Origin of bulklike structure and bond length disorder of Pt37 and Pt6Ru31 clusters on carbon: comparison of theory and experiment

We describe a theoretical analysis of the structures of self-organizing nanoparticles formed by Pt and Ru-Pt on carbon support. The calculations provide insights into the nature of these metal particle systems-ones of current interest for use as the electrocatalytic materials of direct oxidation fuel cells-and clarify complex behaviors noted in earlier experimental studies. With clusters deposited via metallo-organic Pt or PtRu(5) complexes, previous experiments [Nashner et al. J. Am. Chem. Soc. 1997, 119, 7760; Nashner et al. J. Am. Chem. Soc. 1998, 120, 8093; Frenkel et al. J. Phys. Chem. B 2001, 105, 12689] showed that the Pt and Pt-Ru based clusters are formed with fcc(111)-stacked cuboctahedral geometry and essentially bulklike metal-metal bond lengths, even for the smallest (few atom) nanoparticles for which the average coordination number is much smaller than that in the bulk, and that Pt in bimetallic [PtRu(5)] clusters segregates to the ambient surface of the supported nanoparticles. We explain these observations and characterize the cluster structures and bond length distributions using density functional theory calculations with graphite as a model for the support. The present study reveals the origin of the observed metal-metal bond length disorder, distinctively different for each system, and demonstrates the profound consequences that result from the cluster/carbon-support interactions and their key role in the structure and electronic properties of supported metallic nanoparticles.     

On the antioxidant mechanism of curcumin: classical methods are needed to determine antioxidant mechanism and activity

[reaction: see structure] The antioxidant activity of curcumin (1, 7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione) was determined by inhibition of controlled initiation of styrene oxidation. Synthetic nonphenolic curcuminoids exhibited no antioxidant activity; therefore, curcumin is a classical phenolic chain-breaking antioxidant, donating H atoms from the phenolic groups not the CH(2) group as has been suggested (Jovanovic et al. J. Am. Chem. Soc. 1999, 121, 9677). The antioxidant activities of o-methoxyphenols are decreased in hydrogen bond accepting media.     

Energy distributions of gallium nanoclusters

Starting with the heat-capacity data of Breaux et al., [J. Am. Chem. Soc. 126, 8629 (2004)] we use the maximum-entropy method to calculate energy distribution functions for gallium-ion nanoclusters over a wide temperature range (100-1050 K). Specifically, we calculate energy distributions for clusters containing n = 39 and n = 45 gallium atoms. For the case of n = 39 clusters the energy distribution gets systematically broader as a function of temperature with no indication of any marked structural change in the cluster. On the other hand, the energy distribution for the n = 45 cluster first gets broader as a function of temperature but then gets narrower again as the temperature is further increased, indicating that there is some kind of structural transition taking place in this cluster species.     

Interfacial tension model for catalytically driven nanorods

We present an analysis of the interfacial tension model for the movement of the catalytically driven nanorod. The model considers the convective reaction-diffusion equation for the production and diffusion of oxygen around the bimetallic nanorod. We solve the equation and find the concentration difference, which drives the nanorod. We use our expression to calculate the force on the nanorod and find that the result is within 20% of the results found earlier [W. Paxton et al., J. Am. Chem. Soc. 128, 14881 (2006)] by an approximate method. Unlike the earlier results, our results are valid from short to long lengths of the nanorod.     

Density functional analysis of the structural evolution of Gan (n=30-55) clusters and its influence on the melting characteristics

Recent experimental results have reported surprising variations in the shapes of the heat capacity curves and melting temperatures of gallium clusters in the size range of 30-55 atoms [G. A. Breaux et al., J. Am. Chem. Soc. 126, 8628 (2004)]. In the present work, we have carried out an extensive density functional investigation on ten selected clusters in the above mentioned size range. In particular, we have analyzed the ground state geometry and the nature of bonding in these clusters using electron localization function. We demonstrate that the existence or otherwise of a large island of atoms bonded with similar strength (i.e., the local order) in the ground state geometry is responsible for the variation in the shape of the heat capacity curve. We attribute the observed higher melting temperatures of some of the clusters (viz., Ga45-Ga48) to the presence of a distinct core and strong covalent bonds between the core and surface atoms. The present work clearly demonstrates that it is possible to understand the general trends observed in the heat capacity curves across the entire series on the basis of the analysis of their ground state.     

Electronic structure, excited states, and photoelectron spectra of uranium, thorium, and zirconium bis(Ketimido) complexes (C5R5)2M[-NCPh2]2 (M = Th, U, Zr; R = H, CH3)

Organometallic actinide bis(ketimide) complexes (C5Me5)2An[-N=C(Ph)(R)]2 (where R = Ph, Me, and CH2Ph) of thorium(IV) and uranium(IV) have recently been synthesized that exhibit chemical, structural, and spectroscopic (UV-Visible, resonance-enhanced Raman) evidence for unusual actinide-ligand bonding. [Da Re et al., J. Am. Chem. Soc., 2005, 127, 682; Jantunen et al., Organometallics, 2004, 23, 4682; Morris et al., Organometallics, 2004, 23, 5142.] Similar evidence has been observed for the group 4 analogue (C5H5)2Zr[-N=CPh2]2. [Da Re et al., J. Am. Chem. Soc., 2005, 127, 682.] These compounds have important implications for the development of new heavy-element systems that possess novel electronic and magnetic properties. Here, we have investigated M-ketimido bonding (M = Th, U, Zr), as well as the spectroscopic properties of the highly colored bis-ketimido complexes, using density functional theory (DFT). Photoelectron spectroscopy (PES) has been used to experimentally elucidate the ground-state electronic structure of the thorium and uranium systems. Careful examination of the ground-state electronic structure, as well as a detailed modeling of the photoelectron spectra, reveals similar bonding interactions between the thorium and uranium compounds. Using time-dependent DFT (TDDFT), we have assigned the bands in the previously reported UV-Visible spectra for (C5Me5)2Th[-N=CPh2]2, (C5Me5)2U[-N=CPh2]2, and (C5H5)2Zr[-N=CPh2]2. The low-energy transitions are attributed to ligand-localized N p --> C=N pi excitations. These excited states may be either localized on a single ketimido unit or may be of the ligand-ligand charge-transfer type. Higher-energy transitions are cyclopentadienyl pi --> CN pi or cyclopentadienyl pi --> phenyl pi in character. The lowest-energy excitation in the (C5Me5)2U[-N=Ph2]2 compound is attributed to f-f and metal-ligand charge-transfer transitions that are not available in the thorium and zirconium analogues. Geometry optimization and vibrational analysis of the lowest-energy triplet state of the zirconium and thorium compounds also aids in the assignment and understanding of the resonance-enhanced Raman data that has recently been reported. [Da Re et al., J. Am. Chem. Soc., 2005, 127, 682.].     

Modeling the stereoselectivity of the Johnson-Claisen rearrangements in the Danishefsky synthesis of gelsemine

[reaction: see text] The stereochemistries of [3,3] sigmatropic Johnson-Claisen (J-C) rearrangements of six intermediates studied in the synthesis of gelsemine were modeled using DFT methodology. The possible origins of the rearrangement stereoselectivity are explored and compared with the experimentally suggested rationalizations by Danishefsky et al. (J. Am. Chem. Soc. 2002, 124, 9812-9824). In the intermediate used for the J-C rearrangement in the Danishefsky synthesis (3), the closure is inhibited by the repulsive interactions between the enolate terminus and the carbon atoms of the double bond as well as with the hydrogen on C7. The closure is favored by stabilizing interactions between the enolate terminus and the H's of the oxetane ring.     

Voltammetric surface dealloying of Pt bimetallic nanoparticles: an experimental and DFT computational analysis

Voltammetric dealloying of bimetallic platinum-copper (Pt-Cu) alloys has been shown to be an effective strategy to modify the surface electrocatalytic reactivity of Pt bimetallic nanoparticles (S. Koh and P. Strasser, J. Am. Chem. Soc., 2007, 129, 12624). Using cyclic voltammetry and structural XRD studies, we systematically characterize the Pt-Cu precursor compounds as well as the early stages of the selective Cu surface dissolution (dealloying) process for Pt(25)Cu(75), Pt(50)Cu(50), and Pt(75)Cu(25) alloy nanoparticles annealed at both low and high temperature. We also assess the impact of the synthesis conditions on the electrocatalytic reactivity for the oxygen reduction reaction (ORR). To gain atomistic insight into the observed voltammetric profiles, we compare our experimental results with periodic DFT calculations of trends in the thermodynamics of surface Cu dissolution potentials from highly stepped and kinked Pt(854) single crystal surfaces. The modeling suggests a dependence of the electrochemical Cu dissolution potentials on the detailed atomic environment (coordination number, nature of coordinating atoms) of the bimetallic Pt-Cu surfaces. The DFT-predicted shifts in electrochemical Cu dissolution potentials are shown to qualitatively account for the observed voltammetric profiles during Cu dealloying. Our study suggests that metal-specific energetics have to be taken into account to explain the detailed dealloying behavior of bimetallic surfaces.     

Hybrid one-electron/many-electron methods for ionized states of molecular clusters

To describe singly-ionized states of molecular clusters we devised an effective Hamiltonian approach that combines (1) accurate monomer ionization potentials from many-electron wave functions with (2) polarization shifts and (3) effective monomer couplings obtained from a simple one-electron approach (the superposition-of-fragment-states (SFS) method [Valeev et al., J. Am. Chem. Soc., 2006, 128, 9882]). The accuracy of the intermolecular coupling parameters evaluated with SFS Hartree-Fock (HF) and Density-Functional-Theory (DFT) variants was evaluated for several weakly-bound dimers and compared against the state-of-the-art equation-of-motion ionization-potential coupled-cluster singles and doubles (EOM-IP-CCSD) data of Krylov and co-workers. The SFS-HF method produces coupling integrals accurate to a few percent, whereas SFS-DFT predictions are substantially worse. A hybrid approach combining SFS-HF couplings and shifts with EOM-IP-CCSD ionization potentials of monomers (denoted as SFS-EOM-IP-CCSD) was applied to ionized states of two conformers of a benzene dimer and ten representative DNA base pairs. The 16 considered SFS-EOM-IP-CCSD ionization potentials of the benzene dimer differed from the reference EOM-IP-CCSD IPs of Krylov and co-workers [Pieniazek et al., J. Chem. Phys. 2007, 127, 044317; Bravaya et al., Phys. Chem. Chem. Phys. 2010, 12, 2261] by less than 0.1 eV on average, and at most by 0.2 eV. For the DNA base pairs the mean absolute (median) deviation of the SFS-EOM-IP-CCSD IPs was 0.27 (0.23) eV; several deviations for non-Koopmans states were as large as 0.9 eV. The SFS-EOM-IP-CCSD method can be readily applied to large molecular clusters with computational effort scaling cubically with the size of the cluster.     

Variability of the 15N chemical shielding tensors in the B3 domain of protein G from 15N relaxation measurements at several fields. Implications for backbone order parameters

We applied a combination of 15N relaxation and CSA/dipolar cross-correlation measurements at five magnetic fields (9.4, 11.7, 14.1, 16.4, and 18.8 T) to determine the 15N chemical shielding tensors for backbone amides in protein G in solution. The data were analyzed using various model-independent approaches and those based on Lipari-Szabo approximation, all of them yielding similar results. The results indicate a range of site-specific values of the anisotropy (CSA) and orientation of the 15N chemical shielding tensor, similar to those in ubiquitin (Fushman, et al. J. Am. Chem. Soc. 1998, 120, 10947; J. Am. Chem. Soc. 1999, 121, 8577). Assuming a Gaussian distribution of the 15N CSA values, the mean anisotropy is -173.9 to -177.2 ppm (for 1.02 A NH bond length) and the site-to-site CSA variability is +/-17.6 to +/-21.4 ppm, depending on the method used. This CSA variability is significantly larger than derived previously for ribonuclease H (Kroenke, et al. J. Am. Chem. Soc. 1999, 121, 10119) or recently, using "meta-analysis" for ubiquitin (Damberg, et al. J. Am. Chem. Soc. 2005, 127, 1995). Standard interpretation of 15N relaxation studies of backbone dynamics in proteins involves an a priori assumption of a uniform 15N CSA. We show that this assumption leads to a significant discrepancy between the order parameters obtained at different fields. Using the site-specific CSAs obtained from our study removes this discrepancy and allows simultaneous fit of relaxation data at all five fields to Lipari-Szabo spectral densities. These findings emphasize the necessity of taking into account the variability of 15N CSA for accurate analysis of protein dynamics from 15N relaxation measurements.     

Can ferric-superoxide act as a potential oxidant in P450(cam)? QM/MM investigation of hydroxylation, epoxidation, and sulfoxidation

In view of recent reports of high reactivity of ferric-superoxide species in heme and nonheme systems (Morokuma et al. J. Am. Chem. Soc. 2010, 132, 11993-12005; Que et al. Inorg. Chem. 2010, 49, 3618-3628; Nam et al. J. Am. Chem. Soc. 2010, 132, 5958-5959; J. Am. Chem. Soc. 2010, 132, 10668-10670), we use herein combined quantum mechanics/molecular mechanics (QM/MM) methods to explore the potential reactivity of P450(cam) ferric-superoxide toward hydroxylation, epoxidation, and sulfoxidation. The calculations demonstrate that P450 ferric-superoxide is a sluggish oxidant compared with the high-valent oxoiron porphyrin cation-radical species. As such, unlike heme enzymes with a histidine axial ligand, the P450 superoxo species does not function as an oxidant in P450(cam). The origin of this different behavior of the superoxo species of P450 vis-a?-vis other heme enzymes like tryptophan 2, 3-dioxygenase (TDO) is traced to the ability of the latter superoxo species to make a stronger FeOO-X (X = H,C) bond and to stabilize the corresponding bond-activation transition states by resonance with charge-transfer configurations. By contrast, the negatively charged thiolate ligand in the P450 superoxo species minimizes the mixing of charge transfer configurations in the transition state and raises the reaction barrier. However, as we demonstrate, an external electric field oriented along the Fe-O axis with a direction pointing from Fe toward O will quench Cpd I formation by slowing the reduction of ferric-superoxide and will simultaneously lower the barriers for oxidation by the latter species, thereby enabling observation of superoxo chemistry in P450. Other options for nascent superoxo reactivity in P450 are discussed.     

Limited flexibility of lactose detected from residual dipolar couplings using molecular dynamics simulations and steric alignment methods

The conformational flexibility of lactose in solution has been investigated by residual dipolar couplings (RDCs). One-bond carbon-proton and proton-proton coupling constants have been measured in two oriented media and interpreted in combination with molecular dynamics simulations (MD). Two different approaches, known as PALES (Zweckstetter et al., J. Am. Chem. Soc. 2000, 122, 3791-3792) and TRAMITE (Azurmendi et al., J. Am. Chem. Soc. 2002, 124, 2426-2427), have been used to determine the alignment tensor from a shape-induced alignment model with the oriented medium. The steric alignment of the structures from several MD trajectories has provided ensemble averaged RDCs that have been compared with the experimental ones. The obtained results reveal the almost exclusive presence of a major low energy region defined as syn-phi/syn-psi (> 97%), for which sampling occurs in a dynamic manner. This result satisfactorily agrees with that determined by standard NOE-based methods.     

Large-scale synthesis of thiolated Au25 clusters via ligand exchange reactions of phosphine-stabilized Au11 clusters

Phosphine-stabilized Au11 clusters in chloroform were reacted with glutathione (GSH) in water under a nitrogen atmosphere. The resulting Au:SG clusters exhibit an optical absorption spectrum similar to that of Au25(SG)18, which was isolated as one of the major products from chemically prepared Au:SG clusters (Negishi, Y. et al. J. Am. Chem. Soc. 2005, 127, 5261). Rigorous characterization by optical spectroscopy, electrospray ionization mass spectrometry, and polyacrylamide gel electrophoresis confirms that the Au25(SG)18 clusters were selectively obtained on the sub-100 mg scale by ligand exchange reaction under aerobic conditions. The ligand exchange strategy offers a practical and convenient method of synthesizing thiolated Au25 clusters on a large scale.     

Potential energy surface of 1,6-Methano

The potential energy surface of 1,6-methano[10]annulen-11-ylidene and its isomers has been investigated by density functional (BLYP and B3LYP) molecular orbital methods. These calculations indicate the lowest energy annulene structure to be 56.9 kcal mol(-)(1) higher in energy than triplet 1-naphthylcarbene. These calculations, together with calculations on transition states connecting possible rearrangement products derived from this carbene, indicate that the trapping products reported by Carlton et al. J. Am. Chem. Soc. 1976, 98, 6068-6070 arise from rearrangement of the annulene-carbene to a tricyclic isomer.     

Competing mechanisms in the optically activated functionalization of the hydrogen-terminated Si(111) surface

Recent experiments have shown that organic monolayers on silicon surfaces can be formed through the optically activated surface reaction of H-terminated Si surfaces with terminally unsaturated organic molecules (Eves et al. J. Am. Chem. Soc. 2004, 126, 14318; Sun et al. J. Am. Chem. Soc. 2005, 127, 2514). Possible mechanisms for the formation of this monolayer involve the abstraction of a H atom either at the same attachment site of the molecule (Path A) or from a neighboring site (Path B). Using periodic Density Functional Theory calculations together with an efficient method for finding reaction pathways, we examine both optically activated reaction mechanisms for an alkene and an aldehyde reacting with H-Si(111). Our results show that while Path A is energetically more favorable its significant barrier is likely to limit its viability. Path B on the other hand encounters a much lower H atom abstraction barrier and appears to be more viable.     

H2 storage in isostructural UiO-67 and UiO-66 MOFs

The recently discovered UiO-66/67/68 class of isostructural metallorganic frameworks (MOFs) [J. H. Cavka et al. J. Am. Chem. Soc., 2008, 130, 13850] has attracted great interest because of its remarkable stability at high temperatures, high pressures and in the presence of different solvents, acids and bases [L. Valenzano et al. Chem. Mater., 2011, 23, 1700]. UiO-66 is obtained by connecting Zr(6)O(4)(OH)(4) inorganic cornerstones with 1,4-benzene-dicarboxylate (BDC) as linker resulting in a cubic MOF, which has already been successfully reproduced in several laboratories. Here we report the first complete structural, vibrational and electronic characterization of the isostructural UiO-67 material, obtained using the longer 4,4'-biphenyl-dicarboxylate (BPDC) linker, by combining laboratory XRPD, Zr K-edge EXAFS, TGA, FTIR, and UV-Vis studies. Comparison between experimental and periodic calculations performed at the B3LYP level of theory allows a full understanding of the structural, vibrational and electronic properties of the material. Both materials have been tested for molecular hydrogen storage at high pressures and at liquid nitrogen temperature. In this regard, the use of a longer ligand has a double benefit: (i) it reduces the density of the material and (ii) it increases the Langmuir surface area from 1281 to 2483 m(2) g(-1) and the micropore volume from 0.43 to 0.85 cm(3) g(-1). As a consequence, the H(2) uptake at 38 bar and 77 K increases from 2.4 mass% for UiO-66 up to 4.6 mass% for the new UiO-67 material. This value is among the highest values reported so far but is lower than those reported for MIL-101, IRMOF-20 and MOF-177 under similar pressure and temperature conditions (6.1, 6.2 and 7.0 mass%, respectively) [A. G. Wong-Foy et al. J. Am. Chem. Soc., 2006, 128, 3494; M. Dinca and J. R. Long. Angew. Chem., Int. Ed., 2008, 47, 6766]. Nevertheless the remarkable chemical and thermal stability of UiO-67 and the absence of Cr in its structure would make this material competitive.     

First-principles investigation of finite-temperature behavior in small sodium clusters

A systematic and detailed investigation of the finite-temperature behavior of small sodium clusters, Na(n), in the size range of n=8-50 are carried out. The simulations are performed using density-functional molecular dynamics with ultrasoft pseudopotentials. A number of thermodynamic indicators such as specific heat, caloric curve, root-mean-square bond-length fluctuation, deviation energy, etc., are calculated for each of the clusters. Size dependence of these indicators reveals several interesting features. The smallest clusters with n=8 and 10 do not show any signature of melting transition. With the increase in size, broad peak in the specific heat is developed, which alternately for larger clusters evolves into a sharper one, indicating a solidlike to liquidlike transition. The melting temperatures show an irregular pattern similar to the experimentally observed one for larger clusters [Schmidt et al., Nature (London) 393, 238 (1998)]. The present calculations also reveal a remarkable size-sensitive effect in the size range of n=40-55. While Na(40) and Na(55) show well-developed peaks in the specific-heat curve, Na(50) cluster exhibits a rather broad peak, indicating a poorly defined melting transition. Such a feature has been experimentally observed for gallium and aluminum clusters [Breaux et al., J. Am. Chem. Soc. 126, 8628 (2004); Breaux et al., Phys. Rev. Lett. 94, 173401 (2005)].     

Analysis of the conserved P9-G10.1 metal-binding motif in hammerhead ribozymes with an extra nucleotide inserted between A9 and G10.1 residues

Hammerhead ribozymes (Rz) have catalytically important tandem G:A pairs in the core region, and we recently demonstrated that the P9-G10.1 motif (a sheared-type G:A pair with a guanine residue on the 3' side of the adenine residue) with several flanking base pairs is sufficient for capture of divalent cations, such as Mg(2+) and Cd(2+) ions that are important to maintain full activities (Tanaka et al. J. Am. Chem. Soc. 2002, 124, 4595-4601; Tanaka et al. J. Am. Chem. Soc. 2004, 126, 744-752). We also found that mutant hammerhead ribozymes that have an additional G residue inserted between A9 and G10.1 residues (the metal-binding P9-G10.1 motif) have significant catalytic activities. In this study, we demonstrate that the hammerhead ribozymes are capable of maintaining the catalytically competent structure even when the tandem, sheared-type G:A pairs were perturbed by an insertion of an additional nucleotide, whereas the chirality of the phosphorothioate at the P9 position significantly influenced the enzymatic activity for both the natural and G-inserted ribozymes.     

Controlled excitations of the Belousov-Zhabotinsky reaction: Experimental procedures

The purpose of this research was to explore the unstirred, ferroin-catalyzed Belousov-Zhabotinsky (BZ) reaction as an experimental model for the response of excitable media to small perturbations (slightly larger than the threshold for excitations). Following Showalter et al. (Showalter, K.; Noyes, R. M.; Turner, H. J.Am. Chem. Soc. 1979, 101, 7463-69), we used a positively biased silver electrode to release silver ions into a BZ reaction mixture, removing bromide ions and causing an excitation if sufficient bromide was removed. We found (1) a scaling region in which the delay before activation increased linearly as the size of the perturbation decreased, qualitatively consistent with but not fully explained by the Oregonator of Field et al. (Field, R. J.; K?r?s, E.; Noyes, R. M. J. Am. Chem. Soc. 1972, 94, 8649-64); (2) evidence for a 10 s oligomerization time scale; and (3) that activations were always delayed until after the end of a pulse of current, with the delay essentially constant for sufficiently long pulses, an effect not seen in simple ODE models but consistent with the anomalously large current apparently required for activation (Showalter, K.; Noyes, R. M. J. Am. Chem. Soc. 1976, 98, 3730-31) and explainable by bromide transport. Overall, the BZ system appeared to be well-suited as an experimental prototype, despite its complexity.     

Electronic structure of the Mn4OxCa cluster in the S0 and S2 states of the oxygen-evolving complex of photosystem II based on pulse 55Mn-ENDOR and EPR spectroscopy

The heart of the oxygen-evolving complex (OEC) of photosystem II is a Mn4OxCa cluster that cycles through five different oxidation states (S0 to S4) during the light-driven water-splitting reaction cycle. In this study we interpret the recently obtained 55Mn hyperfine coupling constants of the S0 and S2 states of the OEC [Kulik et al. J. Am. Chem. Soc. 2005, 127, 2392-2393] on the basis of Y-shaped spin-coupling schemes with up to four nonzero exchange coupling constants, J. This analysis rules out the presence of one or more Mn(II) ions in S0 in methanol (3%) containing samples and thereby establishes that the oxidation states of the manganese ions in S0 and S2 are, at 4 K, Mn4(III, III, III, IV) and Mn4(III, IV, IV, IV), respectively. By applying a "structure filter" that is based on the recently reported single-crystal EXAFS data on the Mn4OxCa cluster [Yano et al. Science 2006, 314, 821-825] we (i) show that this new structural model is fully consistent with EPR and 55Mn-ENDOR data, (ii) assign the Mn oxidation states to the individual Mn ions, and (iii) propose that the known shortening of one 2.85 A Mn-Mn distance in S0 to 2.75 A in S1 [Robblee et al. J. Am. Chem. Soc. 2002, 124, 7459-7471] corresponds to a deprotonation of a mu-hydroxo bridge between MnA and MnB, i.e., between the outer Mn and its neighboring Mn of the mu3-oxo bridged moiety of the cluster. We summarize our results in a molecular model for the S0 --> S1 and S1 --> S2 transitions.