Helvetica Chimica Acta – Vol. 92 (2009)
33
using graphite-monochromated MoKa radiation (l 0.71073 ꢂ) and an Oxford Cryosystems Cryostream
700 cooler. Data reduction was performed with HKL Denzo and Scalepack [11]. The intensities were
corrected for Lorentz and polarization effects. For 1, an empirical absorption correction was applied
using SORTAV, which is based on an analysis of symmetry-equivalent reflections in the highly redundant
data set [12]. A numerical absorption correction [13] was applied to each of the remaining data sets.
Each structure was solved by direct methods using SIR92 [14], which revealed, at a minimum, the
positions of the Tl- and Br-atoms. All remaining non-H-atoms were located in subsequent difference
Fourier maps.
In 1, the Tl-atom lies on a center of inversion, while the central N-atom of the cation lies on a C2 axis.
In 6, the anion sits across a mirror plane with the Tl- and two Br-atoms on the mirror. The cation lies over
a crystallographic center of inversion and is, therefore, disordered about this symmetry center. The
overlapping disordered orientations of the ions caused instability in the refinement of the atomic
positions. Therefore, bond-length restraints were applied to most of the bonds within the cation in order
to maintain reasonable geometry. Neighboring atoms within and between each disordered orientation of
the cation were also restrained to have similar atomic displacement parameters. Compound 8 contains
two symmetry-independent cations, each of which is disordered across a center of inversion. The disorder
superimposes inverted copies of the cations upon themselves. This arrangement necessitated defining the
amino N-atom in each cation with a site occupation factor of 0.5, while the site for each C-atom adjacent
to the amino N-atom, as well as the sites of the pyridinyl and pyridinium N-atoms, are occupied by 0.5C þ
0.5N, where the atoms used to define a composite atom site were constrained to have identical fractional
coordinates and anisotropic displacement parameters.
The non-H-atoms in each structure were refined anisotropically. For 1, the H-atom of the central
ammonium group of the cation was placed in the position indicated by a difference Fourier map, and its
position was allowed to refine. The remaining H-atoms, as well as those in each of the other structures,
were placed in geometrically calculated positions and constrained to ride on their parent atoms. Each H-
atom was assigned an isotropic displacement parameter with a value equal to 1.2 Ueq of its parent atom
(1.5Ueq for Me and ꢀNHþ3 groups). A circular difference Fourier map was calculated and used to
estimate the orientation of the H-atoms of the symmetry-unique ꢀNHþ3 group in 1 and of the Me groups
in 2 and 7. Although the H-atoms could not be located definitively for 6, in adding the H-atoms to the
model, it was assumed that the cation is protonated at the imine N-atom, N(5). This assumption was
supported by the fact that a logical NꢀH ··· Br H-bond is evident only if this N-atom is protonated. For 8,
the cations could be protonated at the amino N-atom or at a pyridinyl N-atom, or possibly even
disordered between the two pyridinyl N-atoms, but the disorder of the cations made it difficult to locate
the positions of these H-atoms definitively. Protonation at any of the possible sites leads to a reasonable
arrangement of H-bonds. While refinement of the various alternative models left the R-factor virtually
unchanged, a test refinement of the isotropic atomic displacement parameters of the amino and
pyridinium H-atoms led to reasonable values for these parameters only when the cations were defined as
having one amino and one pyridinium H-atom, with the latter not being disordered between the two
pyridinyl N-atoms. Also, the amino NꢀC bond lengths of ca. 1.35 ꢂ are shorter than would be expected if
the amino group was protonated.
The refinement of each structure was carried out on F2 using full-matrix least-squares procedures,
which minimized the function Sw(Fo2 ꢀ Fc2 )2. Complex 3 crystallizes in a polar space group, and the
correct absolute structure has been confirmed by refinement of the absolute structure parameter
[15][16], which converged to a value of ꢀ 0.016(6). Corrections for secondary extinction were applied in
each case. For 5, four reflections, whose intensities were considered to be outliers, were omitted from the
final cycles of refinement. For each structure, the largest peaks of residual electron density were always
within the vicinity of the Tl- or Br-atoms. The scattering factors for non-H-atoms were taken from [17],
and the scattering factors for H-atoms were taken from [18]. For 3, anomalous dispersion effects were
included in Fc [19]; the values for f’ and f’’ were those from [20]. The values of the mass attenuation
coefficients were those from [21]. All calculations were performed using SHELXL97 [22]. The figures
were drawn using ORTEPII [23] and the PLUTON routine in the program PLATON [24].