Pronounced effects of crystal structures on intramolecular electron transfer in mixed-valence biferrocenium cations: Structural, EPR, and 57Fe Mo?ssbauer characteristics

Article abstract of DOI:10.1021/om9908884

Crystallographic properties play an important role in controlling the rate of electron transfer in mixed-valence 1′,1?-disubstituted-biferrocenium triiodide salts. The X-ray structure of neutral 1′,1?-dinaphthylmethylbiferrocene has been determined at 298 K. The corresponding mixed-valence 1′,1?-dinaphthylmethylbiferrocenium triiodide exhibits two crystalline morphologies at 298 K. Dark crystals, formed when a CH₂Cl₂ solution of triiodide salt was allowed to evaporate slowly, crystallize in space group P1. Dark crystals, obtained when a layer of hexane was allowed to slowly diffuse into a CH₂Cl₂ solution of triiodide salt, crystallize in monoclinic space group P2₁/n. The observations of the structural characteristics of 1′,1?-dinaphthylmethylbiferrocenium triiodide are also consistent with our Mo?ssbauer studies. The cation with space group P1 shows two doublets in the variable-temperature Mo?ssbauer spectra at temperatures below 100 K. An increase of temperature causes the two doublets to move together, resulting in an average-valence doublet at 130 K. At 300 K, the spectrum of this sample shows a single doublet which is characteristic of a valence-detrapped cation in which the electron-transfer rate exceeds ～10⁸ s^-1. On the other hand, the cation with P2₁n phase exhibits a Mo?ssbauer spectrum characteric of a valence-trapped cation at 300 K. Obviously, the intramolecular electron-transfer rate is quite sensitive to environment perturbations caused by different crystal packing arrangements. At 77 K, the EPR spectrum of cation with P1 phase is clearly a typical axial-type spectrum with g_∥ = 3.16 and g_⊥ = 1.91. Surprisingly, the EPR spectrum of the cation with P2₁/n phase consists of two g_∥ signals (3.67 and 2.85) and two g_⊥ signals (2.01 and 1.79). We suggest that the origin arises from the interaction of spin-spin exchange resulting from a dipole-dipole interaction that develops between cations. The syntheses, characterizations, and physical properties of mixed-valence 1′,1?-di(4-biphenylmethyl)biferrocenium and 1′,1?-di(9-anthracenylmethyl)biferrocenium triiodide salts are also described.

Full text of DOI:10.1021/om9908884

1096
Organometallics 2000, 19, 1096-1106
P r on ou n ced Effects of Cr ysta l Str u ctu r es on
In tr a m olecu la r Electr on Tr a n sfer in Mixed -Va len ce
Bifer r ocen iu m Ca tion s: Str u ctu r a l, EP R, a n d 57F e
Mo1ssba u er Ch a r a cter istics
Teng-Yuan Dong,*,1 Pei-Har Ho, Xiao-Qian Lai, Zhi-Wei Lin, and
Kuan-J iuh Lin2
Department of Chemistry, National Sun Yat-Sen University, Kaohsuing, Taiwan, and the
Institute of Chemistry, Academia Sinica, Nankang, Taipei, Taiwan
Received November 5, 1999
Crystallographic properties play an important role in controlling the rate of electron
transfer in mixed-valence 1′,1′′′-disubstituted-biferrocenium triiodide salts. The X-ray
structure of neutral 1′,1′′′-dinaphthylmethylbiferrocene has been determined at 298 K. The
corresponding mixed-valence 1′,1′′′-dinaphthylmethylbiferrocenium triiodide exhibits two
crystalline morphologies at 298 K. Dark crystals, formed when a CH2Cl2 solution of triiodide
salt was allowed to evaporate slowly, crystallize in space group P1h. Dark crystals, obtained
when a layer of hexane was allowed to slowly diffuse into a CH2Cl2 solution of triiodide salt,
crystallize in monoclinic space group P21/n. The observations of the structural characteristics
of 1′,1′′′-dinaphthylmethylbiferrocenium triiodide are also consistent with our Mo¨ssbauer
studies. The cation with space group P1h shows two doublets in the variable-temperature
Mo¨ssbauer spectra at temperatures below 100 K. An increase of temperature causes the
two doublets to move together, resulting in an average-valence doublet at 130 K. At 300 K,
the spectrum of this sample shows a single doublet which is characteristic of a valence-
detrapped cation in which the electron-transfer rate exceeds ∼108 s-1. On the other hand,
the cation with P21/n phase exhibits a Mo¨ssbauer spectrum characteric of a valence-trapped
cation at 300 K. Obviously, the intramolecular electron-transfer rate is quite sensitive to
environment perturbations caused by different crystal packing arrangements. At 77 K, the
EPR spectrum of cation with P1h phase is clearly a typical axial-type spectrum with g| )
3.16 and g ) 1.91. Surprisingly, the EPR spectrum of the cation with P21/n phase consists
of two g| signals (3.67 and 2.85) and two g signals (2.01 and 1.79). We suggest that the
origin arises from the interaction of spin-spin exchange resulting from a dipole-dipole
interaction that develops between cations. The syntheses, characterizations, and physical
properties of mixed-valence 1′,1′′′-di(4-biphenylmethyl)biferrocenium and 1′,1′′′-di(9-anthra-
cenylmethyl)biferrocenium triiodide salts are also described.
In tr od u ction
revealed that the environment surrounding a cation is
perhaps the most important factor in determining the
rate of intramolecular electron transfer.8-12 Compounds
1 give unusual temperature-dependent Mo¨ssbauer spec-
The study of intramolecular electron transfer in
mixed-valence complexes has given insight into the
factors that affect rates of electron transfer in solution
redox processes, solid-state materials, and biological
electron-transport chains.3-6 One of the most thoroughly
studied mixed-valence compounds is biferrocenium tri-
iodide salts.7-25 Studies of electron transfer in the series
of mixed-valence biferrocenium compounds 1 (Chart 1)
(10) Iijima, S.; Saida, R.; Motoyama, I.; Sano, H. Bull. Chem. Soc.
J pn. 1981, 54, 1375.
(11) Nakashima, S.; Katada, M.; Motoyama, I.; Sano, H. Bull. Chem.
Soc. J pn. 1987, 60, 2253.
(12) Kai, M.; Katada, M.; Sano, H. Chem. Lett. 1988, 1523.
(13) Dong, T.-Y.; Schei, C. C.; Hsu, T. L.; Lee, S. L.; Li, S. J . Inorg.
Chem. 1991, 30, 2457.
(14) Dong, T.-Y.; Chou, C. Y. J . Chem. Soc., Chem. Commun. 1990,
1332.
(15) Dong, T.-Y.; Schei, C. C.; Hwang, M. Y.; Lee, T. Y.; Yeh, S. K.;
Wen, Y. S. Organometallics 1992, 11, 573.
(1) Department of Chemistry, National Sun Yat-Sen University.
(2) Institute of Chemistry, Academia Sinica.
(3) Day, P. Int. Rev. Phys. Chem. 1981, 1, 149.
(4) Brown, D. Mixed-Valence Compounds, Theory and Applications
in Chemistry, Physics, Geology and Biology; Reidel: Boston, MA, 1980.
(5) Creutz, C. Prog. Inorg. Chem. 1983, 30, 1.
(6) Richardson, D. E.; Taube, H. Coord. Chem. Rev. 1984, 60, 107.
(7) Hendrickson, D. N.; Oh, S. M.; Dong, T.-Y.; Kambara, T.; Cohn,
M. J .; Moore, M. F. Comments Inorg. Chem. 1985, 4, 329.
(8) Dong, T.-Y.; Hendrickson, D. N.; Iwai, K.; Cohn, M. J .; Geib, S.
J .; Rheingold, A. L.; Sano, H.; Motoyama, I.; Nakashima, S. J . Am.
Chem. Soc. 1985, 107, 7996.
(16) Webb, R. J .; Geib, S. J .; Staley, D. L.; Rheingold, A. L.;
Hendrickson, D. N. J . Am. Chem. Soc. 1990, 112, 5031.
(17) Webb, R. J .; Rheingold, A. L.; Geib, S. J .; Staley, D. L.;
Hendrickson, D. N. Angew. Chem., Int. Ed. Engl. 1989, 28, 1388.
(18) Dong, T.-Y.; Kambara, T.; Hendrickson, D. N. J . Am. Chem.
Soc. 1986, 108, 5857.
(19) Dong, T.-Y.; Kambara, T.; Hendrickson, D. N. J . Am. Chem.
Soc. 1986, 108, 4423.
(20) J ang, H. G.; Geib, S. J .; Kaneko, Y.; Nakano, M.; Sorai, M.;
Rheingold, A. L.; Montez, B.; Hendrickson, D. N. J . Am. Chem. Soc.
1989, 111, 173.
(9) Dong, T.-Y.; Hendrickson, D. N.; Pierpont, C. G.; Moore, M. F.
J . Am. Chem. Soc. 1986, 108, 963.
10.1021/om9908884 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 02/24/2000

Biferrocenium Cations
Organometallics, Vol. 19, No. 6, 2000 1097
Ch a r t 1
F igu r e 1. Configurational diagrams. The three-level
system that models the molecule is described by a fictitious
spin S ) 1. The pair of levels in the double well is
associated with Sz ) (1; the single level in the single well
with Sz ) 0.
tra. At temperatures below 77 K they each show two
doublets, one for the Fe(III) site and the other for the
Fe(III) site (electron-transfer rate < ∼108 s-1). In each
case the two doublets move together with increasing the
temperature, eventually, and become a single “average-
valence” doublet. Hendrickson suggested8,9 that the
temperature dependence of the Mo¨ssbauer spectrum is
due to the onset of lattice dynamics associated with the
triiodide counterions and alkyl substituents.
Recently, we reported systematic studies of electron
transfer in the solid state for a series of mixed-valence
biferrocenium cations (2-4), which have led to a wealth
of new experimental results. Major progress has been
made in studying the dependence of electron-transfer
rates on the lattice dynamics and on the cation-anion
interactions, as well as on the structural micromodifi-
cation. We found that the cation-anion van der Waals
interactions between the halide substituent in the
benzyl unit and the neighboring triiodide in the series
of compounds 2 can have a dramatic influence on the
rate of electron transfer.15 More recently, we prepared
a series of polyalkylbiferrocenium triiodide salts (3, 4)
to study the influence of ring tilt between two cyclopen-
tadienyl (Cp) rings on electron transfer and to investi-
gate energetic control of electron-transfer rates.14,22-25
Deviations of the Cp rings from the parallel position
were found to correlate quite well with the critical
temperature for electronic delocalization-localization in
mixed-valence biferrocenium salts.
action may lead to the charge-ordered state of the
mixed-valence crystal and to interesting peculiarities
of the phase transitions. Recently, a Blume-Emery-
Griffith (BEG) Hamiltonian was found to provide a
qualitative explanation for the various thermal behav-
iors of electronic localization exhibited by mixed-valence
biferrocenium derivative salts in the solid state.27,28 The
BEG Hamiltonian uses a fictitious spin S ) 1, associated
with a three-level system which results from the dis-
cretization of the configurational diagram of the biferro-
cenium cation (Figure 1). Such a particular, triple-well-
shaped, diagram is obtained through an extension of the
standard PKS model.29 Most of the Mo¨ssbauer data
concerning biferrocenium salts could be reproduced by
using the S ) 1 model.
Mixed-valence 1′,1′′′-dibenzylbiferrocenium triiodide
was reported to have two distinct crystallographic
phases which show that the electron-transfer rates are
extremely sensitive to changes in crystal lattice.30 Two
different crystalline morphologies (needle and platelike
crystals) of 1′,1′′′-dibenzylbiferrocenium triiodide were
found. The needle crystal in space group P1h has a
Mo¨ssbauer spectrum characteric of a valence-detrapped
electronic structure down to a temperature of 25 K. The
platelike crystal with space group P21/n shows a valence-
trapped Mo¨ssbauer spectrum up to 300 K. However,
because of poor diffraction quality and disorder in the
triiodide anion, a satisfactory refinement of the struc-
ture was not obtained. In our previous communication,31
we reported the first new mixed-valence compound 5a
(Scheme 1) in two crystallographic phases with satisfac-
tory structural refinements. Our observation is an
interesting result that the crystallographic properties
of the structure play an important role in controlling
the rate of electron transfer. The present work will focus
The quantum-mechanical theory of electronic states
of isolated mixed-valence clusters was developed by
Tsukerblat.26 It was shown that the intercluster inter-
(21) Kaneko, Y.; Nakano, M.; Sorai, M.; J ang, H. G.; Hendrickson,
D. N. J . Am. Chem. Soc. 1989, 28, 1067.
(22) Dong, T.-Y.; Lee, T. Y.; Lin, H. M. J . Organomet. Chem. 1992,
427, 101.
(23) Dong, T.-Y.; Chang, C. K.; Huang, C. H.; Wen, Y. S.; Lee, S. L.;
Chen, J . A.; Yeh, W. Y.; Yeh, A. J . Chem. Soc., Chem. Commun. 1992,
526.
(24) Dong, T.-Y.; Huang, C. H.; Chang, C. K.; Wen, Y. S.; Lee, S. L.;
Chen, J . A.; Yeh, W. Y.; Yeh, A. J . Am. Chem. Soc. 1993, 115, 6357.
(25) Dong, T.-Y.; Lee, S. H.; Chang, C. K.; Lin, H. M.; Lin, K. J .
Organometallics 1997, 16, 2773.
(27) Boukheddaden, K.; Linares, J .; Galam, S.; Varret, F. J . Phys.:
Condens. Matter 1993, 5, 1.
(28) Boukheddaden, K.; Linares, J .; Galam, S.; Varret, F. J . Phys.:
Condens. Matter 1993, 5, 5179.
(29) Wong, K. Y.; Schatz, P. N. Prog. Inorg. Chem. 1981, 28, 369.
(30) Webb, R. J .; Dong, T.-Y.; Pierpont, C. G.; Boone, S. R.; Chadha,
R. K.; Hendrickson, D. N. J . Am. Chem. Soc. 1991, 113, 4806.
(31) Dong, T.-Y.; Lai, X. Q.; Lin, K. J . Angew. Chem., Int. Ed. Engl.
1997, 36, 2002.
(26) Koryachenko, A. V.; Klokishner, S. I.; Tsukerbat, B. S. Chem.
Phys. 1991, 150, 295.

1098 Organometallics, Vol. 19, No. 6, 2000
Dong et al.
Ta ble 1. Exp er im en ta l a n d Cr ysta l Da ta for th e
X-r a y Str u ctu r es of 5a a n d 6a
Sch em e 1
5a
5a
6a
formula
mw
cryst syst
space group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
C
42H34Fe2I3
C42H34Fe2I3
1031.1
monoclinic
P21/n
17.286(5)
9.188(7)
23.069(4)
C42H34Fe2
650.40
triclinic
P1h
8.229(2)
8.477(2)
11.202(3)
91.45(1)
103.99(1)
95.74(1)
753.4(3)
1
1.433
0.993
0.70930
22.42
0.0260
0.0703
1031.1
triclinic
P1h
9.836(2)
10.418(2)
10.638(3)
62.83(1)
86.06(1)
69.74(1)
904.5(4)
1
1.893
3.391
0.70930
22.42
0.0354
0.0875
93.44(2)
3658(3)
4
Z
D
calcd, g cm-3
1.873
3.355
0.70930
22.43
0.0348
0.0980
µ, mm-1
λ, Å
θ limits, deg
R1
wR(f 2)
on the spectroscopic measurements for mixed-valence
cations 5a -c. The understanding of electron-transfer
processes in 5a -c provides detailed insight into the
fundamental theoretic aspect of charge ordering and
phase transitions in molecular crystals. Thus the gen-
eral importance of this study is that it is a test case for
the cooperative S ) 1 model. The results of the X-ray
structural work on 5a ,b together with the electrochemi-
cal measurements, variable-temperature Mo¨ssbauer
data, and EPR measurements are presented in this
paper.
F igu r e 2. ORTEP drawing for 6a with 50% thermal
ellipsoids.
the value of 1.65 Å found for ferrocene32 than to the
value of 1.70 Å found for the ferrocenium ion.33 The
dihedral angle between the two least-squares planes of
the Cp rings for the ferrocenyl moiety in 6a is 1.0°, while
the two Cp rings bonded to the iron ion are nearly
eclipsed, with average staggering angle of 1.2°.
Molecu la r Str u ctu r e of 5a . 1′,1′′′-Dinaphthylmeth-
ylbiferrocenium triiodide (5a ) exhibits two crystalline
morphologies at room temperature. Dark crystals in the
triclinic space group P1h were formed when a CH2Cl2
solution of 5a was allowed to evaporate slowly. The
refinement of the structure imposed inversion centers
on both the cation and the triiodide anion. Thus, the
two ferrocenyl moieties are crystallographically equiva-
Resu lts a n d Discu ssion
Molecu la r Str u ctu r e of Neu tr a l Bifer r ocen e 6a .
Details of the X-ray crystal data collections and unit-
cell parameters are given in Table 1. The molecular
structure of 6a is shown in Figure 2, and selected bond
distances and angles are given in Table 2. This neutral
biferrocene exists in a trans conformation with the two
iron ions on opposite sides of the planar fulvalenide
bridge.
In 6a , the average distance from the Fe ion to the C
atoms is 2.042(4) Å. Furthermore, there is no significant
difference between the Fe-Cp distance (1.650(6) Å) and
Fe-fulvalenide distance (1.649(1) Å). Inspection of the
Fe-Cp distances shows that these values are closer to
(32) Seiler, P.; Dunitz, J . D. Acta Crystallogr., Sect. B 1979, 35, 1068.
(33) Mammano, N. J .; Zalkin, A.; Landers, A.; Rheingold, A. L. Inorg.
Chem. 1977, 16, 297.

Biferrocenium Cations
Organometallics, Vol. 19, No. 6, 2000 1099
Ta ble 2. Selected Bon d Dista n ces (Å) a n d An gles
(d eg) of 5a a n d 6a
6a
5a a
5a b
Fe-C(1)
2.060(2)
2.040(2)
2.038(3)
2.040(2)
2.042(2)
2.052(2)
2.037(3)
2.034(3)
2.039(3)
2.034(3)
2.090(5)
2.055(6)
2.039(6)
2.046(6)
2.050(6)
2.074(5)
2.043(5)
2.053(6)
2.064(6)
2.073(6)
Fe-C(2)
Fe-C(3)
Fe-C(4)
Fe-C(5)
Fe-C(6)
Fe-C(7)
Fe-C(8)
Fe-C(9)
Fe-C(10)
Fe(1)-C(1)
Fe(1)-C(2)
Fe(1)-C(3)
Fe(1)-C(4)
Fe(1)-C(5)
Fe(1)-C(6)
Fe(1)-C(7)
Fe(1)-C(8)
Fe(1)-C(9)
Fe(1)-C(10)
Fe(2)-C(11)
Fe(2)-C(12)
Fe(2)-C(13)
Fe(2)-C(14)
Fe(2)-C(15)
Fe(2)-C(16)
Fe(2)-C(17)
Fe(2)-C(19)
Fe(2)-C(20)
I(1)-I(2)
2.053(6)
2.054(7)
2.067(7)
2.056(7)
2.028(7)
2.065(6)
2.046(6)
2.043(6)
2.055(6)
2.043(6)
2.129(6)
2.095(7)
2.047(7)
2.050(7)
2.079(9)
2.126(7)
2.091(7)
2.069(7)
2.110(7)
2.8950(11)
2.9318(10)
2.8936(8)
180.0
I(3)-I(1)
I(2)-I(1)-I(2)c
I(2)-I(1)-I(3)
176.31(2)
a
b
With P1h phase. With P21/n phase. c Symmetry equivalents:
-x, -y, -z.
lent. An ORTEP plot of the molecular structure is given
in Figure 3a, whereas Figure 4a shows an extended
packing arrangement adopted in this P1h phase. Col-
lected in Table 2 are selected bond distances and angles.
Inspection of the iron to Cp ligand bond lengths (Fe-
Cp, 1.668(2) and 1.676(2) Å) shows that both distances
are intermediate between the analogous values for
neutral ferrocene (1.65 Å)32 and the ferrocenium ion
(1.70 Å).33 The average Fe-C distance is 2.059(5) Å,
which is also intermediate between the value for fer-
rocene (2.045(1) Å) and the ferrocenium ion (2.075 Å).
The plane of the Cp ring forms a dihedral angle of 4.4°
with the fulvalenide plane, and the two Cp rings in each
ferrocenyl moiety are not perfectly eclipsed but are
rotated relative to one another by 7.0°. As shown in
Figure 4a, an intermolecular naphthyl-naphthyl (π-
π) interaction (3.648 Å) is observed between neighbor
cations. We believe that this π-π interaction plays a
quite important role in controlling the intramolecular
electron-transfer rate. A detailed discussion will be
presented in the last section.
F igu r e 3. (a) ORTEP drawing for 5a with P1h phase
obtained when a CH2Cl2 solution of 5a was allowed to
evaporate slowly. (b) ORTEP drawing for 5a with P21/n
phase obtained when a layer of hexane was allowed to
slowly diffuse into a CH2Cl2 solution of 5a .
slowly diffuse into a CH2Cl2 solution of 5a . The molec-
ular structure is also shown in Figure 3b, whereas
Figure 4b shows a packing arrangement adopted in this
P21/n phase. Collected in Table 2 are given selected bond
distances and angles for this structure.
In P21/n phase, the two metallocene moieties in the
cation are not equivalent. The average Fe1-C distance
(2.052(4) Å) and the Fe1-Cp distance (1.660(4) Å)
indicate that this metallocene is in the Fe(II) oxidation
state. In contrast, the average Fe2-C distance (2.086(5)
Å) and average Fe2-Cp distance (1.693(4) Å) indicate
that this metallocene is in the Fe(III) oxidation state.
The two Cp rings associated with Fe1 and Fe2 are not
quite parallel, and the dihedral angles are 2.1° and 5.5°,
respectively. Furthermore, the two Cp rings associated
with Fe1 and Fe2 are nearly eclipsed, with average
staggering angles of 3.0° and 2.3°, respectively. The
extended packing arrangement of the P21/n phase of 5a
consists of columns of interacting cations surrounded
by triiodide anions (Figure 4b). The closest cation-
In P1h phase, the I3- anion is at the inversion center,
showing a symmetric structure. The I-I bond distance
is 2.8936(8) Å, which compares well with other deter-
minations of the symmetric triiodide bond distance.34
The triiodide is strictly linear as required by the
inversion symmetry.
Dark crystals in the monoclinic space group P21/n
were obtained when a layer of hexane was allowed to
(34) Runsink, J .; Swen-Walstra, S.; Migchelsen, T. Acta Crystallogr.
B 1972, 28, 1331.

1100 Organometallics, Vol. 19, No. 6, 2000
Dong et al.
Ta ble 3. Cyclic Volta m m etr y for Va r iou s
Bifer r ocen es
d
compound
biferrocene
E1/2(V)a
∆E1/2(V)b
∆ (mV)c
10-5 × Kcom
0.31
0.62
0.33
0.68
0.32
0.67
0.33
0.67
0.31
70
75
61
60
73
58
63
60
1.80
6a
6b
6c
0.35
0.35
0.34
8.6
8.6
5.8
a
All half-wave potentials are referred to the AgCl/Ag electrode.
Peak separation between two waves. c Peak-to-peak separation
b
between the resolved reduction and oxidation wave maxima.
d
Kcom: comproportionation equilibrium constants.
methyl substituents in the cation with P21/n phase show
a cisoid conformation relative to the fulvalenide ligand,
and this is similar to other dialkyl mixed-valence
biferrocenium triiodide salts. In the case of the cation
with P1h phase, the two equivalent naphthylmethyl
substituents show a transoid conformation relative to
the fulvalenide ligand. Consequently, the packing ar-
rangement of the cations and anions of 5a with P21/n
phase is different from that with P1h phase. The packing
arrangement for both phases can be described as
columns of cations and anions. However, it appears the
naphthylmethyl substituents with P21/n phase lead to
further slippage of the cations from the steplike stacks
seen in the P1h phase.
Electr och em ica l Resu lts. Electrochemical data for
neutral biferrocenes 6a -c, as well as those for some
other relevant compounds, are shown in Table 3. These
binuclear biferrocenes all show two successive reversible
one-electron oxidation-reduction waves. Electrochemi-
cal reversibility was demonstrated by the peak-to-peak
separation between the resolved reduction and oxidation
wave maxima and a 1:1 relationship of the cathodic and
anodic peak currents.
In comparison with biferrocene, the substituents on
the Cp rings in 6a -c clearly act as net electron-
withdrawing groups. It has been reported that the peak-
to-peak (∆E1/2) separation can gauge the interaction
between the two Fe sites in a biferrocenium cation.35,36
The separations between the two oxidation-reduction
waves for 6a -c are 0.35, 0.35, and 0.34 V, respectively.
In other words, the magnitudes of the electronic inter-
actions between the two Fe sites in 5a -c are nearly
equivalent in solution state.
F igu r e 4. (a) Packing arrangement of 5a with P1h phase.
(b) Packing arrangement of 5a with P21/n phase.
cation contacts result from intermolecular contacts of
naphthyl substituents. However, the naphthyl rings in
the P21/n phase of 5a are not anywhere to contact to
each other, as found in the P1h phase of 5a . Only the
naphthyl substituent associated with the Fe(III) met-
allocene and the naphthyl substituent on another Fe(III)
metallocene in another cation column contact each
other. They contact in a parallel edge-to-edge fashion
such that the shortest cation-cation contact is 3.789 Å.
In the case of the P21/n phase, the triiodide anion is
asymmetric (I1-I2 2.895(1), I1-I3 2.932(1) Å). In other
words, the I3 atom carries more negative charge,
showing I2-I1‚‚‚‚I3- character. Furthermore, the I3-
anion is not quite linear (176.31(2)°).
57F e Mo1ssba u er Ch a r a cter istics. It has been noted
that the 57Fe Mo¨ssbauer spectral properties of 1′,1′′′-
dialkylbiferrocenium triiodide salts are dependent on
sample history.8-12 To examine this phenomenon more
thoroughly, we prepared samples of 5a -c by three
different methods as described in the Experimental
Section. First, we examined a microcrystalline sample
of 5a , formed by adding a benzene solution of iodine to
a benzene solution of corresponding neutral biferrocene
(35) Morrison, W. H., J r.; Krogsrud, S.; Hendrickson, D. N. Inorg.
Chem. 1973, 12, 1998.
(36) (a) Atzkern, H.; Huber, B.; Ko¨hler, F. H.; Mu¨ller, G.; Mu¨ller,
R. Organometallics 1991, 10, 238. (b) Bunel, E. E.; Campos, P.; Ruz,
P.; Valle, L.; Chadwick, I.; Ana, M. S.; Gonzalez, G.; Manriquez, J . M.
Organometallics 1988, 7, 474. (c) Cowan, D. O.; Shu, P.; Hedberg, F.
L.; Rossi, M.; Kistenmacher, T. J . J . Am. Chem. Soc. 1979, 101, 1304.
(d) Moulton, R.; Weidman, T. W.; Vollhardt, K. P. C.; Bard, A. J . Inorg.
Chem. 1986, 25, 1846. (e) Obendorf, D.; Schottenberger, H.; Rieker,
C. Organometallics 1991, 10, 1293.
A comparison of structural features between the two
different crystallographic phases of 5a is interesting.
The naphthylmethyl substituent on the Cp ring is
situated differently. The two nonequivalent naphthyl-

Biferrocenium Cations
Organometallics, Vol. 19, No. 6, 2000 1101
Ta ble 4. 57F e Mo1ssba u er Lea st-Squ a r es-F ittin g
P a r a m eter s
a
compound T, K
∆EQ
δb
Γc
area ratiod
1.00:3.50
5a e
300
200
80
0.438
1.129
2.077
0.523
1.207
2.127
0.653
1.318
2.063
1.120
1.138
1.223
1.206
1.242
0.956
1.539
0.902
1.516
0.818
1.610
0.407
2.085
0.543
1.906
0.501
1.842
0.634
2.062
0.337
0.782
2.079
0.357
1.134
2.099
1.156
1.222
1.239
0.992
1.544
0.975
1.628
0.916
1.665
0.866
0.438 0.296, 0.247
0.447 0.448, 0.442
0.441 0.282, 0.282
0.511 0.373, 0.391
0.505 0.502, 0.476
0.502 0.292, 0.291
0.540 0.397, 0.389
0.539 0.465, 0.474
0.538 0.374, 0.377
0.451 0.427, 0.441
0.472 0.463, 0.467
0.525 0.576, 0.590
0.519 0.586, 0.604
0.532 0.596, 0.650
0.537 0.496, 0.453
0.524 0.486, 0.472
0.523 0.459, 0.459
0.516 0.456, 0.511
0.528 0.543, 0.498
0.516 0.502, 0.511
0.430 0.285, 0.281
0.441 0.251, 0.267
0.421 0.501, 0.445
0.434 0.337, 0.361
0.416 0.457, 0.457
0.427 0.350, 0.350
0.532 0.418, 0.403
0.523 0.264, 0.270
0.416 0.330, 0.330
0.411 0.402, 0.404
0.462 0.358, 0.335
0.516 0.395, 0.395
0.503 0.552, 0.552
0.524 0.307, 0.307
0.453 0.395, 0.402
0.512 0.530, 0.526
0.521 0.598, 0.604
0.534 0.485, 0.462
0.518 0.581, 0.545
0.529 0.479, 0.479
0.525 0.493, 0.523
0.534 0.487, 0.455
0.518 0.514, 0.513
0.532 0.479, 0.453
1.00:2.39
1.67:1.00
5a f
300
250
150
140
130
120
110
80
5a g
5be
5bf
300
300
300
80
5bg
300
2.00:1.00
3.57:1.00
80
5ce
300
150
120
110
100
90
F igu r e 5. Variable-temperature 57Fe Mo¨ssbauer spectra
of 5a prepared by method 1.
80
1.7231 0.520 0.472, 0.480
6a (method 1). The variable-temperature 57Fe Mo¨ss-
bauer spectra are shown in Figure 5. The various
absorption peaks were fitted to Lorentzian lines. The
resulting fitting parameters are collected in Table 4. The
prominent features in the 80 K Mo¨ssbauer spectrum are
two doublets, one with a ∆EQ ) 2.063 mms-1 (Fe(II)
metallocene) and the other with a ∆EQ ) 0.653 mms-1
(Fe(III) metallocene). Both doublets were fitted to have
the same area. This pattern of two doublets is what is
expected for a mixed-valence biferrocenium cation which
is valence trapped on the Mo¨ssbauer time scale (∼108
s-1). Fitting the 80 K spectrum also clearly shows a third
doublet with ∆EQ ) 1.318 mms-1. This doublet is
associated with a mixed-valence biferrocenium cation
in which the electron-transfer rate is greater than the
Mo¨ssbauer time scale. The ratio of valence-localized to
valence-delocalized species in the 80 K spectrum is 1.67:
1.00. As the temperature is increased, the intensity of
the detrapped signal grows at the expense of trapped
signal. The ratio of valence-localized to valence-delocal-
ized species charges from 1.67:1.00 to 1.00:3.50 as the
temperature is changed from 80 to 300 K.
a
b
Quadrupole-splitting in mms-1
.
Isomer shift referenced to
iron-foil in mms-1
.
c Full width at half-height taken from the least-
squares-fitting program. The width for the line at more positive
d
velocity is listed first for each doublet. The ratio of localized state
to delocalized state. e Microcrystalline sample. f Prepared by evapo-
rating the CH2Cl2 solution of biferrocenium. g Prepared by diffus-
ing the hexane into the CH2Cl2 solution of biferrocenium.
with the P1h phase, prepared by slow evaporation from
a CH2Cl2 solution (method 2), were also collected at
temperatures ranging from 80 to 300 K. Selected spectra
are shown in Figure 6, and the fitting parameters are
listed in Table 4. At temperatures below 100 K, the
cation of 5a shows two doublets in the variable-tem-
perature Mo¨ssbauer spectra, one representing the Fe(II)
site and the other for the Fe(III) site. An increase of
temperature causes the two doublets to move together,
resulting in an “average-valence” doublet at 130 K. At
300 K, the spectrum of this sample shows a single
doublet which is characteristic of a valence-detrapped
cation in which the electron-transfer rate exceeds ∼108
s-1
.
On the other hand, the sample of 5a with P21/n phase,
prepared by slowly diffusing hexane into a CH2Cl2
solution containing 5a (method 3), exhibits a Mo¨ssbauer
spectrum characteric of a valence-trapped cation (Fe(II),
The observations of the structural characteristics of
5a are also consistent with our Mo¨ssbauer studies. 57Fe
Mo¨ssbauer spectra for a more crystalline sample of 5a

1102 Organometallics, Vol. 19, No. 6, 2000
Dong et al.
F igu r e 6. Variable-temperature 57Fe Mo¨ssbauer spectra
of 5a with P1h phase prepared by method 2.
F igu r e 8. Variable-temperature 57Fe Mo¨ssbauer spectra
of 5b prepared by methods 1 (top) and 2 (bottom).
F igu r e 7. Variable-temperature 57Fe Mo¨ssbauer spectra
of 5a with P21/n phase prepared by method 3.
∆EQ ) 2.085 mms-1; Fe(III), ∆EQ ) 0.407 mms-1),
which remains trapped even at 300 K (Figure 7).
Obviously, the intramolecular electron transfer in 5a
is quite sensitive to environmental perturbations caused
by different crystal-packing arrangements.
Three different preparations of 5b were also examined
with variable-temperature Mo¨ssbauer spectroscopy.
Figure 8 illustrates the spectra obtained for samples
prepared by methods 1 and 2. The features in all these
spectra include two doublets, one with a ∆EQ ) ∼2
mms-1 (Fe(II) site) and the other with ∆EQ ) ∼0.5
mms-1 (Fe(III) site). Both doublets have the same area.
This pattern of two doublets is expected for a mixed-
valence biferrocenium cation which is valence-trapped
on the time scale of the Mo¨ssbauer technique (electron-
transfer rate < ∼108 s-1 on the solid state). The sample
of 5b obtained by method 3 gives Mo¨ssbauer spectra
with localized (two doublets) and delocalized (one aver-
age doublet) spectra superimposed. Collected in Table
4 are the resulting fitting parameters.
F igu r e 9. Variable-temperature 57Fe Mo¨ssbauer spectra
of 5c prepared by method 1.
of 5c obtained by method 1 were also collected at
temperatures ranging from 80 to 300 K. Selected spectra
are shown in Figure 9, and the fitting parameters are
listed in Table 4. At temperatures below 110 K, the
cation of 5c shows two doublets (electron-transfer rate
< ∼108 s-1). The two doublets are seen to move together
Owing to the low solubility of 5c in the CH2Cl2
solvent, we could not study the dependency of sample
history. However, 57Fe Mo¨ssbauer spectra for a sample

Biferrocenium Cations
Organometallics, Vol. 19, No. 6, 2000 1103
as the sample temperature is increased, finally becom-
ing a single average doublet at ∼120 K.
Our Mo¨ssbauer results indicate that changing the
substituent in the biferrocenium cations 5a -c leads to
dramatic changes in the electron-transfer rate in the
solid state. The interactions of the cation-cation and
the cation-anion undoubtedly alter the structural pack-
ing arrangement and therefore the potential energy
surface of the mixed-valence molecule. We believe that
the difference in electron-transfer rate is not due to a
difference in the electronic effect of the substituents. The
electrochemical measurements for the corresponding
neutral biferrocenes 6a -c strongly support our sugges-
tion. In the last section, we will provide an explanation
for the difference in electron-transfer rates in the mixed-
valence cations 5a -c.
Electr on P a r a m a gn etic Reson a n ce. For a mono-
nuclear ferrocenium cation with no low-symmetry crys-
tal field distortion, an axial-type spectrum will be
observed with g| ) 6 and g ) 0. As the low-symmetry
crystal field increases, both g| and g approach a value
of 2. In the case of a binuclear mixed-valence biferro-
cenium cation, the value of the g tensor anisotropy (∆g)
can be used as a rough estimate of the nature of the
electronic ground state of the mixed-valence biferro-
cenium cation. Hendrickson reported that systems with
larger g tensor anisotropy (∆g > 1.5) are valence trapped
on the Mo¨ssbauer time scale at all temperatures, while
those with small anisotropy (∆g < 1.1) are valence
detrapped throughout the thermally accessible temper-
ature range.30 Systems with ∆g between 1.1 and 1.5
show Mo¨ssbauer behavior intermediate between these
two extremes; they are valence trapped at low temper-
atures and can become detrapped at higher tempera-
tures. Here, we are interested in studying the nature
of the electronic ground state of the two different
crystalline morphologies of mixed-valence cation 5a by
EPR technique.
F igu r e 10. EPR spectra of 5a (a) with P1h phase and (b)
with P21/n phase.
Ta ble 5. EP R Da ta of 5a a n d Oth er Rela ted
Com p ou n d s
compounda
T/K
g|
g
∆gb
c
biferrocenium I3
4.2
4.2
4.2
3.58
2.98
3.02
1.72
1.92
2.01
1.89
2.01
1.97
1.91
2.01
1.79
1.86
1.06
1.07
c
1′,1′′′-dibutylbiferrocenium I3
1′,1′′′-dibenzylbiferrocenium I3
c
c
1′,1′′′-diiodobiferrocenium I3
4.2
2.75
0.76
5a d
5a e
77
77
3.16
3.67
2.85
1.25
1.66
1.06
X-band EPR spectra were run for the samples of 5a
with P1h phase and P21/n phase at 77 K. Spectra are
shown in Figure 10, and the g values evaluated from
these spectra and other related EPR spectra are col-
lected in Table 5. The 77 K spectrum of 5a with P1h
phase is clearly a typical axial-type spectrum (g| ) 3.16
and g ) 1.91). Surprisingly, the 77 K spectrum of 5a
with P21/n phase consists of two g| signals (3.67 and
2.85) and two g signals (2.01 and 1.79). Furthermore,
it seems that the g| value of 3.16 for the P1h phase lies
midway between the g| values of 3.67 and 2.85 for the
P21/n phase. Similarly, the feature at g ) 1.91 for the
P1h phase lies midway between the features at g ) 2.01
and 1.79 for the P21/n phase. To explain this unusual
observation at 77 K, we cannot simply conclude that the
sample of 5a with P1h phase exhibits a delocalized EPR
spectrum (electron-transfer rate > ∼1010 s-1) and the
sample of 5a with P21/n phase exhibits a localized EPR
spectrum. At 77 K, our Mo¨ssbauer results indicate that
the intramolecular electron-transfer rates are less than
the time scale (∼108 s-1) of Mo¨ssbauer technique for
both phases. This observation can be interpreted in
terms of the S ) 1 model.
a
d
Powder sample. b ∆g ) g| - g . c From ref 29. With P1h phase.
e With P21/n phase.
Heisenberg-type exchange interaction, as well as dipole-
dipole intermolecular interaction. It was found to pro-
vide a qualitative explanation for the various thermal
behaviors of electronic localization exhibited by mixed-
valence biferrocenium salts in the solid state. As shown
in Figure 1, the three-level system is described by a
fictitious spin S ) 1. The pair of levels in the double
well is associated with sz ) (1 and the single level in
the single well with sz ) 0 and a 2-fold degeneracy. The
levels of Sz ) 0, (1 can be considered respectively as
“delocalized” and “localized” states. This model suggests
a coexistence of localized and delocalized components
with progressive conversion. In terms of the S ) 1
model, we would like to suggest that the origin for our
usual EPR observation of 5a with P21/n phase arises
from the interaction of spin-spin exchange resulting
from a dipole-dipole interaction which develops be-
tween trapped cations. Therefore, in an effort to assess
the evidence for spin-spin interactions in 5a , the solid-
state packing arrangements of both phases are exam-
ined in detail. The closest Fe-Fe contacts in 5a appear
in Figure 11. The packing arrangements are displayed
As mentioned in the introduction, the S ) 1 model is
a theory of charge ordering and phase transition in
molecular crystals. This model involves intramolecular

1104 Organometallics, Vol. 19, No. 6, 2000
Dong et al.
F igu r e 11. Geometry of the closest intermolecular con-
tacts between cations of 5a with (a) P1h phase and (b) P21/n
phase.
in Figure 4. The P1h phase of 5a results in cation-cation
interactions throughout the cation column generated by
the inversion center. Furthermore, this Fe-Fe contact-
ing in the P1h phase occurs with parallelogram fashion.
In the P1h phase, the two ferrocenyl moieties in each
cation are crystallographically equivalent. Thus, the
unpaired electron in e2g (dx -y , dxy) orbitals can oscillate
back and forth between the two Fe centers in each
biferrocenium cation. In contrast to the P1h phase, the
unpaired electron in the biferrocenium cation with P21/n
phase is located primarily in the Fe(III) center, since
the two ferrocenyl moieties are not crystallographically
equivalent. The P21/n crystal structure of 5a results in
pairwise cation-cation contact. The closest Fe(III)-
Fe(III) contact in the P21/n phase is 9.793 Å. A stronger
dipole-dipole interaction in the P21/n phase can be
expected to produce a singlet-triplet splitting (Figure
12, D| ) 540 G, D ) 450 G). This explains why two g|
and two g signals were observed in the EPR spectrum
of 5a with P21/n phase.
F igu r e 12. Zero-field and magnetic field splitting of
energy levels.
the metal to adjust their distance to the larger dimen-
sion appropriate for a (Cp)2Fe(III) moiety. At the same
time the dimensions of the (Cp)2Fe(III) moiety contract
to those of a (Cp)2Fe(II) species. In Hendrickson’s model,
the 3d electrons of the Fe ions couple strongly with this
out-of-phase combination of Cp-Fe-Cp breathing modes.
Recently, we also suggested that the Cp ring tilting
must be considered in the Hendrickson micromodulation
picture.22-25 Furthermore, the theoretical S ) 1 model
supports the coexistence of localized and delocalized
states.27,28
2
2
Hendrickson’s model can be applied to explain the
absence of line broadening observed in the variable-
temperature Mo¨ssbauer spectra of 5a with P1h phase.
The observed changes in Mo¨ssbauer spectra are at-
tributable to the onset of lattice dynamics in the P1h
phase occurring at a rate that is fast on the Mo¨ssbauer
time scale at all temperatures.
Con clu d in g Com m en ts. The goal of this section is
to present an explanation for the differences of electron-
transfer rates in the series of mixed-valence biferro-
cenium cations.
The differences in the 57Fe Mo¨ssbauer spectral prop-
erties between two phases can be attributable to the
differences in the molecular conformation of the bifer-
rocenium cation. In the P1h phase, the crystal structure
results show that the two ferrocenyl moieties of the
cation are symmetric. Rapid electron transfer is then
possible in this symmetric arrangement. The cations
with P21/n phase are asymmetric; that is, the two irons
are not in equivalent environments. This asymmetry
results in a zero-point energy difference for intramo-
lecular electron transfer. On the basis of the S ) 1
model, the zero-point energy difference between two
vibronic states for a biferrocenium cation is proportional
to the cation-cation interaction. As mentioned in the
section on EPR, the cations in the P21/n phase exhibit
stronger cation-cation (dipole-dipole) interactions.
Thus, one vibronic state of the mixed-valence cation is
energetically more stable than the other state. This
explains why a localized electronic structure of 5a with
P21/n phase was observed.
To explain the dependence of sample history of
Mo¨ssbauer results, Hendrickson suggested37 that the
factors that are potentially important in controlling the
rate of intramolecular electron transfer in a mixed-
valence cation include (1) the effective barrier for
electron transfer in the cation, (2) the effective barrier
for charge oscillation in the anion, and (3) the intermo-
lecular cation-cation and cation-anion interactions. A
change in valence for a transition metal complex is
generally accompanied by a sizable coordination sphere
reorganization. For an iron metallocene, the distance
between Fe and Cp ring centroid changes from 1.65 Å
for Fe(II) to 1.70 Å for Fe(III). When an intramolecular
electron transfer occurs in a biferrocenium cation, the
two Cp rings bound to the Fe(II) ion move away from
(37) Kambara, T.; Hendrickson, D. N.; Dong, T.-Y.; Cohn, M. J . J .
Chem. Phys. 1987, 86, 2326.

Biferrocenium Cations
Organometallics, Vol. 19, No. 6, 2000 1105
dried three-necked flask (250 mL) and then dried under
vacuum at 2 Torr and 30 °C for 4 h. Dried THF (20 mL),
followed by butyllithium (3.15 mL; 1.6 M in hexane), was
added under nitrogen. The resulting solution was stirred at
-25 °C for 30 min, during which 1-lithio-1′-bromoferrocene
gradually precipitated. 1-Naphthylnitrile (5 mmol) was then
added, and the solution was further stirred at -25 °C for
another 25 min. Water (20 mL) was added, and the resulting
mixture was extracted with CH2Cl2 (50 mL × 2). The combined
extracts were dried over sodium sulfate and evaporated at
reduced pressure. The residue was chromatographed on a
column of neutral alumina (activity II). Elution with hexane
gave eluates containing dibromoferrocene and bromoferrocene.
Continued elution with ethyl acetate/hexane (1:25) gave elu-
ates that yielded the desired compound (20%). The properties
Ch a r t 2
1
are as follows. H NMR (CDCl3, ppm): 8.25 (t, 1H), 7.94 (m,
2H), 7.77 (d, 1H), 7.53 (m, 3H), 4.87 (s, 2H), 4.60 (s, 2H), 4.45
(s, 2H), 4.19 (s, 2H). MS: M+ at m/z 418, 420.
Syn th esis of 1′-(4-Biph en ylcar bon yl)-1-br om ofer r ocen e
(7b). The preparation of 7b (48% yield) was carried out
according to the same procedure as the synthesis of 7a , except
using 4-biphenylcarbonitrile as starting material. The physical
properties are as follows. 1H NMR (CDCl3, ppm): 8.00 (d, 2H),
7.68 (m, 4H), 7.77 (d, 1H), 7.45 (m, 3H), 5.00 (t, 2H), 4.63 (t,
2H), 4.46 (t, 2H), 4.17 (t, 2H). MS: M+ at m/z 444, 446. Mp:
94-96 °C.
We believe that the most important factor in control-
ling the intramolecular electron transfer in the series
of biferrocenium cations is the symmetry of the cation.
We have prepared a series of asymmetric biferrocenium
cations (9-12 in Chart 2) to study the influence of zero-
point energy difference on electron-transfer rates.38 The
magnitudes of zero-point energy difference between the
two vibronic states for asymmetric biferrocenium cations
9-12 are nondetectable by electrochemical measure-
ments. However, compounds 9-12 show localized Mo¨ss-
bauer characterics at 300 K. Up to the present, no
asymmetric biferrocenium cation showing delocalized
Mo¨ssbauer characterics has been reported.
The observations for the mixed-valence cation 5a
definitively show that the crystallographic properties of
the lattice play an important role in controlling the rate
of electron transfer. In chemical and biological systems,
there are several cases known where slowing solvent
molecule motion and lattice dynamics can dramatically
affect the rate of electron transfer. In biological systems,
protein conformational changes and the motion of amino
acid groups have been suggested as providing control
and specificity in electron transfer between proteins.39
Our observation is an interesting result in view of the
fact that the crystallographic properties can turn on or
off the electron transfer.
Syn th esis of 1′-(9-An th r a cen eca r bon yl)-1-br om ofer -
r ocen e (7c). The preparation of 7c (50% yield) was carried
out according to the same procedure as the synthesis of 7a ,
except using 9-anthraldehyde as starting material. The physi-
1
cal properties are as follows. H NMR (CDCl3, ppm): 8.69 (d,
2H), 8.43 (s, 1H), 8.01 (d, 2H), 7.48 (m, 4H), 4.94 (d, 1H), 4.45
(s, 2H), 4.27 (d, 1H), 4.09 (m, 4H), 3.86 (d, 1H), 2.70 (d, 1H).
MS: M+ at m/z 470, 472. Mp: ∼160 °C (decomp).
Red u ction of 7a -c. The reduction reaction was carried out
by carefully adding, with stirring, small portions of AlCl3 to a
mixture of corresponding ferrocene 7 and LiAlH4 in dried
ether. After 1 h, the solution became yellow, an excess of H2O
was added to it, and the ether layer was separated. The ether
layer was washed with H2O and dried over MgSO4. After the
evaporation of the solvent, the crude product was chromato-
graphed on neutral Al2O3, eluting with hexane. The first band
was the desired compound 8.
1
The physical properties of 8a (71% yield) are as follows. H
NMR (CDCl3, ppm): 8.10 (d, 1H), 7.85 (t, 1H), 7.72 (d, 1H),
7.44 (m, 4H), 4.38 (t, 2H), 4.19 (s, 2H), 4.16 (s, 2H), 4.08 (t,
2H). MS: M+ at m/z 404, 406. Mp: 113.5-114.5 °C.
1
The physical properties of 8b (61% yield) are as follows. H
NMR (CDCl3, ppm): 7.24-7.57 (m, 9H), 4.37 (s, 2H), 4.18 (d,
4H), 4.10 (s, 2H), 3.80 (s, 2H). MS: M+ at m/z 430, 432. Mp:
93.0-95.0 °C.
1
The physical properties of 8c (52% yield) are as follows. H
Exp er im en ta l Section
NMR (CDCl3, ppm): 8.41 (t, 3H), 8.01 (d, 2H), 7.51 (m, 4H),
4.72 (s, 2H), 4.40 (s, 2H), 4.25 (s, 2H), 4.06 (d, 4H). MS: M+ at
m/z 454, 456. Mp: 135.0-135.5 °C.
Gen er a l In for m a tion . All manipulation involving air-
sensitive materials were carried out by using standard Schlenk
techniques under an atmosphere of nitrogen (Scheme 1).
Chromatography was performed on neutral alumina (activity
II). Solvents were dried as follows: THF and ether were
distilled from Na/benzophenone; CH2Cl2 was distilled from
P2O5. The sample of dibromoferrocene was prepared according
to the literature procedure.40
Syn th esis of 1′,1′′′-Disu bstitu ted -bifer r ocen e (6a -c). A
mixture of corresponding ferrocene 8 (1 g) and activated copper
(5 g) was heated under N2 at 120-130 °C for 24 h. After cooling
to room temperature the reaction mixture was repeatedly
extracted with CH2Cl2 until the CH2Cl2 extracts appeared
colorless. The extracts were evaporated and chromatographed.
The first band eluted with hexane yielded starting material.
Continued elution with CH2Cl2 afforded the desired compound.
The compound could be recrystallized from hexane/benzene.
Syn th esis of 1′-Na p h th oyl-1-br om ofer r ocen e (7a ). Di-
bromoferrocene (1.72 g, 5 mmol) was placed in a freshly oven-
1
The physical properties of 6a (78% yield) are as follows. H
(38) Dong, T.-Y.; Lin, P. J .; Lin, K. J . Inorg. Chem. 1996, 35, 6037.
(39) (a) Cannon, R. D. Electron-Transfer Reactions; Butterworth:
London, 1980. (b) Haim, A. Comments Inorg. Chem. 1985, 4, 113. (c)
J ortner, J .; Bixon, M. B. J . Chem. Phys. 1988, 88, 167. (d) Brown, D.
B. Mixed-Valence Compounds, Theory and Applications in Chemistry,
Physics, Geology, and Biology; Reidel: Boston, MA, 1980. (e) Cannon,
R. D.; White, R. P. Prog. Inorg. Chem. 1988, 36, 195.
NMR (CDCl3, ppm): 7.94 (d, 2H), 7.81 (t, 2H), 7.66 (d, 2H),
7.48 (m, 4H), 7.31 (m, 2H), 7.08 (d, 2H), 4.42 (s, 4H), 4.30 (s,
4H), 4.01 (s, 8H), 3.79 (s, 4H). Mass spectrum: M+ at m/z 650.
Mp: ∼170.0 °C (decomp).
1
The physical properties of 6b (75% yield) are as follows. H
(40) Dong, T.-Y.; Lai, L. L. J . Organomet. Chem. 1996, 509, 131.
NMR (CDCl3, ppm): 7.14-7.55 (m, 18H), 4.42 (t, 4H), 4.27 (s,

1106 Organometallics, Vol. 19, No. 6, 2000
Dong et al.
4H), 4.00 (s, 8H), 3.48 (s, 4H). MS: M+ at m/z 702. Mp: ∼155.0
ψ rotation. Selected bond distances and angles are given in
Table 2. Tables of the X-ray crystal data, final positional
parameters for all atoms, complete tables of bond distances
and angles, and thermal parameters of these compounds are
given in the Supporting Information.
°C (decomp).
1
The physical properties of 6c (72% yield) are as follows. H
NMR (CDCl3, ppm): 8.28 (t, 5.6H), 7.98 (d, 4H), 7.48 (m, 8.5H),
4.62 (d, 4H), 4.44 (s, 8H), 4.17 (q, 4H), 3.95 (s, 4H). MS: M+ at
m/z 750. Mp: ∼215.0 °C (decomp).
Str u ctu r e Deter m in a tion of 5a w ith P hI P h a se. A dark
crystal (0.28 × 0.20 × 0.16 mm), which was grown by slow
evaporation from a CH2Cl2 solution, was used for data collec-
tion at 298 K. The θ-2θ scan technique was used to record
the intensities for all reflections for which 2.16° < θ < 22.42°.
Of the 2356 unique reflections, there were 2356 reflections
with Fo > 2.0σ(Fo2), where σ(Fo2) were estimated from counting
statistics. Absorption corrections were applied. These data
were used in the final refinement of the structural parameters.
A three-dimensional Patterson synthesis was used to de-
termine the heavy-atom positions, which phased the data
sufficiently well to permit location of the remaining non-
hydrogen atoms from Fourier synthesis. All non-hydrogen
atoms were refined anisotropically. Hydrogen atoms were
calculated at ideal distances. The greatest residual electron
density upon completion of refinement was in the vicinity of
the iodide atoms. Final residuals were R1 ) 0.0354 and wR(f 2)
) 0.0875.
Mixed -Va len ce Com p ou n d s 5a -c. A crystalline sample
of 5 was prepared by adding a benzene/hexane (1:1) solution
containing a stoichiometric amount of iodine to a benzene
solution of 6 at 0 °C. The resulting dark green crystals were
filtered and washed repeatedly with cold hexane (method 1).
A more crystalline sample can be prepared by slowly evaporat-
ing the CH2Cl2 solution of 5 (method 2) or by slowly diffusing
hexane into a saturated CH2Cl2 solution of 5 at -10 °C
(method 3). Microanalyses of the three different preparations
were identical within experimental error.
Anal. Calcd for 5a (C42H34Fe2I3): C, 48.92; H, 3.32. Found
(method 1): C, 48.88; H, 3.25. Found (method 2): C: 48.84;
H, 3.29. Found (method 3): C: 48.93; H, 3.30.
Anal. Calcd for 5b (C46H38Fe2I3): C, 51.01; H, 3.54. Found
(method 1): C, 51.02; H, 3.53. Found (method 2): C: 50.99;
H, 3.55. Found (method 3): C: 51.03; H, 3.53.
Anal. Calcd for 5c (C50H38Fe2I3): C, 53.09; H, 3.38. Found
(method 1): C, 53.38; H, 3.32. Found (method 2): C: 53.37;
H, 3.37. Found (method 3): C, 53.28; H, 3.35.
Str u ctu r e Deter m in a tion of 5a w ith P 21/n P h a se. A
dark black crystal (0.34 × 0.24 × 0.24 mm) was obtained when
a layer of hexane was allowed to slowly diffuse into a CH2Cl2
solution of 5a . Data were collected with θ in the range 1.43-
22.43°. The data were also corrected for absorption with an
empirical ψ rotation. Of the 4764 unique reflections, there were
P h ysica l Meth od s. 57Fe Mo¨ssbauer measurements were
made on a constant-velocity instrument, which has been
previously described. Computer fitting of the 57Fe Mo¨ssbauer
data to Lorentzian lines was carried out with a modified
version of a previously reported program.41 Velocity calibra-
tions were made using a 99.99% pure 10 µm iron foil. Typical
line widths for all three pairs of iron foil lines fell in the range
0.24-0.27 mm s-1. Isomer shifts are reported relative to iron
foil at 300 K but are uncorrected for temperature-dependent,
second-order Doppler effects. It should be noted that the isomer
shifts illustrated in the figures are plotted as experimentally
obtained. Tabulated data are provided.
1H NMR spectra were run on a Varian VXR-300 spectrom-
eter. Mass spectra were obtained with a VG-BLOTECH-
QUATTRO 5022 system.
Electrochemical measurements were carried out with a BAS
100W system. Cyclic voltammetry was perform with a station-
ary glassy carbon working electrode, which was cleaned after
each run. These experiments were carried out with a 1 × 10-3
M solution of biferrocene in dry CH2Cl2/CH3CN (1:1) containing
0.1 M (n-C4H9)4NPF6 as supporting electrolyte. The potentials
quoted in this work are relative to a Ag/AgCl electrode at 25
°C.
2
3289 with Fo > 2.0σ(Fo2). These data were used in the final
refinement of the structural parameters. Structure refinement
was carried out in the same manner as described above. Final
residuals were R1 ) 0.0348 and wR(f 2) ) 0.0980.
Str u ctu r e Deter m in a tion of 6a . An orange crystal (0.32
× 0.28 × 0.26 mm) was obtained by slow evaporation from a
CH2Cl2 solution. Data were collected with 3.74° < 2θ < 44.84°.
Of the 1955 unique reflections, there were 1955 with Fo
>
2.0σ(Fo2). These data were used in the final refinement of
structural parameters. The structure was solved and refined
as described for 5a . Final residuals were R1 ) 0.0260 and
wR(f 2) ) 0.0703.
Ack n ow led gm en ts are made to the National Science
Council (NSC89-2113-M-110-003), Taiwan, ROC, and
Department of Chemistry at Sun Yat-Sen University.
Su p p or tin g In for m a tion Ava ila ble: Complete tables of
crystal data, positional parameters, bond lengths and angles,
and thermal parameters for 5a and 6a . This material is
available free of charge via the Internet at http://pubs.acs.org.
The single-crystal X-ray determinations of compounds 5a
and 6a were carried out on an Enraf Nonius CAD4 diffractom-
eter at 298 K. Absorption corrections were made with empirical
(41) Lee, J . F.; Lee, M. D.; Tseng, P. K. Chemistry 1987, 45, 50.
OM9908884