Spin distribution in electron-rich piano-stool iron(III) pyridylalkynyl radical cations containing [(η2-dppe)(η5-C 5Me5)FeC≡C]+ End Groups

Article abstract of DOI:10.1021/om100173x

This experimental and theoretical contribution is aimed at investigating the electronic structure of cationic electron-rich ethynylpyridyl Fe(III) derivatives of the formula [(η²-dppe)(η⁵-C ₅Me₅)FeC≡C(x-C₅H₄N)][PF ₆] (x = 4, 3, 2; 1a-c[PF₆]) and [(η²-dppe) (η⁵-C₅Me₅)FeC≡C(2,5-C ₅H₃NX)][PF₆] (X = Cl, Br; 2a,b[PF ₆]). The Moessbauer, NMR, and ESR characterization of these paramagnetic species are reported and discussed in connection with DFT results. Special emphasis is put on the electronic effect of the nitrogen atom and of the halogen substituent on the spin distribution within the pyridyl unit. It is shown that ¹H NMR constitutes a straightforward empirical way to investigate the slight changes in spin distribution taking place on the heteroaryl ring.

Full text of DOI:10.1021/om100173x

Organometallics 2010, 29, 2491–2502 2491
DOI: 10.1021/om100173x
Spin Distribution in Electron-Rich Piano-Stool Iron(III) Pyridylalkynyl
Radical Cations Containing [(η2-dppe)(η5-C5Me5)FeCtC]þ End Groups
,†
Frederic Paul,* Floriane Malvolti, Gregory da Costa, Sylvie Le Stang,
†
‡
†
ꢀ ꢀ ꢀ
Frederic Justaud, Gilles Argouarch, Arnaud Bondon,* Sourisak Sinbandhit,
†
†
,‡
§
ꢀ ꢀ
Karine Costuas,† Loic Toupet, and Claude Lapinte*,†
†
ꢀ
Sciences Chimiques de Rennes, UMR CNRS 6226, Universite de Rennes 1, Campus de Beaulieu, Bat. 10C,
F-35042 Rennes Cedex, France, PRISM, UMR CNRS 6026, Universite de Rennes 1, CS 34317, Campus de
^
‡
ꢀ
Villejean, 35043 Rennes Cedex, France, §CRMPO, Universiteꢀ de Rennes 1, Campus de Beaulieu, 35042 Rennes
ꢀ
Cedex, France, and Institut de Physique de Rennes (IPR), UMR CNRS 6251, Universite de Rennes 1,
Campus de Beaulieu, F-35042 Rennes Cedex, France
Received March 3, 2010
This experimental and theoretical contribution is aimed at investigating the electronic structure of
cationic electron-rich ethynylpyridyl Fe(III) derivatives of the formula [(η2-dppe)(η5-C5Me5)FeCt
C(x-C5H4N)][PF6] (x = 4, 3, 2; 1a-c[PF6]) and [(η2-dppe)(η5-C5Me5)FeCtC(2,5-C5H3NX)][PF6]
€
(X=Cl, Br; 2a,b[PF6]). The Mossbauer, NMR, and ESR characterization of these paramagnetic
species are reported and discussed in connection with DFT results. Special emphasis is put on the
electronic effect of the nitrogen atom and of the halogen substituent on the spin distribution within
the pyridyl unit. It is shown that 1H NMR constitutes a straightforward empirical way to investigate
the slight changes in spin distribution taking place on the heteroaryl ring.
Introduction
reported on compounds incorporating “(η2-dppe)(η5-C5Me5)-
FeCtC-” groups.3-5 The simplest structurally among these
organoiron metallo-ligands are the monodentate pyridylalkynyl
complexes 1a-c.6 These Fe(II) compounds were fully character-
ized and shown to behave as functional pyridines toward various
metal centers.5 The corresponding Fe(III) complexes 1a-c[PF6]
were also previously isolated and briefly characterized. These
species were shown to be weaker ligands than the equivalent
Fe(II) complexes.5 More recently, the related Fe(II) compounds
2a,b, featuring a halogen in a position ortho to the nitrogen, were
synthesized (Chart 1) and used as precursors to access new
bipyridine ligands functionalized with two Fe(II) redox-active
groups.3 However, the corresponding Fe(III) complexes 2a,
b[PF6] have not been characterized so far. Thus, in comparison
to functional phenylacetylide Fe(III) complexes such as 3-X[PF6]
(Chart 1), relatively few data are available regarding the pyridyl-
based Fe(III) analogues.7-10 Such a knowledge is, however,
crucial when optimal tuning of a given electronic property of
functional assemblies made from these compounds is sought.
Pyridyl-based mono- or polydentate ligands incorporating
electron-rich organometallic acetylide fragments have attracted
considerable attention recently.1 Upon complexation, such “me-
tallo-ligands” possessing redox-active substituents should pro-
vide an easy access to various kinds of molecular architectures
presenting unique properties for information storage or informa-
tion processing at the molecular level.2 We have previously
*To whom correspondence should be addressed. F.P. and C.L.: tel,
(þ33) 02 23 23 59 62; fax, (þ33) 02 23 23 56 37. A.B.: tel, (þ33) 2 23 23 65
61; fax, (þ33) 2 23 23 46 06.
(1) (a) Koutsantonis, G. A.; Jenkins, G. I.; Schauer, P. A.; Szczepaniak,
B.; Skelton, B. W.; Tan, C.; White, A. H. Organometallics 2009, 28, 2195–
2205. Ge, Q.; Corkery, T. C.; Humphrey, M. G.; Samoc, M.; Hor, T. S. A. Dalton
Trans. 2009, 6192–6200. (b) Ge, Q.; Corkery, T. C.; Humphrey, M. G.; Samoc,
M.; Hor, T. S. A. Dalton Trans. 2008, 2929–2936. (c) Muro, M. L.; Diring, S.;
Wang, X.; Ziessel, R.; Castellano, F. N. Inorg. Chem. 2008, 47, 6796–6803.
(d) Muro, M. L.; Diring, S.; Wang, X.; Ziessel, R.; Castellano, F. N. Inorg. Chem.
2009, 48, 11533–11542. (e) Cifuentes, M. P.; Humphrey, M. G.; Koutsantonis, G.
A.; Lengkeek, N. A.; Petrie, S.; Sanford, V.; Schauer, P. A.; Skelton, B. W.;
ꢀ
Stranger, R.;White, A. H.Organometallics 2008, 27, 1716–1726. (f) Pelerin, O.;
Olivier, C.; Roisnel, T.; Touchard, D.; Rigaut, S. J. Organomet. Chem. 2008,
693, 2153–2158. (g) Xu, H.-B.; Zhang, L.-Y.; Xie, Z.-L.; Ma, E.; Chen, Z.-N.
Chem. Commun. 2007, 2744–2746. (h) Wong, C.-Y.; Tong, G. S. M.; Che,
(4) (a) Ibn Ghazala, S.; Gauthier, N.; Paul, F.; Toupet, L.; Lapinte, C.
Organometallics 2007, 26, 2308–2317. (b) Paul, F.; Goeb, S.; Justaud, F.;
Argouarch, G.; Toupet, L.; Ziessel, R. F.; Lapinte, C. Inorg. Chem. 2007, 46,
9036–9038.
ꢀ
C.-M.; Zhu, N. Angew. Chem., Int. Ed. 2006, 45, 2694–2698. (i) Mery, D.;
Ornelas, C.; Daniel, M.-C.; Ruiz, J.; Rodrigues, J.; Astruc, D.; Cordier, S.; Kirakci,
K.; Perrin, C. C. R. Chim. 2005,8,1789–1797. (j) Ku€hn, F. E.; Zuo, J.-L.; de Biani,
F. F.; Santos, A. M.; Zhang, Y.; Zhao, J.; Sandulache, A.; Herdtweck, E. New J.
Chem. 2004, 28, 43–51. (k) Engratkul, C.; Schoemaker, W. J.; Grzybowski, J. J.;
Guzei, I.; Rheingold, A. Inorg. Chem. 2000, 39, 5161–5163.
(2) (a) Balzani, V.; de Silva, A. P. Electron Transfer in Chemistry; Wiley-
VCH: Weinheim, Germany, 2000; Vol. 5. (b) Norgaard, K.; Bjornholm, T.
Chem. Commun. 2005, 1812–1823. (c) Gianneschi, N. C.; Masar, M. S., III;
Mirkin, C. A. Acc. Chem. Res. 2005, 38, 825–837. (d) Lehn, J.-M. Angew.
Chem., Int. Ed. Engl. 1990, 29, 1304–1319. (e) Lehn, J. M. Supramolecular
Chemistry-Concepts and Perspectives; VCH: Weinheim, Germany, 1995.
(3) Justaud, F.; Argouarch, G.; Gazalah, S. I.; Toupet, L.; Paul, F.;
Lapinte, C. Organometallics 2008, 27, 4260–4264.
(5) Le Stang, S.; Paul, F.; Lapinte, C. Inorg. Chim. Acta 1999, 291,
403–425.
(6) Le Stang, S.; Lenz, D.; Paul, F.; Lapinte, C. J. Organomet. Chem.
1998, 572, 189–192.
(7) Paul, F.; da Costa, G.; Bondon, A.; Gauthier, N.; Sinbandhit, S.;
Toupet, L.; Costuas, K.; Halet, J.-F.; Lapinte, C. Organometallics 2007,
26, 874–896.
ꢀ
(8) Paul, F.; Toupet, L.; Thepot, J.-Y.; Costuas, K.; Halet, J.-F.;
Lapinte, C. Organometallics 2005, 24, 5464–5478.
(9) Costuas, K.; Paul, F.; Toupet, L.; Halet, J.-F.; Lapinte, C.
Organometallics 2004, 23, 2053–2068.
(10) Denis, R.; Toupet, L.; Paul, F.; Lapinte, C. Organometallics
2000, 19, 4240–4251.
r
2010 American Chemical Society
Published on Web 05/14/2010
pubs.acs.org/Organometallics

2492 Organometallics, Vol. 29, No. 11, 2010
Chart 1. Selected Pyridyl- and Aryl-Based Fe(III) Compounds
Paul et al.
Scheme 1. Oxidation of the Fe(II) Complexes 1a-c and 2a,b
In the present contribution, we have therefore performed the
extensive spectroscopic characterization of 1a-c[PF6] and
2a,b[PF6]. A density functional theory (DFT) study was also
undertaken in order to complete our understanding of the
electronic structures of these open-shell derivatives, particu-
lar attention being paid to the influence of the nitrogen atom
and/or of the halogen substituent on the spin distribution.
Figure 1. ORTEP representations of the cation 2bþ at the 50%
˚
probability level. Selected distances (A) and angles (deg) for 2b[PF6]:
Fe1-(Cp*)centroid = 1.779, Fe1-P1 = 2.2515(10), Fe1-P2 =
2.2508(9), Fe1-C37 = 1.889(3), C37-C38 = 1.207(4), C38-
C39=1.440(3), C39-C40=1.389(5), C40-C41=1.378(4), C41-
C42 =1.369(5), C39-C43 =1.395(4), N1-C42 = 1.321(4), N1-
C43=1.330(4), Br1-C42=1.903(3); P1-Fe1-P2=83.51(4), Fe1-
C37-C38 = 175.5(3), C37-C38-C39 = 176.2(3), C43-C39-
Fe1-(Cp*)centroid = -136.7.13
Results
Synthesis of the Fe(III) Complexes. The precursor Fe(II)
compounds 1a-c and 2a,b were synthesized according to
previously published procedures and subsequently oxidized
following the “classic” workup (Scheme 1) to give the desired
corresponding Fe(III) complexes.3,5,6 The purity of the known
Fe(III) samples 1a-c[PF6] was checked by electrochemistry and
IR spectroscopy, while the new compounds 2a,b[PF6] were fully
characterized by the usual means (Experimental Section). The
latter exhibit characteristic spectroscopic signatures very similar
to these previously observed for 1b[PF6].5 Thus, they possess
low-intensity charge transfer bands in the visible range pre-
sumed to be LMCT bands, an absorption of weak intensity in
the near-IR range, presumably corresponding to a symmetry-
forbidden LF transition with a strong d-d character,8 and
also two IR vibrational modes of very low intensity in the
Table 1. 57Fe Mossbauer Fitting Parameters at 80 K for Selected
Complexes
€
IS ((0.01
QS ((0.01
area
(%)
compd
1a
mm s-1
)
mm s-1
)
Γ (mm s-1 a
)
0.24
0.25
0.25
0.25
0.24
0.23
1.95
0.91
1.90
0.87
1.87
0.92
0.11
0.15
0.12
0.16
0.12
0.15
100
100
100
57
100
100
1a[PF6]
1b
1b[PF6]b
1c
1c[PF6]
a Half-width of the peaks of the signal. b Two other doublets present at
isomeric shifts of 0.13 mm s-1 (29%; QS = 1.22 ( 0.03 mm s-1) and of
0.21 mm s-1 (14%; QS = 1.92 ( 0.02 mm s-1).
1940-2010 cm-1 spectral region, one of which (or both in case
˚
bromine bond C42-Br1 of 2b[PF6] is ca. 0.15 A longer than the
carbon-chlorine bond C42-Cl1 of 2a[PF6] (1.903(3) vs
5
.
of Fermi coupling)11 corresponds to νCtC
Solid-State Structures of 2a,b[PF6]. In the course of these
syntheses, the new Fe(III) compounds 2a,b[PF6] could be crys-
tallized, and their solid-state structures were solved by X-ray
diffraction (Figure 1; see also Table 6 in the Experimental
Section). To our knowledge, these structural data are the first
for piano-stool Fe(III) pyridylalkynyl complexes. Crystal data
and space groups are similar for both compounds (Table 6),
as is the conformation adopted by the single molecule in the
asymmetric unit. In line with published data,12 the carbon-
3
˚
1.741(2) A). Relative to their respective Fe(II) parents 2a,b,
the changes in bond lengths upon oxidation are very similar to
these previously observed for related 3-X/3-X[PF6] systems.8,10
Thus, a slight shortening of the Fe-C37 bond and of the
C42-X1 (X = Cl, Br) bonds, along with a slight elongation
of the Fe1-P1, Fe1-P2, and Fe1-(C5Me5) bonds are
˚
(13) Selected distances (A) and angles (deg) for 2a[PF6]: Fe1-
(Cp*)centroid = 1.774, Fe1-P1 = 2.2460(6), Fe1-P2 = 2.2446(5), Fe1-
C37 = 1.8800(19), C37-C38 = 1.216(3), C38-C39 = 1.435(3), C39-
C40=1.396(3), C40-C41=1.378(3), C41-C42=1.374(3), C39-C43 =
1.374(3), N1-C42 = 1.328(2), N1-C43 = 1.328(2), Cl1-C42 = 1.741(2);
P1-Fe1-P2 = 83.40(2), Fe1-C37-C38 = 173.9(2), C37-C38-C39 =
176.7(2), C43-C39-Fe1-(Cp*)centroid = -139.8.
(11) Paul, F.; Mevellec, J.-Y.; Lapinte, C. Dalton Trans. 2002, 1783–1790.
(12) Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen,
A. G.; Taylor, R. J. Chem. Soc., Perkin Trans. 2 1987, S1–S19.

Article
Organometallics, Vol. 29, No. 11, 2010 2493
observed, the rest of the bond lengths remaining fairly constant
within experimental uncertainties (Supporting Information).
€
Mossbauer Characterization of 1a-c and 1a-c[PF6]. The
pyridyl complexes were then characterized by 57Fe Mossbauer
€
spectrometry at 80 K in their neutral Fe(II) (1a-c) and radical
cation Fe(III) (1a-c[PF6]) states. The characteristic doublets
expected for these compounds were always detected as the
sole signal, except for 1b[PF6] (Table 1).14 In this case, two
additional doublets of lower intensities were also detected.
These presumably originate from iron-containing species
which result from partial decomposition of this reactive
Fe(III) sample. The isomeric shifts (IS) of the Fe(II) (1a-c)
and Fe(III) samples (1a-c[PF6]) remain quite close and are
not very sensitive to the nature of the pyridyl ring or to the
redox state of the iron center for a given isomer. In contrast,
the quadrupolar splitting (QS) is clearly diagnostic of the
redox state of the metal center, QS values of around 0.9 mm
s-1 being observed for Fe(III) complexes, while QS values
closer to 1.9 mm s-1 are observed for Fe(II) complexes.15,16
Overall, these results are reminiscent of those previously
reported for the 3-Xþ alkynyl complexes (Chart 1).10 The
similarities of the IS and QS parameters obtained for both
Fe(II) and Fe(III), whatever the position of the nitrogen on
the aromatic ring, clearly indicate that the electronic environ-
ment of the iron nucleus is not sensitive to the structure of the
pyridylalkynyl ligand.
Figure 2. ESR spectrum of 1aþ in a CH2Cl2/1,2-C2H4Cl2 (1/1)
glass at 80 K.
NMR of the Fe(III) Complexes 1a-c[PF6] and 2a,b[PF6].
The 1H NMR spectra of 1a-c[PF6] and 2a,b[PF6] have sub-
sequently been recorded in dichloromethane-d2 (Figure 3).
The 1H NMR signals were detected in ranges similar to those
of the known Fe(III) radicals 3-X[PF6] (Table 3).7
Our first trial was to assign the various 1H signals detected.
For 1a[PF6], this proved possible by monitoring the 1H
NMR shifts of various mixtures of 1a/1a[PF6] as a function
of the Fe(II) composition of the sample (Figure 4). In line
with previous studies on such Fe(III) radical cations,7,17 the
1
electron self-exchange reaction is much faster than the H
NMR chemical shift time scale and an averaged set of signals
is always observed. A progressive increase in the concentra-
tion of the paramagnetic complex 1a[PF6] thereby allows
firm identification of most of the signals detected for the
Fe(III) samples, on the basis of their assignment in the
spectra of the diamagnetic Fe(II) samples.
Table 2. Experimental ESR g Valuesa and Redox Potentialsb,c for
1a-c[PF6] and 2a,b[PF6] and computed g Values and Ionization
Potentials for 1a-c-Hþd
compd
g1
g2
g3
<g>
Δg
Eꢀ (V) IPꢀ (eV)c
As for 3-X[PF6] compounds, rather specific 1H NMR
shifts are observed for the “(η2-dppe)(η5-C5Me5)Fe” fragment
(Figure 3). Thus, the C5Me5 protons (H9) give rise to the most
intense peak of the spectrum near -10 ppm, while the endo
and exo phenyl protons (H1/H2/H3 and H4/H5/H6) of the dppe
ligand appear in the diamagnetic range as two characteristic
sets of signals in a roughly 2/2/1 ratio. The assignment of these
sets of signals was also confirmed by COSY experiments.
Assignment of the endo and exo sets was made by analogy
with 3-X[PF6] compounds.7 Among the methylene protons H7
and H8, the latter come out slightly below -3 ppm, while the
former is often not detected but is believed be in the 5-9 ppm
range, presumably hidden below the aromatic dppe protons.
Finally, the pyridyl protons of 1a[PF6] correspond to the most
shifted signals and were detected around the -23 ppm range
(Ha) and around 25 ppm (Hb). Protons ortho and para to the
acetylide bridge are shifted to high field, while meta protons
are shifted to low field, in line with the observations previously
made for 3-X[PF6] species.7 The assignment of the various
pyridyl protons of 1a[PF6] was further confirmed by polariza-
tion transfer studies.
For the other isomers of 1a[PF6], the signals of the protons
belonging to the “(η2-dppe)(η5-C5Me5)Fe” fragment were
also readily identified on the basis of their characteristic
shifts and intensities. The remaining signals, which are again
the most shifted ones, correspond to the pyridyl protons. On
the basis of the study of 3-X[PF6] complexes,7 the most
shifted signals at low field should correspond to hydrogen
atoms in meta positions relative to the alkynyl bridge,
whereas those at high field correspond to hydrogen atoms
1aþ
1bþ
1cþ
2aþ
2bþ
1.969 2.028 2.494 2.163 0.525 -0.03b
1.967 2.028 2.490 2.162 0.521 -0.11b
1.966 2.024 2.500 2.163 0.534 -0.08b
1.967 2.027 2.503 2.165 0.536 -0.07c
1.972 2.028 2.509 2.169 0.537 -0.06c
1a-Hþd 1.990 2.040 2.162 2.064 0.172
1b-Hþd 1.992 2.036 2.137 2.055 0.145
1c-Hþd 1.990 2.038 2.173 2.067 0.183
6.190
6.008
5.915
a Experimental ESR g values ((0.005) determined at 77 K in CH2Cl2/
C2H4Cl2 (1/1) glass. b Values from ref 5. c Values from ref 3. d Computed
g-tensor values and ionization potentials (IP) for the model compounds
1a-c-Hþ are reported for the sake of comparison.
ESR of the Fe(III) Complexes 1a-c[PF6] and 2a,b[PF6].
The five Fe(III) complexes were characterized by ESR. A
rhombic signal was detected for each of them in solvent
glasses at 80 K (Figure 2 and Table 2).8,15 In accordance with
€
the conclusion drawn from the Mossbauer data, the ESR
data suggest that the main features of the HOMO are not
very sensitive to the structure of the pyridylalkynyl ligand
(see Figure 6). Indeed, the mean g values and the anisotropy
(Δg) of the detected signals remain sensibly similar for
1a[PF6] and 1b[PF6], a slight increase in Δg taking place
for 1c[PF6]. However, the data for 2a,b[PF6 ] reveal that an
increase of the anisotropy takes place relative to 1b[PF6]
when a halogen substituent is appended in a position para to
the acetylide linker.
(14) (a) Greenwood, N. N. Mo€ssbauer Spectroscopy; Chapman and
Hall: London, 1971; (b) G€utlich, P.; Link, R.; Trautwein, A. In Inorganic
Chemistry Concepts. Mo€ssbauer Spectroscopy and Transition Metal
Chemistry; G€utlich, P., Link, R., Trautwein, A., Eds.; Springer-Verlag:
Berlin, 1978; Vol. 3, pp 191-201.
(15) Paul, F.; Lapinte, C. Coord. Chem. Rev. 1998, 178/180, 431–509.
(16) Guillaume, V.; Thominot, P.; Coat, F.; Mari, A.; Lapinte, C.
J. Organomet. Chem. 1998, 565, 75–80.
^
(17) Roger, C.; Hamon, P.; Toupet, L.; Rabaa, H.; Saillard, J.-Y.;
Hamon, J.-R.; Lapinte, C. Organometallics 1991, 10, 1045–1054.

2494 Organometallics, Vol. 29, No. 11, 2010
Paul et al.
allowed us to differentiate the pyridyl protons Hb from Hd at
low field and Ha from Hc at high field (Supporting In-
formation). For 1b[PF6], things were not so straightforward,
because even if the COSY correlation allowed us to distinguish
the signals corresponding to Ha and Hc from that of He in the
set of signals at high field, it did not allow their particular
assignment. The definitive assignment of Ha and Hc was
1
eventually made by comparing the H NMR spectrum of
1b[PF6] with those of 2a,b[PF6], in which Hc has been replaced
by X (X = Cl, Br). Indeed, the spectra of the last two com-
plexes closely resemble that of 1b[PF6] but lack a signal at high
field. The latter was therefore assigned to Hc in 1b[PF6].
The 13C NMR spectra of 1a-c[PF6] and 2a,b[PF6] have
also been recorded, and a tentative assignment of the de-
tected signals is proposed on the basis of the previous
investigation on 3-X[PF6] analogues (Supporting Informa-
tion).7 Likewise, we have assumed that none of the
R-acetylide carbon atoms (Ca in Chart 2) have presently
been detected, and in some cases also not the β atoms (Cb),
which are expected to appear at very low fields. With the
exception of the ipso aryl carbon atoms of the dppe ligand,
which were previously also suspected to escape detection,7
all carbon atoms belonging to the “(η2-dppe)(η5-C5Me5)-
Fe” fragment were always identified from their characteris-
tic shifts, many of these assignments being confirmed by
HMQC correlations. The remaining signals were thus con-
sidered to belong to the pyridyl carbon atoms.18
Temperature Dependence of the 1H NMR Signals. We
1
have then examined the temperature dependence of the H
NMR shifts of 1a-c[PF6] between 300 and 180 K (Figure 5
and the Supporting Information). Similarly to 3-X[PF6]
monoradicals,7 a linear 1/T dependence can be evidenced
for the paramagnetic shifts (δiso) with very good fits for all
nuclei, as apparent in Figure 5 (Curie behavior).19 This
behavior suggests that the ground state (GS) of these orga-
nometallic radicals has no closely lying excited states origi-
nating from spin-orbit coupling. As previously discussed,7
the extrapolated values at infinite T for several of these shifts
differ somewhat from the ideal “diamagnetic” values, which
should be those of the corresponding Fe(II) parents (1a-c).
This possibly happens because the weak pseudocontact con-
tribution to these particular isotropic shifts does not strictly
follow a Curie law, in contrast to the assumption made in eq 2
(see hereafter), and consequently leads to the observed devia-
tion upon extrapolation.
Derivation of Hyperfine Coupling Constants and Spin
Densities for the Primary Carbon Atoms of the Bridge of
1a[PF6], 1b[PF6], 1c[PF6] and 2[PF6]. Likewise to what has
been previously done for the 3-X[PF6] radical cations,7 the
contact hyperfine coupling constants and the spin densi-
ties were derived for selected protons of the pyridylalkynyl
linker from the 1H NMR contact shifts (Supporting In-
formation).
The NMR shift of a paramagnetic compound (eq 1a) is the
sum of a diamagnetic contribution (δdia) and of an isotropic
contribution (δiso), the latter actually originating from the
Figure 3. 1H NMR spectra of (a) 1a[PF6], (b) 1b[PF6], and
(c) 1c[PF6] in CD2Cl2 at 25 ꢀC with proposed assignments for
selected protons according to Chart 2.
(18) However, the modest paramagnetic shifts of several among these
signals call for caution until additional NMR data are available on this
class of compounds. Indeed, some of these signals might equally
correspond to diamagnetic impurities issued from a slow decomposition
of the Fe(III) radicals in the CD2Cl2.
(19) (a) Bertini, I.; Galas, O.; Luchinat, C.; Parigi, G.; Spina, G.
J. Magn. Reson. 1998, 130, 33–44. (b) Golding, R. M.; Pascual, R. O.;
Vrbancich, J. Mol. Phys. 1976, 31, 731–744.
in para and ortho positions. This was what we observed for
1b,c[PF6] and 2a,b[PF6], but no more precise assignment
could be gained from the study of Fe(II)/Fe(III) mixtures.
We had therefore to resort to polarization transfer studies to
assign these signals. For 1c[PF6], an unambiguous assign-
ment proved possible from COSY correlations, since it

Article
Organometallics, Vol. 29, No. 11, 2010 2495
Table 3. Detected (δ) and Isotropic (δiso) 1H NMR shifts (δ (0.1 ppm) Recorded for 1a-c[PF6] and 2a,b[PF6] at 20 ꢀC in CD2Cl2
a
-CtC(C5H3NX)-
dppe
H1
C5Me5
H9
compd
δ (ppm)
Ha
Hb
Hc
Hd
He
H2
7.1
H3
6.2
H4
0.9
H5
H6
H7
H8
1a[PF6] (X = H)
δ
-23.1
-29.7
-31.2
-38.2
-32.0
-39.3
-29.7
-37.0
-29.4
-36.7
25.1
17.0
28.4 -33.2
21.4 -41.3
25.0 -33.3 22.6
17.7 -40.0 14.3
28.6
21.3
28.7
21.4
5.3
3.4 7.9 n.o.b -2.9 -10.7
-4.9 -12.1
3.7 7.8 n.o.b -2.8 -10.7
δiso
δ
δiso
δ
δiso
δ
δiso
δ
-2.3 -0.2 -1.1 -6.4 -3.9 0.6
-34.4 6.0 7.1 6.2 1.2
-42.5 -1.8 -0.2 -1.1 -6.1 -3.6 0.5
5.9
-0.5 -1.2 -6.1 -3.8
-35.2 5.7 7.0 6.2 0.9
-42.0 -2.1 -0.3 -1.1 -6.4 -3.9 0.5
-34.5 5.7 7.1 6.2 0.9
3.5 7.8 n.o.b -3.0 -10.7
-41.3 -2.1 -0.2 -1.1 -6.4 -3.8 0.5 -4.8 -12.2
1b[PF6] (X = H)
1c[PF6] (X = H)
2a[PF6] (X = Cl)
2b[PF6] (X = Br)
-4.8 -12.1
-2.1 -10.0
-4.1 -11.5
6.8
6.1
1.2
3.5 7.8 9.9c
0.5 7.1 7.1
3.4 7.8 n.o.b -3.0 -10.7
-4.8 -12.2
δiso
a Proposed assignment according to Chart 2 (CHDCl2 at 5.35 ppm). b Not observed. c Tentative assignment.
Chart 2. 1H Nuclei Numeration Corresponding to the Proposed Assignment for 1a-c[PF6] and 2a,b[PF6]
presence of the electronic spin of the unpaired electron(s).20
This contribution (eq 1b) is itself the sum of a contact (δc)
and a pseudocontact term (δpc).
Table 4. π-Spin Densities on Primary Pyridyl Carbon Atoms
Derived from the 1H NMR Contact Shifts of the Nearby Protons
using the McConnell Relationship (eq 4) for 1a-c[PF6] and 2a,
b[PF6] at 295 K
δobs ¼ δdia þ δiso
δiso ¼ δc þ δpc
ð1aÞ
(Fπ)Ca for pyridyl C(H) atoms
compd
Cd
Ce
Cf/N
Cg/N
Ch/N
0.025
ð1bÞ
ð1cÞ
1a[PF6]
1b[PF6]
1c[PF6]
2a[PF6]
2b[PF6]
0.018
0.022
0.024
0.022
0.022
-0.008
-0.010
-0.008
-0.010
-0.010
0.023
0.022
M
δpc ¼ ðδpcÞ þ ðδpcÞ
L
-0.007
0.025
0.024
As shown in eq 1c, the pseudocontact shift is also the sum
of two contributions. The second one, called the ligand-
centered term (δpc)L, can be neglected for hydrogen nuclei in
complexes where the unpaired electron mostly resides on the
metal center. The pseudocontact contribution to the isotro-
pic shift therefore results essentially from the first term,
which represents the metal-centered term (δpc)M. This con-
tribution can be roughly evaluated (eq 2), provided the
diagonal values of the g tensor and some geometrical features
of these Fe(III) radicals are known.7 In this equation, μ0 is
the vacuum permittivity, μB is the Bohr magneton, k is the
Boltzmann constant, S is the spin of the radical (S = 1/2), g is
g3, g^ is the average of g1 and g2 (Table 2), rM is the distance
between the metal center and the proton considered, while θ
is the angle between rM and the acetylide axis.
a π-spin densities expressed in electrons (e) for carbon atoms numer-
ated according to Chart 3.
X-ray for 2a,b[PF6],21 and subsequently deduced from the
isotropic shift of each proton to give the corresponding
contact shift (δc). From these contact shifts, the contact
contribution to the isotropic hyperfine constants (AH)con
was in turn computed with eq 3. In this equation, γH is the
proton magnetogyric ratio, p is the reduced Planck constant,
and g is the mean electron g value (Table 2), while other
constants have been previously defined. Finally, the McCon-
nell expression (eq 4) provides a straightforward access to the
spin density on the nearby carbon atom from this contact
1
hyperfine coupling for these S = /2 radicals (Table 4).22
Interestingly, these data reveal that while replacement of Hc
in 1b[PF6] by a halogen atom in 2a,b[PF6] has a sizable
influence on the redox and ESR signatures of the Fe(III)
radical (Table 2), this structural modification apparently
does not modify significantly the spin distribution on the π
M
ðδpcÞ ¼ ðμ0=4πÞðμB2=9kTÞ½SðS þ 1Þꢀ
ðg 2 - g^ Þð3 cos2 θ - 1Þ=rM
ð2Þ
2
3
This contribution was computed for 1a-c[PF6] and2a,b[PF6],
using the geometrical parameters of the corresponding Fe(II)
complexes (1a-c) for 1a-c[PF6] and those determined by
(21) Comparison between the bond distances between 2a,b and 2a,
b[PF6] indicates only modest changes in geometry upon oxidation
(Supporting Information).
(20) Bertini, I.; Luchinat, C. NMR of Paramagnetic Molecules in
Biological Systems; Benjamin/Cummings: Menlo Park, CA, 1986.
(22) (QCH)
H is a constant presently taken as -66 MHz.20

2496 Organometallics, Vol. 29, No. 11, 2010
Paul et al.
Table 5. Calculated Spin Densities (e) on Selected Fragments or Atoms for [(η2-dpe)(η5-C5H5)FeCtC(x-C5H3N)]þ Complexes (dpe
=1,2-Diphosphinoethane; x = 4, 3, 2) in the Perpendicular and Parallel Conformationsa
CtC(x-C6H3N)
C atom
of C5H5
P atom
of dpe
ΔE b(kJ
compd
Fe
Ca
Cb
Cc
Cd
Ce
Cf
Cg
Ch
N
mol-1
)
c
1a-Hþ
(n = 4)
1b-Hþ
(n = 3)
0.799
0.691
0.738
0.654
0.648
0.822
0.673
0.695
-0.003
-0.002
-0.003
-0.002
-0.002
-0.003
-0.003
-0.002
-0.016
-0.015
-0.016
-0.014
-0.014
-0.018
-0.014
-0.016
-0.054
-0.026
-0.041
-0.009
-0.007
-0.067
-0.024
-0.033
0.299
0.246
0.271
0.233
0.234
0.303
0.244
0.240
-0.025
-0.035
-0.018
-0.026
-0.026
-0.025
-0.036
-0.036
0.014
0.083
0.007
0.069
0.070
0.008
0.097
0.095
-0.001
-0.032
0.003
-0.032
-0.033
0.001
0.002
0.075
0.067
-0.030
-0.030
0.010
9.1 e
0
d
^
c
-0.003
0.109
0.109
0.001
0.113
0.108
0.010
0.094
0.094
14.4e
2.1
0
d
^
^
1
d
2
c
1c-Hþ
(n = 2)
0.001
-0.016
-0.015
19.4e
8.9
0
d
^
^
-0.040
-0.038
0.045
0.044
1
d
2
a See Chart 3 for atom numbering. b Relative energy between the conformations. c Parallel conformation. d Perpendicular conformation (^1, nitrogen atom
toward C5H5 ligand; ^2:, nitrogen atom away from C5H5 ligand). e For this compound, see the computational details in the Experimental Section.
Figure 4. Observed 1H NMR shifts for (a) 1a/1a[PF6] mixtures in CD2Cl2 at 25 ꢀC (assignment according to Chart 2) and (b) aromatic
protons of the dppe ligand.
Figure 5. Temperature dependence of selected 1H NMR shifts for 1a[PF6] in CD2Cl2 with proposed assignment according to Chart 2.
manifold of the pyridyl ring.
ðAHÞcon ¼ δcp3γHkT=ðgμBSðS þ 1ÞÞ
Theoretical Results. DFT calculations were performed
on the mononuclear Fe(II) and Fe(III) model complexes
1a-H0/þ, 1b-H0/þ, and 1c-H0/þ, in which the dppe and C5Me5
ligands have been replaced by 1,2-diphosphinoethane (dpe)
and C5H5, respectively. These complexes were optimized in
ð3Þ
ð4Þ
H
ðAHÞcon=h ¼ ðQCHÞ ðFπÞC

Article
Chart 3. Mononuclear Model Complexes Used in the DFT
Calculations and Numbering Scheme Used
Organometallics, Vol. 29, No. 11, 2010 2497
The lowest unoccupied spin orbital or LUSO(β) has, in
each case, a strong metallic dxz metallic character which also
presents a non-negligible π pyridine character in the case
of 1b-Hþ and 1c-Hþ (see Figure 6). The LUSO(R) and
LUSOþ1(R) and the corresponding LUSOþ1(β) and
LUSOþ2(β) are strongly metallic in character and corre-
spond to the well-known “eg*” orbitals of a metallic system
in a pseudo-Oh symmetry. It has to be noted that this
electronic structure is notably affected by a 90ꢀ rotation of
the metallic fragment around the acetylide axis, in line with
the rather large differences in energy found between both
arrangements.
The ionization potentials have been derived for 1a-c-H
(Table 2) considering the most stable conformations for each
redox state. The order for the ionization potentials computed
for 1a-c-H (1a-H . 1b-H > 1c-H) does not follow the order
of the redox potentials experimentally found for 1a-c (1a .
1c > 1b). While specific solvation or electrostatic interac-
tions might explain this discrepancy for the last two isomers,
these data nevertheless clearly confirm that oxidation of 1a
should be significantly more difficult than for the two other
isomers, as experimentally observed.
The atomic spin densities are given in Table 5 for all
studied conformations. Very close spin distributions are
found for the two perpendicular conformations of 1b-Hþ
and 1c-Hþ. In contrast, the spin distributions in the parallel
conformations sensibly differ for all three Fe(III) model
complexes. Nevertheless, this configuration is significantly
higher in energy compared to the perpendicular arrange-
ments (>10 kJ mol-1) and is thus less representative. The
ESR g tensors were calculated for the cationic system in all
arrangements (Supporting Information). They are given in
Table 2 for the most stable structures (perpendicular). The
giso value is in each case smaller than the experimental
measurements (∼2.05-2.07 compared to ∼2.16), and the
anisotropy is less important (∼0.1-0.2 compared to ∼0.5).
These systematic deviations are mainly due to the simplifica-
tion of the ligands in the coordination sphere of the metal,25b
an effect also apparent in the lower atomic spin-density
values found for 1a-Hþ, 1b-Hþ, and 1c-Hþ (see later). The
poor influence of the position of the nitrogen atom in the six-
membered ring on the EPR g tensor experimentally stated
(Table 2) is, however, well reproduced.
several geometries corresponding to the various parallel and
perpendicular conformations that the pyridyl group can
adopt. Two conformations were considered for the 1a-Hþ
complex: one where the aryl plane and the ethane bridge of
the dpe ligand is roughly parallel, 1a -Hþ, and one where the
plane of the functional aryl ring roughly bisects the dpe and
C5H5 ligands, 1a^-Hþ (Chart 3). Three conformations were
considered for 1b-Hþ and 1c-Hþ (the nitrogen atom being
out the aryl axis): a “parallel” conformation (1b -Hþ and 1c -
Hþ) and two “perpendicular” conformations; one with the
nitrogen atom pointing toward the cyclopentadienyl ligand
(1b^1-Hþ and 1c^1-Hþ) and another one with the nitrogen
atom pointing away from it (1b^2-Hþ and 1c^2-Hþ).
The “perpendicular” conformations of these model com-
pounds were optimized without symmetry constraints, while
the parallel conformation was imposed by constraining the
involved dihedral angles, the geometry optimization process
leading to one of the perpendicular structures. The parallel
conformations are actually ca. 0.1-0.2 eV higher in energy
than the most stable perpendicular one (Table 5). Surpris-
ingly, a sizable difference in stability (ca. 0.1 eV) was found
between 1c^1-Hþ and 1c^2-Hþ, revealing a non-negligible
(electronic) influence of the nitrogen position in the ortho
pyridyl complex. Notably, no such difference in stability was
found in the meta isomer.
Energies of the frontier molecular spin-orbitals (SOs) for
1a-Hþ, 1b-Hþ, and 1c-Hþ complexes in their most stable
arrangement are shown in Figure 6. The presence of the
nitrogen atom in the aromatic cycle modifies the usual
energy scheme of the highest occupied spin-orbitals found
in the series of 3-Xþ arylacetylide compounds previously
studied. The lone pair of the nitrogen atom is energetically
close to the “t2g” set of metallic orbitals (the discussion is
based on a spin-restricted scheme for sake of clarity). Thus,
depending on the pseudosymmetry of the nitrogen lone pair
(position of the nitrogen atom in the conjugated ring), it can
be found almost unchanged, as in 1a-Hþ, where it has a axial
symmetry, or in an in-phase and out-of-phase combinations
with the metallic “t2g” orbital of the same pseudosymmetry.
This is illustrated in Table S5 of the Supporting Information
by the plots of the frontiers SOs for each system.
Discussion
Electronic Structures of 1a-c[PF6] and 2a,b[PF6]. Accord-
ing to the present study, 1a-c[PF6] and 2a,b[PF6] constitute
new examples of Fe(III) organometallic metal-centered
radical cations.15 Comparison of their ESR (g, Δg; Table 2),
5
)
redox (Fe(III)/Fe(II) potentials; Table 2), and infrared (νCtC
signatures with those previously determined for their 3-X[PF6]
analogues reveals that the electronic effect of the pyridylalk-
ynyl ligands is comparable to that exerted by 4-phenylalkynyl
ligands functionalized by strongly electron-withdrawing
groups, such as for instance 3-NO2[PF6] or 3-CN[PF6].8,11
The metal-centered character of these radicals evidenced by
€
Mossbauer and ESR spectroscopy is well rationalized by
DFT computations on 1a-Hþ, 1b-Hþ, and 1c-Hþ, which also
indicate that sensibly similar spin densities are present on the
metal center of the three model complexes, regardless of the
appended pyridyl group (Table 4). Experimentally, the metal-
lic character of the unpaired electron is also clearly indicated by

2498 Organometallics, Vol. 29, No. 11, 2010
Paul et al.
Figure 6. Energy diagram of the frontier spin-orbitals of the optimized systems 1a-Hþ, 1b-Hþ, and1c-Hþ in their most stable conformations.
The percentage of metallic character is indicated for the R SOs. The two highest occupied R SOs are plotted (isocontour 0.05 [e/bohr3]1/2).
Scheme 2. Comparison between the Contact Hyperfine Coupling Constants (in G) of Selected Hydrogen Atoms (AH)con of 1a-cþ and
2a,bþ and of 3-NO2þ, 3-CNþ, and 3-Hþ (insert)
the relatively large asymmetry of the rhombic ESR spectra
observed for 1a-c[PF6] (Δg ≈ 0.53 ( 0.01).8 Notably, this
asymmetry is not very sensitive to the position of the nitrogen
in the pyridyl heterocycle (Δg[1aþ] ≈ Δg[1bþ] e Δg[1cþ]), in
spite of the different nature/ordering of some of the frontier
MOs presently evidenced for each isomer 1aþ, 1bþ, and 1cþ. A
more apparent effect of the different electronic structures of
these isomers is that the change in the ESR anisotropy along
the 1a-cþ series is not related to the shifts between the Fe(III/
II) potential redox potentials between the corresponding
compounds (Eꢀ[1aþ] g Eꢀ[1cþ] g Eꢀ[1bþ]), as previously
observed across the 3-Xþ series.8 A similar statement can also
be made when the ESR anisotropies and ionization potentials
computed for the model complexes by DFT are considered
(Table 2). However, between similar isomers such as the meta
pyridyl complexes 1b[PF6] and 2a,b[PF6], this proportionality
is restored and the ESR anisotropy increases in the order of the
oxidation potentials (Table 2). Actually, this electronic sub-
stituent effect is reminiscent ofthatpreviously observedamong
3-H[PF6], 3-Cl[PF6], and 3-Br[PF6], albeit in a slightly less
marked way.8 Likewise, this effect can be related to an increase
in the metallic character of the radical upon depressing the
electron-releasing capabilities of the arylacetylide ligand by
appending an electron-withdrawing substituent to the pyridyl
ring of 1b[PF6] in a para position. Finally, when closer atten-
tion is brought to 1a[PF6], possessing an axially symmetric
pyridyl ligand, strong similarities between the UV-near-IR
spectra of that compound and those of 3-NO2[PF6] or 3-CN-
[PF6] are found.8 This further suggests that the low-lying exci-
ted states of all these Fe(III) complexes have rather similar
energies and exhibit comparable transition moments, as expected
for excited states involving (partly) forbidden LF transitions.
Spin Distribution in 1a-c[PF6] and 2a,b[PF6]. As revealed
by the DFT calculations, most of the spin density resides on
the metal center and a significant part also on the β-acetylide
carbon atom (Table 5). However, a sizable positive spin
density (up to ca. 10% of that present on the metallic center)
is also present on the pyridyl rings. A precise account of the
changes in spin distribution taking place on the pyridyl
fragments between 1a-c[PF6] and 2a,b[PF6] can be gained
1
from H NMR studies (Table 4).23,24 Indeed, according to
McConnell (eq 4), the proton contact hyperfine coupling
constants (Scheme 2) are proportional to the spin densities on
the neighboring carbon atoms. From these experimental data,
€
it appears more plainly than from ESR or Mossbauer data that
€
(23) Kaupp, M.; Kohler, F. H. Coord. Chem. Rev. 2009, 253, 2376–
2386.
(24) Bertini, I.; Luchinat, C.; Parigi, G. Solution NMR of Paramag-
netic Molecules. Application to Metallobiomolecules and Models; Else-
vier: Amsterdam, 2001.

Article
Organometallics, Vol. 29, No. 11, 2010 2499
the positive spin density delocalized on the pyridylethynyl
ligand of 1a-c[PF6] depends primarily on the position of the
nitrogen in the heterocycle. Unsurprisingly, the order found
(meta > ortho > para) corresponds to the electron-releasing
capability of the pyridyl ring previously evidenced by electro-
chemistry (Table 2). The positive spin densities found on the
(hetero)aryl group for 1a-c[PF6] are lower than those pre-
viously found for 3-H[PF6], those for 1a[PF6] closely
resembling the densities previously found for 3-NO2-
[PF6].7 Interestingly, as mentioned above, the presence of
a para halogen substituent on the pyridyl heterocycle does
not seem to modify strongly the spin distribution on that
fragment (Table 4). A marginal decrease of the positive
spin density on Cd, Ce, and Cf is indicated by the isotropic
shifts of Ha, Hb, and Hd, respectively, among the 3-pyridyl
compounds (the order found is 2a[PF6] ≈ 2b[PF6] <
1b[PF6]). This observation indicates that the mesomeric
contribution due to the halogen substituent is not dominant
over the inductive effect of the electropositive nitrogen atom
within the pyridyl ring (Scheme 3). In line with related
studies,8,25 this indicates that the electronic influence of the
nitrogen atom is in part inductively transmitted from the
(functional) pyridyl rings to the metal center.
Figure 7. Correlation of spin densities determined by 1H NMR
for selected carbon atoms of the pyridyl group for 1a[PF6],
1b[PF6]2, and 1c[PF6] vs spin densities computed for the same
atoms by DFT for 1a-Hþ, 1b-Hþ, and 1c-Hþ. Vertical error bars
represent the estimated uncertainty on the calculated spin
densities ((0.01 e).
Valence Bond (VB) Description of 1a-c[PF6]. When the
experimental (Table 4) and theoretical (Table 5) spin dis-
tributions are analyzed in terms of valence bond (VB)
structures (Scheme 3), it appears that the VB mesomers D1
and E3 have a negligible weight in the bonding description of
1a[PF6] and 1c[PF6]. This happens in 1aþ and 1cþ because
the unpaired electron does not like to be strongly localized on
the pz orbital of the nitrogen atom due to the stronger
electronegativity of nitrogen relative to carbon. In contrast,
for 1bþ and 2a,bþ, this radical cannot be located on nitrogen
due to the symmetry of the SOMO. Accordingly, in these
compounds, the spin densities found suggest that the VB
structures C2-E2 participate with quite comparable weights
in the bonding description. The slightly greater spin density
present on the pyridyl group could explain the greater kinetic
instability found for 1b[PF6] relative to 1a[PF6] and 1c[PF6].
Note that according to the available structural data for 2a,
b[PF6] the weight of VB mesomers Bn-En remains very small
in the bonding description. The latter is obviously dominated
by the VB An mesomer (n = 1-3), in line with the strong
localization of the electronic vacancy on the iron atom evi-
denced by DFT. The dominance of A1-3 is also suggested by
Regarding the experimental estimates of the spin densities
found by NMR, a fair qualitative match is found with
computed spin densities for 1a-Hþ, 1b-Hþ, and 1c-Hþ in
their most stable (perpendicular) conformation.23 Indeed,
considering a 0.01 e uncertainty on the DFT values, correct
signs and relative magnitudes are retrieved by DFT for the
atomic spin densities on the carbon atoms of the pyridyl
ring (Figure 7). These spin densities are somewhat over-
estimated in the simplified model compounds used in the
computations, but similar statements have been made
before.7 Considering the significant differences in spin den-
sity predicted by DFT between the parallel and perpendicu-
lar conformations, the experimental spin densities found for
1a-c[PF6] suggest that essentially the perpendicular con-
formation of these Fe(III) radicals is present in solution at
ambient temperature. This is indeed possible, since the latter
corresponds to the most stable conformer for the model
compounds (Supporting Information). Note that previous
computations on 3-NO2þ suggest that these energetic differ-
ences in stability might be less pronounced for the real
compounds, due to the greater steric interactions taking
place between the (hetero)aryl group and the C5Me5 frag-
ment in the perpendicular conformations.7 However, the
€
the infrared and the Mossbauer data. Indeed, while previous
infrared measurements indicated only a modest shift of the
triple-bond stretch to lower wavenumbers upon oxidation of
1a-c in solution (ΔνCtC <60 cm-1),5,11 the present study
1
good Curie dependences stated for the various H NMR
€
reveals that these compounds exhibit a Mossbauer signature
isotropic shifts (Figure 5) further suggest that no conforma-
tional change takes place in the 100-300 K temperature
range for 1a-c[PF6]. Indeed, according to the DFT calcula-
tions, such a phenomenon should result in a marked change
in the spin distribution on the pyridine fragments, which in
turn would induce a departure from the good Curie behavior
observed.26
with a QS characteristic of Fe(III) alkynyl complexes with-
out any indication of a marked cumulenic character.16 This
would not have been the case if the contribution of Bn-Dn
(n = 1-3) was more pronounced.
Conclusion. This work reveals that the pyridylethynyl
Fe(III) radicals 1a-c[PF6] and 2a,b[PF6] overall resemble
their nonheterocyclic congeners 3-Xþ, which bear a strongly
electron-withdrawing substituent in a para position. In 1a-
c[PF6], it is the nitrogen atom present in the pyridyl rings
which qualitatively replaces this substituent, by virtue of its
large electronegativity. However, a closer inspection of the
various spectroscopic and redox signatures of these organo-
metallic radicals reveals that the electronic substituent effect
of the heterocycle cannot be reduced to a single parameter
across the 1a-c[PF6] series. Indeed, the heteroaryl rings
appear to influence in different ways the characteristic
(25) (a) Paul, F.; Ellis, B. J.; Bruce, M. I.; Toupet, L.; Roisnel, T.;
Costuas, K.; Halet, J.-F.; Lapinte, C. Organometallics 2006, 25, 649–665.
(b) Gauthier, N.; Tchouar, N.; Justaud, F.; Argouarch, G.; Cifuentes, M. P.;
Toupet, L.; Touchard, D.; Halet, J.-F.; Rigaut, S.; Humphrey, M. G.; Costuas,
K.; Paul, F. Organometallics 2009, 28, 2253–2266.
(26) See for instance: (a) Walter, M. D.; Berg, D. J.; Andersen, R. A.
Organometallics 2007, 26, 2296–2307. (b) Walter, M. D.; Berg, D. J.;
Andersen, R. A. Organometallics 2006, 25, 3228–3237. (c) Schultz, M.;
Boncella, J. M.; Berg, D. J.; Tilley, T. D.; Andersen, R. A. Organometallics
2002, 21, 460–472.

2500 Organometallics, Vol. 29, No. 11, 2010
Paul et al.
Scheme 3. VB Representation of the Substituent-Dependent Delocalization of the Unpaired Electron in 1a-cþ Complexes vs That in 3-Hþ
([Fe] = (η2-dppe)(η5-C5Me5)Fe)
electronic/bonding parameters probed by these experimental
techniques for each isomer.
Importantly, the present study reveals that beside the many
spectroscopic techniques traditionally used to characterize
1
radical cations, H NMR constitutes an interesting comple-
at the LCC (Toulouse, France). They were obtained by using a
constant acceleration spectrometer previously described
with a 50 mCi 57Co source in a Rh matrix.28 The sample
temperature was controlled by an Oxford MD306 cryostat
and an Oxford ITC4 temperature controller. Computer fit-
€
ting of the Mossbauer data to Lorentzian line shapes was
€
mentary method. Indeed, Mossbauer, ESR, and UV-vis-
carried out with a previously reported computer program.29
The isomer shift values are reported relative to iron foil at
298 K and are not corrected for the temperature-dependent
near-IR are poorly suited to experimentally investigate small
electronic differences in spin distribution on the ligand for
these metal-centered organometallic radicals. In contrast, 1H
NMR proves much more sensitive to the electronic perturba-
tions localized on the pyridyl group and allows unambiguous
distinction between these paramagnetic isomers. Further-
more, this technique provides an accurate picture of the spin
distribution on the heterocyclic ring, which is qualitatively
reproduced by DFT calculations performed on simpler mod-
els. On the basis of these results, NMR spectroscopy appears
perfectly suited to empirically address slight electronic
changes such as those brought by coordination of the nitrogen
lone pair to a Lewis-acidic site and will be used accordingly in
future investigations aimed at examining complexation reac-
tions of these Fe(III) metallo-ligands.
€
second-order Doppler shift. The Mossbauer sample cell
consists of a 2 cm diameter cylindrical Plexiglas holder. The
complexes 1a-c[PF6]25 and 2a,b3 were obtained as previously
reported.
General Procedure for the Synthesis of the Mononuclear Fe-
(III) Alkynyl Complexes 2a[PF6] and 2b[PF6]. A 0.95 equiv
amount of [Fe(η5-C5H5)2][PF6] (0.040 g; 0.120 mmol) was added
to a solution of the corresponding Fe(II) precursor (0.125 mmol)
in 10 mL of dichloromethane. An instantaneous darkening of
the solution was observed. Stirring was maintained for 1 h at
room temperature, and the solution was concentrated in vacuo
to approximately 2 mL. Addition of 25 mL of n-pentane allowed
precipitation of a dark solid. Decantation and subsequent
washing with 3 ꢁ 3 mL portions of toluene followed by 3 ꢁ 3
mL diethyl ether and drying under vacuum yielded the desired
complex. Both compounds were crystallized as dark rhombic
cubes by slow diffusion of n-pentane into a dichloromethane
solution of the compound.
Experimental Section
General Data. All manipulations were carried out under an
inert atmosphere. Solvents or reagents were used as follows:
Et2O and n-pentane, distilled from Na/benzophenone; CH2Cl2,
distilled from CaH2 and purged with argon, opened/stored
under Ar. The [(η5-C5H5)2Fe][PF6] ferrocenium salt was pre-
pared by previously published procedures.27 Transmittance-
FTIR spectra were recorded using a Bruker IFS28 spectrometer
(400-4000 cm-1). UV-visible and near-infrared (near-IR)
spectra were recorded using a a Cary 5000 spectrometer. All
NMR experiments were performed on a Bruker AVANCE 500
instrument operating at 500.15 MHz for 1H and 125.769 MHz
for 13C, with a 5 mm broad-band observe probe equipped with a
z-gradient coil (Supporting Information). ESR spectra were
recorded on a Bruker EMX-8/2.7 (X-band) spectrometer. Cyclic
voltammograms were recorded using a EG&G potentiostat
(M.263) on platinum electrodes referenced to an SCE electrode
and were calibrated with the [FcH]/[FcH]þ couple taken at
[(η2-dppe)(η5-C5Me5)FeCtC{5-(2-C6H3N1Cl1)}][PF6] (2a[PF6]).
Yield: 83%. Color: dark gray. IR (ν,KBr,cm-1): 2009 (vw), 1945 (m,
CtC). UV-Vis-near-IR (CH2Cl2; λmax (ε/103 dm3 M-1 cm-1)):
312 (sh, 13.45); 380 (sh, 2.81); 560 (1.52); 635 (1.46); 1900 (0.08).
[(η2-dppe)(η5-C5Me5)FeCtC{5-(2-C6H3N1Br1)}][PF6] (2b[PF6]).
Yield: 84%. Color: dark gray. IR (ν,KBr,cm-1): 2004 (vw), 1946 (w,
CtC). UV-Vis-near-IR (CH2Cl2; λmax (ε/103 dm3 M-1 cm-1)):
314 (sh, 14.32); 380 (sh, 3.39); 562 (1.77); 636 (1.66); 1900 (0.08).
Crystallography. Crystals of 2a,b[PF6] were obtained as des-
cribed above. The samples were studied on a Sapphire 3
Xcalibur CCD instrument with graphite-monochromated Mo
KR radiation. The cell parameters were obtained with Denzo
and Scalepack with 10 frames (ψ rotation: 1ꢀ per frame).30 The
(28) Boinnard, D.; Boussekssou, A.; Dworkin, A.; Savariault, J.-M.;
Varret, F.; Tuchagues, J.-P. Inorg. Chem. 1994, 33, 271.
(29) Varret, F.; Mariot, J.-P.; Hamon, J.-R.; Astruc, D. Hyperfine
Interact. 1988, 39, 67.
0.46 V in CH2Cl2.27 The 57Fe Mossbauer spectra were recorded
€
(30) Otwinowski, Z.; Minor, W. In Methods in Enzymology; Carter,
C. W., Sweet, R. M., Eds.; Academic Press: London, 1997; Vol. 276,
pp 307-326.
(27) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910.

Article
Table 6. Crystal Data and Data Collection and Refinement
Parameters for 2a[PF6] and 2b[PF6]
Organometallics, Vol. 29, No. 11, 2010 2501
1a-Hnþ, and 1a-Hnþ (n = 0, 1), respectively, were used in order
to reduce computational effort. Electron correlation was treated
within the local density approximation (LDA) in the Vos-
ko-Wilk-Nusair parametrization.37 The nonlocal corrections
of Becke38 and of Perdew39 were added to the exchange and
correlation energies, respectively. The numerical integration
procedure applied for the calculations was developed by te
Velde et al.40 The standard ADF TZP basis set was used: i.e.,
a triple-ζ Slater-type orbital (STO) basis for Fe 3d and 4s and a
single-ζ function for 4p of Fe. A triple-ζ STO basis set was
employed for H 1s, for 2s and 2p of C and N, and for 3s and 3p of
P extended with a single-ζ polarization function (2p for H; 3d for
C, N, and P). Orbitals up to 1s, 2p, and 4p were kept frozen for
C and N, for P, and for Fe, respectively. Full geometry optimi-
zations (assuming C1 symmetry) were carried out on each
complex, using the analytical gradient method implemented
by Versluis and Ziegler.41 Geometry optimization convergence
criteria were more drastic than default criteria (energy change
2a[PF6]
2b[PF6]
formula
C
43H43Cl1N1-
C43H43Br1N1-
P2Fe1,PF6
915.45
150(2)
monoclinic
P21/c
17.5014(5)
12.8073(4)
18.3184(4)
90.0
94.169(2)
90.0
P2Fe1,PF6
870.99
120(2)
monoclinic
P21/c
17.3677(5)
12.6604(4)
18.3312(4)
90.0
93.827(2)
90.0
fw
temp (K)
cryst syst
space group
˚
a (A)
˚
b (A)
˚
c (A)
R (deg)
β (deg)
γ (deg)
3
˚
V(A )
Z
4021.71(19)
4
1.438
4095.1(2)
4
1.485
Dcalcd (g cm-3
)
cryst size (mm)
F(000)
diffractometer
0.30 ꢁ 0.25 ꢁ 0.12
0.24 ꢁ 0.20 ꢁ 0.14
˚
<0.0005 hartree, atomic position displacement <0.005 A).
1796
1868
For the cationic systems, the parallel configuration was ob-
tained by restraining the torsion angle between the plane of
the pyridyl group and the P-Fe-P plane. Spin-unrestricted
calculations were performed for all the open-shell systems
considered. The Cartesian coordinates of the optimized geome-
tries are given as Supporting Information. The EPR g tensor
calculations were calculated as implemented in ADF.42 In this
case, the functional PBE was used,43 and relativistic corrections
were taken into account using the ZORA (zeroth order regular
approximation) spin-orbit Hamiltonian with the appropriate
basis set.44 Representations of the molecular structures, orbi-
tals, and spin densities were done using MOLEKEL4.345 and
ADF-GUI.46
CCD Saphire
3 Xcalibur
Mo KR
CCD Saphire
3 Xcalibur
Mo KR
radiation
abs coeff (mm-1
θ range (deg)
hkl range
)
0.621
0.711
2.75-27.00
-22 to þ22
-15 to þ14
-23 to þ23
51 394
2.59-27.00
-21 to þ21
-16 to þ16
-23 to þ23
54 715
total no. of rflns
no. of obsd rflns
6852
8474
(I > 2σ(I))
no. of
0/496
0/496
restraints/params
final R
Rw
R indices (all data)
Rw (all data)
goodness of fit/F2 (Sw)
0.035
0.098
0.046
0.101
0.988
0.039
0.066
0.101
0.072
0.693
Acknowledgment. A. Mari (LCC, Toulouse, France) is
€
acknowledged for Mossbauer measurements. The
MNERT (Ph.D. grant to F.M.), Region Bretagne (Ph.
D. grant to G.d.C.), CNRS and ATCO (Ph.D. grant to
S.L.S.) are acknowledged for financial support. The
theoretical calculations were performed using HPC
resources from GENCI CINES and GENCI IDRIS
(Grant 2009-[x2010080649]). Dr. Katy Green is acknow-
ledged for editorial help.
data collection31 (Table 6) provided reflections for 2a,b[PF6].
Subsequent data reduction with Denzo and Scalepack30 gave
the independent reflections (Table 6). The structures were solved
with SIR-97, which revealed the non-hydrogen atoms.32 After
anisotropic refinement, the remaining atoms were found in
Fourier difference maps. The complete structures were then
refined with SHELXL9733 by the full-matrix least-squares
technique (use of F2 magnitude; x, y, z, βij for Fe, P, C, N,
and/or O atoms, x, y, z in riding mode for H atoms with
variables N(var), observations and w used as defined in
Table 6). Atomic scattering factors were taken from the
literature.34 ORTEP views of 2a,b[PF6] were obtained with
PLATON98.35
Supporting Information Available: Text, figures, and tables
giving additional NMR data for 1a-c[PF6], derivation of spin
densities for selected aromatic protons based on isotropic shifts,
changes for selected bond lengths and angles between 2a,b and
2a,b[PF6], Cartesian coordinates and total bonding energies of
DFT Calculations. DFT calculations were carried out using
the Amsterdam Density Functional (ADF) program.36 The
model compounds [(η2-dpe)(η5-C5H5)FeCtC(x-C6H3N)]nþ
(37) Vosko, S. D.; Wilk, L.; Nusair, M. Can. J. Chem. 1990, 58, 1200–
1211.
(38) (a) Becke, A. D. J. Chem. Phys. 1986, 84, 4524–4529. (b) Becke,
A. D. Phys. Rev. A 1988, 38, 3098–3100.
(dpe =1,2-diphosphinoethane; x = 4, 3, 2) labeled 1a-Hnþ
,
(39) Perdew, J. P. Phys. Rev. B 1986, 33, 8822–24.
(40) te Velde, G.; Baerends, E. J. J. Comput. Phys. 1992, 99, 84–
98.
(41) Versluis, L.; Ziegler, T. J. Chem. Phys. 1988, 88, 322–328.
(42) (a) Schreckenbach, G.; Ziegler, T. J. Phys. Chem. A 1997, 101,
3388–3399. (b) van Lenthe, E.; van der Avoird, A.; Wormer, P. E. S. J. Chem.
Phys. 1997, 107, 2488–2498.
(43) (a) Hammer, B.; Hansen, L. B.; Norskov, J. K. Phys. Rev. B
1999, 59, 7413–7421. (b) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev.
Lett. 1996, 77, 3865–3868.
(31) Kappa CCD Software; Nonius BV, Delft, The Netherlands, 1999.
(32) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G.;
Giacovazzo, C.; Guagliardi, A.; Moliterni, A. G. G.; Polidori, G.;
Spagna, R. J. Appl. Chem. 1998, 31, 74–77.
(33) Sheldrick, G. M. SHELX97-2: Program for the Refinement of
€
€
Crystal Structures; University of Gottingen, Gottingen, Germany, 1997.
(34) International Tables for X-ray Crystallography; Kynoch Press
(present distributor D. Reidel, Dordrecht, The Netherlands): Birmingham,
U.K., 1974; Vol. IV.
(35) Spek, A. L. PLATON: A Multipurpose Crystallographic Tool;
Utrecht University, Utrecht, The Netherlands, 1998.
(44) (a) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys.
1993, 99, 4597–4610. (b) van Lenthe, E.; Baerends, E. J.; Snijders, J. G.
J. Chem. Phys. 1994, 101, 9783–9792. (c) van Lenthe, E.; Ehlers, A. E.;
Baerends, E. J. J. Chem. Phys. 1999, 110, 8943–8953.
(36) (a) SCM, ADF 2007.01 ed.; Theoretical Chemistry, Vrije Univer-
siteit, Amsterdam, The Netherlands, 2007. SCF, ADF 2008.01; Theoretical
Chemistry, Vrije Universiteit, Amsterdam, The Netherlands, 2008. (b) te
Velde, G.; Bickelhaupt, F. M.; Fonseca Guerra, C.; van Gisbergen, S. J. A.;
Baerends, E. J.; Snijders, J.; Ziegler, T. Theor. Chem. Acc. 2001, 22,
931–967.
€
€
(45) Flukiger, P.; Luthi, H. P.; Portmann, S.; Weber, J. MOLEKEL4.3;
Swiss Center for Scientific Computing (CSCS), 2001.
(46) ADF-GUI; SCM, Amsterdam, The Netherlands, 2008; www.
scm.com.

2502 Organometallics, Vol. 29, No. 11, 2010
Paul et al.
all DFT optimized geometries for 1a-H0/þ, 1b-H0/þ, and
1c-H0/þ, contour plots of the frontier spin-orbitals of 1a-c-
Hþ, and calculated g tensors of 1a-c-Hþ in different conforma-
tions and CIF files giving crystallographic data for 2a,b[PF6].
This material is available free of charge via the Internet at
http://pubs.acs.org. Final atomic positional coordinates, with
estimated standard deviations, bond lengths and angles, and
anisotropic thermal parameters, have also been deposited at the
Cambridge Crystallographic Data Centre and were allocated
the deposition numbers CCDC 730944 and CCDC 745013,
respectively. Ordering information is given on any current
masthead page.