Langmuir 2008, 24, 6683-6692
6683
Adsorption of Organic Matter at Mineral/Water Interfaces: 7. ATR-FTIR and Quantum Chemical Study of Lactate Interactions with Hematite Nanoparticles Juyoung Ha,*,† Tae Hyun Yoon,†,§ Yingge Wang,† Charles B. Musgrave,‡,⊥ and Gordon E. Brown, Jr.†,| Surface & Aqueous Geochemistry Group, Department of Geological & EnVironmental Sciences, and Department of Chemical Engineering Stanford UniVersity, Stanford, California 94305-2115, Department of Chemical Engineering, Stanford UniVersity, Stanford, California 94305-5025, Department of Chemistry, Hanyang UniVersity, Seoul, 133-791, Korea, and Photon Science Department and SSRL, 2575 Sand Hill Road, SLAC, MS 69, Menlo Park, California 94025 ReceiVed January 14, 2008. ReVised Manuscript ReceiVed March 6, 2008 The interaction of the L-lactate ion (L-CH3CH(OH)COO-, Lact-1) with hematite (R-Fe2O3) nanoparticles (average diameter 11 nm) in the presence of bulk water at pH 5 and 25 °C was examined using a combination of (1) macroscopic uptake measurements, (2) in situ attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy, and (3) density functional theory modeling at the B3LYP/6-31+G* level. Uptake measurements indicate that increasing [Lact-1](aq) results in an increase in Lact-1 uptake and a concomitant increase in Fe(III) release as a result of the dissolution of the hematite nanoparticles. The ATR-FTIR spectra of aqueous Lact-1 and Lact-1 adsorbed onto hematite nanoparticles at coverages ranging from 0.52 to 5.21 µmol/m2 showed significant differences in peak positions and shapes of carboxyl group stretches. On the basis of Gaussian fits of the spectra, we conclude that Lact-1 is present as both outer-sphere and inner-sphere complexes on the hematite nanoparticles. No significant dependence of the extent of Lact-1 adsorption on background electrolyte concentration was found, suggesting that the dominant adsorption mode for Lact-1 is inner sphere under these conditions. On the basis of quantum chemical modeling, we suggest that inner-sphere complexes of Lact-1 adsorbed on hematite nanoparticles occur dominantly as monodentate, mononuclear complexes with the hydroxyl functional group pointing away from the Fe(III) center.
1. Introduction The nature of bonding between organic species and metal oxide surfaces can substantially alter the properties of metal oxide substrates and thereby affect the geochemical cycling of metals, the dissolution of redox-sensitive metal oxides, and the aggregation of colloids and nanoparticles. Low-molecular-weight (LMW) organic acids can adsorb onto metal oxide surfaces either by specific chemical interactions (chemisorption) to form innersphere complexes or by nonspecific interactions (physisorption) via hydrogen bonding and/or electrostatic interactions to form outer-sphere complexes.1–11 Such organic materials can also interact to form a surface precipitate,12,13 as well as ternary surface * To whom correspondence should be addressed. E-mail: jyha@ stanford.edu. Phone: 650-723-7513. Fax: 650-725-2199. † Department of Geological & Environmental Sciences, Stanford University. ‡ Department of Chemical Engineering, Stanford University. § Hanyang University. | SLAC. ⊥ Present address: Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309-0215. (1) Hug, S. J.; Sulzberger, B. Langmuir 1994, 10, 3587. (2) Filius, J. D.; Hiemstra, T.; van Riemsdijk, W. H. J. Colloid Interface Sci. 1997, 195, 368. (3) Awatani, T.; Dobson, K. D.; McQuillan, A. J.; Ohtani, B.; Uosaki, K. Chem. Lett. 1998, 849. (4) Axe, K.; Persson, P. Geochim. Cosmochim. Acta 2001, 65, 4481. (5) Johnson, S. B.; Yoon, T. H., Jr. Langmuir 2005, 21, 2811. (6) Yoon, T. H.; Johnson, S. B., Jr. Langmuir 2004, 20, 5655. (7) Cornell, R. M.; Schindler, P. W. Colloid Polym. Sci. 1980, 258, 1171. (8) Davis, J. A. Geochim. Cosmochim. Acta 1982, 46, 2391. (9) Johnson, S. B., Jr.; Healy, T. W.; Scales, P. J. Langmuir 2005, 21, 6356. (10) Rosenqvist, J.; Axe, K.; Sjoberg, S.; Persson, P. Colloids Surf., A 2003, 220, 91. (11) Violante, A.; Gianfreda, L. Soil Sci. Soc. Am. J. 1993, 57, 1235. (12) Hongshao, Z.; Stanforth, R. EnViron. Sci. Technol. 2001, 35, 4753.
complexes with cations.14 Among the LMW organic acids, lactate plays a significant environmental role because of its prevalence in soil environments resulting from its exudation by plant roots, production by fungi, and discharge by microorganisms.15–19 However, only a few studies have investigated the sorption modes of lactate onto metal oxide surfaces. Cornell and Schindler7 performed infrared spectroscopic experiments on lactate adsorption on goethite (R-FeOOH) and amorphous Fe(III) hydroxide. On both iron (hydr)oxide surfaces studied by Cornell and Schindler,7 lactate was suggested to adsorb as monodentate innersphere surface complexes. They concluded that the carboxyl group of lactate is involved in binding to goethite surfaces and that the deprotonated alcoholic hydroxyl groups also participate in binding to amorphous Fe(III) hydroxide surfaces. Filius et al.2 studied the adsorption of lactate along with other LMW organic acids on goethite and fit the uptake data using the CD-MUSIC model. Based on their fits, they suggested that lactate adsorbs onto goethite predominantly as outer-sphere complexes. However, Filius et al.2 also considered and included a number of innersphere species as minor surface complexes at the goethite/water interface in order to improve the goodness of their model fit to the experimental uptake data as a function of pH. It should be (13) Ler, A.; Stanforth, R. EnViron. Sci. Technol. 2003, 37, 2694. (14) Lenhart, J. J.; Bargar, J. R.; Davis, J. A. J. Colloid Interface Sci. 2001, 234, 448. (15) Kato-Noguchi, H. Physiol. Plant. 2000, 109, 28. (16) Hakki, E. E.; Akkaya, M. S. Enzyme Microb. Technol. 2001, 28, 259. (17) Ryan, P. R.; Delhaize, E.; Jones, D. L. Ann. ReV. Plant Phys. Plant Mol. Biol. 2001, 52, 527. (18) Bylund, D.; Samskog, J.; Markides, K. E.; Jacobsson, S. P. J. Am. Soc. Mass Spectrom. 2003, 14, 236. (19) Van Hee, P.; Neels, H.; De Doncker, M.; Vrydags, N.; Schattemann, K.; Uyttenbroeck, W.; Hamers, N.; Himpe, D.; Lambert, W. Clin. Chem. Lab. Med. 2004, 42, 1341.
10.1021/la800122v CCC: $40.75 2008 American Chemical Society Published on Web 06/04/2008
6684 Langmuir, Vol. 24, No. 13, 2008
noted that the suggested outer-sphere lactate surface complexes in the study by Filius et al. were not confirmed spectroscopically. In a related study, Awatani et al.3 examined lactate adsorption on titanium dioxide using ATR-FTIR spectroscopy and concluded that lactate adsorbs in a bidentate fashion involving hydroxyl and carboxyl groups. As shown by these examples, the structures and binding modes of lactate species at metal oxide/water interfaces may vary, depending on the types of metal oxides, experimental conditions and methods used, and are still subjects of debate. Among the Fe oxides present in natural environments, hematite (R-Fe2O3) is one of the most abundant mineral phases.20 It is also a common mineral nanoparticle in the environment, occurring in soils,21 acid mine drainage effluent,22 and on bacterial surfaces23,24 as well as in atmospheric dust.25 From an environmental perspective, one of the more important features of mineral nanoparticles is their high characteristic surface area, which facilitates their role as powerful sequestrants of ions from solution. In the present study, we examined the interaction of lactate ion (Lact-1) with nanoparticles of hematite (R-Fe2O3) in the presence of bulk water at 25 °C using a combination of in situ attenuated total reflectance-Fourier transform infrared (ATRFTIR) spectroscopy and quantum chemical modeling. The mode and extent of lactate adsorption on hematite as a function of lactate concentration and background electrolyte concentration were investigated at pH 5.0. ATR-FTIR spectroscopy is one of the most direct methods by which to distinguish different structures of organic adsorbents at metal oxide/water interfaces under in situ conditions.6,12,13 Results from ATR-FTIR spectroscopy can, in favorable cases, be used to differentiate between inner- and outer-sphere organic species, but they do not always provide definitive information on different coordination geometries for inner-sphere complexes, such as monodentate versus bidentate structures. Here, we have employed quantum chemical methods at the density functional theory level to aid the interpretation of our ATR-FTIR spectra and to provide a more detailed model of the coordination geometry of adsorbed innersphere lactate complexes at the interface between hematite nanoparticles and aqueous solutions.
2. Experimental Section 2.1. Materials. Reagent-grade sodium L-lactate from SigmaAldrich was dissolved in deionized water to make a 0.1 M aqueous stock solution containing 0.01 or 0.5 M NaCl. To remove dissolved CO2 in solution, the deionized (Milli-Q Plus) water used in the experiments was boiled for 60 min prior to use and cooled in a glovebox to avoid the diffusion of CO2 back into solution. Highpurity (>99.95%) hematite (R-Fe2O3) used in this investigation was purchased from Alfa Aesar. It has a reported Brunauer-EmmetTeller (BET) surface area of 250 m2/g and a mean particle diameter of 10 nm. The surface area value was checked in our laboratories by N2 BET measurements using a Micromeritics ASAP 2020 Analyzer, and the mean particle diameter was checked by dynamic light scattering using a Malvern Instruments ZEN3601 Zetasizer Nano-ZS. The hematite particles were further characterized using a FEI XL30 Sirion scanning electron microscope (SEM) at the Stanford Nanocharacterization Laboratory, a Jeol TEM1230 transmission electron microscope (TEM) at the Stanford Cell Sciences (20) Cornell, R. M.; Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses; Wiley: New York, 2003. (21) Penn, R. L.; Zhu, C.; Xu, H.; Veblen, D. R. Geology 2001, 29, 843. (22) Hochella, M. F, Jr. Geochim. Cosmochim. Acta 2001, 66, 735. (23) Banfield, J. F.; Zhang, H. ReV. Mineral. Geochem. 2001, 44, 1. (24) Banfield, J. F.; Welch, S.; Zhang, H.; Ebert, T.; Peen, R. L. Science 2000, 289, 751. (25) Anastasio, C.; Martin, S. ReV. Mineral. Geochem. 2001, 44, 293.
Ha et al. Imaging Facility, and a microdiffraction instrument (µ-XRD) at station 32-ID of the Advanced Photon Source (APS) at Argonne National Laboratory. Stock hematite suspensions were prepared by ultrasonically dispersing the hematite nanoparticles (10 g/L) in aqueous 0.01 or 0.5 M NaCl at pH 5.0 using a Branson model 450 digital sonifier equipped with a 0.5 in. horn. The suspensions were then slowly rotated end over end for 24 h prior to use. 2.2. ATR-FTIR Sample Preparation and Measurements. Wet paste samples for ATR-FTIR analysis were prepared by adding aliquots of the hematite stock suspensions to polypropylene centrifuge tubes, followed by the addition of Lact-1 stock solution and 0.01 or 0.5 M NaCl stock solution, depending on the experimental objective, to achieve the desired concentration of Lact-1. The pH of individual sample tubes was adjusted to the desired value by adding small volumes of 1.0 N HCl and/or 1.0 N NaOH. The tubes were then purged with high-purity nitrogen gas and slowly rotated end over end for 48 h while the pH was readjusted every 12 h. To separate and concentrate the pastes, each sample was centrifuged at 10 000g for 30 to 60 min. The supernatant was decanted and passed through a 20 nm syringe filter to remove any residual hematite particles. A thin layer of the concentrated wet paste of each sample was uniformly applied directly to the germanium (Ge) ATR crystal with a small volume of the decanted supernatant being applied on top of the wet paste layer to measure lactate sorbed on hematite nanoparticle surfaces. The sample holding region was sealed with a lid to prevent evaporation during ATR-FTIR measurements. Solution samples of lactate were directly applied to the Ge ATR crystal, and a lid was placed over the crystal. The Lact-1 and Fe(III)-lactate solutions for ATR-FTIR measurements were prepared in 0.01 M NaCl. (See Supporting Information for the Fe(III)-lactate solution sample preparation.) To ensure that the solutions contained predominantly one species, calculations were performed using reported stability constants and FITEQL 3.2.28–30 The lactate species in solution as a function of pH are summarized in Figure S1 in Supporting Information. Attenuated total reflectance-Fourier transform infrared (ATRFTIR) measurements were made using a Nicolet FTIR spectrometer (NEXUS470) equipped with a mercury cadmium telluride (MCT) detector and a horizontal attenuated total reflectance attachment (germanium crystal). Data collection and spectral calculations were carried out using OMNIC software (version 6.0a, Nicolet Instrument Corp.). Five hundred scans were collected and averaged per sample with a spectral resolution of 4 cm-1. The strong spectral contribution of water was removed from the ATR-FTIR spectrum of each aqueous lactate spectrum by subtracting the spectrum of a 0.01 M NaCl solution at the same pH. All final ATR-FTIR spectra of wet pastes were processed in a manner similar to that of the aqueous samples by subtracting the spectrum of the corresponding filtered suspension supernatants from the spectrum of the measured sample. The normalization and baseline correction of ATR-FTIR spectra were carried out with great care so as not to enhance or diminish the magnitude of any spectral peaks or shift the peak positions. (See Supporting Information for detailed baseline correction and Figure S2 for normalization procedures.) Deconvolution of the normalized spectra into individual components using the program PeakFit4 (PeakFit version 4 for Windows, SYSTAT Software Inc.) makes it possible to estimate the fraction of each type of Lact-1 surface complex for different values of Γlactate. A Gaussian line shape was used in the fitting analysis with linear background fitting within the wavenumber range of 1200-1800 cm-1. (26) Degenhardt, J.; McQuillan, A. J. Langmuir 1999, 15, 4595. (27) McClenny, W. A.; Krost, K. J.; Daughtrey, E. H., Jr.; Williams, D. D.; Allen, G. A Appl. Spectrosc. 1994, 48, 706. (28) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum Press: New York, 1982; Vol. 5. (29) Arnold, R. G.; Olson, T. M.; Hoffmann, M. R. Biotechnol. Bioeng. 1986, 28, 1657. (30) Herbelin, A. L.; Westall, J. C. FITEQL Version 3.2: A Computer Program for Determination of Chemical Equilibrium Constants from Experimental Data; Department of Chemistry, Oregon State University: Corvallis, OR, 1996.
Adsorption of Organic Matter at Interfaces 2.3. Uptake and Dissolution Measurements. The filtered supernatants described in section 2.2 were diluted to appropriate concentrations (if required) and transferred into 0.5 mL filter-capped vials. They were analyzed for residual Lact-1 using a Dionex ion chromatograph equipped with an AS40 autosampler, a GP50 gradient pump, a LC30 chromatography oven, and an ED40 electrochemical detector. Lact-1 adsorption was calculated from the difference between the added lactate concentration and that measured in the supernatant after 48 h of reaction. The dissolved iron concentration was monitored in filtered supernatants using inductively coupled plasma (ICP) spectrometry. A TJA IRIS Advantage/1000 Radial ICAP spectrometer equipped with a solid state CID detector was used for these measurements. 2.4. Quantum Chemical Calculations. Calculations of optimized geometries and IR vibrational frequencies for Lact-1 and iron clusters were performed using the Gaussian 03 program.32 Structures of the modeled molecules and clusters were energy minimized with respect to all atomic coordinates using Hartree-Fock (HF) or B3LYP hybrid DFT functionals with the 6-31+G* or 6-31G* basis set. Modeled structures were optimized without symmetry constraints on any of the molecules. Frequencies were calculated on the basis of the energyminimized structures and scaled by a factor of 0.8929 for HF33 and 0.9613 for B3LYP34 to correct for systematic errors such as the neglect of anharmonicity, basis set limitations, and approximating electron correlations.35 No imaginary frequencies were found for any of the optimized structures, indicating that they represent the true minima in the potential energy surface. Atomic movements corresponding to calculated frequencies were monitored with MOLDEN (version 4.0) software.36 [Fe3+ · (H2O)6]+3 and [Fe3+ · (H2O)5(OH)]+2 molecular clusters were chosen as simplified models of the hematite nanoparticle substrates to calculate the effects of different geometries of the adsorbed Lact-1 on vibrational frequencies. Edge-sharing iron octahedral clusters, [(Fe3+)2(OH)2(H2O)8]+3, were also employed as models for Fe3+ dimer complexes to simulate the effects of a neighboring iron octahedron on the vibrational frequencies of Lact-1 surface complexes. The effect of water molecules on calculated structures and vibrational frequencies was explicitly modeled by including different numbers of water molecules around the Lact-1 molecule in the calculations. The atomic coordinates resulting from the geometry optimization of one such hydrated lactate molecule are given in Table S2 in Supporting Information.
3. Results and Discussion 3.1. Hematite Nanoparticle Characterization. The µ-XRD pattern of the hematite nanoparticles shows good agreement with earlier X-ray diffractograms of hematite as determined by Zachariasen37 (Figure S3 in Supporting Information). The shapes (31) Johnson, S. B.; Yoon, T. H.; Kocar, B. D., Jr. Langmuir 2004, 20, 4996. (32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, reVision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (33) Pople, J. A.; Scott, A. P.; Wong, M. W.; Radom, L. Isr. J. Chem. 1993, 33, 345. (34) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391. (35) Frisch, M. J.; Del Bene, J. E.; Raghavachari, K.; Pople, J. A. Chem. Phys. Lett. 1981, 83, 240. (36) Schaftenaar, G.; Noordik, J. H. J. Computer-Aided Mol. Design 2000, 14, 123. (37) Zachariasen, W. H. Skr. Nor. Vidensk.-Akad., [Kl.] 1: Mat. NaturVidensk. Kl. 1928, 1928, 1.
Langmuir, Vol. 24, No. 13, 2008 6685
and sizes of hematite nanoparticles were characterized using SEM and TEM (Figure S4 in Supporting Information). The nanoparticles are spherical and highly aggregated; their morphology did not change after reaction with Lact-1 (not shown). The average particle diameter, based on measurement of 150 nanoparticles, was found to be 11.2 ( 2.7 nm by SEM and 10.5 ( 1.6 nm by TEM. The specific surface area was determined by the BET method (N2 adsorption) to be 218 ( 6.3 m2/g (vs 250 m2/g reported by Alfa Aesar). The average particle size determined using laser light scattering is 25.7 ( 2.0 nm, which is significantly larger than the values from TEM or SEM observations as a result of the aggregated nature of the particles, as shown in SEM and TEM images (Figure S4 in Supporting Information). The average size chosen for the hematite nanoparticles is 10.8 ( 3.1 nm on the basis of the averaged values of SEM and TEM particle size measurements. The measured total surface area of 218 ( 6.3 m2/g was used for all further calculations. This value is likely to be different from that of the wet nanoparticles because of differences in the extent of aggregation between the wet and dry states of the nanoparticles. Recently, Vikesland et al.38 showed that the actual available reactive surface area in their study of carbon tetrachloride removal by magnetite nanoparticles was greater than that estimated from the dry magnetite particles. In addition, Gilbert et al.39 studied the aggregation of iron oxyhydroxide nanoparticles as a function of solution pH and reported that the formation of nanoclusters occurred between pH 5 and 6.6 without further changes in the sizes of the aggregates during 10 weeks of aging. They reported that these stable nanoclusters still retain a very high surface area and suggested that the reactivity of these aggregates of nanoparticles may be different from that of bulk materials. Others have studied the reactivity of aggregated nanoparticles with respect to different organic and inorganic ligands and reported an enhanced reactivity of nanoparticle aggregates compared to that of microparticles.38,40,41 On the basis of these earlier studies, it is likely that the reported lactate adsorption on aggregated hematite nanoparticles in the present study would be different from that on the surfaces of microparticles or bulk crystals and that the extent of lactate adsorption would depend on the extent of nanoparticle aggregation. Additional detailed experiments are necessary to determine the effect of colloidal aggregation of hematite nanoparticles on lactate adsorption mechanisms. 3.2. ATR-FTIR Measurements and Quantum Chemical Calculations of Aqueous Lactate. Lactate occurs in aqueous solutions as fully deprotonated or singly protonated species, depending on pH. Each of these species is expected to give rise to a distinctively different infrared spectrum as a result of the different point symmetries of the two molecules as a function of pH value (pKa ) 3.6).28 ATR-FTIR spectra of the lactate ion (Lact-1) and lactic acid (HLact) are presented in Figure 1. ATRFTIR spectra collected for different protonation states of lactate agree with those reported earlier.3,42 In the spectrum of Lact-1, a strong peak was observed at 1575 cm-1, corresponding to the asymmetric stretch (νas) of deprotonated carboxyl group. The major peaks at lower frequencies are due either to combinations of the symmetric stretching of the carboxyl group (νs), C-C stretching (νC-C), and CCH bending (δCCH), or to symmetric and (38) Vikesland, P.; Heathcock, A.; Rebodos, R.; Makus, E. EnViron. Sci. Tehcnol. 2007, 41, 5277. (39) Gilbert, B.; Lu, G.; Kim, C. J. Colloid Interface Sci. 2007, 313, 152. (40) Madden, A. S.; Hochella, M. F., Jr. Geochim. Cosmochim. Acta 2005, 69, 389. (41) Madden, A. S.; Hochella, M. F., Jr.; Luxton, T. Geochim. Cosmochim. Acta 2005, 70, 4.95. (42) Strathmann, T. J.; Myneni, S. C. B. Geochim. Cosmochim. Acta 2004, 68, 3441.
6686 Langmuir, Vol. 24, No. 13, 2008
Ha et al.
Figure 1. ATR-FTIR spectra of lactate in aqueous solution as a function of pH. Solution concentration: 0.1 M lactate in 0.01 M NaCl. Table 1. Experimental Frequencies of Lact-1 from the ATR-FTIR Spectrum and Corresponding Peak Assignments Based on the Quantum Chemical Simulationsa observed species -
CH3CH(OH)COO CH3CH(OH)COOCH3CH(OH)COO-
(vas)c
(vs)d
∆v
reference
1575 1575 1566
1417 1417 1422
153 153 144
this study Strathmann and Myneni42 Awatani et al.3
calculation parameters
(vasco)c
(vsco)d
∆v
state
HF/6-31 G* HF/6-31 G** B3LYP/6-31 G* B3LYP/6-31 G** HF/6-31 G* HF/6-31 G** B3LYP/6-31 G* B3LYP/6-31 G** B3LYP/6-31 G** B3LYP/6-31 G** B3LYP/6-31 G**
1433 1399 1567 1456 1443 1385 1558 1530 1530 1550 1525
1343 1311 1418 1352 1346 1325 1420 1466 1409 1416 1408
90 88 149 104 97 60 138 64 121 134 117
vacuum vacuum vacuum vacuum 5 watersb 5 watersb 5 watersb 2 watersb 4 watersb 5 watersb,e SCRF (PCM)
a HF ) Hartree-Fock; SCRF ) self-consistent reaction field; PCM ) polarized continuum model. b Solvation effect simulated by adding a specific number of water molecules in the vicinity of the carboxylic or alcoholic functional group of lactate. c Refers to the asymmetric stretching vibration of carbonyl. d Refers to the symmetric stretching vibration of carbonyl. e See Supporting Information for Cartesian coordinates of the water molecules with respect to lactate.
asymmetric stretching of C-OH (νRC-OH). The major absorption peaks for Lact-1 are summarized in Table S1 with assignments to different vibrational modes based on quantum chemical calculations carried out in this study as well as by others.3,42,43 In performing the quantum chemical calculations for the aqueous Lact-1 species, the optimum levels of theory and basis sets were selected based on comparison between the observed vibrational frequencies and the calculated vibrational frequencies of the asymmetric (νas) and symmetric (νs) stretching of Lact-1 (Table 1). This comparison shows significant differences depending on the level of theory, the basis set selected, and the environmental conditions (i.e., vacuum vs hydrated conditions). The experimental and theoretical results match reasonably well at the B3LYP/6-31+G* level, with the effect of solvation modeled by explicitly adding five water molecules around the Lact-1 molecule. It is necessary to add the effect of solvation in an aqueous environment, particularly for molecules with polar (43) Pike, P. R.; Sworan, P. A.; Cabaniss, S. E. Anal. Chim. Acta 1993, 280, 253.
Figure 2. ATR-FTIR spectra of (i) aqueous Lact-1 at 0.1 M and pH 5.0 in 0.01 M NaCl; (ii) filtered supernatant of sorption sample at Γ ) 3.52 µmol/m2; (iii) aqueous Fe(III)-lactate complex at 5 mM lactate and 10 mM Fe(III) and pH 2.5 in 0.01 M NaCl; (iv) Lact-1 adsorbed at the hematite/water interface at Γ ) 3.52 µmol/m2 and pH 5.0 in 0.01 M NaCl.
functional groups, because hydrogen bonds between these functional groups and water molecules are likely to have a significant influence on vibrational frequencies. Water molecules were placed in the vicinity of the carboxyl and alcohol functional groups of the Lact-1 molecule, and the modeled structures were optimized without symmetry constraints on any of the molecules. (See Table S2 in Supporting Information). The B3LYP/6-31+G* level of theory and basis set, coupled with the simulated hydrous environment, were used in all subsequent calculations of geometric structures and vibrational frequencies for adsorbed Lact-1 on hematite surfaces. 3.3. ATR-FTIR Measurements of Adsorbed Lactate at 0.01 M NaCl. The experimental ATR-FTIR spectrum of Lact-1 adsorbed on hematite nanoparticles at pH 5 (Figure 2(iv)) differs significantly from that of aqueous Lact-1 at pH 5 (Figure 2(i)) including differences in peak widths and peak frequencies, especially for peaks occurring at 1700-1200 cm-1. However, the number of peaks in the ATR-FTIR spectrum for a wet paste of Lact-1 adsorbed on hematite nanoparticles is the same as that in the ATR-FTIR spectrum of aqueous Lact-1 at pH 5.0. The relationship between the symmetry of organic molecules and their infrared spectra is well established,44,45 and it is possible to suggest the types of surface complexes when the symmetry of the adsorbed molecule is significantly different from that of the solution species on the basis of the number and position of the ATR-FTIR peaks. For example, the ATR-FTIR spectrum of an outer-sphere complex often resembles that of the aqueous species because the outer-sphere complex is expected to retain its waters of hydration and to form no direct chemical bonds with the underlying metal oxide surface. In contrast, the ATRFTIR spectrum of an inner-sphere complex may differ significantly from that of the aqueous species as a result of the distortions of the molecular structure upon adsorption on the substrate surface. Peak shifts or splits in the ATR-FTIR spectra of LMW organic acids are often observed as a result of specific carboxylate-metal ion interactions when an inner-sphere complex forms.6 However, (44) Biber, M. V.; Stumm, W. EnViron. Sci. Technol. 1994, 28, 763. (45) Peak, D.; Ford, R. G.; Sparks, D. L. J. Colloid Interface Sci. 1999, 218, 289.
Adsorption of Organic Matter at Interfaces
no change was observed in the number of peaks in the experimentally obtained ATR-FTIR spectra of Lact-1 adsorbed on hematite surfaces in the present study. Aqueous Lact-1 has C1 point symmetry, which is the lowest point symmetry possible, and thus the overall symmetry of the molecule cannot be lowered upon adsorption to hematite surfaces. It is possible that the overall symmetry of Lact-1 may increase when it adsorbs on hematite surfaces, leading to a decrease in the number of observed ATRFTIR peaks; however, a decrease in the number of ATR-FTIR peaks was not observed. The significant differences in peak positions and shapes between the ATR-FTIR spectra for adsorbed Lact-1 and aqueous Lact-1 suggest a change in the coordination environment of Lact-1 upon the adsorption on hematite nanoparticles. As shown in Figure 2, we observed shifts in the νas peak at 1575 cm-1 to 1591 cm-1 and the νs peak at 1417 cm-1 to 1376 cm-1 when Lact-1 was adsorbed on hematite surfaces at a surface coverage of Γ ) 3.52 µmol/m2. Similar shifts in the νas and νs peak positions have been observed when Lact-1 was adsorbed on TiO23. The νas peak also becomes broader when Lact-1 adsorbs on hematite. For example, the full width at half maximum (fwhm) of the peak representing the νas vibrational mode changed from 46 cm-1 in the aqueous Lact-1 spectrum to 63 cm-1 in the adsorbed Lact-1 spectrum. Similar findings of peak broadenings and shifts have been reported previously by a number of other authors investigating the adsorption modes of LMW organic acids on metal oxide surfaces. In these studies, the observed increases in fwhm and in the frequency of νas peaks are often interpreted as being due to the formation of an outer-sphere adsorption complex.5,10,31 For example, Rosenqvist et al.10 suggested that the broadening of νas peaks is due to the replacement of one or several of the waters of hydration around the ligand by the proton donors at the surface, and hence ligands are adsorbed to the first surface layer of hydroxyl groups and form an outer-sphere complex. Axe and Persson4 reported similar observations of peak broadening and peak shifts and attributed these changes to an asymmetric environment around the ligand. They concluded that the sorbate is hydrogen bonded both to protons at the surface and protons in the surrounding water, hence causing the asymmetric environment around the sorbate. In the present study, we also suggest that the experimentally observed increase in fwhm of the Lact-1 νas peak and the shift of this peak to higher frequency, relative to that of aqueous Lact-1, are in part due to the complex hydrogen bonding environment around the sorbed Lact-1 at the hematite/water interface. It should also be noted that there is a significant similarity between the νas peak position in the spectra of aqueous L-act-1 and the aqueous Fe(III)-lactate complex (cf. Figure 2(i) and (iii)). The νas peak position in the spectrum of aqueous Lact-1 occurs at 1575 cm-1, whereas it occurs at 1577 cm-1 for Lact-1 complexed to Fe3+ in solution. We found that distinguishing between Fe(III)-lactate aqueous complexes and aqueous Lact-1 is difficult on the basis of only observed differences in νas peak in the ATR-FTIR spectra. However, there is a change in the νas peak shape between the aqueous Fe(III)-lactate complex spectrum and the spectrum of Lact-1 adsorbed on hematite. We also observed a significant shift in the νas peak from 1577 cm-1 for aqueous Fe(III)-lactate complexes to 1591 cm-1 for Lact-1 adsorbed on hematite nanoparticles. These observations lead us to suggest that the νas peak broadening and shift are most sensitive to changes in hydrogen bonding around the Lact-1 molecule when Lact-1 adsorbs to hematite nanoparticles and that these changes are likely due to the altered hydrogen bonding environment around the carboxylates. Because of the different
Langmuir, Vol. 24, No. 13, 2008 6687
Figure 3. Adsorption isotherm of Lact-1 on hematite nanoparticles at 25 °C and 0.01 and 0.5 M of NaCl at pH 5.0. The inset shows the concentration of Fe(III) dissolved from the hematite surface as a function of Lact-1 loading.
acidities of different surface groups on hematite, it is possible that a broad distribution of hydrogen bond strengths occurs when Lact-1 adsorbs to hematite surfaces. If this is true, then it follows that Lact-1 is likely to adsorb to the first layer of hydroxyl groups on hematite surfaces (e.g., -OH2 in >FeOH20.5+). In previous studies, the change in the νs peak located near 1375-1385 cm-1 was attributed to adsorbed Lact-1 species forming specific (inner-sphere) chemical bonds with metal oxide surfaces or metal cations in aqueous solutions.3,7,42 In the present study, we also found similar shifts of aqueous Lact-1 ATR-FTIR peaks from 1417 to 1376 cm-1 upon the reaction of Lact-1 with hematite nanoparticles. This particular ATR-FTIR peak at 1376 cm-1 was not present in the aqueous Lact-1 spectrum (Figure 2). Furthermore, the strong similarity of this νs peak position in the spectrum of Lact-1 adsorbed on hematite surfaces with respect to the one in the spectrum of the aqueous Fe(III)-lactate complex suggests that some of the Lact-1 is coordinated to Fe(III) via inner-sphere bonding. Hence, we have assigned the peak at 1376 cm-1 to an inner-sphere surface species of Lact-1 bonded to Fe(III)O5OH octahedra. This similarity also suggests that the reactive sites at the hematite surfaces are likely to possess one exchangeable hydroxyl group or water molecule. 3.4. ATR-FTIR and Macroscopic Uptake Measurements of Adsorbed Lactate at Different Γlactate Values and 0.01 M NaCl. To confirm the attachment modes of Lact-1 surface complexes proposed above and relate them to the observed ATRFTIR peak positions and shapes, ATR-FTIR spectra of Lact-1 adsorbed on hematite nanoparticles as a function of total Lact-1 concentration were collected and examined. Macroscopic observations indicate that increasing the concentration of aqueous Lact-1 increases the extent of Lact-1 uptake by hematite nanoparticles during the 48 h of reaction (Figure 3). We did not conduct a kinetic study on the dissolution of hematite nanoparticle, and hence it is not possible to comment on lactate-promoted hematite dissolution rates. However, as shown in the inset of Figure 3, there is a strong correlation between the amount of Lact-1 adsorbed and the amount of Fe(III) released from the surfaces of hematite nanoparticles, with a steady-state concentration of Fe(III) in the aqueous solution at Γlactate G 1 to 2 µmol/m2. The relevant questions here are (1) what lactate surface coverage would constitute a monolayer on the hematite surfaces and (2) does the lactate coverage at which steady-state hematite dissolution occurs correspond to monolayer lactate coverage. To
6688 Langmuir, Vol. 24, No. 13, 2008
Ha et al.
Figure 4. (a) ATR-FTIR spectra of Lact-1 species at the hematite/water interface at 25 °C, pH 5.0, and 0.01 M NaCl for Lact-1 surface coverages of (i) 0.52, (ii) 1.21, (iii) 2.12, (iv) 3.52, and (v) 5.21 µmol/m2, and (vi) deprotonated aqueous lactate. Solid lines indicate the peak maxima of aqueous deprotonated lactate. Arrows indicate the peak maxima of adsorbed Lact-1. (b) Plot of ATR-FTIR spectra of Lact-1 species as in (a) in the wavenumber range of 1650-1350 cm-1 to indicate the progressive shift of νas and νs frequencies as a function of Γ˜ lactate.
estimate an effective monolayer coverage, the number of reactive surface sites on hematite particles must be estimated. Literature values for reactive surface sites on hematite particles range from 0.6 to 10 sites per nm2, and 4.5 reactive sites per nm2 of bulk hematite particles have been recently reported on the basis of potentiometric titrations.46,47 Using 4.5 sites/nm2 as the estimated value for reactive sites on the hematite nanoparticles, a Γlactate value corresponding to one effective monolayer of Lact-1 on hematite nanoparticles was calculated and found to be 2.8 µmol/ m2. This effective monolayer value is larger than that (∼2 µmol/ m2) at which the Fe(III) solution concentration reaches a steadystate value. On the basis of this reasoning, we infer that sites on hematite nanoparticle surfaces have different reactivities, with Lact-1 first binding and saturating surface sites on hematite nanoparticles that are most reactive with a concomitant release of Fe(III), followed by the binding of additional Lact-1 to less reactive sites that do not release Fe(III). No experimental study has been performed to determine the number of reactive sites on hematite nanoparticles to our knowledge, and hence it is difficult to relate the dissolution of Fe(III) to Γlactate values in a more quantitative fashion at present. However, recent crystal truncation rod diffraction studies of single-crystal hydrated hematite (0001) and (1-102) surfaces48,49 have revealed at least three different types of binding sites on these surfaces, which should have different binding affinities based on observed differences in coordination environments. As can be seen in Figure 4, a continuous broadening and shift of the νas peak to higher frequency were observed as Γlactate increased at pH 5.0 and 0.01 M NaCl. The aqueous Lact-1 peak maximum at 1575 cm-1 progressively shifted to 1591, 1591, 1592, 1595, and 1597 cm-1 at Lact-1 coverages of 0.52, 1.21, 2.12, 3.52, and 5.21 µmol/m2, respectively. To test for the possibility of experimental artifacts (e.g., instrumental drift) in obtaining the peak maximum for each ATR-FTIR spectrum, calibration of the instrument was performed using Lact-1 as the (46) Cromieres, L.; Moulin, V.; Fourest, B.; Giffaut, E. Colloid Surf., A 2002, 202, 101. (47) Jarlbring, M.; Gunneriusson, L.; Hussmann, B.; Forsling, W. J. Colloid Interface Sci. 2005, 285, 212. (48) Trainor, T. P.; Chaka, A. M.; Eng, P. J.; Newville, M.; Waychunas, G. A.; Catalano, J. G., Jr. Surf. Sci. 2004, 573, 204. (49) Tanwar, K. S.; Lo, C. S.; Eng, P. J.; Catalano, J. G.; Walko, D., Jr.; Waychunas, G. A.; Chaka, A. M.; Trainor, T. P. Surf. Sci. 2007, 601, 460–474.
Figure 5. Fit of ATR-FTIR spectra of the aqueous deprotonated lactate species, the aqueous Fe(III)-lactate complex, and the Lact-1 species at the hematite/water interface at 0.01 or 0.5 M NaCl for different Lact-1 surface coverages. Underlying lines represent components fit to the experimentally obtained spectra. The goodness-of-fit values are given in Tables S3 and S4 in Supporting Information.
standard calibrant solution. No changes in the ATR-FTIR spectrum for Lact-1 were observed over a period of 6 months, thus the observed peak shifts as a function of Lact-1 surface coverage are considered to be real. (See Figure S5 in Supporting Information.) Similar findings of peak broadening and shifts of νas as a function of surface coverage for other LMW organic ligands are reported and suggest that those ligands sorb to mineral surfaces as outer-sphere complexes.5,31,50,51 We also ascribe this progressive shift of νas to higher frequency and the broadening of this peak to an increasingly asymmetric environment (i.e., a (50) Nakamoto, K. Infrared and Raman Specra of Inorganic and Coordination Compounds; Wiley & Sons: New York, 1986. (51) Bargar, J. R., Jr.; Parks, G. A. Geochim. Cosmochim. Acta 1997, 61, 2639.
Adsorption of Organic Matter at Interfaces
Figure 6. Plot of the integrated area of different IR peaks at different surface coverages of Lact-1 on hematite nanoparticles at pH 5.0 and 0.01 M NaCl. Peaks centered around 1575 cm-1 represent an outer-sphere Lact-1 species (b), whereas peaks centered around 1376-1380 cm-1 represent inner-sphere Lact-1 species (9). Refer to the text for the assignment of peaks centered at 1593-1598 cm-1 (1).
broader distribution of hydrogen bond lengths) around Lact-1 as the surface coverage of Lact-1 increases. To verify spectral features attributed to inner-sphere versus outer-sphere complexes and to measure the fraction of different possible surface species responsible for the observed changes in νas peaks, ATR-FTIR spectra of Lact-1 sorption samples at different surface loadings were fit with individual Gaussian components. Although the 1575 cm-1 feature in the aqueous Lact-1 spectrum was successfully fit with a single component, it was impossible to fit the spectra of Lact-1 species adsorbed on hematite surfaces at different surface coverages without one additional component within the range of 1650 to 1550 cm-1 as shown in Figure 5. Including this additional component also resulted in a slightly better fit for the spectrum of the aqueous Fe(III)-lactate complex. The position of this additional peak was allowed to vary, and indeed its centroid shifted from 1593 to 1595 cm-1 or 1598 cm-1 depending on Lact-1 coverage (Table S3). In our fitting, we fixed the position of the peak centered at 1575 cm-1, which represents an outer-sphere Lact-1 surface complex that retains its hydration shell. Integrating the areas of the fitted peaks ranging in frequency from 1550 to 1650 cm-1 indicates that a decrease in area of the peak fixed at 1575 cm-1 occurs throughout the range of Lact-1 concentrations examined, suggesting a decrease in the number of outer-sphere Lact-1 complexes with increasing Lact-1 surface loading (Figure 6). In contrast, an increase in the area of the peak centered at 1593-1598 cm-1 was observed with increasing Lact-1 coverage. This change could result from the presence of the following species: (1) dissolved Fe(III)-lactate complexes, (2) nonadsorbed excess aqueous Lact-1, (3) a different type of outer-sphere complex (i.e., one in which at least one of the waters in the first hydration shell is also bonded to Fe(III) in the hematite surface), and/or (4) inner-sphere complexes of Lact-1 on hematite nanoparticles. The position and fwhm of the νas peak in the spectrum of the filtered supernatant at a surface coverage of 3.52 µmol/m2 (Figure 2(ii)) and the spectrum of aqueous Lact-1 (Figure 2(i)) were compared and were found to be very similar. The νas peak position in the spectrum of aqueous Lact-1 is centered at 1575 cm-1, and it is also located at 1575 cm-1 in the spectrum of the filtered supernatant solution. The fwhm of the νas peak for aqueous Lact-1 is 46 cm-1, whereas it is 42 cm-1 for the filtered supernatant
Langmuir, Vol. 24, No. 13, 2008 6689
solution. Therefore, quantifying a fraction of aqueous Lact-1 and Fe(III)-lactate aqueous complexes in the filtered supernatant solution is difficult on the basis of the observed νas peak in the ATR-FTIR spectrum. Strong similarities were also found between the spectrum of the filtered supernatant (Figure 2(ii)) and the spectrum of the aqueous Fe(III)-lactate complex (Figure 2(iii)) within the wavenumber range of 1420-1380 cm-1 where νs peak of Lact-1 occurs. This observation suggests that some of the dissolved Fe(III) in the filtered supernatant solution is indeed complexed with Lact-1. However, even if the dissolved Fe(III)-lactate species (species 1) in the supernatant solution were present within the hematite/water interfacial region, it is unlikely that the additional feature of the fitting peak centered at 1593-1598 cm-1 in the spectrum of Lact-1 adsorbed on hematite nanoparticles originates from the dissolved Fe(III)-lactate species. We draw this conclusion because the centroid and fwhm of the νas peak in the ATR-FTIR spectrum of the filtered supernatant solution, which would contain the dissolved Fe(III)-lactate species, and the spectrum of aqueous Lact-1 are very similar. It is also unlikely that the nonadsorbed excess aqueous Lact-1 species (species 2) accounts for this component at 1593-1598 cm-1 because the spectra in Figure 4 were obtained after the subtraction of the spectrum of the filtered supernatant for each of the wet paste samples. Unless such aqueous species were present within the electrical double layer at hematite/water interfaces in our systems in concentrations significant enough to be detected by ATR-FTIR spectroscopy, our subtraction procedure using the spectrum of the supernatant solution should have removed any spectral features due to dissolved aqueous Lact-1 or Fe(III)-lactate species present within the interfacial region. Therefore, it is likely that species 3 or 4 (or both) is present at the hematite nanoparticle/water interface and accounts for the additional spectral component at 1593-1597 cm-1. However, we suggest that different types of outer-sphere complexes (species 3) in these samples are unlikely because most of the lactate is deprotonated in solution at pH 5 and hence the dominant species involved in forming the outer-sphere complex via electrostatic interaction and/or hydrogen bonding should resemble that of aqueous deprotonated lactate with the complete hydration shell intact. Results from quantum chemical calculations also indicate that the calculated vibrational frequencies shift to higher wavenumber when deprotonated lactate bonds to Fe(III)O6 octahedra as innersphere complexes (Table 2). The calculated νas peaks for aqueous Lact-1 were at 1550 cm-1 whereas the νas peak for adsorbed Lact-1 occurs at 1591 and 1654 cm-1 for the S1 and SOH1 geometries, respectively. (See Figure 7 for details on the geometry of the surface structure.) This result further supports the observed increase in the area of the peak centered at 1593-1598 cm-1, and its shift to higher wavenumbers as Lact-1 surface coverage increases is likely due to an increase in the number of innersphere complexes. The position of νs for the aqueous Lact-1 complex is 1417 cm-1, whereas it ranges from 1376 to 1380 cm-1 for Lact-1 adsorbed on hematite nanoparticles with increasing Γlactate (Figure 4 and Table S3 in Supporting Information). It is likely that these observed changes in the spectra of Lact-1 adsorbed on hematite nanoparticles are due to the presence of the inner-sphere complex of Lact-1 as discussed in section 3.3. However, we cannot completely rule out the possibility of dissolved excess Fe(III)-lactate species within the hematite nanoparticle surface/ water interfacial region. Because of the strong similarities in the νs peak positions and shapes between the spectrum of aqueous
6690 Langmuir, Vol. 24, No. 13, 2008
Ha et al.
Table 2. Quantum Chemical Calculated Frequencies and Assigned Peaks of Adsorbed Lact-1 with Different Surface Complex Structuresa structures S1 S2 S3 SOH1 SOH2 SOH3 SS1 SSOH1 SS3
νas (calcd)
νs (calcd)
∆ν (calcd)
1591 1512 1516 1654 1577 1602 1555 1533 1722
1382 1426 1429 1454 1453 1258 1423 1386 1420
209 86 87 200 124 344 132 147 302
experimental value[µmol/m2]
νas
νs
∆ν
Γ ) 0.52, 1.21, 2.12 Γ ) 3.52 Γ ) 5.21
1591 1591 1595
1390 1382 1382
201 209 213
a Refer to Figure 7 for the specific geometries considered. All structures are calculated with B3LYP/6-311+G* under the specified hydrous environment with five water molecules placed in locations as listed in Supporting Information. Experimental values were obtained from the IR spectra of Lact-1 adsorbed on hematite surfaces at pH 5.0 and 0.01 M NaCl.
Fe(III)-lactate species and the spectra of Lact-1 adsorbed to hematite nanoparticles, distinguishing the dissolved Fe(III)-lactate species from the inner-sphere complex is difficult on the basis of the observed spectral features occurring in the ATR-FTIR spectrum range of 1376-1380 cm-1. Unlike fitting the νas feature as discussed above, the νs feature in the spectrum of aqueous Fe(III)-lactate and in the spectra of Lact-1 adsorbed onto hematite nanoparticle was successfully fit with a single component. Therefore, it was also not possible to differentiate or measure the fraction of the dissolved Fe(III)-lactate species within the interfacial region in a quantitative fashion. At present, we suggest that the changes in the spectral features in this ATR-FTIR spectral range (1375-1380 cm-1) are caused by the inner-sphere complex of Lact-1 because of the lack of definitive evidence of the presence of Fe(III)-lactate species. As shown in Figure 6, the integrated area under the νs peak increases as a function of Γlactate. A shift in the νs peak as a function of Γlactate from 1380 to 1376 cm-1 as Γlactate increases could be due to (1) slight changes in the coordination environment of Lact-1 bound to Fe(III) cations because the metal oxide substrate undergoes surface modification due to increased ligand-promoted dissolution of hematite as Lact-1 surface coverage increases and/or (2) cooperative effects from the neighboring adsorbed Lact-1 molecules, leading to alterations of the relative strengths and frequencies of the νs vibrational mode. We propose that Lact-1 is present as both outer-sphere and inner-sphere complexes when reacted with hematite nanoparticles under the experimental conditions of this study. Recent ATRFTIR spectroscopic studies have also revealed that carboxylic ligands can be simultaneously present in the form of outer-sphere and inner-sphere complexes on Al oxides and Al oxyhydroxides.4,52 We propose that spectral features in the range of 1575-1600 cm-1 are due to a combination of outer-sphere and inner-sphere Lact-1 species, whereas spectral features in the range of 1376-1380 cm-1 are due to inner-sphere Lact-1 species based on peak shift comparisons with the ATR-FTIR spectrum of aqueous Fe(III)-lactate complexes and results reported by others.3,4,10,42 3.5. Quantum Chemical Calculations of Adsorbed Lactate. To further examine the geometry or mode of binding of innersphere Lact-1 to hematite nanoparticle surfaces, we employed a (52) Boily, J. F.; Fein, J. B. Chem. Geol. 2000, 168, 239.
quantum chemical approach for adsorbed Lact-1 in various coordination geometries. As described in section 3.2, a B3LYP/ 6-31+G* level of theory, coupled with the addition of water molecules to explicitly model solvation effects, was used in these calculations. Figure 7 presents nine different optimized coordination geometries considered in the present study, and Table 2 shows the corresponding calculated vibrational frequencies for each structure. (See Supporting Information for details on coordination geometries.) As shown in Table 2, the frequency calculations using the optimized structure with the hydroxyl group pointing away from the Fe(III) moiety (S1) most closely match experimental observations. In contrast, the other models, including the monodentate structure with the hydroxyl groups facing toward the Fe(III) and the bidentate structure with a five-membered ring, show large deviations from experimental frequency values regardless of the coordination environment of Fe(III). Awatani et al.3 suggested that Lact-1 binds dominantly in a mononuclear bidentate structure on TiO2 surfaces involving the hydroxyl and carboxylate groups. Considering the significant differences in electronegativity (χ) values of Ti4+ versus Fe3+ [χ(Ti4+) ) 1.730 and χ(Fe3+) ) 1.556],53 it is possible that the mode of attachment of Lact-1 could differ depending on the type of metal oxide surface. On the basis of our results, we suggest that inner-sphere Lact-1 adsorbed on hematite nanoparticles occurs dominantly as monodentate mononuclear complexes with the hydroxyl functional group pointing away from the Fe(III) center. We have tested the effect of only one hydroxyl molecule in the first coordination shell of Fe(III) in octahedral coordination and the effect of a Fe(III) dimer structure, so we cannot generalize the effects of the Fe(III) coordination environment on the calculated vibrational frequencies of adsorbed Lact-1. According to Chen et al.,53 it is possible that the surface Fe sites on hematite nanoparticles are distorted from the octahedral configuration to undercoordinated Td or square-pyramidal configurations. However, the crystal truncation rod diffraction study of hydrated hematite (0001) and (1-102) surfaces of bulk crystals48,49 showed that the coordination shells of surface Fe3+ ions were composed of the full complement of six oxygens or OH- groups. Additional quantum chemical studies on more realistic systems (including possible undercoordinated iron sites on the hematite surface, larger cluster sizes representing the hydrated hematite surface, and an adequate representation of water in the electrical double layer) are needed. 3.6. ATR-FTIR and Macroscopic Measurements of Adsorbed Lactate at Different Γlactate Values and 0.5 M NaCl. To further confirm the structure of the Lact-1 surface complex proposed above and relate it to the observed ATR-FTIR peaks, the ATR-FTIR spectra of Lact-1 adsorbed on hematite at a different electrolyte concentration were collected and examined. Indirect conclusions about the types of surface complexes formed by anions or cations (i.e., inner sphere vs outer sphere) are often based on changes in adsorption edges as a function of ionic strength. For example, a significant ionic strength effect is typically observed when outer-sphere-bonded surface complexes are the dominant species, yielding distinctively separate adsorption edges.55,56 In contrast, the adsorption edge for inner-sphere complexes does not usually vary as a function of ionic strength. (53) Li, K.; Xue, D. J. Phys. Chem. A 2006, 110, 11332. (54) Chen, L. X.; Liu, T.; Thurnauer, M. C.; Csencsits, R.; Rajh, T. J. Phys. Chem. B 2002, 106, 8539. (55) Suarez, D. L.; Goldberg, S.; Su, C. In EValuation of Oxyanion Adsorption Mechanisms on Oxides Using FTIR Spectroscopy and Electrophoretic Mobility; Sparks, D. L., Grundl, T. J., Eds.; American Chemical Society Symposium Series: Washington, DC, 1998; Vol. 715, p 136.
Adsorption of Organic Matter at Interfaces
Langmuir, Vol. 24, No. 13, 2008 6691
Figure 7. Schematic drawing of possible inner-sphere coordination structures for Lact-1 adsorbed on: (a) a monomer cluster unit of [(Fe3+)(H2O)6]+3, (b) a monomer cluster unit of [(Fe3+)(OH)(H2O)5]+3, and (c) a dimeric cluster unit of [(Fe3+)2(OH)2(H2O)8]+3 or [(Fe+3)2(OH)3(H2O)7]+3. Refer to the text for specific details on each model structure.
However, such a generalization must be made with caution because adsorption edge data alone cannot distinguish between different sorption modes quantitatively.57 Lact-1 uptake measurements show that Lact-1 adsorption on hematite nanoparticle surfaces does not have a significant ionic strength dependence (Figure 3), suggesting that inner-sphere complexes may be the dominant adsorption mode under these experimental conditions. In addition, the ATR-FTIR spectra of Lact-1 adsorbed on hematite nanoparticles at 0.5 M NaCl did not differ from the ones at 0.01 M NaCl, suggesting that the surface complex structures of Lact-1 are similar for these two background (56) Huang, Q.; Zhao, Z.; Chen, W. Chemosphere 2003, 52, 571. (57) Sposito, G. In Geochemical Processes at Mineral Surfaces; Davis, J. A., Hayes, K. F., Eds.; American Chemical Society Symposium Series: Washington, DC, 1986; p 217.
electrolyte concentrations (Figure 5). A similar spectral fitting procedure was performed as described in section 3.3. (See Table S4 for fitting results in Supporting Information.) Integrating the areas of the νas peaks centered at 1575 cm-1 and the νs peaks centered in the range of 1380-1376 cm-1 indicates that increasing the ionic strength of the solution from 0.01 to 0.5 M resulted in a decreased area for the νas peak but an increased area for the νs peak for Γlactate e 2.25 µmol/m2 (Figure 8). This analysis further confirms that the peak at 1575 cm-1 represents outersphere Lact-1 surface complexes, whereas the peak in the range of 1380-1376 cm-1 indicates inner-sphere Lact-1 surface complexes on the hematite nanoparticles. The increase in the peak area for the νs peak is likely due to the increased amount of inner-sphere complex, possibly caused by the conversion of
6692 Langmuir, Vol. 24, No. 13, 2008
Ha et al. Lact-1
surface coverage. Increasing the ionic strength of the solution should compress the electrical double layer, and hence we speculate that this change should facilitate the exchange of one or several of the waters of hydration surrounding Lact-1 with proton donors at the hematite surfaces, leading to the formation of inner-sphere Lact-1 complexes. However, because the molecular geometry and attachment mode responsible for the ATRFTIR peak in the range of 1593-1597 cm-1 is not yet definitively understood on the basis of quantum chemical calculations, this assignment should be considered tentative.
4. Conclusions The present study has shown that Lact-1 adsorbs to hematite nanoparticle surfaces over a wide range of Lact-1 concentrations at pH 5.0. We found no significant dependence of the extent of Lact-1 adsorption on background electrolyte concentration, suggesting that the dominant adsorption mode for Lact-1 is inner sphere under these conditions. However, the ATR-FTIR spectroscopic data and the results from fitting the spectra reveal that both inner-sphere and outer-sphere Lact-1 surface complexes are simultaneously present, with both occurring as major species. Further work on Lact-1 adsorption on hematite nanoparticles, such as adsorption experiments as a function of solution pH at various Γlactate values, is needed to make a quantitative assessment of the amounts of each type of lactate species present within the metal oxide/water interface. On the basis of spectral fitting, an increase in inner-sphere surface complexes and a decrease in outer-sphere surface complexes of Lact-1 were observed with increasing Γlactate at pH 5.0 and 0.01 M NaCl. Because of a strong similarity between the spectrum of dissolved Fe(III)-lactate species and that of Lact-1 adsorbed to hematite nanoparticles in the wavenumber range of 1376-1380 cm-1, it was difficult to quantify the amount of dissolved Fe(III)-lactate species within the hematite nanoparticle surfaces/water interfacial region. However, the presence of Fe(III)-lactate species in this region was confirmed by spectral features in the wavenumber range of 1420-1380 cm-1. The attachment mode of inner-sphere Lact-1 species was predicted using quantum chemical geometry optimization and IR vibrational frequency calculations. A model for adsorbed Lact-1 consisting of a mononuclear monodentate structure with the hydroxyl functional group of Lact-1 pointing away from the Fe(III) moiety provides the best agreement between calculated and experimentally observed vibrational frequencies. Finally, the molecular-level information presented above provides a basis for the development of surface complexation models of Lact-1 adsorption on hematite nanoparticles at pH 5.0 and ionic strength ranging from 0.01 to 0.5 M NaCl. Figure 8. Comparison of integrated areas of different IR peaks at different surface coverages of Lact-1 on hematite nanoparticles at 0.5 and 0.01 M NaCl and at pH 5.0: (a) νs peaks centered at 1376-1382 cm-1 representing an inner-sphere Lact-1 species. (b) νas peaks centered at 1575 cm-1 representing an outer-sphere Lact-1 species;. (c) Refer to the text for the assignment of peaks centered at 1593-1598 cm-1.
a fraction of the outer-sphere complexes to inner-sphere complexes with increasing background electrolyte concentration. Interpreting the change in peak areas in the range of 1593-1598 cm-1 is not straightforward because no specific trend was observed in peak area at an ionic strength of 0.5 M NaCl as a function of
Acknowledgment. We thank Guangchao Li for help in performing ICP measurements. This work was supported by the NSF-NIRT (grant BES-0404400) and the NSF-EMSI (grant CHE0431425 s Stanford Environmental Molecular Science Institute). Supporting Information Available: Details of the sample preparations, ATR-FTIR spectra baseline correction, calibration procedures, quantum chemical modeling, and the coordinates of water molecules. This material is available free of charge via the Internet at http://pubs.acs.org. LA800122V
