Kinetics and Mechanism of Birnessite Reduction by Catechol

and hydroxides containing Mn(III) and Mn(IV) can oxidize organic ligands more rapidly than O2. Aromatic The complex interactions of oxidizable organic ligands with soil and nonaromatic ligands have been reported to dissolve Mn(III,IV) (hydr)oxide minerals have received little study by in situ spectroscopic techniques. We used a combination of an in situ electron reductively Mn(III,IV) (hydr)oxides (Stone and Morparamagnetic resonance stopped-flow (EPR-SF) spectroscopic techgan, 1984a,b) with the formation of polymeric reaction nique and stirred-batch studies to measure the reductive dissolution products that resemble soil humic substances (Shindo kinetics of birnessite (d-MnO2), a common Mn mineral in soils, by and Huang, 1982, 1984). Sorbed contaminants can be catechol (1,2-dihydroxybenzene). The reaction was rapid, indepenreleased into solution following reductive dissolution of dent of pH, and essentially complete within seconds under conditions solid Mn(III,IV) (hydr)oxides by organic ligands and of excess catechol at pH 4 to 6. The overall empirical second-order reactive compounds may be sorbed or coprecipitated rate equation describing the reductive dissolution rate was d[Mn(II)]/ as a consequence of hydrolysis of the Mn(II) that is dt 5 k[CAT]1.0[SA]1.0 where k 5 4 (60.5) (1023 L m22 s21 and [CAT] released (Xyla et al., 1992; Godtfredsen and Stone, 1994; and [SA] are the initial concentrations in molarity and meters square per liter. In the process, catechol was oxidized to the two-electron oLarsen and Postma, 1997). Thus, redox cycling of Mn quinone product. The energy of activation (Ea) for the reaction was is dynamic and coupled to geochemical cycling of other 59 (67) kJ mol21 and the activation entropy (S‡) was 278 6 22 J mol21 metals, C turnover in soils, and N transformations in K21, suggesting that the reaction was surface-chemical controlled and soils and marine sediments (Hem, 1978; Bartlett, 1981; occurs by an associative mechanism. Rates of catechol disappearance Wang and Huang, 1987, 1992; Laha and Luthy, 1990; from solution with simultaneous Mn(II) and o-quinone production Luther et al., 1997; Naidja et al., 1998). were comparable. These data strongly suggest that precursor surfacePivotal early investigations into oxidizable organic complex formation is rate-limiting and that electron transfer is rapid. ligand reactions with solid Mn(III,IV) (hydr)oxides posThe rapid reductive dissolution of birnessite by catechol has significant implications for C and Mn cycling in soils and the availability of Mn tulated that the formation of a precursor surface comto plants. plex is requisite to electron transfer and that innersphere electron transfer or reactions involving surface species were rate-limiting (Stone and Morgan, 1984a,b; B of Mn in soils and geochemical sediments Xyla et al., 1992). No direct measurements of reductant is generally assumed to be mediated by redox reacsorption can be made in most reductive dissolution studtions. Data based on frontier molecular orbital theory ies because of the rapid reduction of Mn(III,IV) (hy(FMOT) considerations (Luther, 1990) show that oxides dr)oxides; however, dissolution rates are directly proportional to the concentration of the surface complex C.J. Matocha, Dep. of Agronomy, Univ. of Kentucky, N-122 Ag. Sci. (Xyla et al., 1992). The only direct line of evidence in Ctr-North, Lexington, KY 40546-0091; D.L. Sparks, Dep. of Plant support of an inner-sphere mechanism was the study and Soil Sciences, 147 Townsend Hall, Univ. of Delaware, Newark, conducted by Gordon and Taube (1962) using 18O-laDE 19717-1303; J.E. Amonette and R.K. Kukkadapu, William R.Wiley Environmental Molecular Sciences Lab. (EMSL), Pacific Northbeled water. These authors found that the UO21 2 prowest National Lab., Richland, WA 99352. Received 20 Aug. 1999. duced by the reaction of MnO2 with U(IV), derived *Corresponding author ([email protected]). both oxygen atoms from MnO2. A previous EPR spectroscopic measurements showed Published in Soil Sci. Soc. Am. J. 65:58–66 (2001). MATOCHA ET AL.: BIRNESSITE REDUCTION BY CATECHOL 59 Table 1. Selected physical and chemical properties of birnessite. that soluble Mn(II) release from birnessite after reaction with catechol was rapid (McBride, 1989a,b). CateProperty Birnessite chol dissolved MnO2 more rapidly than hydroquinone Total Mn, % 49.1 6 0.1† because it is a bidentate ligand able to chelate surface Total K, % 16.0 6 0.1 Total H2O, % 12.4 Mn metal centers (Stone and Morgan, 1984a,b; Mean Oxidation State 3.4 6 0.1 McBride, 1989a,b). However, few rate coefficients deBET Surface Area, m2 g21 40.0 PZC 1.81 6 0.04 scribing Mn(II) release were measured in these systems. P2O 7 Mn(III), % 10.0 The disequilibrium in aqueous redox reactions comXRD, nm 0.732, 0.362, 0.246, 0.142 monly observed in the field (Lindberg and Runnels, † Standard error of the mean. 1984) underscores the need for kinetic studies (Sparks, 1989; Bartlett and James, 1993; Stone et al., 1994). dried. Total concentrations of Mn and K were determined by Recent investigations have demonstrated that dydissolving a known solid weight with 12 M HCl and acidified namic redox reactivity of Mn(III,IV) (hydr)oxide minNH2OH·HCl. Powder X-ray diffraction (XRD) analysis reeral surfaces catalyzes abiotic degradation of organic vealed that diagnostic d-spacings for birnessite agreed with pollutants (Ukrainczyk and McBride, 1993a,b; Dec and published values (Table 1). Point of zero charge (PZC) was Bollag, 1994; Pizzigallo et al., 1995; Cheney et al., 1996). estimated for birnessite by microelectrophoretic mobility meaIn fact, oxidation of catechol to CO2 by birnessite has surements as a function of pH in 1 mM NaCl and 3.2 m L2 been reported and could contribute to significant abiotic surface area concentration (Murray, 1974). Increased negative degradation of organic contaminants in natural settings electrophoretic mobilities were observed for unreacted birnessite as pH increased and extrapolation to zero revealed a PZC (Wang and Huang, 1992; Naidja et al., 1998). This findof 1.81 6 0.04 (Table 1). These results are consistent with ing is surprising because degradation of organic pollutthose reported in the literature for birnessite (Balistrieri and ants has generally been attributed to biological activity. Murray, 1982; Murray, 1974). Oxidation state was measured From this discussion, it is evident that oxidizable oriodometrically by a starch end point and standardized thiosulganic ligand interactions with Mn(III,IV) (hydr)oxide fate solution (Murray et al., 1984; Amonette et al., 1994). minerals in soils are complex. Soil Mn oxides are the Pyrophosphate (P2O 7 )-extractable Mn(III) was measured by principal source of plant available Mn(II) and, thus, reacting large excesses of P2O 7 (50 mM) with birnessite pathways of reduction influence plant toxicity and dis(|2.7 mM MnT) at pH 4 and 238C or 338C by published proceease resistance (Schulze et al., 1995a). Better underdures (Diebler and Sutin, 1964; Davies, 1969; Kostka et al., 1995). High resolution thermogravimetric analysis (HRTGA, standing of the chemistry of these interactions improves TA Instruments) was used to measure both structural and our ability to make sound predictions about contamiadsorbed H2O contents. High-resolution ramp mode at a heatnant fate, cleanup, and plant available Mn(II). Most of ing rate of 108C min2 under flowing N2 was used with Pt the available reductive dissolution kinetics data have crucibles to measure total H2O content, which is operationally been macroscopic measurements (Stone et al., 1994) defined as the weight loss after heating to 3008C (Moore et with few in situ spectroscopic studies. al., 1990). To understand adequately the complex reactivity of Birnessite was further characterized by x-ray absorption Mn(III,IV) (hydr)oxides with organic ligand reductants spectra recorded at beamline X-11A at the National Synchroin soils, however, it is best to begin with simpler model tron Light Source, Brookhaven National Laboratory, Upton, NY. Beam energy was calibrated to the K-absorption edge systems and combine in situ molecular-level spectroof Mn metal foil (6539 eV). The spectra were collected in scopic techniques with macroscopic investigations to fluorescence mode with a Lytle detector and compared with provide mechanistic kinetic information (Sparks, 1995). a standard Mn(IV) dioxide phase pyrolusite (b-MnO2). CorunIn this regard, catechol is a suitable model organic ligand dum (a-Al2O3) was used as a diluent to minimize self-absorpbecause ortho-type semiquinones have been identified tion effects (Schulze et al., 1995b). Comparison of pre-edge in soil humic substances (Steelink, 1964), catechols are XANES spectra of birnessite to reference pyrolusite (iodomecommon intermediates in pesticide degradation pathtric oxidation state of 3.92) showed agreement with published ways (Alexander, 1977), and they have been identified pre-edge energies reported by Manceau et al. (1992). Recent on siderophore molecules (Hersman et al., 1995). Birnstructural refinements of this mineral phase have shown that structural Mn(III) is present in birnessite, often substituted essite is a suitable model mineral because it is one of for Mn(IV) in the lattice (Drits et al., 1997; Silvester et al., the most commonly identified Mn oxide minerals in 1997). External surface area was measured by the BET procesoils and geochemical environments (McKenzie, 1989). dure using N2 (g) as the adsorbate (Brunauer et al., 1938). Our approach to the problem, then, involved a study of the kinetics and mechanism of reductive dissolution Stirred-Batch Experiments of well-characterized birnessite by catechol using in situ EPR-SF and stirred-batch techniques. Stirred-batch reductive dissolution rate experiments were conducted by a pH-stat (60.1 pH units) technique while purgMATERIALS AND METHODS ing with purified N2 at 10, 15, 18, and 238C. Temperature was controlled with a circulating water bath. Birnessite mineral Birnessite Preparation and Characterization suspensions were dispersed by sonification and pretreated at the experimental pH value and ionic strength prior to catechol Birnessite was synthesized according to procedures outlined by McKenzie (1971) by reduction of boiling KMnO4 with addition. Reactions were initiated by addition of a weighed aliquot of catechol stock solution buffered to the desired pH concentrated HCl. The precipitate was vacuum filtered, dialyzed against deionized water to remove salts, and freezewith dilute 0.1 M NaOH or 0.1 M HCl. Catechol stock solutions 60 SOIL SCI. SOC. AM. J., VOL. 65, JANUARY–FEBRUARY 2001 were freshly prepared prior to each experimental run. Sample flat aqueous cell located in the EPR cavity. Before starting the reaction, the magnetic field was swept to ensure no residual aliquots were removed at increasing time intervals and passed through a 0.2-mm pore-size membrane filter into a tared plastic Mn(II) remained in the flat cell between experimental runs. Kinetic studies were conducted under pseudo-first-order tube and acid-quenched. Soluble Mn in the filtrates was measured with flame atomic absorption spectrometry (AAS). The conditions with either catechol or birnessite in excess. Catechol concentrations ranged from 200 to 10 000 mM and birnesUV-VIS spectra of the filtrates in 1-cm cuvettes were recorded to follow formation of products during the reaction sequence site suspension surface area concentrations ranged from 0.45 to 4.5 m L2. Total Mn concentrations as birnessite (100–500 with an HP 8452A Diode Array UV-VIS spectrophotometer. Literature values for the o-quinone monomer were taken from mM) when in deficit (catechol in excess) reflect typical environmental concentrations of Mn. Experiments were conMentasti et al. (1975) as 1460 M2 cm2 at 390 nm. For catechol, Beer’s law was obeyed at concentrations #0.5 mM cateducted at initial pH values of 4, 4.5, 5, and 6 as maintained by a 50 mM K-acetate buffer. Additional experiments indichol at 276 nm and yielded ε 5 2380 M2 cm2. At longer times (.2 min), competing reactions involving the unstable cated that K-acetate buffer did not significantly influence Mn(II) release rates when compared with the pH-stat method. o-quinone monomer with water and unreacted catechol are operative (Dawson and Nelson, 1938; Mason, 1949). CorrecAdjustments of the buffer to the desired pH were made with aliquots of concentrated HCl and pH was measured with a tions were made for removal of suspension volume from the batch reactor during sampling in calculating acid consumption. model 25 pH/ion meter (Fisher Scientific Co., Pittsburgh, PA). Small drifts in pH values (60.05) during the reactions were Mn(II) that was added concurrently with catechol at the beginning of the kinetic run had a negligible influence on the overall due to the finite capacity of the K-acetate buffer. The low ionic strength coupled with the dilute solids concentration was reductive dissolution rate. No significant differences were observed between the presence and absence of a N2 purge. sufficient to keep birnessite dispersed during the time period of the experimental runs. In addition, birnessite suspensions EPR-SF Kinetic Studies Reductive dissolution of birnessite by catechol was followed in situ by an EPR-SF technique (Klimes et al., 1980; Stach et al., 1985; Fendorf et al., 1993) to measure Mn(II) release and detect possible semiquinone intermediates. The six-line diagnostic EPR spectra confirmed that birnessite was reduced to Mn(II) during the reaction with catechol (Fig. 1a). The hyperfine spectrum of high spin Mn(II) is due to coupling of the electron spin (S 5 5/2) to the nuclear spin (I 5 5/2) of the Mn ion. Spectra were recorded at room temperature (238 6 0.58C) and a microwave frequency of 9.55 GHz (X-band) with a Bruker ESP 300E spectrometer. Mixed samples were injected by the stopped-flow (SF) unit (Update Instruments, Inc., Madison, WI) into a flow-through quartz flat aqueous cell (Wilmad Glass Co., Buena, NJ) inserted in a TE102 resonator. A microwave power attenuation of 30 dB and modulation amplitude of 102 T were employed and this power rating was checked to ensure that no saturation effects occurred for the highest Mn(II) concentration (Luca and Cardile, 1989). When the entire six-line spectrum for Mn(II) was desired, field-sweep mode was used (0.05 T). Data processing was done with BRUKER software (WINEPR). Stopped-flow kinetic measurements were made with ms time resolution in time-sweep mode by centering on the increase in intensity of the fourth downfield resonance peak (Ho in Fig. 1a) at 0.3435 T (g 5 1.98) during the reaction sequence. This peak is the most suitable one for quantitative measurements of Mn(II) (Carpenter, 1983). This approach has been used by others in EPR-SF studies (Fendorf et al., 1993). A parallel experiment centering on the “valley” at 0.346 T (H1 in Fig. 1a) showed a decrease in intensity with time which confirmed that H0 intensities corresponded to release of Mn(II) following reductive dissolution of birnessite. Signal intensities were converted to concentration using MnCl2 standards prepared in the same suspension matrix as the samples. The same MnCl2 standards were analyzed with flame AAS to intercorrelate the methods. The EPR technique yielded a linFig. 1. (A) Representative room temperature EPR spectra of 100 ear response to Mn(II) between 5 and 200 mM (Fig. 1b). mM MnT as birnessite (0.45 m2 L21) reacted with catechol (5 3 The SF system consisted of a syringe ram (model 1019), 1023 M ) depicting the characteristic six-line spectra of the Mn(II) ram controller (model 715), 2-mL syringes, and a Wiskindproduct. H0 and H1 indicate the peak and valley used to quantify grid mixing cell. The mixing cell had a dead volume of 1.6 L. Mn(II) concentrations. (B) Intercalibration standard curve relating Details of this system can be found elsewhere (Hubbell et al., EPR signal intensity at the 0.3435 T (H0) peak to flame AAS Mn(II) concentration. 1987). The mixing cell was attached by tubing to the quartz MATOCHA ET AL.: BIRNESSITE REDUCTION BY CATECHOL 61 were sonified prior to filling the inport syringes. Sweep times and c are the reaction orders with respect to the given were set to either 20 or 83s, depending on the experimental reactants. The initial rate approach (Lasaga, 1981) was conditions, with a temporal resolution of 4096 time steps. employed to determine the empirical rate equation. The The data collection software was configured to automatically value for d[Mn(II)]/dt was taken from the initial linear trigger a series of field-sweeps of the reacted suspensions slopes when ,20% of the reaction had occurred. immediately after time-sweep mode kinetic measurements to An initial rate plot of log [Mn(II)] release versus log ensure that there were no shifts in the resonance peak and to [CAT] was linear (r 2 5 0.95) with a slope (a) of 1.06, assist in quantifying final Mn(II) concentrations. The method indicating a first order dependence of the reaction rate of initial rates (Lasaga, 1981) was used to ensure that the on [CAT] at pH 4 (Fig. 4a). The pseudo-first-order rate forward reaction predominated and competing reactions with catechol oxidation products were not operative (Stone, 1987; coefficient, k9, was shown from the y-intercept to be Sparks, 1995). 5 3 1023 s21. By considering the nonvaried reactant, the rate constant kgraph was calculated by the following relationship RESULTS AND DISCUSSION k9 5 kgraph [SA]b [3] EPR-SF and Stirred-Batch Kinetic Studies resulting in a value of 5.55 3 1023 L m22 s21. This value Reductive dissolution of birnessite (E8 5 1.29 V, agreed with the average kcalc determined from the empirBricker, 1965) by catechol is favored thermodynamically ical second-order equation and relevant data in Table between pH 4 (DGr 5 292 kJ mol21) and pH 6 (DGr 5 2 (average kcalc 5 4 3 1023 L m22 s21 for [CAT] of 1023269 kJ mol21) for [Mn21], [CAT], and [o-quinone] of 1022 M, [SA] 5 0.9 m2 L21, [H] 5 1024 M). Similiarly, 1, 1000, and 1 mM. On the basis of the predicted stoichithe initial rate plot for [SA] was linear (r 2 5 0.95) with ometry, slightly more protons were consumed per mole a first-order dependence on [SA] (Fig. 4b). The kgraph Mn(II) released (Fig. 2): value obtained from the y-intercept of 2.6 3 1023 L m22 d-MnO1.7(s) 1 0.7C6H4(OH)2 1 2H → Mn(II)(aq) s21 was in good agreement with the average kcalc of 2.6 3 1023 L m22 s21 (Table 2). The initial rate of Mn(II) 1 0.7C6H4O2(aq) 1 1.7H2O [1] release from birnessite was virtually independent of pH and the appearance of the two-electron o-quinone prodwith a slope near 0 in the pH range 4 to 6 (Fig. 4c). uct (see discussion below). Representative EPR-SF kiApparently, the [H] term did not directly participate netics of the reductive dissolution of birnessite by catein the rate equation (see Eq. [2]) despite the observed chol revealed that the reaction was rapid and essentially consumption of H in Eq. [1]. In their studies with complete in ,10 s when catechol is in excess (Fig. 3). feitknechtite, Stone and Morgan (1984a) reported a On the basis of a previous study with feitknechtite first-order dependence on both hydroquinone and sus(Stone and Morgan, 1984a), the rate of Mn(II) producpension loading, but a reaction order of 0.46 for [H]. tion as a function of birnessite concentration, catechol The overall empirical second-order rate equation deconcentration, and pH, can be described by scribing the reductive dissolution of birnessite by catechol can be written as d[Mn(II)]/dt 5 kobs[CAT][SA][H] [2] d[Mn(II)]/dt 5 k[CAT]1.0 [SA]1.0 [4] where d[Mn(II)]/dt is the rate of Mn(II) production (M s21), [CAT] is the initial concentration of catechol (M), where k 5 4 (6 0.5) 3 1023 L m22 s21 and [CAT] and [SA] is the surface area concentration of the birnessite [SA] are the initial concentrations in M and m2 L21. The suspension (m2 L21), [H] is the hydrogen ion concentrareasonable agreement between the apparent secondtion (M), kobs is the apparent rate constant, and a, b, Fig. 3. Representative EPR-SF kinetics of reductive dissolution of Fig. 2. Typical stoichiometry of H consumed versus dissolved Mn(II) birnessite by catechol at pH 4 as a function of [CAT] conducted at 238C and 0.45 m2 L21 [SA]. produced during birnessite reduction by catechol. 62 SOIL SCI. SOC. AM. J., VOL. 65, JANUARY–FEBRUARY 2001 Table 2. Representative experimental conditions used to measure EPR-SF kinetic data. [CAT] pH (60.05) [SA] kcalc 31022 M m2 L21 31023 L m22 s21 1.00 4.0 0.91 5.3 6 0.5† 0.50 4.0 0.91 3.2 0.25 4.0 0.91 3.4 0.1 4.0 0.91 4.3 6 0.1 1.00 6.0 0.91 4.3 6 0.3 0.50 6.0 0.91 5.6 0.25 6.0 0.91 4.4 0.1 6.0 0.91 4.5 6 0.1 0.50 4.0 0.45 2.2 6 0.2 0.50 4.0 1.42 2.8 0.50 4.0 1.78 2.2 6 0.2 0.50 4.0 2.31 2.6 † Standard error of the mean. (Stone and Morgan, 1984a; Stone, 1987; Ukrainczyk and McBride, 1992; Xyla et al., 1992) that changes in pH might cause changes in the reductive dissolution rate of solid Mn-(III,IV) minerals by oxidizable organic ligands through (i) readsorption of Mn(II) at higher pH; (ii) proton-promoted dissolution at lower pH; (iii) a change in the thermodynamic driving force for the reaction; (iv) increased sorption of catechol on the pH-dependent charged surface of birnessite; and (v) speciation of catechol due to dissociation of the phenolic hydroxyl groups. The general lack of a pH effect may be explained by a combination of the low point of zero charge of the birnessite surface with the following pKs (Balistrieri and Murray, 1982): .MnIVOH1 2 ↔ .MnIVOH 1 H, pKa1 5 2.3 [5] .MnIVOH ↔ .MnIVO2 1 H, pKa2 5 3.3 [6] and the low fraction of dissociated catechol (0.02%) calculated at the highest experimental pH (pH 5 6) on the basis of the pKa1 5 9.7 and pKa2 5 13.7. The protonated form, the dominant catechol species under our experimental conditions, is generally considered less reactive toward oxidation than the deprotonated catecholate (McBride et al., 1988). These findings imply that the sorption affinity of birnessite for catechol to form a precursor surface complex should remain constant over the pH range employed. Stability constants for sorption of organic ligands on metal (hydr)oxide surfaces are often proportional to stability constants for the metal-ligand complex in solution (Kummert and Stumm, 1980). Tris-MnIV(cat)22 3 and bis-MnIII(cat)2 2 complexes form in solution but are better described as Mn(semi)2(cat) and MnII(semi)(cat)2 because of internal metal-ligand redox reactions that convert catechol ligands to semiquinone ligands (Richert et al., 1988). The catechol to Mn(IV) ligand-to-metalcharge-transfer (LCMT) band was reported at 585 nm Fig. 4. Initial rate of dissolved Mn(II) release measured by the EPR(Hartman et al., 1984) and the catechol to Mn(III) SF technique as a function of: (A) [CAT] at pH 4 and 0.90 m2 L21 LCMT band between 550 and 750 nm (Magers et al., [SA]; (B) [SA] at pH 4 and 5 (1023 M [CAT]; and (C) [H] at 1978). The chemical structure of catechol sorbed onto 0.90 m2 L21 [SA]. The dotted lines represent the 95% confidence Al oxide and TiO2 has been proposed to involve 1:1 interval bands. (bidentate surface chelate complex) and 2:1 metal:ligand (binuclear) ratios (Kummert and Stumm, 1980; order rate constants kgraph and kcalc provides further support for the proposed rate equation. McBride and Wesselink, 1988; Rodriguez et al., 1996) and may resemble the surface complex between cateIt was anticipated on the basis of previous reports MATOCHA ET AL.: BIRNESSITE REDUCTION BY CATECHOL 63 Table 3. Activation parameters describing the reductive dissoluchol and surface Mn in this study. Five-membered ring tion of birnessite by catechol. chelates are favored in coordination for steric and enParameter Birnessite/catechol system tropic reasons (Shriver et al., 1994). The “bite” (O-O) distance in the catecholate ligand of 2.77 Å may be Ea†, kJ mol21 58.7 6 6.6‡ ln A 21.1 6 2.7 compatible with surface Mn-Mn distances (assuming a DH‡, kJ mol21 56.2 6 6.6 110 cleavage plane) in birnessite of 2.85 Å, suggesting DS‡, J mol21 K21 277.9 6 22.4 that catechol may also form a binuclear precursor surDG‡, kJ mol21 32.9 6 13.3 face complex. The ability of catechol to bind two surface † Ea (energy of activation), ln A (pre-exponential factor), DH‡ (enthalpy Mn metal centers may explain the observed rapid reducof activation), DS‡ (entrophy of activation), and DG‡ (free energy of activation) were derived from Arrhenius and Eyring plots. tive dissolution rate. It should be pointed out that there ‡ Standard error of the mean. is no spectroscopic evidence confirming the structure of the catechol-birnessite precursor surface complex, parameters (Table 3) derived by Arrhenius and Eyring but past infrared spectroscopic studies have shown that plots (Fig. 6). The activation energy (Ea) of 59 (67) kJ it is difficult to identify (McBride, 1987). Currently, mol21 suggests that the reaction was surface-chemical work in our laboratory is being conducted to observe controlled (Lasaga, 1981; Sparks, 1989, 1995) and agrees the LCMT bands on the reacted birnessite using diffuse with previous reductive dissolution studies (Stone and reflectance spectroscopy. Despite unfavorable sorption Morgan, 1984a). Other investigators have proposed that interactions between uncharged catechol (pKa1 5 9.7, electron transfer was the rate-limiting step and have pKa2 5 13.7) and the highly negatively charged birnessite treated precursor complex formation as a preequilibsurface (PZC |2), in the pH range of most soils, catechol rium step (Stone and Morgan, 1984a; Stone and Morgan, is extremely efficient at solubilizing Mn. 1987). The high positive value for DH‡ (56 6 7 kJ mol21) Catechol underwent oxidative transformation during and negative value for DS‡ (278 6 22 J mol21 K21) reaction with an excess of birnessite (4.5 m2 L21) at pH suggests the possibility of a bimolecular step, which is 4 to form o-quinone with concomitant release of Mn(II) (Fig. 5). Near complete mass balance indicates that the amount of catechol that reacted with birnessite can be accounted for by the production of the o-quinone monomer, the oxidation product expected to form via a twoelectron transfer. First order rate plots (data not shown) were linear for catechol disappearance from solution (r 2 . 0.99) and yielded a pseudo-first order rate coefficient kc for catechol disappearance of 0.70 (60.18) min21. This compares well with the kinetics of o-quinone and Mn(II) appearance (ko 5 0.61 min21 and km 5 0.55 min21) indicating that electron transfer was rapid. These results strongly suggest that catechol sorption to form a precursor surface complex was rate-limiting and product sorption was not affecting the reaction rate. The suggestion that precursor surface complex formation was rate-limiting is supported by the activation Fig. 5. Kinetics of catechol oxidation to o-quinone by birnessite and Fig. 6. Activation parameters derived from (A) Arrhenius and (B) dissolved Mn(II) release at 238C. Experimental conditions were: [CAT] 5 0.2 mM, [SA] 5 4.5 m2 L([MnT] 5 1 mM ), pH 4, I 5 Eyring plots describing the reductive dissolution of birnessite by catechol. 0.01 M NaCl. 64 SOIL SCI. SOC. AM. J., VOL. 65, JANUARY–FEBRUARY 2001 common to second-order reactions (Espenson, 1995). of s symmetry to the empty eg (s*) orbitals of the Mn(IV) metal centers. The Mn-catechol bond could be The observed second-order rate constant measured from EPR-SF studies indicated a more rapid reaction formed with electron transfer before the Mn-O bond cleaves, which is characteristic of associative mechathan would be expected from the high activation energy. However, the high activation energy can be compennisms (Shriver et al., 1994). Frontier-molecular-orbitaltheory (FMOT) predicts a large driving force for elecsated for by the large negative DS‡ value to yield a fast reaction (Espenson, 1995). tron transfer because of the vacant s* orbitals in Mn(IV) (Luther, 1990; Luther et al., 1998). On the basis of our data, no distinction can be made between two one-elecProposed Reaction Mechanism tron transfers or a single concerted two-electron transfer These results provide further support for the general nor whether the electrons are transferred as a hydrogen reaction mechanism initially proposed by Stone and atom (H•) or hydride (H-) ion. The rapid formation of Morgan (1984a) for reductive dissolution of Mn oxides. o-quinone, however, would suggest a concerted twoFormally, the reaction scheme includes the following electron transfer step. Perez-Benito et al. (1996) resteps. ported a possible direct two-electron transfer from oxalate to Mn(IV) in soluble MnO2. Further experiments Precursor Surface Complex Formation are being conducted to substantiate the proposed reaction mechanism. C6H4(OH)2 (aq) 1 .d-MnO2 (s) ↔ k1