Full Paper
spectra from these experiments, one at the end of titration
+53 equiv of 4-papy) and the other measured at 257 K seem
of magnetic moment in variable-temperature and titration ex-
periments, 2) the changes in electronic absorption spectra ob-
served on cooling or on addition of 4-papy, and 3) the assign-
ment of a temperature-independent S=1/2 signal detected by
EPR spectroscopy as arising from 1. Furthermore, we recorded
variable-temperature EPR spectra in the presence of excess 4-
papy (ca. 15 equiv). The obtained spectra resemble closely
those measured on a pure complex and further confirm our
model (Figure S9 in the Supporting Information).
(
to be identical (cf. Figures 9 and 5). This provides strong evi-
dence that in both titration and cooling experiments, we ob-
serve spectral changes associated with the same process,
namely, the coordination of free 4-papy ligand to five-coordi-
nate 2 forming six-coordinate 1 [Eq. (2)]. Since subsequent dis-
sociation to form four-coordinate species can be excluded (see
above), titration data were modeled by exclusively considering
2
ꢀ1
the equilibrium process [Eq. (2)] to yield K =3(1)ꢂ10 Lmol
The proposed equilibrium [Eq. (2)] was further supported by
TD-DFT calculations. Allowing for the limitations of TD-DFT to
calculate electronic spectra for spin-coupled species, we ach-
ieved good qualitative agreement between experiment and
theory. The calculated spectrum for 2 in the S=5/2 state
shows several intense bands in the visible region due to metal-
to-ligand charge transfer transitions (Figure S26 in the Support-
ing Information). In contrast, the S=1/2 state of 1 reveals
weaker transitions in the visible region, whereby the promi-
nent band at 500 nm shows a significant intraligand n!p*
character of 4-papy (Figure S27 in the Supporting Information).
Therefore, the 2 (S=5/2)!1 (S=1/2) transition should be ac-
companied by decreasing absorption in the visible region. This
can be best visualized by calculated gradual changes in the
electronic absorption spectrum (Figure 14). Indeed, the predict-
ed evolution of the spectrum resembles the experimental vari-
able-temperature (Figure 5) and titration (Figure 9) data. More-
a
(
Figure 10). Consequently, the degree of dissociation a for 1 in
ꢀ
4
the investigated solutions (c =5ꢂ10 m) at RT is estimated to
0
be 88%. Hence, 2 is the dominant species in solution at RT. On
addition of 4-papy to or cooling of the parent solution, the
equilibrium [Eq. (2)] is shifted to the right, generating 1. The
isosbestic points in the spectra confirm that only two species
are involved in equilibrium.
Similarly, the evolution of the effective magnetic moment
meff measured by the Evans NMR method during titration
(Figure 11) resembles closely the changes observed in the vari-
able-temperature experiment on cooling (Figure 3). Both addi-
tion of 4-papy and lowering the temperature decrease the
magnetic moment. The titration experiment provided a mini-
mum meff of 2.78(19) m in the presence of a large excess of 4-
B
papy, whereas the variable-temperature data revealed a plateau
of m =2.73 m at low temperatures. Astonishingly, these two
eff
B
different experiments converge to an identical magnetic
moment. This provides strong evidence that the same process
is operative in both experiments. The evolution of meff on titra-
tion can successfully be modeled for 1 and 2 in equilibrium
over,
3774 cm ),
qualitatively reproduced in our calculations for the S=1/2
a
very intense IVLLCT band (exptl: 2650 nm=
ꢀ
1
[29,34]
III
[5,25]
characteristic of a ls-Co (Cat)(SQ) state,
was
ꢀ
1
state of 1 (calcd: 1270 nm=7877 cm ; Figure S28 in the Sup-
porting Information). No similarly intense NIR bands were ob-
served in the calculated spectrum for the S=5/2 state of 2.
2
ꢀ1
[
Eq. (2)] to yield K =5(3)ꢂ10 Lmol , very similar to the value
a
obtained by electronic spectroscopy. Thus, according to the
Evans NMR method the degree of dissociation a for 1 in solu-
ꢀ
4
tion (c =5ꢂ10 m) at RT is estimated to be 83%.
0
Since the parent solution without addition of 4-papy con-
tains about 85% of 2 and 15% of 1 with m =3.86 m (see
eff
B
II
above), 2 must have a hs-Co (SQ)2 electronic configuration
with three uncoupled (at RT) paramagnetic centers. The spin-
II
only value for a pure fully uncoupled hs-Co (SQ) state would
2
be 4.58 m . At low temperatures or in the presence of a large
B
excess of 4-papy, 1 is exclusively formed, and must have
a smaller magnetic moment. The ultimate low-temperature
values of m =2.73 mB obtained from variable-temperature
eff
measurements (Figure 3) and m =2.78(19) m from the titra-
eff
B
III
tion experiment (Figure 11) are ascribed to a ls-Co (Cat)(SQ)
Figure 14. Predicted evolution of the electronic absorption spectrum accom-
II
III
S=1/2 electronic structure for 1. These values significantly
exceed the expected moment of about 1.7 m for an S=1/2
panying the [hs-Co (SQ)
2
(4-papy)] (2)![ls-Co (Cat)(SQ)(4-papy)
2
] (1) conver-
sion; spin-unrestricted B3LYP-TD-DFT calculations with COSMO; see Experi-
mental Section for further details.
B
ligand-based paramagnet. We ascribe the observed deviation
to some decomposition of the complex in solution leading to
species with S>1/2. The decomposition product does not in-
terfere with the equilibrium [Eq. (2)], as confirmed by clean iso-
sbestic points in electronic absorption spectra. Note that an
S=1 species that would give m ꢁ3 m cannot be constructed
Although VT interconversion due to coordination of solvent
[50]
molecules in solution is known,
valence tautomerism in-
duced by coordination/dissociation of ligands in solution is
nearly unexplored. In this context, we mention a recent work
by Minkin and co-workers, who reported six-coordinate com-
eff
B
_
III
for an odd-spin species 1 or 2.
plex ls-[Co (SQ*)(NO) ] featuring one redox-active phenoxa-
2
III
Assuming the presented equilibrium [Eq. (2)], ls-Co (Cat)(SQ)
zine-1-one ligand (Q*; SQ*=one-electron-reduced form of Q*;
_
II
[51]
III
electronic structure for 1, and hs-Co (SQ) configuration for 2,
NO=bidentate ligand). Although the complex has a ls-Co -
2
we can account for all experimental findings: 1) the evolution
(SQ*) electronic structure in the solid state, a neutral ligand Q*
Chem. Eur. J. 2014, 20, 11149 – 11162
11158
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Witt, Alexander
