Host-Guest Interactions between 2,4-Dichlorophenol and Humic

Oct 21, 2008 - Corresponding author: phone: +39 081 253 91 60; fax: +39 081 253 91 ... and thermodynamic parameters for host−guest complexes between...
6 downloads 0 Views 151KB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 8440–8445

Host-Guest Interactions between 2,4-Dichlorophenol and Humic Substances As Evaluated by 1H NMR Relaxation and Diffusion Ordered Spectroscopy ´ AND D A N I E L A Sˇ M E J K A L O V A ALESSANDRO PICCOLO* Dipartimento di Scienze del Suolo, della Pianta, dell’Ambiente e delle Produzioni Animali, and Centro Interdipartimentale di Ricerca per la Spettroscopia di Risonanza Magnetica Nucleare (CERMANU), Universita` di Napoli Federico II, Via Universita` 100, 80055 Portici, Italy

Received June 30, 2008. Revised manuscript received September 10, 2008. Accepted September 17, 2008.

1H NMR measurements of spin-lattice (T ) and spin-spin 1 (T2) relaxation times and diffusion ordered spectroscopy (DOSY) were applied to investigate the association of 2,4-dichlorophenol (DCP) with a soil fulvic (FA-VICO) and humic acid (HA-VICO), and a lignite humic acid (HA-LIG). The 1H T1 and T2 values of DCP were found to decrease with increasing humic concentration, indicating reduction in molecular mobility due to formation of noncovalent interactions. The increased shortening of relaxation times observed upon addition of HA suggested more extensive association of DCP with HA than with FA. The extent of binding was inferred from diffusion coefficients (D) which showed slower diffusion for bound DCP. At 1 mg mL-1 DCP was completely bound by 4.1 and 5.8 mg mL-1 of HA-VICO and HA-LIG, respectively, while full DCP association was not observed even up to 20 mg mL-1 of FA. This was reflected by association constants (Ka): 3.1 ( 0.3 M-1 for FA-DCP, and 15.5 ( 3.1 M-1 and 11.0 ( 1.2 M-1 for HA-VICO and HA-LIG DCP complexes, respectively. The stronger binding to HA is attributed to their larger hydrophobic character enabling formation of stable hydrophobic domains to which DCP becomes associated in host-guest complexes. DCP complexation within humic hydrophobic domains was confirmed by upfield chemical shifts and signal line broadenings observed in 1H NMR spectra. Similar chemical shift variations for the three DCP aromatic protons further indicated π-π interactions, rather than H-bonding, as the main driving force for noncovalent association between DCP and dissolved humic substances. Relaxation and diffusion 1H NMR techniques provide rapid and accuratemeasurementsofbindingconstantsandthermodynamic parameters for host-guest complexes between environmental contaminants and natural organic matter.

Introduction Chlorophenols are highly toxic compounds commonly released in the environment as waste incineration products, wood preservatives, and agricultural pesticides. In addition, * Corresponding author: phone: +39 081 253 91 60; fax: +39 081 253 91 86; e-mail: [email protected]. 8440

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008

they are also formed as byproducts during chlorine pulp bleaching industry and chlorine water disinfection (1, 2). Their movement and transport in the environment, as well as their bioavailability, biotoxicity, and biodegradation are greatly affected by dissolved natural organic matter or humic substances (3, 4). Therefore, both the prediction of the environmental fate of these xenobiotics and the development of efficient strategy for their remediation depend upon a clear understanding of the chlorophenols associations with dissolved humic substances. The interactions between aromatic hydrocarbons containing polar functional groups and dissolved humic substances are believed to have a predominantly noncovalent character (5, 6). Noncovalent molecular interactions can be directly and accurately evaluated by solution nuclear magnetic resonance (NMR) spectroscopy (7, 8). In fact, a noncovalent association between a small sorbate molecule and the sorbent leads to a reduction of the sorbate’s translational and rotational motion, and is shown by changes in spin-lattice (T1) and spin-spin (T2) relaxation times. As for studies with humic substances, NMR relaxation measurements were previously used to study the interactions between fulvic acids and phenol (9), benzene, pyridine (10), acenaphthenone (11), and acetonaphtone (12). T1 values were also used to investigate the interaction between a humic acid and naphtalene, 1-naphtol, and quinoline (13). In addition to the reduced molecular motion of sorbate, the formation of noncovalent interactions varies the effective molecular size of the bonded molecule and thus affects its diffusion behavior. Changes in diffusion can be followed by diffusion ordered 2D NMR spectroscopY (DOSY) that provides a direct correlation of translational diffusion to the chemical shift in the second dimension and has been proved to be a powerful tool for detection of intermolecular interactions, including formation of π-π complexes (14) and hydrogen bondings (15). In order to avoid the overlapping of humic and contaminant signals and directly observe the changes in the contaminant chemical environment, either fluorinated (12, 16), or 2H- (4, 10) or 13C-labeled (9, 13) contaminants were used to study interactions with humic matter. However, the use of such substituted or labeled compounds provides binding information only for the specific nuclei introduced in the contaminant molecule. This limitation may be overcome, in case of aromatic contaminants, by following the aromatic 1H atoms in their interactions with humic molecules (13). In fact, humic matter solutions at low concentrations generally exhibit very low resonances in the aromatic region of 1H NMR spectra, thereby are hardly interfering with the intensive proton resonances of aromatic contaminants. The aim of this work was to apply 1H NMR spectroscopy to investigate and quantify the association between dissolved fulvic and humic acids and 2,4-dichlorophenol, that is one of the top 65 priority pollutants listed by U.S. Environmental Protection Agency (USEPA). The extent of binding was assessed by measuring both T1 and T2 relaxation times and diffusion coefficients as a function of humic concentration. The affinity and spontaneity of humic-2,4-dichlorophenol associations was also estimated by the magnitude of binding constants and Gibbs free energy of noncovalent associations.

Materials and Methods Reagents. All reagents were purchased from Aldrich (Italy). 2,4-dichlorophenol (DCP) had 99% purity and was used 10.1021/es801809v CCC: $40.75

 2008 American Chemical Society

Published on Web 10/22/2008

without further purification. Deuterated solvents were purchased from Eurisotop (France). Humic Substances. A fulvic acid (FA) and two humic acids (HA) were isolated, as previously described (17), respectively, from a volcanic soil (Typic Fulvudand, Lazio, Italy), FA-VICO and HA-VICO, and a North Dakota leonardite (Mammoth Int. Chem. Co.), HA-LIG. Briefly, 200 g of each parent material were shaken overnight with 500 mL of a 1 M NaOH and 0.1 M Na4P2O7 solution. Humic acids (HA) were precipitated from the extracting solution by adding 6 M HCl until pH 1. Ashes were removed first by three cycles of dissolution in 0.5 M NaOH followed by flocculation in 6 M HCl, and then by shaking HA twice in a 0.25 M HF/HCl solution for 24 h. HA were redissolved in 0.5 M NaOH and passed through a strong cation-exchange resin (Dowex 50) to eliminate remaining di- and trivalent metals. The eluate was precipitated at pH 1, dialyzed, and freeze-dried. After homogenization, 30 mg of HA were suspended in H2O, titrated to pH 7 and freezedried again. FA were adsorbed on a XAD-8 column to eliminate soluble hydrophilic impurities (carbohydrates and proteins), and then eluted out of column by a 1 M NaOH solution, immediately neutralized, dialyzed against water, and freeze-dried. Solution-State 1H NMR Spectroscopy. Solution-state NMR spectroscopy was carried out on a Bruker Avance 400 MHz instrument operating at a proton frequency of 400.13 MHz and equipped with a 5 mm Bruker inverse broadband probe. All spectra were acquired by Bruker Topspin 1.3 software. DCP (1 mg) together with progressively larger amount (0-20 mg) of freeze-dried humic material was first added with 10 µL of CD3OD to ensure complete DCP dissolution, and then further dissolved in 1 mL of D2O, and transferred to NMR tubes. Since 1H NMR spectrum of DCPhumic solutions did not change after standing for several hours or several days, it was assumed that the equilibrium was rapidly reached and all samples were analyzed within 2 h after preparation. All NMR experiments were performed at 25.0 ( 0.1 °C. 1H NMR spectra were referenced to the chemical shift of the solvent, resonating at 4.7 ppm, and were performed under saturation of the HOD signal achieved with 54 dB attenuation of a 60W amplifier for a period of 2 s. Prior to the relaxation measurements, the paramagnetic oxygen was removed by bubbling with nitrogen gas for 5 min, followed by a sonication for 15 min. Longitudinal (spin-lattice) relaxation time constants (T1) were obtained by using the inversion-recovery pulse sequence (180°-τ90°-t)n, where n is the number of scans. A recycle time t of 200 s, and a set of τ delays ranging from 100 to 0.01 s were used to acquire T1 of the DCP standard, while for DCP samples added with FA and HA, the t was 50 s (for FA and low-concentrated HA) and 5-2 s (medium- and highconcentrated HA), and τ delays ranged from 40 to 0.005 s. Transverse (spin-spin) relaxation time constants (T2) were measured using a Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. A set of echo delays varying from 0.04 to 32 s were used for DCP standard and DCP added with FA (0.5-3.0 mg mL-1) which showed longer T2, while echo delays from 0.001 to 1 s were applied to the remaining DCP-humic samples. Both T1 and T2 relaxation time constants were obtained using a nonlinear least-squares fit of a single exponential to the resonance intensity. The fittings were calculated by the Origin version 6.1 (MA, U.S.) software. 2D-NOESY spectra were conducted by applying a phase sensitive NOESY sequence with HOD presaturation (54 dB attenuation of a 60W amplifier) during a relaxation delay (50-2 s) and a mixing time ranging from 3s (DCP standard) to 20-100 ms (FA- and HA-DCP samples). All spectra were acquired using 80 scans and 256 increments in f1. 2D-DOSY diffusion-ordered spectra were obtained using a stimulated echo pulse sequence with bipolar gradients

FIGURE 1. 1H chemical shifts of 2,4-dichlorophenol in the absence and presence of fulvic FA and humic acid HA. [DCP] ) 1 mg mL-1, VICO refers to soil extract, LIG to lignite extract. (STEBPGP), and a water gate 3-9-19 pulse train for water suppression. Scans (32-160) were collected using 2.0-2.5 ms sine-shaped pulses (4-5 ms bipolar pulse pair) ranging from 0.674 to 32.030 G cm-1 in 32 increments, with a diffusion time of 5-150 ms, and 8 K time domain data points. Diffusion coefficients of seven standard compounds of known molecular weight were measured in order to express diffusion as a function of molecular weight, from which diffusion data were approximated to molecular sizes. The selected standards CH3OH (32.0 Da), catechol (110.1 Da), caffeic acid (180.2 Da), catechin (290.0 Da), bromocresol green (698.0 Da), and two polystyrene sulfonates having 1100 and 6780 Da were dissolved and acquired in the same way as DCP samples described above. Diffusion coefficients were elaborated using Bruker Topspin 1.3 software.

Results and Discussion Humic Sample Characteristics. The relative percentage of 13C distribution in the different NMR spectral regions from CPMAS-13C NMR spectra of humic samples used in this work is reported and discussed in Supporting Information (SI) Table S1. 1H Chemical Shifts. The 1H NMR spectrum of DCP is shown in Figure 1 and consists of a doublet at 7.38 ppm (H3, J ) 2.4 Hz), a doublet of doublets at 7.14 ppm (H5, J ) 8.8, and 2.4 Hz) and another doublet at 6.89 ppm (H6, J ) 8.8 Hz). Upon addition of humic materials, the DCP proton signals (H3, H5, and H6) were shifted upfield, and, as the concentration of FA and HA increased, the shift was more significant, while the proton signals became progressively broader (Figure 1). The extent of signal shift and broadening was different for the three humic materials. In fact, while well resolved coupling pattern and unchanged chemical shifts VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8441

TABLE 1. T1 and T2 Relaxation and Correlation Time (τc) Values Calculated for H3, H5, and H6 Protons of 2,4-Dichlorophenol (1 mg mL-1) as a Function of Fulvic (FA) and Humic Acid (HA) Concentration in MeOD/D2O at 25°C humic concentration (mg mL-1) DCP, standard FA-VICO 0.5 1.0 2.0 3.0 5.0 7.0 20 HA-VICO 0.5 1.0 2.0 3.0 5.0 7.0 HA-LIG 0.5 1.0 2.0 3.0 5.0 7.0 a

T1 (s)

T2 (s)

H5

H6

H3

H5

H6

H3

H5

H6

22.07

6.93

5.25

2.277

2.288

1.063

1.3

0.5

0.8

3.23 1.18 0.76 0.50 0.26 0.16 0.09

1.99 0.80 0.50 0.32 0.17 0.10 0.06

1.54 0.58 0.37 0.23 0.13 0.08 0.05

0.687 0.223 0.155 0.097 0.051 0.044 0.022

0.581 0.178 0.119 0.064 0.030 0.027 0.016

0.533 0.175 0.107 0.058 0.029 0.022 0.014

0.8 0.9 0.8 0.9 0.8 0.7 0.7

0.6 0.8 0.7 0.8 0.9 0.7 0.7

0.5 0.6 0.6 0.7 0.8 0.7 0.6

0.36 0.14 0.12 0.07 0.04 0.03

0.26 0.11 0.09 0.06 0.03 NDa

0.22 0.10 0.08 0.05 0.03 ND

0.034 0.014 0.011 0.004 0.002 0.001

0.025 0.012 0.009 0.003 0.001 ND

0.023 0.011 0.009 0.003 ND ND

1.4 1.3 1.4 1.8 1.9 2.0

1.4 1.3 1.3 1.9 2.1 ND

1.3 1.2 1.2 1.8 ND ND

0.44 0.16 0.12 0.08 0.06 0.05

0.38 0.13 0.08 0.07 0.04 0.04

0.36 0.12 0.07 0.05 0.03 0.03

0.064 0.018 0.012 0.009 0.006 0.004

0.054 0.016 0.008 0.007 0.005 0.003

0.052 0.012 0.007 0.005 0.003 ND

1.1 1.2 1.3 1.2 1.2 1.5

1.1 1.2 1.4 1.3 1.2 1.4

1.1 1.3 1.4 1.3 1.4 ND

ND, not determined.

were still observed for H3, H5, and H6 after addition of 1 mg FA to DCP, the same amount of HA-LIG and HA-VICO already produced significant peak broadening and upfield shift of 0.07 ppm and 0.05 ppm, respectively. In the case of FA-VICO, comparable DCP broadening and upfield shift were observed only when 20 mg of FA was added (Figure 1). Furthermore, the presence of just 7 mg of HA-VICO caused a complete collapse of the DCP aromatic protons in one large and very broad signal, whereas the same amount of HA-LIG led only to a partial alteration of proton signals. No such upfield shift and broadenings were observed for the signals of HA, FA, and HOD. The peak broadening observed at increasing humic concentrations indicates an enhanced proton relaxation rate due to development of dipolar interactions arising from restricted mobility of DCP molecules (18). On the other hand, the upfield shift of aromatic proton signals is a result of an increased electron shielding around proton nuclei. This can be attributed to the formation of π-π interactions, where the center of magnetic anisotropy is shifted out of the DCP aromatic plane toward the center of newly formed π-π complexes, thereby increasing the electron shielding of all resonating proton nuclei (14). The π-π complexes may be formed either among DCP molecules themselves or between DCP and the aromatic components of humic materials. Another explanation for the upfield shift may reside in the formation of H-bonds between hydroxyl and chlorine substituents in DCP, and complementary groups in humic molecules (19). However, this should have caused a different chemical shift variation for H3, H5, and H6, which should be larger for the hydrophilic FA than for more apolar HA. Since this was not the case, the interaction between DCP and humic molecules should not be primarily related to H-bonding. Nevertheless, a contribution of H bonding to the effect of π-π binding forces might not be totally excluded. 1H Relaxation Times. The 1H spin-lattice (T ) and 1 spin-spin (T2) relaxation times measured for DCP as a function of humic concentration are reported in Table 1. Without humic materials, the T1 relaxation time for DCP proton signals was much greater for H3 (22 s) than for H5 8442

τc (ns)

H3

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008

(7 s), and H6 (5 s). This behavior dramatically changed in the presence of humic molecules. Upon addition of 0.5 mg of FA-VICO, the T1 values were still depending on the position of protons on the aromatic ring but were reduced to 3.2 (H3), 2.0 (H5), and 1.5 s (H6). The decline in T1 values continued up to 7 mg mL-1 of FA, where an equivalent spin-lattice relaxation rate was observed for all DCP protons. Conversely, already at the lowest HA concentration (0.5 mg mL-1), T1 was considerably shortened, becoming 0.44 s and less than 0.36 s for HA-LIG and HA-VICO, respectively (Table 1). Similarly to FA, the 1H T1 values were further decreased at larger HA concentrations. A decreasing tendency with increasing humic concentration was also observed for T2 relaxation times of DCP (Table 1). Again, similarly to T1, the decline in T2 was more significant in the presence of HA, in particular HA-VICO, than for FA. The observed decrease of T1 and T2 values as a function of humic concentration suggests an overall reduction in DCP molecular mobility, that, again, may be attributed to either DCP self-aggregation or formation of weak hydrophobic interactions with humic molecules. Specific interactions shorten T1 values of protons that are near to the interacting group (13), and, thus, the fast initial decline of T1 for H5 and H6 in respect to H3 upon FA-VICO addition could be attributed to specific noncovalent interactions. Nevertheless, the observed T1 and T2 values, inasmuch as chemical shifts discussed above and diffusion coefficients discussed below, are the weight averages for DCP in both the free- and boundstate. In the case of FA-VICO additions (0.5-5 mg mL-1), the different shortening of T1 values for H5, H6, and H3 was still similar to the T1 differences observed for DCP alone, thereby suggesting that the free DCP was still more abundant than the bound one, and only incipient noncovalent interactions were formed with FA-VICO. Conversely, already extensive interactions occurred between DCP and HA at the same concentration, being the DCP weakly bound to HA predominant over the DCP free state. Both T1 and T2 values in solution are dependent on the correlation time (τC), that can be defined as the effective average time needed for the molecule to rotate through one

radian (18, 20), and, thus, a larger τC corresponds to a slower motion. Molecular τC were calculated according to Carper and Keller (21) and are listed in Table 1. Slightly different τC values observed for H3 (1.3 ns), H5 (0.5 ns), and H6 (0.8 ns) for DCP alone indicated a complex anisotropic molecular motion. Upon addition of 0.5 mg of HA-LIG, τC increased to 1.1 ns for H5 and H6, while it remained unchanged for H3. A slight τC increasing tendency was further noted for all protons with enhancing HA-LIG concentration. A similar but more significant τC increase was found for HA-VICO. Since the τC indicates both local mobility and effective molecular size (18), the increased values for τC suggested a decreased mobility of DCP molecules within the humic system, as well as an increased molecular size of the DCP-humic complex. Hence, addition of HA to DCP resulted in the formation of noncovalent but rigid binding to HA. Conversely, shortened or unchanged correlation times were observed upon addition of FA-VICO. A slight increase with increasing FA concentration was noted only for H5, thereby suggesting that this proton was more affected by the binding process than H3 and H6. The weaker effect of FA on τC supports the assumption that DCP associated less with FA than with HA, and, as discussed for signal broadenings (Figure 1), makes the hydrogen-bonding contribution to reduced DCP mobility less likely than π-π interactions. A NOESY experiment was conducted to confirm the formation of noncovalent interactions upon the addition of humic matter. This experiment is based on dipolar interactions and provides information about nuclei separated in space by less than 5 Å (22). In respect to positive diagonal, the 1H-1H NOESY spectrum of DCP standard (SI Figure S1A) reveals a negative correlation between H6 and H5. This correlation is attributed to COSY artifacts due to scalar coupling occurring between these two ortho-coupled protons. Otherwise, there are no cross peaks related to NOE correlations. Thus, even though DCP is a hydrophobic aromatic molecule, no signs of self-aggregation were observed at the concentration of 1 mg mL-1 in CD3OD/D2O solution. Conversely, clear cross-peak signals among DCP proton resonances were observed upon addition of both FA (>2 mg mL-1) and HA (example shown in SI Figure S1C). The fact that NOESY cross peaks had the same phase sign as diagonal peaks may be explained with either formation of larger association complexes or exchange between bound and mobile DCP (23). However, the latter should be unlikely, since only one set of signals was observed on the NMR scale. In addition, even at short mixing times (40 ms), the NOE signals were accompanied by relevant spin-diffusion effect, that is known to occur only in slow tumbling regimes. Such slow tumbling is explained with the occurrence of large macromolecular systems (20, 23) and, here, can be attributed to association complexes of DCP with FA and HA. 1H Diffusion Coefficients (DOSY-NMR). DOSY experiments are widely used to detect molecular aggregation, since diffusion coefficient of a molecular system is inversely related to its hydrodynamic volume and, thus, molecular size (14, 24). However, the possibility of an artificially decreased diffusion due to an increased solution viscosity must be taken into account (25), and, therefore, diffusion coefficients were corrected for changes in solution viscosity. Moreover, the diffusion coefficients were also measured as a function of FA and HA concentration to exclude the effect on relaxation times of paramagnetic species possibly remaining in humic materials. The raw and viscosity-corrected diffusion coefficients (D) of DCP, FA-VICO, HA-VICO, and HA-LIG are presented in SI Figure S2. No significant influence was observed on diffusion constants by changes in samples viscosity. In all cases, the self-diffusion of DCP (DDCP,free ) 7.6 × 10-10 m2s-1) was decreasing with enhancing concentration of humic

matter (SI Figure S1B and D and Figure S2), indicating formation of larger molecular systems. Moreover, while D values were constantly decreasing with increasing FA-VICO concentration, a plateau of constant D values were instead reached for HA-VICO and HA-LIG starting from 5 and 7 mg mL-1, respectively (SI Figure S2). After reaching the plateau, D did not decrease further, thus indicating that DCP was already totally bound and no additional complexes occurred with HA. The observed D values of DCP (SI Figure S2) were related to apparent molecular sizes (Mw), according to a calibration curve relating molecular weight and diffusivity of known standards (D × 1010 ) 81.647 Mw-0.4911, R2 ) 0.9874, curve not shown). Such relation revealed that DCP (163 Da) increased its apparent molecular size to 610 ( 30, 3828 ( 200, and 3500 ( 100 Da, upon addition of 20 mg mL-1of FA-VICO and at the diffusion plateau of HA-VICO and HA-LIG, respectively. Although 610 Da value might be explained with a self-aggregation of DCP into tetramers, it would be very unlikely that an aggregate of larger dimensions than an eicosamer of about 4000 Da was formed upon addition of 20 mg of any HA. Moreover, the molecular sizes detected for DCP after reaching the diffusion plateau coincided with the molecular sizes of dissolved HA (SI Figure S2). Therefore, these results are more plausibly explained with the establishment of complexes between hydrophobic DCP molecules and HA hydrophobic domains. HA would then serve as a host for the DCP guest molecules, and once DCP was bound strongly enough, the whole complex would diffuse at an equal rate. Moreover, the occurrence of such host-guest complexes agrees with the noted upfield chemical shift discussed above, and with the reported observation that the nuclei complexed within hydrophobic cavities experience an increased magnetic shielding (26). Conversely, the lack of diffusion plateau for FA-VICO suggests that fulvic acids interacted less strongly with DCP than HA. This is explained by the prevalently hydrophilic character of FA (SI Table S1), that fails to offer sufficient hydrophobic domains, whereby DCP is preferentially associated. The values of diffusion coefficients were used to calculate the fraction of bound DCP (F), as outlined by Wimmer et al. in 2002 (27). Provided that a fast exchange occurs between free and humic-bound DCP molecules on the NMR scale: F)

DDCP,obs - DDCP,free Dcomplex - DDCP,free

(1)

where DDCP,obs is the measured apparent (weight averaged) diffusion constant of both free and humic-associated DCP (i.e.: DCP diffusion in SI Figure S2), Dcomplex is the diffusion constant of the DCP-FA or DCP-HA complexes when DCP is fully bound, and DDCP,free is the diffusion constant of the DCP molecule in the absence of humic matter. Since DCP was not fully complexed by FA, the Dcomplex for DCP-FA was not experimentally known. However, being DCP small in size in respect to FA, it was assumed that DcomplexDCP-FA ≈ DFA. This is in agreement with diffusion results for the DCP-HA complex, where D values for HA alone represented the upper boundary for DCP-HA binding (SI Figure S2). The calculated amount of DCP associated to humic matter (Figure 2) indicated that 1 mg mL-1 of DCP was completely bound by 4.1 ( 0.2 mg mL-1 of HA-VICO, 5.8 ( 0.8 mg mL-1 of HA-LIG, and 24.9 ( 1.0 mg mL-1 of FA-VICO. The amount of associated DCP to humic matter calculated from diffusion data was compared to that computed from T1 values in SI Figure S3, whereby it is clear that bound DCP derived from T1 values is largely overestimated. Since the evaluation of molecular interactions is usually related to the determination of an association (binding) constant, Ka, the fraction of bound DCP (F), as calculated VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8443

FIGURE 2. Percentage of humic-associated 2,4-dichlorophenol as a function of fulvic (FA) and humic acid (HA) concentration. [DCP] ) 1 mg mL-1, VICO refers to soil extract, LIG to lignite extract.

TABLE 2. Binding Constant (Ka), and Free Gibbs Energy (∆G0) for the Association of 2,4-Dichlorophenol with Fulvic (FA) and Humic (HA) Acids at 25°C Calculated Using Two Different Fitting Procedures: (A) a = 1.00, (B) a = 1.25 a ) 1.00 humic sample FA-VICO HA-VICO HA-LIG

a ) 1.25

Ka (M-1)

∆G ° (kJmol-1)

Ka (M-1)

∆G ° (kJmol-1)

4.5 ( 0.6 24.9 ( 7.6 18.7 ( 3.9

-3.7 ( 0.5 -8.0 ( 2.4 -7.3 ( 1.5

3.1 ( 0.3 15.5 ( 3.1 11.0 ( 1.2

-2.8 ( 0.2 -6.8 ( 0.8 -5.9 ( 0.4

from diffusion values, was further fitted versus the FA and HA concentrations, according to the following equation: Ka[host]free F)a 1 + Ka[host]free

(2)

where [host]free is the concentration of unbound humic matter, and was expressed as moles of organic carbon. A detailed derivation and application of eq 2 can be found elsewhere (27). Two different fitting procedures were used here. In the first approach, similarly as done by Wimmer et al. in 2002 (27), all binding sites were assumed to be identical and the global fits were performed for a ) 1. The experimental points (data not shown) were rather poorly fitted and the calculated Ka of 4.5, 18.7, and 24.9 M-1 were estimated with 13, 21, and 31% of error for FA-VICO, HA-LIG, and HA-VICO, respectively. Despite the large deviation, the magnitude of Ka indicated a generally extensive complexation of humic materials with DCP. The smallest Ka value was obtained for FA-VICO and implied that this material formed the least strong interactions with DCP. In the second approach, the heterogeneity of humic materials was taken into account and the constant a was left “unlocked” during fitting procedures. The best fittings were experimentally found for a ) 1.25. While the order of Ka values: FA-VICO < HA-LIG < HA-VICO, did not change (Table 2), the corresponding Ka were somewhat lower than for a ) 1, and in particular for HA-VICO, whose Ka value was decreased by 40%. However, more noticeably, the estimation errors were decreased down to 10, 11, and 20%, respectively. These findings agree with previous literature, where strong complexations result in Ka subjected to errors of 20% or more (27, 28). The derived Ka allowed calculations of Gibbs free energy of transfer (∆G°transfer ) -RTlnKa) of DCP to FA and HA (Table 2). In all cases, a negative value of ∆G°transfer was found, and 8444

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008

hence, even though a decrease in motion implies a decrease in entropy, the association of DCP with humic matter is spontaneous and overall energetically favorable. In this work we showed that the combination of 1H NMR relaxation measurements and 2D DOSY is a direct and accurate method to quantify the binding behavior of aromatic contaminants toward humic substances. The upfield chemical shifts, longer correlation times, and slower diffusion of aromatic proton signals of DCP in the presence of humic molecules were prevalently attributed to the formation of hydrophobic π-π complexes. Although the values of binding constants indicated extensive interaction between all humic molecules and DCP, this was less strongly complexed by FA than by HA. This phenomenon was attributed to the predominant hydrophobic character of HA, that favored a more effective trapping of the hydrophobic DCP molecules, thereby establishing associations similar to host-guest interactions. In addition, DOSY spectra indicated that it was not HA-LIG rich in aromatic C to form the strongest associations with DCP, but rather HA-VICO rich in hydrophobic alkyl C. It can be thus inferred that hydrophobic humic materials possessing large amounts of alkyl C can be expected to have a more significant impact on the environmental fate and bioavailability of phenolic contaminants, as also previously reported (29, 30) in the case of nonpolar aromatic pollutants.

Acknowledgments This work was partially conducted within the project FISRMESCOSAGR. D.S. gratefully acknowledges the fellowship received within this project. The collaboration of Prof. Domenico Acierno from the Faculty of Engineering of the Universita` di Napoli Federico II, in viscosimetry analyses is appreciated.

Supporting Information Available Investigation on CPMAS-NMR spectroscopy, viscosity, raw and viscosity corrected diffusion coefficients for DCP-FA and DCP-HA complexation, NOESY and DOSY spectra, notes for diffusion behavior of FA and HA with increasing concentration, and comparison of bound DCP estimated from T1 and diffusion measurements. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Ahlborg, U. G.; Thunberg, T. M. Chlorinated phenols: occurrence toxicity metabolism and environmental impact. CRC Crit. Rev. Toxicol. 1980, 7, 1–35. (2) Okawa, K.; Nakano, Y.; Nishijima, W.; Okada, M. Effects of humic substances on the decomposition of 2,4-dichlorophenol by ozone after extraction from water into acetic acid through activated carbon. Chemosphere 2004, 57, 1231–1235. (3) Weissenfels, W. D.; Klewer, H. J.; Langhoff, J. Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: influence on biodegradability and biotoxicity. Appl. Microbiol. Biotechnol. 1992, 36, 689–696. (4) Nanny, M. A.; Maza, J. P. Noncovalent interactions between monoaromatic compounds and dissolved humic acids: A deuterium NMR T1 relaxation study. Environ. Sci. Technol. 2001, 35, 379–384. (5) Gauthier, T. D.; Seitz, W. R.; Grant, C. L. Effects of structural and compositional variations of dissolved humic materials on pyrene Koc values. Environ. Sci. Technol. 1987, 21, 243–248. (6) Engebretson, R. R.; von Wandruszka, R. M. A. Microorganization in dissolved humic acids. Environ. Sci. Technol. 1994, 28, 1934– 1941. (7) So¨derman, O.; Stilbs, P. NMR studies of complex surfactant systems. Prog. Nucl. Magn. Reson. Spectrosc. 1994, 26, 445–482. (8) Shirzadi, A.; Simpson, M. J.; Xu, Y.; Simpson, A. J. Application of saturation transfer double difference NMR to elucidate the mechanistic interactions of pesticides with humic acid. Environ. Sci. Technol. 2008, 42, 1084–1090. (9) Bortiatynski, J. M.; Hatcher P. G.; Minard, R. D. The development of 13C labeling and 13C NMR spectroscopy techniques to study

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17) (18) (19)

the interaction of pollutants with humic substances. In Nuclear Magnetic Resonance Spectroscopy in Environmental Chemistry; Nanny, M. A., Minear, R. A., Leenheer, J. A., Eds.; Oxford University Press: New York, 1997. Nanny, M. A. Deuterium NMR characterization of noncovalent interactions between monoaromatic compounds and fulvic acids. Org. Geochem. 1999, 30, 901–909. Nanny, M. A.; Bortiatynski, M.; Hatcher, P. G. Noncovalent interactions between acenaphtenone and dissolved fulvic acid as determined by 13C NMR T1 relaxation measurements. Environ. Sci. Technol. 1997, 31, 530–534. Dixon, A. M.; Mai, A. M.; Larive, C. K. NMR investigation of the interactions between 4′-fluoro-1′-acetonaphtone and the Suwanee river fulvic acid. Environ. Sci. Technol. 1999, 33, 958– 964. Simpson, M. J.; Simpson, A. J.; Hatcher, P. G. Noncovalent interactions between aromatic compounds and dissolved humic acid examined by nuclear magnetic resonance spectroscopy. Environ. Toxicol. Chem. 2004, 23, 355–362. Viel, S.; Mannina, L.; Segre, A. Detection of π-π complexes by diffusion ordered spectroscopy (DOSY). Tetrahedron Lett. 2002, 43, 2515–2519. Kapur, G. S.; Cabrita, E. J.; Berger, S. The qualitative probing of hydrogen bond strength by diffusion-ordered NMR spectroscopy. Tetrahedron Lett. 2000, 41, 7181–7185. Chien, Y.; Bleam, W. F. 19F NMR study of atrazine in humic and sodium dodecyl sulfate micelles swollen by polar and nonpolar solvents. Langmuir 1997, 13, 5283–5288. Piccolo, A.; Spiteller, M. Electrospray ionization mass spectrometry of terrestrial humic substances and their size fractions. Anal. Bional. Chem. 2003, 377, 1047–1059. Bakhmutov, V. I. Practical NMR Relaxation for Chemists; Wiley: West Sussex, UK, 2004. Beer, P. D.; Tite, E. L.; Ibbotson, A. Novel benzo crown ether cavitand and benzo crown ether-ferrocenyl host molecules that bind bipyridinium and sodium guest cations. J. Chem. Soc., Chem. Commun. 1989, 24, 1874–1876.

(20) Lambert, J. B.; Mazzola, E. P. Nuclear Magnetic Resonance Spectroscopy. An Introduction to Principles, Applications, Amd Experimental Methods; Pearson Education: Princeton, NJ, 2003. (21) Carper, W. R.; Keller, C. E. Direct determination of NMR correlation times for spin-lattice and spin-spin relaxation times. J. Phys. Chem. A 1997, 101, 3246–3250. (22) Mitchell, T. N.; Costisella, B. NMR - From Spectra to Structures. An Experimental Approach; Springer: Berlin, 2004. (23) Sanna, C.; LaMesa, C.; Mannina, L.; Stano, P.; Viel, S.; Segre, A. New class of aggregates in aqueous solution: an NMR, thermodynamic, and dynamic light scattering study. Langmuir 2006, 22, 6031–6041. (24) Morris, K. F.; Johnson, C. S. Resolution of discrete and continuous molecular size distributions by means of diffusion-ordered 2D NMR spectroscopy. J. Am. Chem. Soc. 1993, 115, 4291–4299. (25) Sˇmejkalova´, D.; Piccolo, A. Aggregation and disaggregation of humic supramolecular assemblies by NMR diffusion ordered spectroscopy (DOSY-NMR). Environ. Sci. Technol. 2008, 42, 699– 706. (26) Ahn, D.-R.; Kim, T. W.; Hong, J.-I. Water-soluble resorcin (4)arene: complexation of anionic aromatic guests by cooperativity of electrostatic and hydrophobic interactions. Tetrahedron Lett. 1999, 40, 6045–6048. (27) Wimmer, R.; Aachmann, F. L.; Larsen, K. L.; Petersen, S. B. NMR diffusion as a novel tool for measuring the association constant between cyclodextrin and guest molecules. Carbohydr. Res. 2002, 337, 841–84. (28) Cabaleiro-Lago, C.; Nilsson, M.; So¨derman, O. Self-diffusion NMR studies of the host-guest interaction between β-cyclodextrin and alkyltrimethylammonium bromide surfactants. Langmuir 2005, 21, 11637–11644. (29) Chefetz, B.; Deshmukh, A.; Hatcher, P. Pyrene sorption by natural organic matter. Environ. Sci. Technol. 2000, 34, 2925–2930. (30) Peuravuori, J. Binding of pyrene on lake aquatic humic matter: the role of structural descriptors. Anal. Chim. Acta 2001, 429, 75–89.

ES801809V

VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8445