Environ. Sci. Technol. 2007, 41, 185-191
Sorption of Aromatic Organic Contaminants by Biopolymers: Effects of pH, Copper (II) Complexation, and Cellulose Coating XILONG WANG,† KUN YANG,† S H U T A O , ‡ A N D B A O S H A N X I N G * ,† Department of Plant, Soil and Insect Sciences, University of Massachusetts, Amherst, Massachusetts 01003, and College of Environmental Sciences, Peking University, Beijing 100871, P. R. China
Sorption of hydrophobic organic compounds (HOCs) (i.e., pyrene, phenanthrene, naphthalene, and 1-naphthol) by original and coated biopolymers was examined. Lignin yielded nonlinear isotherms due to its glassy character. Except for pyrene, cellulose showed linear isotherms for other compounds, indicating a partitioning dominant mechanism. Sorption of 1-naphthol by lignin decreased with increasing pH, attributed to both the increased πeΘ-πeΘ repulsion and weakened hydrogen bonds, while the affinity reduction of cellulose for 1-naphthol with increasing pH resulted from only the decrease in H-bonding due to its absence of benzene ring. Complexation of lignin with Cu2+ increased the sorption affinity for phenanthrene (2.6 times) and slightly enhanced its isotherm nonlinearity. For 1-naphthol, lignin-Cu2+ complex had a much higher sorption capacity (7 times) than the original lignin, accompanied by the increased isotherm nonlinearity. Cellulose-coated lignin showed increased sorption affinity and more pronounced nonlinearity for 1-naphthol than the lignin-Cu2+ complex. In comparison, cellulose coating exhibited little effect on sorption affinity for phenanthrene relative to the lignin-Cu2+ complex. Isotherm nonlinearity of coated lignins increased with increasing cellulose coating, indicating more condensed domains produced, supported by an increase (from 68.9 °C for the original lignin to 82.4 °C for the highest cellulose coating level) in glass transition temperature (Tg). Results of this study highlight the importance of structure, polarity, surface O-containing functional groups, and surface charges of biopolymers in controlling HOC sorption.
Introduction Sorption of HOCs to soil/sediment organic matter (SOM) is a main factor that governs their fate in the environment. But, exact sorption mechanisms have not been well understood because of the extreme complexity of SOM molecular structures and compositions (1). Hence, sorption mechanism has been a major focus in environmental and geochemical researches. SOM is a complex, heterogeneous composite of diagenetically altered biopolymers of plant and animal origins (2). * Corresponding author phone: (413) 545-5212; fax: (413) 5453958; e-mail:
[email protected]. † University of Massachusetts. ‡ Peking University. 10.1021/es061389e CCC: $37.00 Published on Web 11/17/2006
2007 American Chemical Society
Biopolymers include polysaccharides (e.g., celluloses, hemicelluloses, starches), lignins, lipids, and cuticular materials (2). These biomolecules undergo gradual decomposition and diagenesis to form SOM (e.g., humic acid), and eventually kerogen (3). Polysaccharides receive increasing attention in various research fields because they have unique structures and characteristics that are different from synthetic polymers. Among many kinds of polysaccharides, cellulose is synthesized primarily in plants and is the most abundant organic compound on earth (4). Lignin is a natural component of plants, usually serving as a binding agent for cellulose and other biomolecules. Its structure is complex, consisting of aldehyde, keto, hydroxy, methoxy, and phenolic groups (5). Lignin and cellulose are major components of paper, which is the largest component of municipal solid waste (MSW) (6). Hence, lignin and cellulose account for most of the organic matter in landfills. Sorption of HOCs by biopolymers in municipal solid waste is one of the factors that limit both their biodegradation and mobility in landfills (7). Cellulose is a preferentially degraded fraction of MSW. However, not all cellulose in MSW can be degraded under landfill conditions (8). Biodegradation of cellulose is incomplete in the presence of lignin (6) due to lignin’s protection, thus preserving cellulose in lignocellulosic materials (9). Lignin has much higher sorption affinity for HOCs than cellulose (10, 11). It is also reported that both hydrophobic (i.e., alkyl and aromatic) carbons and carbohydrate components can be important for sorption and conformation of SOM (1). Additionally, the physical structure of biopolymers, such as cellulose acetate butyrate, and lignin esters, can be partially altered when they interact with each other (12). Conformational and structural changes of biopolymers induced by their interactions, and, in turn, their influence on sorption of HOCs, are unknown. Therefore, the main objectives of this work were to (i) determine the composition/ polarity and functionality changes of lignin after cellulose coating and their impacts on HOC sorption, and (ii) elucidate the role of Cu (II) in altering the conformation and structure of lignin and cellulose since a mixture of CuSO4 and ammonium solution was used to dissolve cellulose to coat it on lignin.
Materials and Methods Sorbates and Sorbents. Lignin (organosolv) (LIG) was purchased from Sigma-Aldrich Co., and cellulose (CEL) was from Fisher Scientific Co. Lignocellulosic materials are produced from industrial processes at high quantity. They are composed primarily of LIG and CEL (13). In lignocellulosic materials, LIG is partially covalently linked to CEL that cannot be readily degraded by cellulolytic organisms (6), and they have been used for removal of metallic cations through complexation reactions (13). Although CEL has low sorption for HOCs, it is unclear whether the association of LIG with CEL would affect sorption of HOCs, thus influencing their bioavailability and degradability (6). As an initial step in addressing this void, cellulose was dissolved in a CuSO4amonium solution to make CEL-LIG complex. Such solution may facilitate the hydrolyzation of CEL or LIG. However, to our knowledge, it is the only available procedure to dissolve CEL. To make CEL-LIG complex samples, cellulose was dissolved in solution that contained 20% ammonium and 0.2 M CuSO4 (14) to make its concentrations at 200, 300, and 500 mg/L, respectively. 160 mL of each solution were added to a bottle containing 8 g lignin. The bottles were placed in a shaker to mix for 24 h, after which they were centrifuged VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
185
at 3000 rpm for 30 min and the supernatant was decanted. To have more cellulose coated, the same procedure was repeated four times for a given CEL solution. Finally, the coated samples were rinsed with deionized water five times, freeze-dried, and ground to pass through a 250 µm sieve. Due to low solubility of CEL in the CuSO4-amonium solution, CEL coating was performed four times. Accurate determination of the coated CEL could not be achieved due to some mass loss of LIG during sample shaking and supernatantsolid separation. To examine the influence of CuSO4 and ammonium solution on composition and structure of LIG, this solution-treated LIG was also used for sorption experiments. The solution-treated LIG was labeled as CLIG and the CEL-coated LIG samples were labeled as 2LCEL-LIG, 3LCEL-LIG, and 5LCEL-LIG, respectively. Pyrene, phenanthrene (Phen), naphthalene (Naph), and 1-naphthol (1-Naph) were used as sorbates. 14C labeled and unlabeled compounds were purchased from Sigma-Aldrich Chemical Co. Characterization of Sorbents. A Carlo Erba 1110 CHN Elemental Analyzer was employed to determine C, H, and N contents of biopolymers. Oxygen content was calculated by mass difference. Elemental composition was determined in duplicate and the averaged data were reported. Diffuse reflectance infrared Fourier transform (DRIFT) spectra of the samples were obtained using a Perkin-Elmer spectrometer (Spectrum One) with a diffuse reflectance sampling accessory (Perkin-Elmer, MA) after Kang and Xing (15). Solid-state crosspolarization magic angle-spinning and total-sideband-suppression 13C NMR spectra were obtained using a Bruker DSX300 spectrometer (Karlsruhe, Germany). The NMR running parameters and chemical shift assignments were reported elsewhere (16). Due to the interference of Cu with NMR analysis, reliable NMR spectra of coated-lignin could not be obtained. Surface Acidity. Surface O-containing acidic groups (i.e., carboxyl and phenolic groups) of the biopolymers were determined using the Boehm titration method (17). Briefly, the biopolymers were equilibrated with dilute HCl solution at pH 2 for 4 d, after which the samples were rinsed with deionized water until no Cl- was detected by 0.1 M AgNO3. The acid-equilibrated biopolymers were freeze-dried for 5 d. Approximately 100 mg dried biopolymers were mixed with 15 mL 0.05 M NaHCO3 for 24 h. After mixing, the vials were centrifuged at 3000 rpm for 30 min, and 8 mL supernatant was withdrawn, filtered, and diluted five times. The excess base in the supernatant was determined with 0.002 M HCl. Surface acidity was calculated based upon the assumption that NaHCO3 only neutralizes relatively strong acid groups (e.g., carboxyl and phenolic groups). ζ-Potential Measurements. ζ-potential of LIG and CEL at different pH values was measured with a ZetaSizer Nano Series instrument (Malvern Instrument Ltd.). Sodium chloride (0.01 M) was used as background electrolyte and 0.01 M HCl was used to adjust pH for LIG and CEL (18). A concentration of 70 mg/L was used for LIG and CEL. ζ-potential values at various pH points were measured in duplicate and the averaged data were reported. Differential Scanning Calorimetry (DSC). To determine glass transition temperature (Tg), calorimetric measurements were performed on original and coated biopolymers using a DSC 2910 differential scanning calorimeter (DuPont Instruments). To keep the samples dry, all samples were freeze-dried and stored in a desiccator before experiment. A heating rate of 5 °C/min was used and nitrogen was employed as purge gas. Sorption Experiments. Sorption isotherms were obtained using a batch equilibration technique. Solid-to-solution ratios were adjusted to attain 30-80% solute uptake by sorbents. The background solution was prepared, containing 0.01 M CaCl2 to maintain a constant ionic strength and 200 mg/L 186
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
NaN3 to minimize biological activity. Due to low aqueous solubility of pyrene, Phen, and Naph, their stock solutions were made in methanol at high concentrations. While nonlabeled stock solutions of 1-Naph were directly prepared in aqueous solution. Test solutions of various solute concentrations were prepared by adding 14C labeled and non-labeled stock solution of the compounds to the background solution. They were mixed for 1 h before adding to the sorption vials with predetermined amounts of biopolymers until the minimum headspace was achieved. All vials were immediately sealed with aluminum foil-lined Teflon screw caps and then placed on a rotary shaker to mix for 5 days at room temperature (23 ( 1 °C). Apparent equilibrium was reached within this time period as shown by preliminary tests. To avoid a cosolvent effect, methanol content in the test solution was controlled below 0.1% in volume. After mixing, the vials were centrifuged at 3000 rpm for 30 min and 0.8 mL supernatant was added to ScintiVerse cocktail (8 mL) (Fisher Scientific Co.) for scintillation counting. All isotherms were obtained using duplicate samples for each concentration including blank. Since the mass loss of solute (e.g., sorption on vials and biodegradation) was negligible, uptake of solutes by biopolymers was determined by mass balance. pH-Dependent Sorption. To examine the effect of pH on sorption of polar and nonpolar compounds by biopolymers, uptake of Phen and 1-Naph by LIG and CEL at various pHs was determined. Since LIG and CEL were partially soluble in alkaline condition, sorption of Phen and 1-Naph in alkaline range was not performed. To see whether LIG and CEL dissolve or not in the experimental pH range, 400 mg of LIG or CEL were added to the vials which contained 16 mL solution with pH of 4.1 and 6.9, respectively, and mixed for 5 d. After mixing, the vials were centrifuged and the precipitate was freeze-dried. Elemental composition of the samples was determined using the same instrument as described above. The C, H, and N contents of both CHT and CEL at two pH points were practically same. Hence, dissolution of CHT and CEL was insignificant. A relatively high concentration of Phen (1.044 mg/L; sum of 14C labeled and non-labeled components) and 1-Naph (600 mg/L) was used, and all samples at various pH points along with blanks were run in duplicate. The averaged uptake of 1-Naph and Phen was used for discussion. All other experimental parameters including the solidsolution ratio, mixing time and centrifugation were identical to the ones employed for the sorption experiments. Sorption Isotherm Models. The logarithmic form of the Freundlich equation and linear model were used for data fitting in this work. Linear model:
Q ) K d Ce
(1)
where Kd is linear partitioning coefficient (L/kg). Q and Ce are equilibrium solid- (mg/kg) and liquid- (mg/L) phase concentrations, respectively.
log Q ) log Kf + n log Ce
(2)
Kf is sorption coefficient ((mg/kg)/(mg/L)n), and n is often used as an indicator of isotherm nonlinearity.
Results and Discussion Elemental Composition. Organic carbon (C) content of LIG was much higher than CEL but its O content was much lower than CEL (Table 1). The C content of CEL-coated biopolymer (e.g., 2LCEL-LIG) was lower than CLIG, indicating the occurrence of coating. Increase in N content of 2LCEL-LIG relative to CLIG and LIG was attributed to ammonium in the solution. Organic carbon content of the CEL-coated biopolymers was practically same. In contrast, O content of the coated
TABLE 1. Elemental Composition, Atomic Ratios, Abundance of Acidic Surface Groups, and Integrated Solid-State 13C NMR Data of Original and Coated Ligninsa
sample
C (%)
Elemental Composition and Abundance of Acidic Surface Groups of Biopolymers acidic surface H (%) O (%) N (%) ash (%) H/C O/C (O+N)/C groups (mmol/g)
LIG CEL CLIG 2LCEL-LIG 3LCEL-LIG 5LCEL-LIG
65.7 41.4 61.1 56.1 55.9 55.7
5.58 6.62 5.50 5.51 5.48 5.48
28.6 52.0 23.9 25.4 25.8 26.1
0.17 0 2.08 2.57 2.55 2.58
0 0 7.4 10.4 10.2 10.1
1.02 1.92 1.08 1.18 1.18 1.18
0.326 0.942 0.294 0.339 0.347 0.351
0.328 0.942 0.323 0.378 0.386 0.391
6.22 0 6.17 6.11 6.04 5.32
Integrated Results of Solid-State 13 C NMR Spectra distribution (%) of C chemical shift, ppm sample
0-50
50-61
LIG CEL
13.5 0
14.0 1.6
aliphatic
aromatic
phenolic
polar
aliphatic polar
61-96
96-109
109-145
145-163
163-190
190-220
C (%)
C (%)
C (%)
C (%)
C (%)
9.1 75.3
6.6 17.6
34.0 0
15.3 0
3.3 2.8
4.3 2.7
43.2 94.5
34.0 0
15.3 0
52.6 100.0
29.7 94.5
a Aliphatic C: total aliphatic C region (0-109 ppm); aromatic C (109-145 ppm); phenolic C (145-163 ppm); polar C: total polar carbon region (50-109 ppm and 145-220 ppm); aliphatic polar C: polar carbon in aliphatic region (50-109 ppm). Uncertainties of triplicate measurements for the elemental composition of selected biopolymers were less than 1%.
TABLE 2. Parameters of Linear and Freundlich Models-Based Sorption Isotherm Fitting for Pyrene, Phen, Naph, and 1-Naph on Original and Coated Ligninsa Koc sorbents
LIG
CEL CLIG 2LCEL-LIG 3LCEL-LIG 5LCEL-LIG CLIG 2LCEL-LIG 3LCEL-LIG 5LCEL-LIG
compounds Pyrene Phen Naph 1-Naph Pyrene Phen Naph 1-Naph Phen
1-Naph
Kd
R2
logKf
n
4.848 ( 4.148 ( 0.027 2.794 ( 0.007 2.526 ( 0.022 2.847 ( 0.041 2.239 ( 0.023 0.891 ( 0.010 0.863 ( 0.033 4.506 ( 0.025 4.451 ( 0.022 4.414 ( 0.027 4.396 ( 0.026 4.064 ( 0.022 4.561 ( 0.026 4.646 ( 0.019 4.676 ( 0.014
0.940 ( 0.956 ( 0.015 0.942 ( 0.008 0.868 ( 0.011 0.863 ( 0.028 1.068 ( 0.019 0.992 ( 0.010 0.991 ( 0.017 0.920 ( 0.019 0.853 ( 0.017 0.837 ( 0.021 0.827 ( 0.021 0.545 ( 0.012 0.529 ( 0.014 0.465 ( 0.010 0.435 ( 0.007
0.056c
156 8 7
0.990 0.982 0.994
0.026d
R2
0.05 Sw
0.2 Sw
0.5 Sw
0.988 0.994 0.999 0.997 0.982 0.995 0.998 0.995 0.992 0.993 0.988 0.989 0.992 0.989 0.992 0.995
144800 24300 923 311 3370 377 19 17 65400 76600 73900 73200 3390 11000 10550 10100
133300 23000 853 259 2790 377 19 17 59000 62500 59000 57600 1820 5730 5030 4630
126100 21900 808 229 2460 377 19 17 54400 54600 50800 49200 1190 3720 3080 2760
a K ((mg/kg)/(mg/L)n); K (L/kg); S : aqueous solubility (mg/L); c, d standard errors of logK and n, respectively. K values (L/kg) were used because f oc w f d of linear isotherms.
samples slightly increased with increasing CEL loading level, which can be attributed to the high O content of CEL. Higher H/C ratio of the CEL-coated biopolymers compared to CLIG suggested their reduced aromaticity because CEL does not have aromatic carbons. CEL coating increased the polarity of LIG as shown by the higher O/C and (O + N)/C ratios of the coated lignins as compared to CLIG. (Table 1). DRIFT. The broad peak at 3335 cm-1 was assigned to OH stretching, in the range of intermolecular H-bonding that gives rise to polymeric association. The bands at 2900 and 2879 cm-1 were attributed to symmetric and asymmetric stretching of CH3 and CH2 groups and the peak at 1428 cm-1 was attributed to CH2 deformation. The peak at 1608 cm-1 was assigned to CdC vibrations of aromatic rings, and CdO stretching of quinone or ketones. The band at 834 cm-1 was from aromatic C components. Peaks at 1320, 1218, 1118, and 1055 cm-1 were attributed to the ring and bridge C-O-C vibrations and C-O stretching. Intensity of these peaks for CLIG, 2LCEL-LIG, and 3LCEL-LIG were similar and lower than 5LCEL-LIG, suggesting more O-containing groups (i.e., OH groups in CEL) in the highest CEL coated LIG (Figure S1-A, see the Supporting Information).
13 C NMR. LIG had both long and short aliphatic chains, which appeared at 29 and 14 (also 22.7) ppm, respectively (Figure S1-B). No peaks were observed in the chemical shift range of 0-50 ppm for CEL, illustrating the absence of alkyl C. Strong peaks within 50-109 ppm in CEL spectra indicated the high abundance of O-substituted aliphatic C. LIG had high aromatic (109-145 ppm) (34%) and phenolic (145-163 ppm) (15.3%) carbons, but these carbons were not observed for CEL (Table 1), in agreement with the absence of benzene ring in its structure. Peaks at 163 and 220 ppm in LIG spectra showed the presence of carboxyl and carbonyl groups. However, no clear peaks were observed in this range for CEL, which implied that CEL had negligible amounts of carbonyl and carboxyl groups. Sorption. Lignin had nonlinear isotherms for pyrene, Phen, Naph, and 1-Naph, with n ranging from 0.868 to 0.956 (Table 2) because LIG is a glassy biopolymer (19). Except for pyrene, CEL had linear isotherms for the other three chemicals (Table 2), indicating a partitioning dominant mechanism, consistent with the previous study on sorption of atrazine, prometon, and trichloroethene (20). X-ray observations indicated that CEL can be crystalline and the
VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
187
FIGURE 1. Sorption isotherms of pyrene, Naph, Phen, and 1-Naph on cellulose, original and treated lignins. structure of this fraction cannot be disrupted by water adsorption because of low accessibility (21). However, the immobile amorphous regions in CEL can be plasticized by water molecules (22), thus making the sorption sites more homogeneous. To interpret the difference in sorption behavior of pyrene, Phen, Naph, and 1-Naph by cellulose, we offer a following mechanism. Most of CEL matrix could be softened by water molecules. However, a small fraction adjacent to the crystalline domain may not be swollen by water alone because of the strong attractive force from the crystalline region and low accessibility. Water and solute molecules may have cooperative softening effect on this small fraction of localized immobile amorphous phase in CEL, and the presence of solute molecules may facilitate softening of this fraction depending on molecular size and concentration levels. Molecular volume of pyrene is larger than Phen, Naph, and 1-Naph, but its aqueous solubility is lower (Table S1, see the Supporting Information). Pyrene concentration in the test solution was much lower than the other three compounds. Water and Phen, Naph, or 1-Naph molecules together might be able to swell the localized immobile amorphous domains, leading to linear isotherms. It has been reported that the coated wax can be swollen by small molecules such as Naph but not larger molecules such as pyrene (23). Due to isotherm nonlinearity, Koc values were derived from 0.2 Sw and compared. LIG and CEL had Koc values in the following order: pyrene > Phen > Naph >1-Naph, consistent with their hydrophobicity (Kow). This emphasizes the importance of solute hydrophobicity in HOC sorption. For a given solute, Koc values by LIG was much higher than CEL, which was attributed to its much lower polarity than CEL (Figure 1, Tables 1 and 2). Biopolymers with low-polarity would have high compatibility with HOCs (24). High aromatic and alkyl C contents in LIG compared to zero abundance of these carbons in CEL provided more hydrophobic domains (sites) for HOCs (Figure S1, Table 1). pH Effect. All ζ-potential values of LIG and CEL were negative and they consistently increased with increasing pH (Figure S2). At low pH LIG and CEL were less negatively charged at their surface because excess protons would neutralize a fraction of negative charges. 1-Naph was present mostly in molecular form in the test solution due to the experimental pH range (E7). For LIG, the πeΘ-πeΘ repulsion (14, 25) between LIG and 1-Naph or Phen would be relatively low due to the low negative charge of LIG (Figure S2) and strong H-bonding between LIG and 1-Naph at low pH. 188
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
FIGURE 2. Effect of pH on Phen and 1-Naph sorption by LIG and CEL. Open circles refer to Phen and the closed ones represent 1-Naph. Y-axis is sorbed concentration of 1-Naph and Phen. However, there was no πeΘ-πeΘ repulsion between CEL and 1-Naph due to the absence of benzene ring in CEL. Because polarity of CEL is much higher than LIG (Table 1), H-bonding of 1-Naph with CEL would be more pronounced than with LIG. Therefore, 1-Naph would be sorbed more to LIG at low pH than high pH due to both weak πeΘ-πeΘ repulsion and strong H-bonding. In comparison, high sorption of 1-Naph by CEL at low pH was due to only the strong H-bonding. At high pH range, more acidic functional groups in LIG and CEL would dissociate as shown by their increased negative ζ-potentials (Figure S2). Therefore, LIG and CEL would be more ionized and negatively charged (Figure S2) and the benzene rings in LIG would be richer in electron, while 1-Naph was still un-ionized due to its high pKa value (9.34) (Table S1). Hence, the πeΘ-πeΘ repulsion between LIG and 1-Naph would be stronger and the H-bonding between them would be less pronounced as compared to that at low pH, which reduced sorption of 1-Naph to LIG (Figure 2). Also, H-bonding between CEL and 1-Naph became weaker with increasing pH because more CEL molecules would be present in ionized form (26), leading to the declined sorption. Phen has more π-electrons than 1-Naph, but its π-electron density would be lower than 1-Naph since all electrons are delocalized in the whole molecule (25). In addition, the OH on the ring can contribute to π-electrons of 1-Naph due to the resonance effect though the OH is not a strong electrondonor group (14). As a result, the πeΘ-πeΘ repulsion between
FIGURE 3. Schematic illustration of LIG structure (A), interaction mechanism between LIG and Cu (II) (B), and LIG and CEL in the presence of Cu (II) (C). OMe: methoxyl group. The red lines represent CEL molecules. LIG and 1-Naph would be stronger than between LIG and Phen. Meanwhile, there is no H-bonding between Phen and LIG or CEL at any pH value. Hence, sorption of 1-Naph by LIG and CEL decreased with increasing pH, while sorption of Phen was not quite affected (Figure 2). Another point is that the acidic groups in LIG and CEL would be protonated at low pH, thus showing higher hydrophobicity. While at high pH, LIG and CEL would be more hydrophilic. Such a change in hydrophobicity of biopolymers may partly explain the reduced sorption of 1-Naph with increasing pH, but not for the nearly invariable Phen sorption in the experimental pH range. Hydrophobicity change of biopolymers as induced by pH variation may affect their sorption of polar HOCs more than apolar HOCs, which needs further investigation. Cu2+ and CEL-Coating Effects. Sorption change of HOCs by CEL-coated LIG relative to the original LIG was examined in this work. The Koc of CEL-coated LIG (2LCEL-LIG) for Phen and 1-Naph was 2.7 and 22 times higher than that of the original LIG, respectively (Table 2). Because a mixture of CuSO4 and ammonium solution was used to dissolve CEL to coat on LIG, the influence of this solution on LIG structure and in turn the effect on HOC sorption were determined. The Koc of CLIG for Phen was 2.6 times and for 1-Naph 7 times higher than the original LIG (Table 2). Polarity of CLIG (0.323) was slightly lower than the original LIG (0.328) (Table 1). The negative correlation between Koc values of HOCs and polarity of SOM was observed in previous studies (24). However, such a small decrease in polarity between LIG and CLIG would unlikely increase the Koc values to such a large extent, especially for 1-Naph. We believe that the structure and surface functionality of LIG has changed after treating with the CuSO4 and ammonium solution. It was reported that wheat straw cell wall residue, which contains both LIG and CEL, is highly interactive and readily forms complexes with paramagnetic metal ions such as copper (II). Electron spin resonance (ESR) and X-ray absorption spectroscopy revealed that Cu2+ could form a relatively stable surface complex on acidic sites (e.g., carboxylic and phenolic moieties). ESR parameters also showed that Cu2+ interacts with natural organic matter through inner-spheretype complexation (27). For our case, the complexation of LIG with Cu (II) was shown by the higher ash content of CLIG (7.4%) relative to the original LIG (Table 1). To visually demonstrate the role of Cu (II) in altering the structure of LIG and the interaction mechanisms of LIG with Cu (II) and CEL, a schematic illustration is given in Figure 3. LIG has multiple O-containing groups such as carboxyl, carbonyl,
hydroxyl, C-O, C-O-C, and phenolic groups as shown in 13C NMR and DRIFT spectra (Figure S1, Table 1). Complexation of these functional groups (e.g., phenolic moieties) with Cu2+ would link LIG molecules or its branch chains together to form larger molecules, thus increasing its molecular weight (Figure 3B). A positive correlation between molecular weight of SOM and polymers and their sorption affinity for HOCs and gas has been observed by many investigators (28-30). Gauthier et al. (31) further suggested that molecular weight of SOM can be used as an indicator of its ability to bind HOCs. Complexation of LIG with Cu2+ could change its conformation, driving a fraction of O-containing groups to the interfaces where they interact (Figure 3B). As a result, O-containing groups on the external surfaces of CLIG would be less abundant than the original LIG (Figure 3B), supported by the decreased peaks at 1608, 1320, 1218, 1118, and 1055 cm-1 in DRIFT spectra and the reduced acid groups on the surface of CLIG (Figure S1-A, Table 1). Therefore, overall effective polarity (11) of CLIG would be lower than LIG, and relatively more hydrophobic domains (alkyl carbons and aromatic cores) in CLIG would be exposed for HOC sorption. Higher Koc values of Phen and 1-Naph by CLIG than the original LIG indicate the marked impact of Cu (II) on surface functionalities and spatial arrangement of hydrophobic domains. Linkage of SOM molecules by metal cations such as Al3+ and Ca2+ through complexation was observed, but its effect on HOC sorption was somehow inconsistent for polar and nonpolar compounds. For instance, Al3+ or Ca2+-SOM (i.e., humic acid) complex had much higher Koc for 1-Naph relative to the untreated SOM, while Al3+ complexation slightly reduced the sorption of Naph by SOM (32-33). After coating CEL on LIG, sorption affinity of the coated lignins for 1-Naph substantially increased as compared to CLIG (Figure 1, Table 2). However, no large difference for Phen was observed. Higher Koc of 1-Naph by the coated lignins could be attributed to the conformational and structural changes of LIG and CEL induced by Cu (II). Secondary interactions between LIG and CEL components were observed by Rials et al. (34), and the lignin-carbohydrate complexes were detected using CPMAS NMR as well (35). When LIG particles come in contact with dissolved CEL molecules, they would interact in three ways. (i) A fraction of polar regions such as hydroxyl, carboxyl, and phenolic groups in LIG would interact with the polar groups (i.e., OH) of CEL at the interfaces by polar interactions such as intermolecular H-bonding (Figure 3C); (ii) LIG is able to react with cellulose or hemicellulose via covalent bonds (36); (iii) VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
189
complexation between LIG and CEL through Cu (II) would link LIG and CEL molecules together. In this process, Cu (II) serves as a “linking bridge” and enhances the interactions between LIG and CEL (Figure 3C). CEL itself had low Koc for Phen, however, sorption of Phen by the CEL-coated LIG (2LCEL-LIG) did not decrease but showed a slight increase (6%) compared to CLIG. This indicates that the magnitude of Phen Koc increase caused by spatial rearrangement of hydrophobic domains in CLIG due to Cu2+ and CEL linkages was greater than sorption reduction resulting from CEL coating, due to its high polarity. Association of LIG with CEL can prevent the internal bonding in CEL molecules, thus decreasing the extent of cross-linking between adjacent mobile chains of CEL (37). Then, “loose” conformation of CEL would form on the LIG surface, which can facilitate the access of 1-Naph to OH groups to have specific interactions with CEL. However, the conformation changes of CEL would have lower effect on sorption affinity for Phen because of its lack of H-bonding interaction. Therefore, the increase in Koc of Phen by 2LCEL-LIG relative to CLIG was much lower than 1-Naph. Both complexation of LIG with Cu (II) and further CELcoating enhanced the sorption isotherm nonlinearity of Phen and 1-Naph (Table 2), showing that sorption became more of adsorption and less partitioning process. LIG had a Tg value of 68.9 °C, comparable with the literature data (70.4 °C) (19). Complexation of LIG with Cu2+ increased its Tg to 74.0 °C (Figure S3), suggesting that a fraction of mobile chains in LIG would be “fixed” after complexation with Cu2+, thus showing decreased mobility. As reported in the stereochemistry of biopolymers, majority of hydroxyl groups in LIG and CEL would form intrachain or interchain H-bonding within continuous molecules, creating more cross-links and condensed domains (38). The condensed domains produced at the interfaces between LIG and CEL were demonstrated by the increased Tg of 2LCEL-LIG (77.9 °C) relative to CLIG (74.0 °C) (Figure S3). These newly formed condensed domains would be responsible for the enhanced sorption nonlinearity of 2LCEL-LIG (Table 2). 3LCEL-LIG exhibited comparable Tg value (78.4 °C) with 2LCEL-LIG, however, 5LCEL-LIG had higher Tg (82.4 °C) than 3LCEL-LIG, but lower than CEL. This revealed that more condensed domains were produced with increasing CEL coating. Hence, the isotherm nonlinearity became more pronounced (Table 2). Concluding Remarks and Environmental Significance. The results of this study reveal that sorption of HOCs by biopolymers is regulated by polarity/composition, conformation, structures, and the polar groups on the surface of biopolymers. Solution pH can change surface charge property, significantly affecting sorption of polar HOCs such as 1-Naph. Sorption of apolar and polar organic compounds by lignin would be greatly enhanced once it complexes with Cu2+. CEL has very low sorption affinity for HOCs, once it is coated on LIG, sorption capacity of coated-LIG for 1-Naph was highly increased while Phen showed a slight increase relative to the LIG-Cu2+ complex. However, the isotherm nonlinearity for both apolar (e.g., Phen) and polar (e.g., 1-Naph) compounds became more pronounced due to formation of new condensed domains. This suggests that association of biopolymers could change their conformation and sorption behavior of HOCs, and CEL-coated LIG would be able to reduce the mobility and transport of HOCs more significantly than the original LIG, thus increasing the persistence of HOCs in the environment.
Acknowledgments This project was supported in part by the CSREES, USDA National Research Initiative Competitive Grants Program (2005-35107-15278), the Massachusetts Agricultural Experiment Station (MAS90), and the NSF of China (40428005). 190
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
Supporting Information Available Selected physicochemical properties of the chemicals in Table S1; DRIFT spectra (A) of LIG, CEL and CEL coated-lignins and 13C NMR spectra (B) of LIG and CEL in Figure S1; Zeta potential versus pH of LIG and CEL in Figure S2; and calorimetric analysis of the glassy transitions of original and coated biopolymers in Figure S3. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Gunasekara, A. S.; Simpson, M. J.; Xing, B. S. Identification and characterization of sorption domains in soil organic matter using structurally modified humic acids. Environ. Sci. Technol. 2003, 37, 852-858. (2) LeBoeuf, E. J.; Weber, W. J. Jr. A distributed reactivity model for sorption by soils and sediments. 8. Sorbent organic domain: discovery of a humic acid glass transition and an argument for a polymer-based model. Environ. Sci. Technol. 1997, 31, 16971702. (3) Ran, Y.; Xing, B. S.; Rao, P. S. C.; Fu, J. M. Importance of adsorption (hole-filling) mechanism for hydrophobic organic contaminants on an aquifer kerogen isolate. Environ. Sci. Technol. 2004, 38, 4340-4348. (4) Kurita, K. Controlled functionalization of the polysaccharide chitin. Prog. Polym. Sci. 2001, 26, 1921-1971. (5) Crist, R. H.; Martin, J. R.; Crist, D. R. Heavy metal uptake by lignin: comparison of biotic ligand models with an ion-exchange process. Environ. Sci. Technol. 2002, 36, 1485-1490. (6) Chen, Y.; Knappe, D. R. U.; Barlaz, M. A. Effect of cellulose/ hemicellulose and lignin on the bioavailability of toluene sorbed to waste paper. Environ. Sci. Technol. 2004, 38, 3731-3736. (7) O ¨ man, C. Comparison between the predicted fate of organic compounds in landfills and the actual emissions. Environ. Sci. Technol. 2001, 35, 232-239. (8) Suflita, J. M.; Gerba, C. P.; Ham, R. K.; Palrnisano, A. C.; Rathje, W. J.; Robinson, J. A. The world’s largest landfill. Environ. Sci. Technol. 1992, 26, 1486-1495. (9) Stinson, J. A.; Ham, R. K. Effect of lignin on the anaerobic decomposition of cellulose as determined through the use of a biochemical methane potential method. Environ. Sci. Technol. 1995, 29, 2305-2310. (10) Garbarini, D. R.; Lion, L. W. Influence of the nature of soil organics on the sorption of toluene and trichloroethylene. Environ. Sci. Technol. 1986, 20, 1263-1269. (11) Xing, B. S.; McGill, W. B.; Dudas, M. J. Cross-correlation of polarity curves to predict partition coefficients of nonionic organic contaminants. Environ. Sci. Technol. 1994, 28, 19291934. (12) Ghosh, I.; Jain, R.; Glasser, W. G. Multiphase materials with lignin. XV. Blends of cellulose acetate butyrate with lignin esters. J. Appl. Polym. Sci. 1999, 74, 448-457. (13) Dupont, L.; Guillon, A. E. Removal of hexavalent chromium with a lignocellulosic substrate extracted from wheat bran. Environ. Sci. Technol. 2003, 37, 4235-4241. (14) Xing, Q. Y.; Xu, R. Q.; Zhou, Z.; Pei, W. W. Fundamentals of organic chemistry. Higher education press. Beijing, China, 1998. (15) Kang, S.; Xing, B. S. Phenanthrene sorption to sequentially extracted soil humic acids and humins. Environ. Sci. Technol. 2005, 39, 134-140. (16) Wang, X. L.; Sato, T.; Xing, B. S. Competitive sorption of pyrene on wood chars. Environ. Sci. Technol. 2006, 40, 3267-3272. (17) Chun, Y.; Sheng, G. Y.; Chiou, C. T. Evaluation of current techniques for isolation of chars as natural adsorbents. Environ. Sci. Technol. 2004, 38, 4227-4232. (18) Taboada-Serrano, P.; Vithayaveroj, V.; Yiacoumi, S.; Tsouris, C. Surface charge heterogeneities measured by atomic force microscopy. Environ. Sci. Technol. 2005, 39, 6352-6360. (19) Leboeuf, E. J.; Weber, W. J., Jr. Macromolecular characteristics of natural organic matter. 1. Insights from glass transition and enthalpic relaxation behavior. Environ. Sci. Technol. 2000, 34, 3623-3631. (20) Xing, B. S.; Pignatello, J. J.; Gigliotti, B. Competitive sorption between atrazine and other organic compounds in soils and model sorbents. Environ. Sci. Technol. 1996, 30, 2432-2440. (21) Mackay, A. A.; Gschwend, P. M. Sorption of monoaromatic hydrocarbons to wood. Environ. Sci. Technol. 2000, 34, 839845. (22) Wu, B. Y.; Taylor, C. M.; Knappe, D. R. U.; Nanny, M. A.; Barlaz, M. A. Factors controlling alkylbenzene sorption to municipal solid waste. Environ. Sci. Technol. 2001, 35, 4569-4576.
(23) Chen, B. L.; Xing, B. S. Sorption and conformational characteristics of reconstituted plant cuticular waxes on montmorillonite. Environ. Sci. Technol. 2005, 39, 8315-8323. (24) Wang, X. L.; Sato, T.; Xing, B. S. Sorption and displacement of pyrene in soils and sediments. Environ. Sci. Technol. 2005, 39, 8712-8718. (25) Sykes, P. A guidebook to mechanism in organic chemistry (sixth edition). Jupiter offset press, Shahdara, Delhi, India, 1989. (26) Severtson, S. J.; Banerjee, S. Sorption of chlorophenols to wood pulp. Environ. Sci. Technol. 1996, 30, 1961-1969. (27) Merdy, P.; Guillon, E.; Dumonceau, J.; Aplincourt, M. Spectroscopic study of copper(II)-wheat straw cell wall residue surface complexes. Environ. Sci. Technol. 2002, 36, 1728-1733. (28) Chin, Y. P.; Weber, W. J., Jr. Estimating the effects of dispersed organic polymers on the sorption of contaminants by natural solids. I. A predictive thermodynamic humic substance organic solute interaction model. Environ. Sci. Technol. 1989, 23, 978984. (29) Kopinke, F. D.; Georgi, A.; Mackenzie, K. Sorption of pyrene to dissolved humic substances and related model polymers. 1. Structure-property correlation. Environ. Sci. Technol. 2001, 35, 2536-2542. (30) Toit, K.; Paul, D. R. Effect of polystyrene molecular weight on the carbon dioxide sorption isotherm. Macromolecules 1982, 15, 1104-1107. (31) 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. (32) Yuan, G. S.; Xing, B. S. Effects of metal cations on sorption and desorption of organic compounds in humic acids. Soil. Sci. 2001,
166, 107-115. (33) Lu, Y. F.; Pignatello, J. J. Sorption of apolar aromatic compounds to soil humic acid particles affected by aluminum(III) ion crosslinking. J. Environ. Qual. 2004, 33, 1314-1321. (34) Rials, T. G.; Glasser, W. G. Multiphase materials with lignin. 4. Blends of hydroxypropyl cellulose with lignin. J. Appl. Polym. Sci. 1989, 37, 2399-2415. (35) Kolodziejski, W.; Frye, J. S.; Maciel, G. E. Carbon-13 nuclear magnetic resonance spectrometry with cross polarization and magic-angle spinning for analysis of lodgepole pine wood. Anal. Chem. 1982, 54, 1419-1424. (36) Tong, X. G; Smith, L. H.; McCarty, P. L. Methane fermentation of selected lignocellulosic materials. Biomass 1990, 21, 239255. (37) Oksanen, T.; Buchert, J.; Viikari, L. The role of hemicelluloses in the hornification of bleached kraft pulps. Holzforschung 1997, 51, 355-360. (38) Zakaria, S.; Hamah, H.; Murshidi, J. A.; Deraman, M. Chemical modification on lignocellulosic polymeric oil palm empty fruit bunch for advanced material. Adv. Polym. Tech. 2001, 20, 289295.
Received for review June 10, 2006. Revised manuscript received September 20, 2006. Accepted September 22, 2006. ES061389E
VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
191
