1 Introduction

Polyhalogenated aromatic compounds, such as chlorophenols, constitute a major group of environmental pollutants having been extensively used as wood preservatives, pesticides, and herbicides (Puhakka and Järvinen 1992; Armenante et al. 1999). Chlorophenols undergo several natural environmental processes such as biodegradation, chemical and photochemical degradation, volatilization, dispersion, and adsorption (Czaplicka 2004). The transformations of environmental contaminants are often mediated by humic substances (HS), the major component of natural organic matter (Piccolo and Celano 1994; Piccolo et al. 1996; Simpson 2006; Celano et al. 2008). HS are ubiquitous in waters, soils, and sediments and are progressively regarded as supramolecular structures of self-associated small biomolecules held together by hydrophobic forces in metastable conformations (Piccolo 2001, 2002; Peravuori 2005; Simpson 2002), which can be disrupted by the action of organic acids (Piccolo et al. 1999).

Oxidative coupling reactions catalyzed by natural oxidoreductase enzymes, such as peroxidase, are believed to degrade chlorophenols in soil. Enzymes catalyze production of highly reactive free radicals by removing an electron and a hydrogen ion from chorophenol hydroxyl groups (Park et al. 1999; Kobayashi et al. 2001). The final products are dechlorinated and oligomerized/polymerized chlorophenols which are less bioavailable, less mobile, and less toxic. Due to the high cost and poor environmental activity of these enzymes, no serious attempts have been conducted to use oxidative enzymes in environmental remediation practices.

An efficient alternative is represented by biomimetic catalysts, such as synthetic metal-porphyrins (Sheldon 1994; Nam et al. 2000), which exert the same enzymatic activity without incurring in the same pitfalls. Water-soluble iron-porphyrins are nontoxic compounds which have been successfully applied as biomimetic catalysts to dechlorinate and detoxify chlorophenols (Fukushima et al. 2003a, b; Hahn et al. 2007) and other organic compounds (Traylor et al. 1984; Song et al. 1997; Meunier and Sorokin 1997). Metal-porphyrins were also found to catalyze polymerization of HS by coupling the humic phenolic moieties present in HS by either chemical oxidation (Piccolo et al. 2005; Šmejkalová and Piccolo 2006; Šmejkalová et al. 2006) or photo-oxidation (Šmejkalová and Piccolo 2005). However, the role of HS in the catalyzed transformation of chlorophenols was so far explained with enhanced oxidation and dechlorination reaction rates (Fukushima et al. 2003a, b; Hahn et al. 2007) and stabilization of iron-porphyrin catalyst on HS (Fukushima and Tatsumi 2006; Fukushima et al. 2007).

In this work, we aimed to show that oxidative reactions catalyzed by synthetic metal-porphyrins in the presence of humic structures inactivate and detoxify chlorophenols through co-polymerization in HS. The study was conducted with pentachlorophenol (PCP) and its fluorinated homologue pentafluorophenol (PFP). The co-polymerization of PCP with humic molecules was assessed by evaluating the molecular dimension of humic matter and the amount of PCP remained free after the catalyzed oxidative reaction with PCP. Due to the large natural abundance and NMR detectability of the 19F nucleus, PFP was selected to obtain more resolved and sensitive 19F-NMR spectra of co-polymerized products than those observable by less abundant 13C nuclei in 13C-NMR spectra of PCP.

2 Materials and methods

2.1 Materials

Pentachlorophenol and pentafluorophenol were acquired from Sigma Aldrich (Germany) at 99% purity and used without further purification. Information on the synthesis, structure, and characterization of the water-soluble meso-tetra-(2,6-dichloro-3-sulfonatophenyl)-porphyrinate of Fe(III) [Fe(TDCPPS)Cl] used here as a biomimetic catalyst (FeP) are reported elsewhere (Piccolo et al. 2005).

The humic acids (HA) were extracted from an oxidized coal (HA-COX) provided by Eniricerche S.p.A. (Italy) and from a North Dakota lignite (HA-LIG) provided by Mammoth, Int. Chem. Co. (USA). The original material was shaken overnight in a 0.5-M NaOH and 0.1-M Na4P2O7 solution under N2 atmosphere. The HA was precipitated from the alkaline extract by adding 6 M HCl until pH 1, purified by three cycles of dissolution in 0.1 M NaOH and precipitation in 6 M HCl, treated with a 0.5% (v/v) HCl–HF solution for 48 h, dialyzed against deionised water until chloride-free (Spectrapore, 3500 Mw cut-off), and freeze-dried. An aliquot of each HA was then suspended in 100 ml of deionised water and titrated (VIT 90 Videotitrator, Radiometer, Copenhagen, Denmark) with a CO2-free solution of 0.1 M KOH to pH 7. The resulting potassium humate was filtered through a Millipore 0.45 μ, freeze-dried, and homogenized in agate mortar. The chemical and spectroscopic characteristics of both HA-COX and HA-LIG were previously reported (Cozzolino et al 2001).

2.2 Co-polymerization reaction

Control humic solutions of both HA-LIG and HA-COX were prepared by re-dissolving 2.0 mg of potassium humate in 10 ml of deionised water. These solutions were modified as follows: (1) addition of 1 mg of PCP, to obtain a second control; (2) addition of 0.1 mg of FeP (100 μL of a 0.0072 mM FeP water solution) and 50 μL of a freshly prepared 5% (w/v) H2O2 solution as oxidizing agent, to obtain a third control of the reaction without PCP; (3) addition of the same amount of FeP catalyst, H2O2 and PCP as above, to induce the oxidative co-polymerization reaction. Further HA control solutions were prepared by adding either FeP or H2O2 alone, in the amounts described above. The solutions were vigorously stirred and then incubated in daylight. Preliminary high-performance size-exclusion chromatography (HPSEC) experiments indicated that oxidative polymerization was negligible after 24 h at room temperature. Therefore, after 24 h of reaction time, all samples were filtered through a 0.45-μm glass filter (Whatman GF/C) and injected in the HPSEC system. Glacial acetic acid (Carlo Erba, Italy) was added (about 60–70 μL) to both control and reaction solutions until pH-meter readings passed from about 6.5–7 to 3.5, and the solutions were re-analyzed by HPSEC.

PFP was also co-polymerized with HA to acquire 19F-NMR spectra of products. The reaction solution contained 10 mg of either COX or LIG potassium humate, 50 mg of PFP, 5 mg of FeP, and 250 μL of 5% (w/v) H2O2, in 50 mL of deionised water. After 24 h of incubation in daylight, the solution pH was brought to 1 by adding 6 N HCl and the precipitated HA was isolated by centrifugation, washed twice with deionised water, and freeze-dried. The freeze-dried reaction products were dissolved in deuterated DMSO for NMR spectral acquisition.

2.3 HPSEC

The HPSEC system was composed by a Shimadzu LC-10-ADVP pump equipped with a Rheodyne rotary injector and a 100-μL sample loop. A Polysep-GFC-P-3000 600 × 7.5 mm i.d. column and a Polysep-GFC-P-3000 75 × 7.5 mm i.d. pre-column (Phenomenex, Inc., CA, USA) were used. The elution flow rate was set to 0.6 mL min−1 for an eluting solution made of 0.1 M NaH2PO4, buffered at pH 7 with 0.5 M NaOH, filtered through Millipore 0.45 μm, and degassed with He. Detection was achieved at 30°C constant temperature by using a Gilson 116 UV-detector operating at 280 nm and a Unipoint Gilson Software to record each run automatically. The column total (V t = 13.47 mL) and void volume (V 0 = 6.29 mL) were measured with water and a Blue dextran (M w = 2,000 kDa) aqueous solution, respectively, and calibrated with polystyrene sulphonates of known molecular weights (130,000, 32,000, 16,800, 6,780, and 4,300 Da, Polymer Standard Service, Germany). The apparent weight-averaged molecular weight (M w) was calculated from triplicate HPSEC chromatograms, as previously reported (Piccolo et al. 2005). The M w are apparent values because not only molecular associations were measured instead of single molecules, but also because there are no standards which may precisely represent complex humic structures. The relative standard deviation of calculated values varied to a maximum of 7%.

2.4 Quantification of co-polymerized PCP

Different PCP concentrations (0.1, 0.5, 1, 5, and 10 μg mL−1) were prepared in 1 mL of deionised water and added with 0.2 mg of either COX or LIG potassium humates, 1 μg of FeP (100 μL of a 7.2 × 10−5 mM FeP water solution), and 5 μL of a 5% (w/v) H2O2 solution, and incubated for 24 h. Control solutions were prepared with either PCP alone or added with the same HA amount as above. Each reaction and control solution was loaded on a Strata-X (Sigma Aldrich, Germany) SPE column to separate PCP from HA. The adsorbed PCP was then eluted with methanol, dried under N2, re-suspended in acetonitrile, and sylilated with BSTFA [N,O-bis(trimethylsilyl)trifluoroacetamide, Sigma Aldrich, Germany] at 80°C for 30 min. Sylilated samples were dried under N2 flux and then re-dissolved in 1 mL of n-hexane (MS grade, Merck, Germany) before determination.

Silylated PCP samples were determined with an RtX-5 30-m long, 0.25-mm i.d., fused silica capillary column (Restek Corporation, Bellefonte, PA) using a Clarus 500 gas chromatograph (PerkinElmer, Palo Alto, CA) equipped with an electron capture detector (ECD) and set with a split ratio of 20:1. Helium was the carrier gas (1 mL min−1) and a 96% argon/4% methane gas mixture was the make-up gas (30 mL/min). The injector temperature was set at 250°C and the oven temperature went from the initial 80°C to the final 200°C at 15°C min−1 rate. This temperature was kept for 5 min, then at 15°C min−1 the oven temperature reached 250°C, which was kept for 3 min more. PCP was determined by an external calibration curve, with relative standard deviation among triplicates of a maximum of 5%.

2.5 19F-NMR spectroscopy

19F-NMR spectra were acquired with a Bruker UltraShield 500 MHz magnet operating at a proton frequency of 500.13 MHz and at a fluorine frequency of 470.59 MHz, equipped with a Quadruple (QXI) Resonance Probe (5 mm PAQXI Z-grad), and using a Carr-Purcell-Meiboom-Gill sequence to suppress broad lines with proton decoupling. The 19F-NMR chemical shift scale was calibrated using 0.05% trifluorotoluene in CDCl3, resonating at −63.72 ppm.

3 Results and discussion

3.1 Analytical HPSEC

The effects of the catalyst-assisted oxidative co-polymerization of PCP into selected HA were assessed by evaluating the apparent weight-average molecular weights (M w) of resulting reaction mixtures for both HA-LIG and HA-COX. The co-polymerized PCP-HA solutions were injected into the HPSEC system either before or after addition of acetic acid that was used to lower the mixture pH to 3.5. It has been shown that the hydrophobically stable HA supramolecular conformations are altered when added with organic acids to lower the pH to 3.5, as new and stronger intermolecular hydrogen bonds are formed between complementary humic functions (Piccolo et al. 1999, 2002; 2003; Cozzolino et al. 2001). However, if humic supramolecular conformations are firmly stabilized by intermolecular covalent bonds induced by catalytic oxidative polymerization, their alteration by pH lowering is significantly reduced (Piccolo et al. 2005; Šmejkalová and Piccolo 2005).

The HPSEC chromatograms for both HA-LIG and HA-COX illustrate the molecular size distribution of two humic solutions subjected to different treatments and after acetic acid addition (Fig. 1). Elution profiles are relative to control HA solutions at pH 6.5 (Fig. 1a), and humic solutions undergone the catalyzed oxidative coupling reaction without and with PCP (Fig. 1b and c, respectively), whereas the modifications induced by acetic acid addition in the respective solutions are shown by chromatograms in Fig. 1a′, b′, and c′. For both control solutions, the diffuse peak at about 9 mL of elution volume in control chromatograms was reduced considerably after acetic acid addition, though more extensively for HA-COX than for HA-LIG. The absorption was shifted to larger elution volumes, whereby two additional peaks appeared in the case of HA-COX, whereas a single peak was shown for HA-LIG at about the column total volume. This behavior further confirms earlier reports that indicated HA signals in HPSEC as due to apparently large, though weakly associated molecular sizes, which are easily disrupted by pH change (Piccolo et al. 1999, 2002; 2003).

Fig. 1
figure 1

HPSEC chromatograms of HA-COX and HA-LIG solutions under different treatments: (a) HA control solution before and (a′) after acetic acid addition; (b) HA + FeP reaction solution without PCP and (b′) after acetic acid addition; (c) HA + FeP + PCP reaction solution and (c′) after acetic acid addition

Full size image

The elution profiles of humic solution varied substantially when the materials were treated with FeP catalyst and H2O2 oxidant (HA + FeP; Fig. 1b). For both HA, the first peak at about 9 mL increased significantly in absorbance as compared to control, and, in the case of HA-LIG, it even appeared at slightly lower elution volumes. The concomitant presence of PCP in the reaction solution (HA + FeP + PCP) increased further the first peak intensity (Fig. 1c), thereby implying a contribution of PCP to the peak absorbance. Contrary to control, when the HA + FeP and HA + FeP + PCP solutions were treated with acetic acid (Fig. 1b′ and c′), the first peak maintained a similar, if not larger, absorbance, though its maximum was shifted to larger elution volumes. Moreover, while the HPSEC profile of HA-COX added with acetic acid was similar to control, the HA-LIG solution showed two more chromatographic peaks which were not present in the AcOH-treated control solution. The limited action of acetic acid to modify the chromatographic behavior of both HA + FeP and HA + FeP + PCP solutions to the same extent as that of the control suggests that intermolecular covalent bonds must have been formed among humic molecules and/or between these and PCP. This means that the oxidative biomimetic catalysis was successful in stabilizing the previously loose humic conformation, thus implying the occurrence of HA polymerization (Piccolo et al. 2005; Šmejkalová and Piccolo 2005) and PCP-HA co-polymerization.

The M w values obtained from size-exclusion chromatograms are shown in Table 1. The humic solutions added with PCP alone (not shown in Fig. 1) revealed a decrease in the M w value, in respect to control solution, more for HA-COX than for HA-LIG. This should be interpreted with the ability of the poorly water-soluble PCP to interact with the hydrophobic domains of humic superstructures, and thus weaken the humic conformational stabilities (Cozzolino et al. 2001). Nevertheless, it appears that PCP established stronger hydrophobic interactions with HA-COX than with HA-LIG and thus limited the conformational disruption of HA-COX following AcOH addition (positive ΔM w values in respect to control; Table 1). When the humic solution was subjected to the catalyzed oxidative reaction the behavior observed in chromatograms either without or with PCP (Fig. 1) was reflected by M w values, which were significantly larger than control (Table 1). This may be, respectively, explained with either a self-polymerization among humic molecules or a co-polymerization of PCP with humic molecules. Moreover, the M w values for both HA-COX and HA-LIG remained significantly larger than for control even after acetic acid addition. This confirmed that humic conformations had been stabilized by the occurrence of covalent bonds either among humic molecules or between PCP and humic molecules (Piccolo et al. 2005; Šmejkalová and Piccolo 2005). However, it must be noted that M w values for the oxidative coupling either with or without PCP did not show a substantial difference, thus implying that self-polymerization of humic molecules was equally conformationally determinant as PCP co-polymerization.

Table 1 Apparent weight-averaged molecular weight (M w) and their standard deviation (±SD) of humic solutions under different treatments, as measured from HPSEC chromatograms before and after acetic acid (AcOH) addition
Full size table

3.2 GC–ECD

In order to verify whether PCP was really co-polymerized in humic matter, the PCP remaining free in solution after the co-polymerization reaction was isolated and quantified by gas-chromatography/electron capture detector (GC–ECD; Table 2). Free PCP recovered from solution with HA was slightly different from that obtained from the blank solution without HA, the deviations growing larger with greater PCP concentration. The differences should be attributed to a physical adsorption of PCP on the humic structures. Conversely, when PCP was subjected to the catalyzed oxidative reaction in the presence of both HA materials, the contaminant remaining free in solution at all concentrations significantly decreased (Table 2), thus implying that PCP was mostly bound to humic molecules by a covalent coupling promoted by the FeP catalyst.

Table 2 Added and recovered PCP (mg L−1), as detected by GC/ECD, when present alone in solution (blank), in addition to HA (PCP + HA), and in the catalyzed oxidative co-polymerization reaction (PCP + HA + FeP)
Full size table

Oxidized PCP reaction products were not detected by GC–ECD (Fig. 2), thus implying that they, if formed, may have also been co-polymerized in humic matter. Moreover, it was noted that the PCP covalent binding to both humic materials increased with the progressively larger PCP concentration in the reaction mixture. In fact, at the greatest PCP concentration (10 mg L−1), the disappearance of free PCP from solution was nearly 100% for both HA, thereby suggesting a large potential of humic matter to co-polymerize this contaminant. While the two HA did not generally behaved substantially different towards PCP catalyzed co-polymerization, PCP addition in no-catalyst conditions seemed to show a trend of larger adsorption on HA-LIG than on HA-COX (Table 2). This may be attributed to a greater presence of flexible long-chain alkyl molecules in HA-LIG (Cozzolino et al 2001), which may be conformationally more efficient in weakly interacting with PCP.

Fig. 2
figure 2

GC–ECD chromatogram of products after catalyzed co-polymerization reaction between HA-COX and PCP (10 mg L−1) under daylight for 24 h. Attenuation was raised from 0 to 10 after 7 min when sylilating byproducts peaks were not longer eluted

Full size image

3.3 19F-NMR spectroscopy

Both liquid- and solid-state 13C-NMR spectroscopy were previously used to determine whether 13C-labeled contaminants were weakly or covalently bound to humic substances (Hatcher et al. 1993; Guthrie et al. 1999). However, unless the contaminant is adequately labeled with 13C nuclei, it is hard to obtain 13C-NMR spectra sufficiently sensitive to show the contaminant 13C signals above those of HA. In this work, we employed pentafluorophenol, a homologue of PCP (Šmejkalová et al. 2009), to exploit the larger detectability of 19F nuclei in liquid-state 19F-NMR spectroscopy than that of natural abundant 13C nuclei in 13C-NMR spectra of PCP. In 19F-NMR spectra of HA solutions added with PFP, only the 19F nuclei are visible and it is thus immediate to verify whether or not PFP is covalently bound to HA. In fact, if PFP did not covalently couple with humic molecules, one would expect the 19F-NMR spectra to be similar to those of PFP alone. Conversely, if the catalyzed oxidative reaction was capable to covalently bind PFP to humic molecules and, thus, change the magnetic environment of PFP molecules, 19F-NMR spectra should show as many additional 19F signals at different chemical shifts as the new PFP different covalent bonds to humic molecules.

The 19F-spectra obtained after different experiments entailing PFP reactions were almost identical for both HA-COX (not shown) and HA-LIG (Fig. 3). Spectrum A in Fig. 3 was obtained with a solution containing HA, PFP, and H2O2, but without the FeP catalyst. This 19F-spectrum shows only three signals at −161.6, −165.3, and −171.8 ppm for the five fluorine atoms in PFP. Since it is identical to that of PFP alone (not shown), it indicates that no structural modification was induced in PFP by the concomitant presence of humic matter and H2O2.

Fig. 3
figure 3

19F-NMR spectra of reaction solutions: a HA-LIG + PFP + H2O2, b PFP + FeP + H2O2 (no HA), and c HA-LIG + PFP + FeP + H2O2 under exposition to daylight for 24 h

Full size image

Conversely, the reaction conducted without HA, but with PFP, H2O2, and the FeP catalyst, provided a different signals distribution in the 19F-spectra (spectrum b in Fig. 3). Though the signals of PFP as such were still visible, a number of additional resonances appeared in spectrum B, mainly in the downfield spectral region. These signals should be attributed to a de-halogenation and a self-oligomerization of PFP under oxidative catalytic conditions, based on signals attribution of similar NMR systems observed under enzymatic catalysis (den Besten et al. 1993; Cnubben et al. 1995; van der Bolt et al. 1997). Though the precise assignment of these signals cannot be provided here, the three signals at −138.1, −149.5, and −160.4 in spectrum b (Fig. 3) may be temptatively attributed to a PFP dimeric structure (Ando and Matsuura 1995). These results clearly show that only the action of the biomimetic catalyst was capable to produce a structural modification of PFP under oxidative conditions.

Furthermore, when the FeP-catalyzed oxidative reaction was conducted in the presence of humic matter, the 19F-spectra showed an even more complex spectral region between −140 and −160 ppm (spectrum C in Fig. 3) with many additional signals. However, some of the downfield signals observed in spectrum B for the reaction mixture without HA were not visible in spetrum C. These differences suggest that the presence of HA induced a diversification in the PFP reaction pathways leading to a larger number of 19F signals and chemical shift spread.

In line with previous results using unlabeled and 13C-labeled chlorophenols in catalytic reactions with humic substances (Hatcher et al. 1993; Hahn et al. 2007), the multitude of signals and the chemical shifts complexity in the 19F-spectrum, arising from the FeP oxidative catalysis in the presence of HA, indicates that PFP was mostly covalently bound to humic molecules. This finding combined to earlier results, which have shown that phenolic constituents of humic materials were polymerized in the same oxidative catalytic conditions of this study (Piccolo et al. 2005; Šmejkalová and Piccolo 2005), suggest that such catalytic reactions have led to the co-polymerization of PFP with both HA.

In conclusion, this work has shown that PCP may be covalently incorporated into humic matter upon an oxidative coupling reaction catalysed by a biomimetic iron-porphyrin. The catalytic co-polymerization of PCP with humic substances was indicated by: 1. the increase in molecular dimension of reaction products as compared to control HA, 2. the nearly complete disappearance of free PCP isolated from reaction solutions, and, 3. the occurrence of multiple 19F-NMR signals whose 19F nuclei become magnetically different due to formation of covalent bonds between humic molecules and PFP. All these results point out that the oxidative biomimetic catalysis can be an effective technology to co-polymerize polyhalogenated phenols in humic matter and detoxify environmental compartments such as waters, soils, and sediments.