even fixed by the central double bond. The resulting planar surface area of the diphenoquinone molecules
2838 Analytical Chemistry, Vol. 70, No. 14, July 15, 1998
is similar to that of the structurally isomeric dioxine,7 emphasizing the need for further toxicological investigations. The analyses show that, in matrixes obtained under thermally oxidizing conditions, dihydroxybiphenyls/diphenoquinones appear side by side with dioxins and analogous compounds, which could originate from competing formation reactions. The concentrations of both groups of contaminants in all cases were shown to be of comparable orders of magnitude. The analytical method presented in this paper provides a further analytical tool to characterize halogenated contaminants as they occur in thermal processes. For this reason, we propose to use the chlorinated diphenoquinones as marker substances for chlorinated dibenzo-p-dioxins. Phenols can also be coupled by enzymes.6,18 Therefore, the formation of diphenoquinones in the human body after consumption of phenol-rich food should be taken into consideration. Oxidative conditions which are caused by enzymatic activity could produce diphenoquinones from phenolic compounds without the competitive formation of dioxins or chlorinated dioxins. These conditions are possible in a variety of different matrixes. Furthermore, analyses should focus on biological samples, such as urine and blood, and on beverages, such as tea or coffee, as well as on smoked fish and meat. Received for review December 30, 1997. Accepted April 14, 1998. AC971390J (18) Booth, H.; Saunders, B. C. J. Chem. Soc. 1956, 940-948.
