Nickel(II)-molybdenum(III)-cyanide clusters: synthesis and magnetic behavior of species incorporating [(Me(3)tacn)Mo(CN)(3)

The substitution of Mo(III) for Cr(III) in metal-cyanide clusters is demonstrated as an effective means of increasing the strength of the magnetic exchange coupling and introducing magnetic anisotropy. Synthesis of the octahedral complex [(Me(3)tacn)Mo(CN)(3)] (Me(3)tacn = N,N',N"-trimethyl-1,4,7-triazacyclononane) is accomplished with the addition of precisely 3 equiv of LiCN to a solution of [(Me(3)tacn)Mo(CF(3)SO(3))(3)] in DMF. An excess of LiCN prompts formation of a seven-coordinate complex, [(Me(3)tacn)Mo(CN)(4)](1)(-), whereas less LiCN produces multinuclear species such as [(Me(3)tacn)(2)Mo(2)(CN)(5)](1+). In close parallel to reactions previously performed with [(Me(3)tacn)Cr(CN)(3)], assembly reactions between [(Me(3)tacn)Mo(CN)(3)] and [Ni(H(2)O)(6)](2+) or [(cyclam)Ni(H(2)O)(2)](2+) (cyclam = 1,4,8,11-tetraazacyclotetradecane) afford face-centered cubic [(Me(3)tacn)(8)Mo(8)Ni(6)(CN)(24)](12+) and linear [(Me(3)tacn)(2)(cyclam)NiMo(2)(CN)(6)](2+) clusters, respectively. Generation of the former involves a thermally induced cyanide linkage isomerization, which rapidly leads to a low-spin form of the cluster containing diamagnetic Ni(II) centers. The cyclic voltammagram of this species in DMF reveals a sequence of six successive reduction waves spaced approximately 130 mV apart, suggesting class II mixed-valence behavior upon reduction. The magnetic properties of the aforementioned linear cluster are consistent with the expected ferromagnetic coupling and an S = 4 ground state, but otherwise vary slightly with the specific conformation adopted (as influenced by the packing of associated counteranions and solvate molecules in the crystal). Magnetization data indicate an axial zero-field splitting parameter with a magnitude falling in the range [D] = 0.44-0.72 cm(-1), and fits to the magnetic susceptibility data yield exchange coupling constants in the range J = 17.0-17.6 cm(-1). These values represent significant increases over those displayed by the analogous Cr(III)-containing cluster. When perchlorate is used as a counteranion, [(Me(3)tacn)(2)(cyclam)NiMo(2)(CN)(6)](2+) crystallizes from water in a dimeric form with pairs of the linear clusters directly linked via hydrogen bonding. In this case, fitting the magnetic susceptibility data requires use of two coupling constants: one intramolecular with J = 14.9 cm(-1) and another intermolecular with J' = -1.9 cm(-1). Reacting [(Me(3)tacn)Mo(CN)(3)] with a large excess of [(cyclam)Ni(H(2)O)(2)](2+) produces a [(Me(3)tacn)(2)(cyclam)(3)(H(2)O)(2)Ni(3)Mo(2)(CN)(6)](6+) cluster possessing a zigzag structure that is a simple extension of the linear cluster geometry. Its magnetic behavior is consistent with weaker ferromagnetic coupling and an S = 6 ground state. Similar reactions employing an equimolar ratio of reactants afford related one-dimensional chains of formula [(Me(3)tacn)(cyclam)NiMo(CN)(3)](2+). Once again, the ensuing structure depends on the associated counteranions, and the magnetic behavior indicates ferromagnetic coupling. It is hoped that substitutions of the type exemplified here will be of utility in the design of new single-molecule magnets.     

A one-step conversion of isoquinolinium salts into naphthalene derivatives1

Berberine chloride (1) upon treatment with NaOAc-Ac₂O yields naphthalene derivatives 4 and 5 In like fashion, 8 gives naphthalene 10; isoquinoline methiodide leads to β-naphthyl acetate; and 12 provides 14. The reaction has been extended to the 3-carbonyl pyridinium series where 3-acetylpyridine methiodide and 3-pyridinecarboxaldehyde methiodide furnish lactones 16 and 17, respectively. All these transformations proceed by initial nucleophilic attack of the acetic anhydride anion on the immonium carbon atom.     

Conservative methods for the Toda lattice equations

We are concerned with the numerical integration of the Todalattice equations by using different conservative methods. Numericalexperiments suggest that the global error for isospectral schemesdecreases exponentially with time but it is almost constantfor either symplectic or more general integrators. We providea theoretical explanation for these experimental findings.