Environ. Sci. Technol. 2001, 35, 379-384
Noncovalent Interactions between Monoaromatic Compounds and Dissolved Humic Acids: A Deuterium NMR T1 Relaxation Study M A R K A . N A N N Y * ,†,‡ A N D J E S U S P . M A Z A † School of Civil Engineering and Environmental Science and Institute for Energy and the Environment, 202 West Boyd Street, Room 334, University of Oklahoma, Norman, Oklahoma 73019
Deuterium nuclear magnetic resonance spectroscopy (2H NMR) spin-lattice relaxation (T1) experiments were used to examine solution-phase, noncovalent interactions between deuterated monoaromatic compounds (phenold5, pyridine-d5, benzene-d6) and Suwannee River, soil, and peat humic acids. Noncovalent interactions, in aqueous solution, were examined as a function of solution pH, monoaromatic hydrocarbon functional groups, and humic acid identity. Benzene interacted with dissolved humic acids at all pH values; however, these interactions increased with decreasing pH and generally were proportional with the humic acid percent aromaticity. Pyridine behaved similarly as benzene; however, two modes of interaction between pyridine and humic acids were detected as a function of pH and humic acid type: bonding with the lone pair of electrons of pyridine’s nitrogen and π-π interactions between the aromatic ring of pyridine and aromatic components of humic acid. The latter interaction was favored by increasing humic acid percent aromaticity and decreasing solution pH. On the other hand, because of its strong capacity for hydrogen bonding, phenol interacted preferentially with water, except at pH values 5 or lower and with humic acids with 45% or greater aromaticity. Under these conditions, strong interactions between phenol and humic acids were observed. These results demonstrate that solution-phase, noncovalent interactions between monoaromatic compounds and humic acids are a function of solution pH, percent aromaticity, and the monoaromatic functional group.
Introduction The role of dissolved natural organic matter (NOM) in the chemical and biological fate of organic compounds is coming to be understood as environmentally significant. NOM enhances the apparent solubility of hydrophobic organic compounds (1-6), while decreasing their bioavailability, thereby slowing biodegradation and lowering biotoxicity (7-9). NOM can act as an electron-transfer agent, thereby accelerating the rate of reductive degradation of organic halogens (10-12) and nitrobenzenes and chlorobenzenes * Corresponding author phone: (405)325-4234; fax: (405)325-4217; e-mail:
[email protected]. † School of Civil Engineering and Environmental Science. ‡ Institute for Energy and the Environment. 10.1021/es0012927 CCC: $20.00 Published on Web 12/14/2000
2001 American Chemical Society
(13) in anaerobic environments. Therefore, a detailed understanding of the chemical interactions between aromatic hydrocarbons and humic and fulvic acids is crucial to understanding the environmental chemistry of aromatic compounds in aqueous solutions. This is especially important for aromatic hydrocarbons containing polar functional groups and/or are fairly soluble and therefore can interact with humic and fulvic acids via other mechanisms besides hydrophobic sorption. These interactions are believed to be predominantly noncovalent interactions such as hydrogen bonding, π-π interactions, electrostatic attraction, van der Waals interactions, and dipole-dipole interactions. Nuclear magnetic resonance spectroscopy (NMR) can monitor noncovalent interactions between organic molecules in solution by measuring the spin-lattice (T1) relaxation time. Within the extreme narrowing limit, T1 is a function of the overall molecular translational and rotational motion; therefore, differences in the relaxation time can indicate changes in the motion of the sorbate molecule (14-16). Interactions with other molecules cause the small sorbate molecules to decrease their tumbling frequency and experience a decreased T1. Anisotropic motion of rigid aromatic molecules in solution, such as pyridine (17) and phenol and aniline (18), have been examined in solution by comparing relative T1 values of the NMR-sensitive nuclei within the examined molecule. T1 relaxation measurements have been used to characterize environmentally pertinent organic compound humic and fulvic acid interactions. Bortiatynski et al. (19) characterized and quantified noncovalent interactions between 13Clabeled phenol and Suwannee River humic acid. Likewise, Nanny et al. (20) used T1 relaxation measurements to examine noncovalent interactions between 13C-labeled acenaphthenone and Suwannee River fulvic acid dissolved in a methanol/D2O solvent. Three interactions were identifed: a weak sorption of acenaphthenone to fulvic acid, an enhanced solubilization of acenaphthenone by fulvic acid, and an interaction between the solvent and acenaphthenone. The enhanced solubilization was hypothesized to arise from acenaphthenone becoming encapsulated in three-dimensional fulvic acid structures predominantly solvated with methanol. Using a similar strategy, Nanny (21) used deuterium (2H) NMR to examine the noncovalent interactions between phenol-d5, pyridine-d5, and benzene-d6 and fulvic acids from the Suwannee River and Big Soda Lake. These monoaromatic compounds are structurally similar, i.e., monoaromatic hydrocarbons but differ in polarity and proton-donating or -accepting capability. Use of 2H NMR is advantageous because deuterium is a relatively sensitive nuclei for NMR experiments and also because deuterium relaxes predominantly via a quadrupolar relaxation mechanism, thereby greatly simplifying interpretation of T1 data. As well, 2H NMR is advantageous since these monoaromatic compounds lack a strong fluorescence signal and cannot be examined using fluorescence methods typically employed in the past to examine interactions between polycyclic aromatic hydrocarbons such as pyrene and phenanthrene and humic and fulvic acids. This paper continues this research by examining interactions of benzene-d6, pyridine-d5, and phenol-d5 with three humic acids of varying aromaticity and origin.
Experimental Section Materials. Suwannee River, soil, and peat humic acids were purchased from the International Humic Substance Society (IHSS). Deuterated solvents benzene-d6, pyridine-d5, and VOL. 35, NO. 2, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
379
phenol-d5 (Cambridge Isotope Laboratories) were used as received. Hydrochloric acid (Fisher Scientific) was trace metal grade and sodium hydroxide (EM Science) contained no more than 0.001% iron. Water was treated to a resistance of 18 MΩ-cm with a Sybron/Barnstead system containing, in series, an anion-exchange cartridge, a cation-exchange cartridge, and an activated carbon cartridge. Sample Preparation. Humic acid solutions (∼1 mg of humic acid/mL) were prepared by dissolving a precisely weighed amount of humic acid in ultrapure water and sonocating for 20 min to ensure complete dissolution. NMR samples containing 0.1% (v/v) deuterated monoaromatic compound (either pyridine-d5 or benzene-d6) were prepared in 5 mm NMR tubes (Wilmad Glass) by adding 1 µL of deuterated solvent to 1 mL of humic acid solution. A 0.12 M stock solution of phenol-d5 was made, from which aliquots were added to the NMR tubes to produce a phenol-d5 concentration of 0.012 M. Humic acid solutions and deuterated solvents were measured and transferred using a calibrated adjustable pipet with acid-washed plastic tips (Oxford). Sample pH was measured with a 3.5 × 183 mm glass electrode (Cole-Parmer, model 5990-30) and a model 12 research pH meter (Corning Scientific Instruments). pH was adjusted in the NMR tube by adding 1-µL aliquots of concentrated trace metal HCl (∼12 M) or NaOH stock solution (∼0.3 M). All glassware, NMR tubes, and plastic pipet tips were washed with trace metal HCl prior to use. Samples were made approximately 1-2 h prior to NMR analysis and kept in the dark at ambient temperature. Viscosity Experiments. Viscosity was measured in triplicate with an Ostwald-type viscometer (ASTM kinematic viscometer; size 25) at 20.5 ( 0.5 °C in a constant temperature water bath. Sample conditions (humic acid concentration, pH) were the same as with samples used in NMR experiments, except for the absence of the deuterated monoaromatic compounds in the viscosity measurement samples. NMR Experiments. Deuterium NMR spectra were recorded on a Varian 300 MHz operating at a resonance frequency of 46.044 MHz and a sweep width of 1000 Hz, utilizing the lock channel on a 5-mm standard 1H,13C probe (D300-5 probe, Nalorac Cryogenic Corporation). Temperature was maintained within a constant range of 20.7 ( 0.3 °C. T1 was measured with a standard inversion-recovery sequence using gated decoupling. A recycle delay of at least 5 T1 was employed, and 9 τ values per experiment with 120 scans per τ value were used. The data were processed with an exponential multiplication corresponding to a 1.0 Hz signal width. Chemical shifts were referenced to deuterium oxide sealed in glass capillary tubes inserted into the NMR tube. The spectrometer was initially shimmed and locked on neat benzene-d6, then the observe cable was connected to the lock channel of the probe and the spectrometer run unlocked while recording 2H spectra. Magnetic field stability was demonstrated by the lack of change in the signal width of neat benzene-d6 before and after each run. Replicate T1 measurements were made for several different phenol-d5-, pyridine-d5-, and benzene-d6-humic acid samples, and in all cases, there was less than a 10% difference. The uncertainty for each measurement was calculated as part of the computer processing of T1.
Results & Discussion Effect of Viscosity and Interaction with Pure Water. Solvent viscosity affects the solute correlation time, which in turn affects the T1 of the solute. The solute correlation time is the time it takes for a solute molecule to return to its original state after moving through its allowed rotational and vibrational states. Equation 1 (22) expresses the relationship between solvent coefficient of viscosity, η, and the correlation 380
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 2, 2001
TABLE 1. Calculated Solute Correlation Time for Benzene and Pyridine Molecules and Corresponding T1 Relaxation Time Based on Solvent Coefficient of Viscositya solvent
η (cP)
τc (s)
T1 (s) benzene
T1 (s) pyridine 0.735 (R) 0.711 (β) 0.724 (γ) 0.742 (R) 0.717 (β) 0.730 (γ) 0.755 (R) 0.730 (β) 0.742 (γ) 0.722 (R) 0.698 (β) 0.711 (γ) 0.752 (R) 0.727 (β) 0.740 (γ)
pure water
1.018 2.7069 × 10-12
0.646
soil humic acid
1.010 2.6847 × 10-12
0.652
peat humic acid
0.993 2.6385 × 10-12
0.663
Suwannee River 1.037 2.7567 × 10-12 humic acid
0.635
Suwannee River 0.996 2.6475 × 10-12 fulvic acid
0.661
a Humic or fulvic acid concentration ) 1 mg/ mL for all solvents; temperature ) 20.5 °C.
time, τc, of the monoaromatic molecule:
τc ) (4π fa3/3kT)η
(1)
where f is the microviscosity factor of 1/6 that accounts for the fact that diffusing molecules do not experience a continuous medium (23), a is the molecular radius (2.49 × 10-10 m for benzene), k is the Boltzman constant, and T is the absolute temperature. In the limit of rapid isotropic motion, the relationship between τc and the T1 of a quadrulpole nucleus, i.e., deuterium, is given in eq 2 (24):
1/T1 ) (3π2/10)((2I + 3)/(I 2(2I - 1)))(e2qQ/h)2τc (2) where I is the spin quantum number (I ) 1 for deuterium), e is the elementary charge of an electron, q is the electric field gradient, Q is the nuclear quadrupole moment, and h is Plank’s constant. The term e2qQ/h is the quadrupole coupling constant, which for benzene is 196.5 ( 1.3 kHz (25). Thus, solvent viscosity determines the upper limit for the solute correlation time and, subsequently, the T1 for a quadrupolar nuclei such as deuterium. The coefficient of viscosity for each humic acid sample was constant over the pH range of ∼3-∼11; depending upon the sample, the average η varied between 0.993 and 1.037 cP with a standard deviation of ( 0.002 cP. Table 1 lists η, the calculated solute correlation time for a benzene molecule, and the corresponding T1 values, based only on the solvent coefficient of viscosity for each sample. Figures 1 and 2 show the measured T1 values for benzene and phenol, respectively, in pure water as a function of pH (only R and β deuterium data are presented for phenol; the T1 of the γ deuterium nuclei could not be measured accurately since it appeared as a shoulder of the β deuterium nuclei signal). The calculated and measured T1 values of benzene are similar, indicating that in pure water, solvent viscosity is the primary factor influencing the molecular motion of benzene. The measured T1 of phenol (0.10-0.25 s) are reduced nearly 3-6-fold as compared to the calculated T1 benzene values, demonstrating that τc of phenol has increased beyond that expected based just on solvent viscosity. This increase is presumably due to hydrogen bonding between phenol and water molecules. Interaction between phenol and water is further supported by the fact that the T1 of phenol is constant up to a pH of ∼9.5, after which it declines,
FIGURE 1. T1 relaxation time, as a function of pH, for benzene-d6 in the presence of water, soil humic acid, Suwannee River (SR) humic acid, and peat humic acid.
FIGURE 3. T1 relaxation time, as a function of pH, for pyridine-d5 in the presence of water, soil humic acid (A), peat humic acid (B), and Suwannee River humic acid (SR; C).
FIGURE 2. T1 relaxation time, as a function of pH, for phenol-d5 in the presence of water, soil humic acid (A), peat humic acid (B), and Suwannee River humic acid (SR; C). indicating a decrease in the motion of the phenol molecule. At pH values above the pKa of phenol (pKa ) 9.5), the phenolic anion has a strong electrostatic attraction to the hydrogen atoms of water. Comparison of benzene and phenol data demonstrate that, in addition to solvent viscosity, the chemical behavior of the monoaromatic functional groups with the aqueous solvent have a significant effect on the correlation time and T1. Figure 3 shows the change of T1 for the R, β, and γ deuterium nuclei of pyridine in pure water as a function of pH. The measured T1 of pyridine are less than the calculated T1 values for pyridine based on solvent viscosity (Table 1), using a molecular radius of 2.49 × 10-10 m and quadrupole coupling constants of 184.2, 187.3, and 185.7 kHz for the
R, β, and γ deuterium nuclei, respectively (26). However, the measured T1 values of pyridine are not as low as for phenol, indicating that pyridine does not interact as strongly with water as does phenol. The large difference between the measured T1 values of the γ deuterium nucleus and the R and β nuclei as compared to the calculated T1 values demonstrates that pyridine is experiencing anisotropic motion as well as that the differences of the quadrupolar coupling constants for each of the deuterium nuclei is not great enough to cause the observed differences in measured T1. At a pH of 3.4, pyridine displays distinctive anisotropic motion, i.e., rotation about the C2 axis, as evidenced by the similar but higher T1 values for the R and β deuterium nuclei (0.58-0.62 s) as compared to the lower T1 of the γ deuterium nuclei (0.33 s). As pyridine rotates about its C2 axis, the R and β deuterium nuclei take longer to relax because their motion is greater relative to that of the γ deuterium nuclei. At a pH below 5.4, pyridine is a protonated cation, readily forming hydrogen bonds with the partial negative charge of the oxygen atom of water. Above a pH of 5.4, pyridine is electrostatically neutral, but the lone pair of electrons of nitrogen can still form weak hydrogen bonds with water. This is demonstrated at pH >6 by the overall reduction in the T1 of the R and β deuterium nuclei as well as the decreased difference between the R and β and the γ T1. This illustrates that the molecular motion of pyridine is becoming more isotropic as solution pH increases. At a pH of 11.4, the T1 of the R, β, and γ deuterium nuclei are nearly equivalent showing that hydrogen bonding between water and pyridine is no longer prominent and that the molecular motion of pyridine is mostly isotropic. VOL. 35, NO. 2, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
381
TABLE 2. Carbon Distributions in Humic and Fulvic Acids As Determined from Quantitative Carbon-13 NMR Spectroscopy (34, 35) sample
% ketone (220-190 ppm)
% carboxyl (190-165 ppm)
% aromatic (165-110 ppm)
% acetal/aromatic (110-90 ppm)
% heteroaliphatic (90-60 ppm)
% aliphatic (60-0 ppm)
Suwannee River humic acid peat humic acid soil humic acid Suwannee River fulvic acid Big Soda Lake fulvic acida
8 5 6 7 2
19 20 18 20 20
37 47 50 24 12
9 4 4 5 -
7 5 6 11 14
21 19 16 33 52
a
Percent aromatic region is from 165 to 90 ppm.
TABLE 3. NMR Signal Width (Hz) at Half-Height sample peat humic acid + benzene peat humic acid + benzene peat humic acid + benzene
pH 8.9 5.4 3.4
signal width (Hz) 7.34 6.32 14.63
peat humic acid + phenol peat humic acid + phenol peat humic acid + phenol
10.1 6.2 2.7
12.75 (R) 35.41 (R) 33.54 (R)
soil humic acid + phenol soil humic acid + phenol soil humic acid + phenol
9.4 5.1 3.2
17.11 (R) 75.36 (R) 200 (R)
10.1 5.5 3.4
4.73 (R) 6.90 (R) 8.55 (R)
Suwannee River humic acid + pyridine Suwannee River humic acid + pyridine Suwannee River humic acid + pyridine
Interactions with Humic Acids. Table 2 compares structural attributes of the three humic acids used in this study as well as two fulvic acids used in a similar previous study (21). Paramagnetic interferences inherent in the humic acids, e.g., metal ions or free radicals, that could reduce T1 values were addressed by maintaining a constant humic acid concentration over the examined pH range. In this manner, any paramagnetic impurities originating from the humic acid remained constant through out the measurements; therefore, any changes in T1 resulted solely from changes in the overall molecular motion of the examined monoaromatic hydrocarbon. Changes in the chemical shift position of the deuterated monoaromatic solutes were not observed upon addition of the humic acids, indicating that covalent bonding was not occurring. This is similar to Anderson (27) and Herbert and Bertsch (28), who used 1H and 19F NMR to measure chemical shift changes caused by noncovalent interactions between pesticides or substituted fluorobenzenes with soluble humic acids. Although both studies detected chemical shift changes, these chemical shift changes were quite small and subtle, thereby making interpretation difficult. Table 3 shows that small to moderate increases in the signal width measurements illustrate that monoaromatics are interacting with the dissolved humic acids. T1 Relaxation Time: Benzene. Figure 1 shows the T1 data for benzene in the presence of humic acids as a function of pH. The most striking observation is that all the T1 values for benzene in the presence of humic acid are less than those of benzene in pure water, indicating that benzene is interacting with the humic acids. If benzene T1 values were only a function of solvent viscosity, they would be similar to that of benzene in pure water. This is supported further by the fact that the signal width of benzene increased in the presence of peat humic acid to 7.34 Hz from 2.48 Hz in pure water. Herbert and Bertsch (28) observed similar increases in the 2H NMR signal width of fluorobenezene-d5, from 2.25 Hz in pure water to 12.1 Hz at a pH of 7 and 18.2 Hz at a pH of 4 in the presence of Lakeland humic acid. However, possible paramagnetic contamination from the humic acid 382
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 2, 2001
was not accounted for and could be the cause for the increased signal width (29). The extent of interaction between benzene and the humic acids (Suwannee River e peat < soil) at pH g8 can be ranked by examining T1. Figure 1 illustrates that over the entire examined pH range, interactions between benzene and soil humic acid are independent of pH, while interactions with Suwannee River humic acid increase upon decreasing pH as indicated by the ∼10% decrease in T1 at a pH of 3.2. At pH 9, pyridine interacts with all three humic acids similarly; at the high pH, the T1 values are lower than those for pure water showing that interaction with humic acids is occurring. This interaction is anisotropic, as seen by the similar but greater T1 values for R and β deuterium nuclei, relative to the γ deuterium nucleus. However, the fact that the difference in T1 values between these nuclei is less in the presence of humic acids than in pure water shows that isotropic interactions are also occurring with humic acids, i.e., π-π interactions between the aromatic ring of pyridine and aromatic components of humic acid. At pH