Humic Acid Fractionation upon Sequential Adsorption onto Goethite

Feb 9, 2008 - For instance, the fractionation of aquatic natural organic matter (NOM) occurs upon sorption onto iron oxide mineral surfaces, and this ...
20 downloads 11 Views 173KB Size
Download PDF
Langmuir 2008, 24, 2525-2531

2525

Humic Acid Fractionation upon Sequential Adsorption onto Goethite Seunghun Kang and Baoshan Xing* Department of Plant, Soil, and Insect Sciences, UniVersity of Massachusetts, Amherst, Massachusetts 01003 ReceiVed September 19, 2007. In Final Form: December 3, 2007 Mineral-humic complexes are commonly distributed in natural environments and are important in regulating the transport and retention of hydrophobic organic contaminants in soils and sediments. This study investigated the structural and conformational changes of humic acid (HA) and mineral-HA complexes after sequential HA adsorption by goethite, using UV-visible spectroscopy, high performance size exclusion chromatography (HPSEC), Fourier transform infrared (FT-IR) spectroscopy, and solid-state 13C nuclear magnetic resonance (NMR) spectroscopy. The HA remaining in the solution after adsorption showed low polarity index values ((N + O)/C), which indicates that polar functional moieties are likely to adsorb on the goethite surface. In addition, we observed decreased E4/E6 and E2/E3 ratios of unbound HA with increasing number of coatings, implying that aliphatic rich HA fractions with polar functional moieties readily adsorb to the goethite surface. According to IR spectra, carbohydrate carbon would be the important fractions associated with goethite. NMR spectra provided evidence for HA fractionation during adsorption onto the mineral surface; that is, aliphatic fractions were preferentially adsorbed by goethite while aromatic fractions were left in solution. Relatively small molecular weight (MW) HA fractions had a greater affinity for the goethite surface based on analyses of the HPSEC chromatograms, which differs from the results reported in the literature. Finally, our results suggest that the polar aliphatic fractions of HA were mainly adsorbed to goethite via electrostatic attraction and/or ligand exchange reactions.

1. Introduction Interactions between mineral surfaces and soil organic matter (SOM) are of significant importance in a variety of natural environments.1,2 Minerals comprising the major components of soils, suspended solids, and sediments commonly possess at least a partial surface coating with SOM.3 The sorption of SOM can alter many physicochemical properties of minerals,4,5 including the rate and extent of dissolution6,7 and the physical stability of colloidal-sized particles.8,9 Consequently, SOM sorption on minerals plays important roles in the detoxification of hazardous compounds and in the transport and binding of organic and inorganic contaminants.5 Moreover, sorption by clay minerals can protect SOM against biodegradation, hence affecting biogeochemical carbon cycling and carbon sequestration processes.10,11 SOM consists of heterogeneous components with a wide range of molecular weight (MW) and diverse functionality ranging from nonpolar polymethylene chains to highly polar carboxylic groups. Due to the heterogeneities of SOM, various modes of SOM sorption are observed with different minerals. For instance, * To whom correspondence should be addressed. Telephone: 1-413545-5212. Fax: 1-413-545-3958. E-mail: [email protected]. (1) Au, K. K.; Penisson, A. C.; Yang, S. L.; O’Melia, C. R. Geochim. Cosmochim. Acta 1999, 63, 2903-2917. (2) Kleber, M.; Mikutta, R.; Torn, M. S.; Jahn, R. Eur. J. Soil Sci. 2005, 56, 717-725. (3) Mayer, L. M.; Xing, B. Soil Sci. Soc. Am. J. 2001, 65, 250-258. (4) Angove, M. J.; Fernandes, M. B.; Ikksan, J. J. Colloid Interface Sci. 2002, 247, 282-289. (5) Wang, K.; Xing, B. J. EnViron. Qual. 2005, 39, 342-349. (6) Johnson, S. B.; Yoon, T. H.; Kocar, B. D.; Brown, G. E. Langmuir 2004, 20, 4996-5006. (7) Yoon, T. H.; Johnson, S. B.; Musgrave, C. B.; Brown, G. E. Geochim. Cosmochim. Acta 2004, 68, 4505-4518. (8) Huygens, D.; Boeckx, P.; Van Cleemput, O.; Oyarzun, C.; Godoy, R. Biogeosciences 2005, 2, 159-174. (9) Jobbagy, E. G.; Jackson, R. B. Ecol. Appl. 2000, 10, 423-436. (10) Mikutta, R.; Kleber, M.; Tom, M. S.; Jahn, R. Biogeochemistry 2006, 77, 25-56. (11) Wagai, R.; Mayer, L. M. Geochim. Cosmochim. Acta 2007, 71, 25-35.

the fractionation of aquatic natural organic matter (NOM) occurs upon sorption onto iron oxide mineral surfaces, and this likely contributes to the large adsorption-desorption hysteresis.12-14 In addition, mineral surfaces with different surface properties have particular affinities for specific functional moieties of SOM. For example, preferential sorption of high MW hydrophobic fractions was observed on kaolinite and montmorillonite surfaces.5 Sorption of NOM on iron and aluminum oxide minerals mostly occurs via the carboxylic functionality of NOM.1,11 Therefore, the sorption mechanisms of SOM by minerals are very diverse, including electrostatic interaction, ligand exchange-surface complexation, and entropy-driven hydrophobic interaction. With respect to the fractionation of SOM upon adsorption to minerals, several groups have suggested that hydrophobic and/ or higher MW SOM fractions and more aromatic components of SOM preferentially adsorb to iron oxide surfaces, leaving behind lower MW, more aliphatic components in solution.15-23 However, others reported that goethite has higher sorption affinity for carboxylic functional moieties over hydrophobic and/or higher MW SOM such as carbohydrate and aliphatic fractions.17,24 SOM, (12) McKnight, D. M.; Hornberger, G. M.; Bencala, K. E.; Boyer, E. W. Water Resour. Res. 2002, 38, Art. No. 1005. (13) Ochs, M.; Cosovic, B.; Stumm, W. Geochim. Cosmochim. Acta 1994, 58, 639-650. (14) van de Weerd, H.; van Riemsdijk, W. H.; Leijnse, A. EnVion. Sci. Technol. 1999, 33, 1675-1681. (15) Davis, J. A. Geochim. Cosmochim. Acta 1982, 46, 2381-2393. (16) Tipping, E. Chem. Geol. 1981, 33, 81-89. (17) McKnight, D. M.; Bencala, K. E.; Zellweger, G. W.; Aiken, G. R.; Feder, G. L.; Thorn, K. A. EnVion. Sci. Technol. 1992, 26, 1388-1396. (18) Gu, B. H.; Schmitt, J.; Chen, Z.; Liang, L. Y.; McCarthy, J. F. Geochim. Cosmochim. Acta 1995, 59, 219-229. (19) Kaiser, K.; Zech, W. Soil Sci. Soc. Am. J. 1997, 61, 64-69. (20) Meier, M.; Namjesnik-Dejanovic, K.; Maurice, P. A.; Chin, Y. P.; Aiken, G. R. Chem. Geol. 1999, 157, 275-284. (21) Namjesnik-Dejanovic, K.; Maurice, P. A.; Aiken, G. R.; Cabaniss, S.; Chin, Y. P.; Pullin, M. J. Soil Sci. 2000, 165, 545-559. (22) Zhou, Q. H.; Cabaniss, S. E.; Maurice, P. A. Water Res. 2000, 34, 35053514. (23) Hur, J.; Schlautman, M. A. J. Colloid Interface Sci. 2003, 264, 313-321. (24) Chorover, J.; Amistadi, M. K. Geochim. Cosmochim. Acta 2001, 65, 95-109.

10.1021/la702914q CCC: $40.75 © 2008 American Chemical Society Published on Web 02/09/2008

2526 Langmuir, Vol. 24, No. 6, 2008

obtained from oxidative decomposition of lignocellulose upon sorption to aluminum and iron oxyhydroxides, formed strong complexes between surface metals and acidic organic ligands.25 The sorption of natural organic material on goethite surfaces was attributed to their polar carboxylic and phenolic functionality.18 For instance, goethite has higher sorption affinity for high polar carboxylic functional moieties over low polarity carbohydrate and nonpolar aliphatic fractions.24 In addition, the aromatic components of SOM preferentially sorbed to the mineral surfaces have a high content of carboxylic moieties.26 Furthermore, sorption would increase with an increase in the number of carboxyl groups per molecule as well as with the acidity of the carboxyl group.27-29 Although the adsorption studies of well-defined polymers or small organic acids are useful to explain SOM adsorptive fractionation, SOM sorption is undoubtedly more complicated than in these model systems because of their inherent complexities. Until now, mechanisms remain unclear for SOM adsorption onto goethite, one of the common oxide minerals in soils and sediments. Therefore, a molecular-level understanding of SOM sorption by goethite requires additional studies with complementary analyses. The objective of this study was to investigate the fractionation behavior of humic acid (HA) on goethite upon sequential sorption and to examine HA sorption mechanisms on goethite surface by employing UV-visible spectroscopy, high performance size exclusion chromatography (HPSEC), Fourier transform infrared (FT-IR) spectroscopy, and solid-state 13C nuclear magnetic resonance (NMR) spectroscopy. 2. Experimental Section 2.1. Mineral and HA. Goethite was purchased from Fluka and used as received because of its high purity (g98%).30 The N2 Brunauer-Emmett-Teller (BET) method with a Beckman Coulter (Foullerton, CA) SA3100 surface analyzer was used to measure the specific surface area of goethite at 77 K. Soil HA, extracted from the soil of Mount Toby located in the Connecticut River valley in central western Massachusetts, was used for preparation of humicoxide complexes. The HA preparation and chemical characteristics were described in a previous research.31 2.2. Sorption Isotherms. All sorption isotherms were obtained using a batch equilibration technique. The background solution was 0.01 M NaCl in deionized water with 200 mg/L NaN3 to reduce microbial activity. HA concentrations ranged from 10 to 100 mg/L. The sorption experiments were conducted with eight concentration points; each point including the blank solution was run in triplicate and carried out in borosilicate glass vials at room temperature. To equilibrate goethite suspensions, 15 mL of background solution was added into a vial containing 75 mg of goethite and then the vial was rotated in an end-over-end shaker (150 rpm) for 24 h without any pH adjustment. After 24 h of equilibration, the suspension was centrifuged for 20 min at 7600g and the supernatant was discarded. A volume of 15 mL of HA solution was then added to the centrifuged goethite paste. The vials were sealed with aluminum foil-lined Teflon screw caps and placed on the shaker (150 rpm) for 72 h at room temperature. The suspension pHs were adjusted with diluted NaOH or HCl using an Accumet model 510 combination pH electrode and readjusted every 24 h. A preliminary study showed that continued mixing beyond 48 h did not significantly affect the sorption. After (25) Kaiser, K.; Guggenberger, G. Org. Geochem. 2000, 31, 711-725. (26) Kaiser, K. Org. Geochem. 2003, 34, 1569-1579. (27) Davis, J. A.; Gloor, R. EnViron. Sci. Technol. 1981, 15, 1223-1229. (28) Edwards, M.; Benjamin, M. M.; Ryan, J. N. Colloids Surf., A 1996, 107, 297-307. (29) Wang, L. L.; Chin, Y. P.; Traina, S. J. Geochim. Cosmochim. Acta 1997, 61, 5313-5324. (30) Perez-Ramırez, J.; Mul, G.; Kapteijn, F.; Moulijn, J. A.; Overweg, A. R.; Domenech, A.; Ribera, A.; Arends, I. W. C. E. J. Catal. 2002, 207, 113-126. (31) Kang, S.; Xing, B. EnVion. Sci. Technol. 2005, 39, 134-140.

Kang and Xing equilibration, the vials were removed from the shaker and then centrifuged at 7600g for 20 min. For additional separation of fine particles from the supernatant, the supernatant was withdrawn with a plastic syringe and filtered with a 0.2 µm polycarbonate membrane filter. The filtered supernatant was diluted with the background solution, and then the concentration was measured with total dissolved organic carbon analysis (Shimadzu TOC-V/TN). A calibration curve between the HA concentrations and the carbon content of the adsorbate was obtained using a wide range of adsorbate concentrations. The amount of sorbed HA by goethite was calculated by mass difference because of little sorption by the vials and no biodegradation. The adsorption data were fit with both Langmuir and Freundlich isotherm models using Sigma Plot at p < 0.05. Langmuir equation is represented as x/m ) Q0bCe/(1 + bCe), where x/m (mg C/g) is the amount of HA sorbed per mass of goethite, Ce (mg C/L) is the equilibrium aqueous-phase solute concentration, Q0 (mg C/g) is the maximum sorption capacity, and b is a constant as a solute-surface interaction energy-related parameter. Freundlich model is represented as x/m ) KfCne , where x/m (mg C/kg) and Ce (mg C/L) are the same as described above, Kf (mg C(1-n)Ln kg-1) is indicative of the sorbatesorbent binding capacity, and n is the isotherm linearity parameter, an indicator of site energy heterogeneity. The fit of both Freundlich and Langmuir models was tested with an F-test to determine which model best described the data.32 For all isotherms, the F-test indicated that the data were best fitted by the Langmuir model (data not shown). In addition, the F-test was used for determining the statistical significance of the differences for adsorption parameters (Q0 and b) among the treatments. Statistical significances were evaluated at P < 0.01. 2.3. Goethite-Humic Complex Preparation. HA sorption to goethite was conducted with short-term batch experiments. A known amount of HA was dissolved in a minimum volume of 0.5 M NaOH to prepare a 5 g/L solution with 200 parts per million (ppm) sodium azide (NaN3) in 0.01 M NaCl as a background solution to avoid any microbial degradation. The HA solution was adjusted to pH 5 with 0.1 M NaOH and HCl, and then the HA solution was filtered with a membrane filter (0.45 µm pore size). The humic-goethite complex was prepared at a HA to goethite ratio of 1:5 (w/w), and the HAgoethite suspension was gently shaken on a shaker. The suspension pHs were adjusted with diluted NaOH or HCl using an Accumet model 510 combination pH electrode and readjusted every 24 h. After shaking for 72 h, the suspension was centrifuged at 7600g for 15 min. The precipitate was washed with the background solution several times to remove the unbound HA fraction until there was no color in the washing water. The washed precipitate was freezedried, ground, and then stored for subsequent uses. The supernatant HA was saved for further experiments. This unbound HA solution was acidified with 6 M HCl followed by washing with a 0.1 M HF/0.3 M HCl solution three times to remove any associated mineral fraction and then centrifuged at 7600g for 20 min to obtain precipitated HA. The precipitated HA was freeze-dried, ground, and stored for further analysis. For the sequential adsorption of HA to goethite, fresh goethite was added to the HA supernatant for further complexation. This process was repeated four times (see the Supporting Information, Figure S1). A standard goethite without HA was prepared by the same procedures as those for the goethite with HA coating. For the preparation of a standard HA, the source HA (SHA) was subjected to the same procedures as those used for precipitated unbound HA samples. 2.4. Characterization of HA. 2.4.1. Elemental Analysis and UVVis Spectroscopy. The C, H, and N contents of the HA were determined using a Carlo Erba 1110 CHN elemental analyzer. Oxygen content was calculated by the mass difference. Ash content was measured by heating the samples at 800 °C for 4 h. Using an Agilent 8453 UV-visible spectrometer, the ratios of absorbance for the supernatant (unbound) HA from each adsorption step at 465 and 665 nm (E4/E6 ratio) and 254 and 365 nm (E2/E3 ratio) were determined in the sequential coating process to explore the fractionation of HA. (32) Salloum, M. J.; Dudas, M. J.; McGill, W. B.; Murphy, S. M. EnVion. Toxicol. Chem. 2000, 19, 2436-2442.

HA Fractionation upon Adsorption onto Goethite

Langmuir, Vol. 24, No. 6, 2008 2527

The HA solution (or supernatant) after each coating step was adjusted to pH 5.0 with diluted HCl and NaOH before measurement. 2.4.2. IR Spectroscopy. The diffuse reflectance Fourier transform (DRIFT) spectra were recorded with a Perkin-Elmer Spectrum One FT-IR spectrometer equipped with a lithium tantalate (LiTaO3) detector and a DRIFT accessory (Shelton, CT). The freeze-dried complex or unbound HA (5 mg) was mixed gently with 95 mg of KBr using a pestle and mortar and analyzed with DRIFT. The KBr powder was used as a background. DRIFT spectra were recorded from 450 to 4000 cm-1 at 4 cm-1 resolution over 200 coaveraged scans. All absorption spectra were converted to the Kubelka-Munk function using Spectrum software (Perkin-Elmer). The DRIFT HA spectrum of the HA-goethite complex was obtained by subtracting the spectrum of the standard goethite prepared without any adsorbed HA. 2.4.3. 13C Nuclear Magnetic Resonance Spectroscopy. Solid-state NMR was employed to characterize the structural composition of the source HA and the unbound HA. The NMR spectrum was acquired at a frequency of 75.48 MHz for carbon on a Bruker (Rheinstetten, Germany) Advance DSX-300 spectrometer. The 13C spectrum of HA was obtained using a total sideband suppression pulse program (cross-polarization magic-angle spinning with total sideband suppression) in a 7 mm zirconia rotor with a Kel-F cap. The instrument was run under the following conditions: contact time, 1 ms; spinning speed, 5 kHz; 90° 1H pulse, 5 µs; acquisition delay, 4 s; and the number of scans, from 5000 to 10000.

13C

2.4.4. Molecular Weight of Adsorbed HA by HPSEC. The characterization of HA based on MW was investigated with high performance size exclusion chromatography (HPSEC). A detailed description of the HPSEC method is provided elsewhere.22,33 Briefly, the mobile phase consisted of degassed 0.1 M NaCl buffered with 2 mM phosphate at pH 7.2 with a flow rate of 0.5 mL/min. A TSKGEL column (Tosoh Bioscience, G3000PWXL, 7.8 mm × 30 cm, and particle size 6 µm) and TSK-GEL guard column (Tosoh Bioscience, 6.0 mm × 4 cm, and particle size 12 µm) were used with a Perkin-Elmer 200 LC diode array detector (254 nm) and a fluorescence detector (467/548 nm, Ex/Em) subsequently. All HA samples were filtered through 0.20 µm polycarbonate membrane filters prior to sample injection. In preliminary tests for HA interaction with the column matrix, the recovery rate within analysis time (40 min) was 98-103% of the initial injection amount of HA, analyzed with an organic carbon analyzer and UV-vis spectroscopy. To calculate the MW of HA, sodium polystyrene sulfonate (PSS) of nominal MW values 35, 18, 4.6, and 1.8 kDa (Polysciences, Inc.) was employed as standard molecules.22,33,34 As reported by the supplier, the actual MW values for these PSS samples were 32.9, 14.9, 5.18, and 1.43 kDa. Acetone (58 Da, HPLC grade) was used as a low MW standard. The concentration of the standards was 200 mg/L. A semilog linear calibration curve (see the Supporting Information, Figure S2) was used to calculate the MW of the HA samples. Mn (number average molecular weight) and Mw (weight average molecular weight) values were determined using the following equations.22 N

∑fM i

Mn )

∑f

i

Mw )

2 i

i)1

(2)

N

∑fM i

i

i)1

where fi is the frequency of a characteristic molecular weight, Mi is the weight of a characteristic molecular weight of i fraction, and N is the number of molecular fractions according to the molecular weight. The polydispersity (F) of HAs is calculated by Mw/Mn, which would be 1 for a pure substance, whereas, for a mixture of molecules, Mn < Mw and F > 1. The MW distribution of adsorbed HA on goethite was calculated from the HPSEC chromatogram difference between the previous unbound HA and immediately next unbound HA.34

3. Results and Discussion 3.1. HA Sorption to Goethite. The sorption isotherm of HA by goethite showed a steep initial slope at low concentrations and then reached a plateau as equilibrium carbon concentration increased, which implies a high affinity of binding sites for HA (Figure 1). The isotherm conformed to the Langmuir adsorption equation. The maximum sorption observed in this study was approximately 3.75-5.61 mg C/g at different pHs. The adsorption of HA on goethite was strongly pH-dependent (Figure 1): the sorption increased with decreasing pH, which has been commonly observed for SOM on goethite.4,27,35 The increase in adsorption at low pH and the decrease in adsorption at high pH can be explained by considering the electrostatic interactions between the acidic functional moieties of HA and the goethite surface. Although the electrostatic contribution to the free energy of adsorption is relatively small,35 electrostatic interactions can be regarded as one of the surface adsorption modes. Surface complexation between the acidic functional groups of HA and the surface hydroxyl groups on minerals plays an important role in HA sorption.36-38 When the pH increases, the surfaces become negatively charged, which causes the oxygen atoms on the mineral surface to be tightly bound and thus less likely to react with acidic functional groups in solution.38 As the pH decreases, neutral and positively charged surface sites are formed, and the metaloxygen bond is weakened due to the decreased electron density of the bond. Therefore, complexation by ligand exchange at acidic pH may be one of the binding mechanisms for HA by goethite in this study. Hydrophobic interaction between the goethite surface and protonated HA is another possible sorption mechanism; however, this entropy-driven process might not be of major importance among the adsorption processes of SOM by minerals.28 Unlike the case of SOM, hydrophobic effects have been found to be most important for neutral organic compounds such as polycyclic aromatic hydrocarbons, which are highly hydrophobic and thus extremely surface reactive.40 3.2. Structural Characterization and Fractionation of HA Adsorbed to Goethite. In our study, the E4/E6 ratio of the unbound HAs after sequential coating on goethite (GHAs) significantly decreased with increasing adsorption (Table 1). The E4/E6 ratio

i

i)1

(1)

N

N

∑fM

i

i)1

(33) Chin, Y. P.; Aiken, G.; O’Loughlin, E. EnViron. Sci. Technol. 1994, 28, 1853-1858. (34) Zhou, Q. H.; Maurice, P. A.; Cabaniss, S. E. Geochim. Cosmochim. Acta 2001, 65, 803-812.

(35) Evanko, C. R.; Dzombak, D. A. EnViron. Sci. Technol. 1998, 32, 28462855. (36) Nordin, J.; Persson, P.; Laiti, E.; Sjoberg, S. Langmuir 1997, 13, 40854093. (37) Specht, C. H.; Frimmel, F. H. Phys. Chem. Chem. Phys. 2001, 3, 54445449. (38) Goldberg, S.; Forster, H. S.; Godfrey, C. L. Soil Sci. Soc. Am. J. 1996, 60, 425-432. (39) Arnarson, T. S.; Keil, R. G. Mar. Chem. 2000, 71, 309-320. (40) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. EnVironmental Organic Chemistry; Wiley: New York, 2003.

2528 Langmuir, Vol. 24, No. 6, 2008

Figure 1. Adsorption isotherm of HA on goethite. Maximum sorption capacity (Q0) and parameter b were calculated from the Langmuir sorption model. Q0 and b values among the treatments are significantly different at the p ) 0.01 level using the F-test.

of the SHA was 8.74 and decreased from 7.91 to 6.38 from the first to fourth sorption, which indicates that fractionation of the HA occurred. Generally, the E4/E6 ratio is highly related to the humification degree (decomposition of organic matter) and MW of SOM.41 Well-decomposed SOM has a relatively low E4/E6 ratio, low polarity, and high MW. Therefore, the SOM fractions remaining in the solution have low functional groups and high MW, meanwhile polar moieties with relatively low MW seemed more likely to adsorb on goethite surfaces. This finding is contrast to previous observations that high MW organic matter adsorbed preferentially to iron(III) oxyhydroxide surfaces.23,34 However, it is in good agreement with reports that high contents of acidic functional groups of organic matter favored intense interactions with goethite surfaces.42 The E2/E3 ratio is related to SOM aromaticity.43 The E2/E3 ratios of GHAs were lower than that of the SHA (Table 1), showing that the adsorbed HA fractions had a lower aromaticity than the SHA. The polarity index ((N + O)/C)) of unbound HA decreased after adsorption by goethite (Table 1). The reduction of the polarity index of the GHAs suggests preferential sorption of polar fractions on goethite surfaces with sequential adsorptions, which is in agreement with another report.26 A key factor in the interaction of strongly sorbing organic matter with oxide minerals is the acidity of the adsorbing compounds.26,44 Sorption increases with increasing number of carboxyl groups per molecule and acidity of carboxyl groups.15,29 Furthermore, the binding of carboxyl groups in organic matter to oxide mineral surfaces is generally strong due to complexation reactions with surface metals.15,45,46 Thus, the desorption is minimal.19,47 Therefore, we surmise that the HA rich in carboxylic functionality seems to anchor to the goethite surface through ligand exchange with the surface hydroxyl groups of the goethite surface or through carboxylate bridging with Fe ions present on the surface by the electron donor-acceptor mechanism. The results of the NMR analysis of the SHA and unbound HA fractions after adsorption on goethite (GHAs) appear in Figure (41) Chen, Y.; Senesi, N.; Schnitzer, M. Soil Sci. Soc. Am. J. 1977, 41, 352358. (42) Kaiser, K.; Guggenberger, G. Eur. J. Soil Sci. 2007, 58, 45-59. (43) Guo, M. X.; Chorover, J. Soil Sci. 2003, 168, 108-118. (44) Filius, J. D.; Meeussen, J. C. L.; Lumsdon, D. G.; Hiemstra, T.; Van Riemsdijk, W. H. Geochim. Cosmochim. Acta 2003, 67, 1463-1474. (45) Wershaw, R. L.; Leenheer, J. A.; Sperline, R. P.; Song, Y. A.; Noll, L. A.; Melvin, R. L.; Rigatti, G. P. Colloids Surf., A 1995, 96, 93-104. (46) Kubicki, J. D.; Schroeter, L. M.; Itoh, M. J.; Nguyen, B. N.; Apitz, S. E. Geochim. Cosmochim. Acta 1999, 63, 2709-2725. (47) Avena, M. J.; Koopal, L. K. EnViron. Sci. Technol. 1998, 32, 2572-2577.

Kang and Xing

2 and Table 2. The NMR data show a quantitative distribution of different carbon moieties in the SHA and GHAs. The following major spectral bands were identified and integrated:31 0-50 ppm, mainly aliphatic or paraffinic carbons; 50-60 ppm, methoxy groups; 60-96 ppm, -CH2O- groups; 96-108 ppm, anomeric groups; 108-145 ppm, aromatic groups; 145-162 ppm, phenolic groups; and 162-220 ppm, carboxylic groups and carbonyl groups. Integration results showed that paraffinic C content did not differ significantly between SHA and GHAs, which revealed that aliphatic structures alone probably have little influence on the sorption of HA to mineral surfaces. Preferential exclusion of alkyl C has been observed for the sorptive interaction of NOM with hydrous oxides.12,26 However, the 13C NMR spectra of GHAs (Figure 2 and Table 2) reveal that the amount of aliphatic carbon groups with oxygen (50-108 ppm), including the methyl ether and carbohydrate carbon, diminished in the supernatant HA after adsorption (Figure 2 and Table 2). This finding indicates that heteroaliphatic carbons, aliphatic alcohols, and carbohydrate moieties preferentially adsorbed to goethite, which agrees with previous reports that the preferentially adsorbed heteroaliphatic fractions can be more reactive and form more stable complexes with metal cations.17,18 In addition, aromatic C fractions (108145 ppm) of HA are likely to remain in the solution after adsorption, while the content of phenolic C fractions (145-162 ppm) derived from lignin apparently decrease in the supernatant after HA sorption to goethite. Aromatic structures alone probably have little influence on the sorption of organic matter to the mineral surface;12,19,25 however, aromatic moieties with carboxyl and other functional groups have been recognized as structural units strongly involved in the sorption of organic matter. In our study, the contents of carboxylic and carbonyl groups, possibly attached to aromatic and/or aliphatic moieties, decreased after sorption to goethite removed these groups. The sorption of benzene carboxylic acids to goethite is known to increase with the number of carboxyl functional groups.35 Therefore, the adsorbed HA fractions likely contained molecules with high numbers of acidic groups attached per aromatic and aliphatic moiety, which are involved in the ligand exchange that may be one of the important surface complexation mechanisms of HA on goethite. Therefore, we propose that the acidic functional moieties may first complex to the goethite surfaces, and then hydrophobic interactions through association of the long aliphatic chains and/or larger aromatic ring structures lead to additional sorption.48 The DRIFT spectra emphasize speciation and adsorption modes of HA upon sorption on mineral surfaces. The band features of the spectra of SHA are at 2928 and 2847 cm-1 for aliphatic C-H stretching and at 1707 and 1578 cm-1 for carboxylic CdO stretching (Figure 3).18,49 The absorption bands at 1442 and 1371 cm-1 are assigned to the coupled C-O stretching and the OH in plane bending vibration, respectively, of the carboxylic acid moieties of HA. Phenolic and aliphatic alcohol fractions were evidenced by the presence of strong shoulders around 1250 and 1070 cm-1, respectively.49 Noticeably, by comparison with SHA, the stretching band of carboxylate COO- of the adsorbed HA fractions onto goethite (GHCs) is shifted from 1707 and 1578 cm-1 to 1664 and 1552 cm-1, respectively, indicating that COO- functional groups of HA may have complexed with goethite surfaces (Figures 3 and 4). Furthermore, a new band around 1400 cm-1 appeared due to the complexation between COO- of HA and Fe on goethite (48) Kang, S.; Xing, B. Langmuir 2007, 23, 7024-7031. (49) Wander, M. M.; Traina, S. J. Soil Sci. Soc. Am. J. 1996, 60, 1087-1094.

HA Fractionation upon Adsorption onto Goethite

Langmuir, Vol. 24, No. 6, 2008 2529

Table 1. Elemental Compositions, Atomic Ratios, Polarity Index, and E2/E3 and E4/E6 Ratios of the Source HA (SHA) and the Unbound HAs (GHAs) after Coating by Goethitea sample

C (%)b

H (%)b

N (%)b

O (%)b

ash (%)

H/C

O/C

(N + O)/C

E2/E3c

E4/E6 c

SHA GHA1d GHA2d GHA3d GHA4d

53.1 55.2 55.9 55.4 56.9

4.5 5.3 5.7 5.1 5.4

2.8 2.3 1.9 2.1 2.4

39.6 37.2 36.5 37.4 35.3

1.2 2.6 2.9 3.1 3.2

1.02 1.15 1.22 1.10 1.14

0.56 0.51 0.49 0.51 0.47

0.60 0.54 0.52 0.54 0.50

7.51 ( 0.12 6.79 ( 0.07 6.69 ( 0.02 6.56 ( 0.09 6.12 ( 0.05

8.74 ( 0.21 7.91 ( 0.08 7.03 ( 0.08 6.71 ( 0.12 6.38 ( 0.07

a C/H: atomic ratio of carbon to hydrogen. C/O: atomic ratio of carbon to oxygen. (N + O)/C: atomic ratio of the sum of nitrogen and oxygen to carbon. E2/E3 of humic acid is the ratio of the absorbance at 254 nm to that at 365 nm. E4/E6 of humic acid is the ratio of the absorbance at 465 nm to that at 665 nm. SHA: source humic acid. GHA: unbound humic acid after coating by goethite. b Values are on an ash-free basis. c Mean of ten replicates ( standard deviation. d Number is the sequence of adsorption.

Figure 2. Solid-state 13C NMR spectra for the source HA (SHA) and the two representative unbound HAs (GHAs) after coating on goethite.

surfaces (Figure 4), which was supported by the fact that a strong absorption band at 1400 cm-1 appeared when organic matter interacted with Fe to form complexes with goethite.18,50 The peak at 1552 cm-1 of GHCs in Figure 4 most likely represents an inner-sphere adsorption mode for carboxylic acids by goethite.51 A previous study reported that the outer-sphere adsorption mode of organic acid sorption onto clay minerals may change to inner-sphere complexation during dehydration,48 which was investigated with slight peak shifts of the asymmetric vibration of COO- from 1590 to 1550 cm-1. Therefore, we suggest that the electrostatic attraction or the outer-sphere adsorption of the HA on the goethite surface occurs first, followed by relatively strong bonding and inner-sphere complexation between the carboxylic groups of the HA and the surface. Another important observation is that the peak intensity of the bands at 1250 and 1070 cm-1 significantly diminished after HA sorption to goethite (Figure 3), which clearly indicates high affinity for the phenolic OH and C-O functional groups of HA.18,50 In the case of GHCs, we observed a shoulder around 1250 cm-1 and a clear peak at 1610 cm-1 representing the CdC of aromatic moieties, which confirms the fact that aromatic moieties with oxygen may be an important component to adsorb on goethite surfaces. Compared with the SHA, the GHAs showed a relatively weak band around 1070 cm-1, representing carbohydrates or polysaccharide-like substances. The intensity reduction of this peak reveals the complexation between the C-O functional groups of HA and goethite surfaces and/or is due to the contribution of the H-bonding mechanism. Similarly, many (50) Fu, H. B.; Quan, X. Chemosphere 2006, 63, 403-410. (51) Persson, P.; Axe, K. Geochim. Cosmochim. Acta 2005, 69, 541-552. (52) Jardine, P. M.; Weber, N. L.; McCarthy, J. F. Soil Sci. Soc. Am. J. 1989, 53, 1378-1385.

reported that the linear threadlike polysaccharides could be strongly adsorbed by oxide and aluminosilicate mineral surfaces.51,53,54 Finally, due to the adsorption of aliphatic carbon fractions, GHAs showed slightly decreased peak intensity at 2928 cm-1 with increasing sorption, and a weak peak at 2928 cm-1 was observed in GHCs. Figure 5 shows the MW distribution of HA during the adsorption process. The apparent MW (MWAP) of the SHA was 7.2 kDa and the MWAP of the GHAs was around 15.2 kDa. Meanwhile, the GHCs included both relatively small MWAP (4.5 kDa) and high MWAP (35 kDa) fractions (Figure 5). The adsorbed amount of the small MWAP fraction was greater than that of the high MWAP fraction, which implies that the small MWAP HA fraction preferentially adsorbed onto the goethite surface. The case of adsorbed HA with relatively low MWAP (4.5 kDa) differed from previous reports that high MWAP organic materials preferentially adsorbed onto clay minerals. However, the MWAP ranges of the preferentially adsorbed organic materials in their studies were between 1 and 10 kDa.20,23,24,29,52 Thus, the 4.5 kDa HA is within these MWAP ranges. The Mn and Mw of GHAs were calculated from an absorbance at 254 nm of the UV detector (Table 3). The Mw of the GHAs was very similar to that of the SHA because the change of Mw may result from a distribution of high MW fractions. Thus, the adsorption of high MWAP HA fractions (>20 kDa) may be quite low in comparison to relatively small MWAP HA fractions (around 5 kDa). Meanwhile, the Mn of the GHAs was higher than that of SHA due to high adsorption of the relatively small MWAP fractions onto goethite. The change of Mw and Mn resulted in the reduction of the polydispersity (F) of the GHAs, which implies the fractionation of different MWAP HA fractions (Table 3). In addition, the polydispersity of the GHCs increased from 4.25 of SHA to around 6-8, which reveals that goethite provided adsorption sites of HA with various MW ranges. Fluorescence detection at a specific excitation/emission (Ex/ Em) wavelength has been employed in HPSEC separation and characterization of organic matter.23,55 Generally, fluorescence signals in organic matter can be classified into three types of fluorophoric groups. One of the three groups usually has maximum Ex < 305 nm and maximum Em < 350 nm, which is related to aromatic amino acids, and is often referred to as proteinlike fluorescence.56,57 The other two groups with Ex and Em wavelengths ranging from 300 to 600 nm are attributed to humic substances, with humic acid fluorescing at longer Ex/Em wavelengths and fulvic acid at shorter wavelengths, referred to as humic-like and fulvic acid fluorescence, respectively.55-58 In (53) Gu, B. H.; Doner, H. E. Clays Clay Miner. 1992, 40, 151-156. (54) Theng, B. K. G. Clays Clay Miner. 1982, 30, 1-10. (55) Wu, F. C.; Evans, R. D.; Dillon, P. J. EnVion. Sci. Technol. 2003, 37, 3687-3693. (56) Mopper, K.; Schultz, C. A. Mar. Chem. 1993, 41, 229-238. (57) Coble, P. G. Mar. Chem. 1996, 51, 325-346. (58) Baker, A. EnViron. Sci. Technol. 2002, 36, 1377-1382.

2530 Langmuir, Vol. 24, No. 6, 2008

Kang and Xing

Table 2. Integration Results of the Solid-State 13C NMR Spectra of the Source HA (SHA) and the Unbound HAs (GHAs) after Coating by Goethitea distribution of C chemical shift (ppm), % sample

0-50

50-60

60-96

96-108

108-145

145-162

162-220

aliphatic C, %

aromatic C, %

aliphaticity (%)

SHA GHA1 GHA2 GHA3 GHA4

15.80 15.61 15.36 15.33 15.27

5.85 5.30 4.96 5.02 4.79

14.32 14.05 13.76 13.53 12.61

5.79 5.51 5.33 5.32 4.96

28.38 30.71 32.07 33.19 35.41

12.58 12.24 12.27 12.04 11.81

17.28 16.59 16.25 15.57 15.15

41.76 40.47 39.41 39.20 37.63

40.96 42.95 44.34 45.23 47.22

50.5 48.5 47.1 46.4 44.3

a GHA: unbound humic acid after coating by goethite. Aliphatic C ) total aliphatic carbon region (0-108 ppm). Aromatic C ) total aromatic carbon region (108-162 ppm). Aliphaticity ) aliphatic C (0-108 ppm)/sum of aliphatic C and aromatic C (0-162 ppm).

Figure 4. DRIFT spectra of goethite-HA complexes (GHCs) obtained from sequential sorption. The spectral range was from 1000 to 2000 cm-1 due to interference from Fe-O stretching peaks below 1000 cm-1 and weak peaks over 2000 cm-1.

Figure 3. DRIFT spectra of the source HA (SHA) and unbound HAs (GHAs) after coating on goethite.

our study, we employed a specific fluorescence at Ex/Em 467/ 548 nm, which is representative for humic-like fluorescence wavelengths,55 to separate HA and calculate the MW of HA. It was observed that both the Mn and Mw of the HAs obtained by fluorescence were significantly smaller that those obtained by using a UV detector (Table 3). HPSEC chromatograms from the fluorescence detector had a distribution of molecular Ex/Em of a humic-like fluorophore at 467/548 nm. For instance, a recent study reported that the UV mode can detect wide ranges of organic matter fractions including large MWAP fractions, such as proteins, which have a very strong signal at Ex/Em 280/325 nm and a negligible signal at Ex/Em 350/470 nm.59 They also showed that the MWAP of humic associated phenolic compounds was dominant around 7.6 kDa, which supports our results. Thus, humic-like fractions prefer to adsorb onto goethite in comparison with large size protein fractions (>20 kDa). Banaitis et al.60 reported that the sorption extents of tryptophan-like and tyrosine-like components onto goethite are distinctively lower than that of humiclike fractions. The abundance of acidic functional groups and aromatic carbon was higher with relatively small size fractions, whereas a higher (59) Maie, N.; Scully, N. M.; Pisani, O.; Jaffe, R. Water Res. 2007, 41, 563570. (60) Banaitis, M. R.; Waldrip-Dail, H.; Diehl, M. S.; Holmes, B. C.; Hunt, J. F.; Lynch, R. C.; Ohno, T. J. Colloid Interface Sci. 2006, 304, 271-276.

Figure 5. Comparison of the molecular weight distribution of HA before adsorption (SHA), remaining in solution after adsorption (GHAs, black curve), and (by difference) adsorbed to the surface (gray curve) with absorbance at 254 nm of UV and Ex/Em 467/548 nm of fluorescence.

content of aliphatic structures was observed for larger size fractions.61 Therefore, preferential sorption of relatively small size fractions (2 kDa), observed in the fluorescence mode, may (61) Shin, H.; Monsallier, J. M.; Choppin, G. R. Talanta 1999, 50, 641-647.

HA Fractionation upon Adsorption onto Goethite

Langmuir, Vol. 24, No. 6, 2008 2531

Table 3. Mw, Mn, and Polydispersity (G) of the Source HA (SHA) and the Unbound HAs after Coating by Goethite (GHAs), and the Goethite-HA Complexes (GHCs) at 254 nm of UV and Ex/Em 467/548 nm of Fluorescence Humic Acid (dalton)a SHA

GHA1

GHA2

GHA3

GHA4

at 254 nm of UV

Mn Mw F

5610 ( 310 23830 ( 540 4.25 ( 0.15

10760 ( 210 23070 ( 410 2.14 ( 0.19

7400 ( 260 22960 ( 540 3.10 ( 0.21

6910 ( 310 23260 ( 410 3.37 ( 0.16

7320 ( 230 23690 ( 410 3.24 ( 0.18

at Ex/Em 467/548 nm of fluorescence

Mn Mw F

1270 ( 120 6390 ( 190 5.02 ( 0.23

2770 ( 130 7600 ( 240 2.74 ( 0.25

1740 ( 90 7580 ( 210 4.35 ( 0.09

1600 ( 110 6900 ( 240 4.31 ( 0.11

1660 ( 70 6780 ( 190 4.08 ( 0.07

Goethite-HA Complex (dalton)a

a

GHC1

GHC2

GHC3

GHC4

at 254 nm of UV

Mn Mw F

2880 ( 130 19150 ( 330 6.65 ( 0.31

2670 ( 160 19980 ( 510 7.48 ( 0.45

2570 ( 220 18310 ( 540 7.12 ( 0.39

2230 ( 120 17120 ( 310 7.68 ( 0.24

at Ex/Em 467/548 nm of fluorescence

Mn Mw F

1080 ( 180 3730 ( 210 3.45 ( 0.34

1190 ( 110 3990 ( 170 3.35 ( 0.21

1050 ( 170 3710 ( 290 3.53 ( 0.41

920 ( 90 3980 ( 290 4.33 ( 0.32

Mean of five replicates ( standard deviation.

be associated more with ligand exchange and/or electrostatic attraction through acidic functional groups of aromatic fractions, whereas adsorption of larger size fractions (>5 kDa) from the UV mode may be more dependent on hydrophobic interactions. Since a greater contribution of ligand exchange and/or electrostatic attraction is expected in the goethite system, the adsorption of small MW fractions having more acidic functional groups seems to surpass the adsorption of higher MW fractions primarily governed by hydrophobic interactions.

4. Conclusions HA sorption by goethite in this study significantly depends on solution pH. Acidic conditions yielded higher sorption capacity, and sorption decreased with increasing pH. HA molecules underwent fractionation during adsorption on the goethite surface. To date, it has been reported that high MW and hydrophobic fractions prefer to adsorb to goethite. However, the decreased polarity of the HA fractions remaining in supernatants after coating implies adsorption of hydrophilic fractions of the original HA. DRIFT showed that goethite provided sorption sites for carboxylic carbons. NMR results suggested that carbohydrate carbons were the main fractions to adsorb by goethite. We suggest that the process for the adsorption of HA occurred in the following steps:

(1) electrostatic attraction or outer-sphere complexation of the polar functional moieties on relatively small and intermediate MW HA fractions (MWAP ) 4.5 kDa) on the goethite surface; (2) ligand exchange of the carboxylic groups of the HA to yield inner-sphere complexes; and (3) hydrophobic interactions on the HA-goethite complex that can provide new hydrophobic sorption sites for high MW HA fractions (MWAP ) 35 kDa) with condensed aromatic or long chain aliphatic moieties. Adsorbed SOM may influence the sorption characteristics of minerals and inhibit dissolution and weathering of minerals.5,7,62 In addition, SOM chemically adsorbed on mineral surfaces can resist chemical/biological degradation,11,42 which has significant implications for global carbon cycling. Acknowledgment. This research was supported by the Massachusetts Agricultural Experiment Station (MA 8532). Supporting Information Available: Flow chart of the preparation of the humic acid-goethite complex and semilog linear calibration curve used to calculate the molecular weights of the humic acid samples. This material is available free of charge via the Internet at http://pubs.acs.org. LA702914Q (62) Golubev, S. V.; Bauer, A.; Pokrovsky, O. S. Geochim. Cosmochim. Acta 2006, 70, 4436-4451.