DFT Studies of Solvation Effects on the Nanosize Bare, Thiolated, and Redox Active Ligated Au 55 Cluster † Ganga Periyasamy, † Engin Durgun, ‡ Jean-Yves Raty, ‡ and F. Remacle* ,† Department of Chemistry, Bat. B6c, UniVersity of Lie `ge, B4000 Liege, Belgium, and Department of Physics, Bat. B5, UniVersity of Lie `ge, B4000 Liege, Belgium ReceiVed: December 18, 2009; ReVised Manuscript ReceiVed: March 11, 2010 The structural and electronic properties of the bare Au 55 cluster, and of the model thiol passivated Au 55 (SCH 3 ) 42 , the redox active ligated Au 55 S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy · 2Cl and Au 55 (SCH 3 ) 41 (S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy) · 2Cl (bpy ) N-methyl-4,4′-bipyridinium) complexes are studied at the DFT level in the gas phase and with an explicit water layer. For all complexes, neutral, positive, and negative charge states are investigated. The thiol ligation distorts the outer layer of the approximate icosahedral geometry of the bare cluster and induces a charge transfer from the gold core to the ligand shell. The anchoring of a single redox active ligand on the bare Au 55 leads to the formation of a cavity around the S-Au bond. We show that this cavity formation is prevented by the thiol ligands in Au 55 (SCH 3 ) 41 (S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy) · 2Cl. The vertical addition of one electron to the [Au 55 S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy · 2Cl] 0 cluster is followed by a charge transfer from the Au 55 core to the bpy 2+ ligand, which is accompanied by a mechanical motion of the redox active bpy arm driven by electrostatic interactions. The presence of a solvent shell does not alter the structure but significantly decreases the computed charging energies of the clusters, making them comparable with experimental values. The computed redox potential differences are in good agreement with the experimental values. Introduction Gold clusters exhibit a rich array of interesting electronic, optical, chemical, and catalytic properties, typically with a sharp size threshold, that has sparked a huge interest in several areas. For example, gold particles are catalytically active when their diameter falls below 2 nm and completely inactive above. Among small size gold nanoparticles (NP), Au 55 clusters possess an ideal size (1.4 nm) for catalytic activity, a full shell geometry, and oxidation-resistant properties. 1-5 Gold Au 55 nanoclusters are not stable in the bare form and need to be stabilized by ligand layers such as in Au 55 (PPh 3 ) 12 - Cl 6 , 6 Au 55 (BSH) 12 2+ , 7 Au 55 (SC n ) 32 clusters (chain length n ) 12, 18), 8 Au 55 (T 8 -O-SS-SH) 12 C l6 , 9 and Au 55 (Ph 2 PC 6 H 4 SO 3 Na) 12 - Cl 6 , 6 etc. They can self-assemble into one-, two-, and three- dimensional arrays, 4,10-12 and because of their large charging energies (0.3-0.5 eV) they can operate as single electron devices at room temperature, 1 which makes them suitable for applica- tions in nanoelectronics. 4 In addition, one can also anchor redox active molecules in the ligand shell to increase their functionalities. 13-15 When used in catalysis or electrocatalysis, as chemical sensors or electronic components, these hybrid nanoclusters are often surrounded by solvent molecules. 16,17 The water layer as well as the chemical nature of the ligands were shown to strongly affect their electronic properties. 3,15,18,19 We report on a systematic computational study of these effects at the DFT level of theory. We first investigate the structural and electronic properties of the bare Au 55 cluster in neutral, positive, and negative charge states and the effect of an explicit water layer of 54 solvent molecules. We then explore the effect of a densely packed outer layer of passivating SCH 3 ligands, with and without a water layer, on the three charge states. We finally replace one of the SCH 3 by the redox active S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy 2+ .2Cl and compare its properties with those of a Au 55 cluster bonded to a single S(CH 2 ) 2 CO 2 - (CH 2 ) 10 bpy 2+ · 2Cl, to quantify the stabilization effect of the thiol layer on the geometry and to estimate how this layer affects the redox properties of the complex. The properties of the two complexes, Au 55 (SCH 3 ) 41 (S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy 2+ ) · 2Cl and Au 55 S(CH 2 ) 2 CO 2 (CH 2 ) 10 bpy 2+ · 2Cl were also studied in the presence of a partial water layer of 21 molecules. Our choice of Au 55 (SCH 3 ) 42 that has not been observed experimentally is motivated by the fact that the thiol layer is densely packed and is a good model of the densely packed short self-assembled monolayer that was used in the CV measurements of the redox properties of the bpy ligand by Willner et al. 13 We also report on computations Au 55 (SCH 3 ) 32 , a model for the Au 55 (SC 18 ) 32 cluster experimentally identified 8 where the dilution of the ligand shell leads to the formation of Au-S-Au bridges. The formation of such bridges has been reported at the experimental and computational levels for other cluster sizes, for example, Au 102 (SR) 44 20,21 and Au 25 (SR) 18 -22-24 for which crystallographic data are available, and on the EI-MS resolved Au 38 (RS) 24 25-29 and Au 144 (RS) 60 . 30,31 On the theoretical side, there is an intense research effort to characterize the properties of gold clusters of various sizes, ligand shells, and charge states using both quantum chemistry and solid-state DFT methods 12,21,23,28,32-39 (for recent reviews, see refs 29, 40, and 41 and references therein). For large cluster sizes, larger than a few dozen atoms, parameter-free electronic- structure methods cannot be applied to perform an unbiased structure optimization because of the large computational resources needed. Instead one has to resort to approximate methods. It turned out that in the case of gold clusters the resulting structures depend on the chosen approximation in a † Part of the special issue “Protected Metallic Clusters, Quantum Wells and Metallic Nanocrystal Molecules”. * To whom correspondence should be addressed, fremacle@ulg.ac.be. † Department of Chemistry, Bat. B6c. ‡ Department of Physics, Bat. B5. J. Phys. Chem. C 2010, 114, 15941–15950 15941 10.1021/jp9119827  2010 American Chemical Society Published on Web 03/30/2010