Pronounced effects of ring tilting on rates of intramolecular electron transfer in polyalkyl-substituted mixed-valence biferrocenium triiodides

Article abstract of DOI:10.1021/om961032r

Relatively minor perturbations caused by Cp ring substituents in a series of mixed-valence biferrocenium cations have pronounced effects on the electronic structure and rate of intramolecular electron transfer. The X-ray structures of 1′,2′,1?,2?-tetrapropylbiferrocene, l′,2′,1?,2?-tetrabenzylbiferrocene, l',3',l',3'-tetrapropylbiferrocene, 1′,3′,1?,3?-tetrabenzylbiferrocene, and their corresponding mixed-valence biferrocenium triiodide salts have been determined at 298 K. The rates of intramolecular electron transfer in the mixed-valence cations were estimated by variable-temperature ⁵⁷Fe M?ssbauer spectroscopy (time scale ～10⁷ s^-1). The features in all of the 80 K spectra include two doublets, one with a quadrupole splitting (ΔE_Q) of ～2 mm s^-1 (Fe(II) site) and the other with ΔE_Q = ～0.6 mm s^-1 (Fe(III) site). This pattern of two doublets is expected for a mixed-valence biferrocenium cation which is valence-trapped on the time scale of the M?ssbauer experiment (electron-transfer rate < ～10⁷ s^-1 in the solid state). Increasing the sample temperature in l′,2′,1?,2?-tetrapropylbiferrocenium triiodide, 1′,2′,1?,2?-tetrabenzylbiferrocenium triiodide, and 1′,3′,1?,3?-tetrabenzylbiferrocenium triiodide causes the two doublets to move together with no discernible line broadening, eventually becoming a single average-valence doublet at temperatures of 295, 265, and 190 K, respectively. In the case of 1′,3′,1?,3?-tetrapropyl-biferrocenium triiodide, the two doublets at low temperatures do not coalesce into an average-valence doublet at 300 K (electron-transfer rate < ～10 s^-1 at 300 K). Thus, the micromodification of the relative positions of the alkyl substituents and the nature of the alkyl substituents has a dramatic influence on the rate of intramolecular electron transfer. The deviations of the Cp rings from the parallel position correlate quite well with the M?ssbauer critical temperature for electronic delocalization-localization in mixed-valence cations. Qualitatively, we suggest that the metal nonbonding orbitals (d_x²-y², d_xy) start to interact with the ligand π orbitals as the Cp rings tilt from the parallel position. On the basis of the density functional calculations, a theoretical explanation of the influence of bending back the Cp rings on the electron-transfer rates is presented. The electrochemical measurements, IR data, and near-IR spectra for these mixed-valence biferrocenium cations are also presented.

Full text of DOI:10.1021/om961032r

Organometallics 1997, 16, 2773-2786
2773
Articles
P r on ou n ced Effects of Rin g Tiltin g on Ra tes of
In tr a m olecu la r Electr on Tr a n sfer in
P olya lk yl-Su bstitu ted Mixed -Va len ce Bifer r ocen iu m
Tr iiod id es
Teng-Yuan Dong,*,1 Shu-Hwei Lee,1 Chung-Kay Chang,1 Hsiu-Mei Lin,1 and
Kuan-J iuh Lin2
Department of Chemistry, National Sun Yat-Sen University, Kaohsiung, Taiwan, and
the Institute of Chemistry, Academia Sinica, Nankang, Taipei, Taiwan
Received December 6, 1996X
Relatively minor perturbations caused by Cp ring substituents in a series of mixed-valence
biferrocenium cations have pronounced effects on the electronic structure and rate of
intramolecular electron transfer. The X-ray structures of 1′,2′,1′′′,2′′′-tetrapropylbiferrocene,
1′,2′,1′′′,2′′′-tetrabenzylbiferrocene, 1′,3′,1′′′,3′′′-tetrapropylbiferrocene, 1′,3′,1′′′,3′′′-tetraben-
zylbiferrocene, and their corresponding mixed-valence biferrocenium triiodide salts have been
determined at 298 K. The rates of intramolecular electron transfer in the mixed-valence
cations were estimated by variable-temperature 57Fe Mo¨ssbauer spectroscopy (time scale
∼107 s-1). The features in all of the 80 K spectra include two doublets, one with a quadrupole
splitting (∆EQ) of ∼2 mm s-1 (Fe(II) site) and the other with ∆EQ ) ∼0.6 mm s-1 (Fe(III)
site). 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 experiment (electron-transfer
rate < ∼107 s-1 in the solid state). Increasing the sample temperature in 1′,2′,1′′′,2′′′-
tetrapropylbiferrocenium triiodide, 1′,2′,1′′′,2′′′-tetrabenzylbiferrocenium triiodide, and
1′,3′,1′′′,3′′′-tetrabenzylbiferrocenium triiodide causes the two doublets to move together with
no discernible line broadening, eventually becoming a single “average-valence” doublet at
temperatures of 295, 265, and 190 K, respectively. In the case of 1′,3′,1′′′,3′′′-tetrapropyl-
biferrocenium triiodide, the two doublets at low temperatures do not coalesce into an average-
valence doublet at 300 K (electron-transfer rate < ∼10 s-1 at 300 K). Thus, the
micromodification of the relative positions of the alkyl substituents and the nature of the
alkyl substituents has a dramatic influence on the rate of intramolecular electron transfer.
The deviations of the Cp rings from the parallel position correlate quite well with the
Mo¨ssbauer critical temperature for electronic delocalization-localization in mixed-valence
2
2
cations. Qualitatively, we suggest that the metal nonbonding orbitals (dx -y , dxy) start to
interact with the ligand π orbitals as the Cp rings tilt from the parallel position. On the
basis of the density functional calculations, a theoretical explanation of the influence of
bending back the Cp rings on the electron-transfer rates is presented. The electrochemical
measurements, IR data, and near-IR spectra for these mixed-valence biferrocenium cations
are also presented.
In tr od u ction
trated on the extent of electron delocalization, the shape
The rich chemistry of bridged mixed-valence dimers,
especially those of ruthenium dimers3 and biferrocenium
cations,4-23 has captured considerable attention. Stud-
ies on these mixed-valence compounds have concen-
(5) Dong, T.-Y.; Cohn, M. J .; Hendrickson, D. N.; Pierpont, C. G. J .
Am. Chem. Soc. 1985, 107, 4777.
(6) Cohn, M. J .; Dong, T.-Y.; Hendrickson, D. N.; Geib, S. J .;
Rheingold, A. L. J . Chem. Soc., Chem. Commun. 1985, 1095.
(7) Dong, T.-Y.; Hendrickson, D. N.; Iwai, K.; Cohn, M. J .; Rheingold,
A. L.; Sano, H.; Motoyama, S. J . Am. Chem. Soc. 1985, 107, 7996.
(8) Iijima, S.; Saida, R.; Motoyama, I.; Sano, H. Bull. Chem. Soc.
J pn. 1981, 54, 1375.
X Abstract published in Advance ACS Abstracts, May 15, 1997.
(1) National Sun Yat-Sen University.
(2) Academia Sinica.
(9) Nakashima, S.; Masuda, Y.; Motoyama, I.; Sano, H. Bull. Chem.
Soc. J pn. 1987, 60, 1673.
(10) Kai, M.; Katada, M.; Sano, H. Chem. Lett. 1988, 1523.
(11) Dong, T.-Y.; Kambara, T.; Hendrickson, D. N. J . Am. Chem.
Soc. 1986, 108, 5857.
(12) Dong, T.-Y.; Schei, C. C.; Hsu, T. L.; Lee, S. L.; Li, S. J . Inorg.
Chem. 1991, 30, 2457.
(3) For reviews, see: (a) Day, P. Int. Rev. Phys. Chem. 1981, 1, 149.
(b) Mixed-Valence Compounds, Theory and Applications in Chemistry,
Physics, Geology and Biology; Brown, D. B., Ed.; Reidel Publishing,
Co.: Boston, MA, 1980. (c) Creutz, C. Prog. Inorg. Chem. 1983, 30,
1-73. (d) Richardson, D. E.; Taube, H. Coord. Chem. Rev. 1984, 60,
107.
(4) Hendrickson, D. N.; Oh, S. M.; Dong, T.-Y.; Moore, M. F. Comm.
Inorg. Chem. 1985, 4, 329.
(13) Webb, R. J .; Geib, S. J .; Staley, D. L.; Rheingold, A. L.;
Hendrickson, D. N. J . Am. Chem. Soc. 1990, 112, 5031.
S0276-7333(96)01032-1 CCC: $14.00 © 1997 American Chemical Society

2774 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
Ch a r t 1
of the corresponding intervalence transfer (IT) band,
and the role of the bridge in the electron-transfer
process. Their physical properties have been discussed
in light of Hush’s model24 and the PKS model.25
The series of mixed-valence biferrocenium triiodide
salts (1-14, Chart 1), fulvalenide-bridged mixed-valence
compounds, has been studied extensively in the solid
state regarding the nature of the electronic structure
in the ground state, various structural factors, and
lattice dynamics,6-10 including the electronic and vi-
bronic coupling between two metal ions, the nature of
the counterion,11-14 and the cation-anion interactions.26
Compounds 3-5 and 9 give unusual temperature-
dependent Mo¨ssbauer spectra.4-8 At temperatures
below 200 K, they show two doublets, one for the FeII
and the other for the FeIII sites. Increasing the sample
temperature in each case causes the two doublets to
move together with no discernible line broadening,
eventually becoming a single “average-valence” doublet
at temperatures of 275, 245, 275, and 260 K. The
Mo¨ssbauer results indicate that compound 2 has a
valence-trapped electronic structure (electron-transfer
rate less than ∼107 s-1 in solid state).27 Hendrickson
suggested4 that there is an onset of dynamics associated
with the counterions in the dialkylbiferrocenium cation,
(14) Webb, R. J .; Rheingold, A. L.; Geib, S. J .; Staley, D. L.;
Hendrickson, D. N. Angew. Chem., Int. Ed. Engl. 1989, 28, 1388.
(15) Nakashima, S.; Katada, M.; Motoyama, I.; Sano, H. Bull. Chem.
Soc. J pn. 1987, 60, 2253.
(16) Dong, T.-Y.; Kambara, T.; Hendrickson, D. N. J . Am. Chem.
Soc. 1986, 108, 4423.
(17) Sorai, M.; Nishimori, A.; Hendrickson, D. N.; Dong, T.-Y.; Cohn,
M. J . J . Am. Chem. Soc. 1987, 109, 4266.
(18) Kambara, T.; Hendrickson, D. N.; Dong, T.-Y.; Cohn, M. J . J .
Chem. Phys. 1987, 86, 2326.
(21) Dong, T.-Y.; Schei, C. C.; Hwang, M. Y.; Lee, T. Y.; Yeh, S. K.;
Wen, Y. S. Organometallics 1992, 11, 573.
(22) Dong, T.-Y.; Chou, C. Y. J . Chem. Soc., Chem. Commun. 1990,
1332.
(23) Dong, T.-Y.; Lee, T. Y.; Lin, H. M. J . Organomet. Chem. 1992,
427.
(24) Hush, N. S. Prog. Inorg. Chem. 1967, 8, 391.
(25) Wong, K. Y.; Schatz, P. N. Prog. Inorg. Chem. 1981, 28, 369
and references therein.
(19) Konno, M.; Hyodo, S.; Iijima, S. Bull. Chem. Soc. J pn. 1982,
55, 2327.
(20) Dong, T.-Y.; Hwang, M. Y.; Schei, C. C.; Peng, S. M.; Yeh, S. K.
J . Organomet. Chem. 1989, 369, C33.
(26) Kambara, T.; Hendrickson, D. N.; Dong, T.-Y.; Cohn, M. J . J .
Chem. Phys. 1987, 86, 2362.
(27) Nakashima, S.; Katada, M.; Motoyama, I.; Sano, H. Bull. Chem.
Soc. J pn. 1987, 60, 2253.

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2775
and this probably influences the rate of intramolecular
electron transfer. Recently, we found20,21 that the
interactions between the cation and anion in 10-14
undoubtedly alter the potential energy surface of the
mixed-valence molecule and, therefore, the rate of
intramolecular electron transfer. Mixed-valence cations
of 10 and 12 are localized on the time scale of the
Mo¨ssbauer experiment. Furthermore, compound 11 has
a valence-trapped electronic structure at temperatures
below 150 K and becomes a valence-detrapped electronic
structure at a temperature of ∼200 K. In comparison
with para-substituted compounds (10-12), a valence-
detrapped electronic structure is seen even at 77 K for
ortho-substituted mixed-valence compounds of 13 and
14. We suggested that relatively minor perturbations
caused by interactions between neighboring cations
and anions have pronounced effects on electron trans-
fer.22,28-30
A recent interesting finding is that there is a signifi-
cant influence on the electron-transfer rate in the mixed-
valence biferrocenium salts (15-18) when the cyclopen-
tadienyl (Cp) rings in each ferrocenyl moiety are tilted
from a parallel relation between the two Cp rings
around the Fe ion. Such a structural modification would
lead to greater metal-ligand interactions as the rings
tilt.
geometric influences of the Cp ring tilting on the rates
of electron transfer in the series of mixed-valence
biferrocenium cations.
Molecu la r Str u ctu r es of 24-27. Details of the
X-ray crystal data collections and unit-cell parameters
of 24b,c, 25b,c, 26b,c, and 27b,c are given in Tables 1
and 2, and their molecular structures are shown in
Figures 1-8. Selected bond distances and angles are
given in Tables 3 and 4. These neutral and mixed-
valence compounds exist in a trans conformation with
the two iron ions on opposite sides of the planar
fulvalenide bridge observed for most biferrocenes and
biferrocenium cations. The two Cp rings in the ful-
valenide bridge are crystallographically coplanar for
these compounds. Thus, the π interaction between the
two ferrocenyl units in each molecule is not destroyed.
We believe that the degree of coplanarity between the
two Cp rings in the fulvalenide bridge plays a most
important role in determining the magnitude of the
electron-transfer rate.
In the neutral complexes 24b,c and 25b,c, the aver-
age distance from the iron atom to the two Cp rings are
1.650(3), 1.645(5), 1.646(3), and 1.651(5) Å, respectively.
Furthermore, there is no significant difference between
the Fe-Cp distance and the Fe-fulvalenide distance.
Inspection of these Fe-Cp distances shows that these
values are closer to the value of 1.65 Å found for
ferrocene31 than to the value of 1.70 Å found for the
ferrocenium ion.32
The respective dihedral angles between the two least-
squares planes of the Cp rings for a given ferrocenyl
moiety in 24b,c and 25b,c are 1.2(2), 0.7(2), 0.8(2), and
0.3(7)°, while the two Cp rings bonded to each iron ion
are nearly eclipsed, with respective average staggering
angles of 2.2(5), 6.5(5), 6.8(4), and 17.9(5)°. Each C-C
bond length in the Cp ring agrees well with that in
ferrocene (average value of 1.42 Å).31 Mean bond
distances from the Fe atom to the ring carbon atoms in
24b,c and 25b,c are 2.044(5), 2.043(4), 2.041(5), and
2.043(9) Å, respectively. These values agree well with
the value of 2.045 Å observed for ferrocene.31
In our previous paper,29 we also suggested that the
difference in the electron-transfer rates for the series
of dialkylbiferrocenium cations is mainly due to the
degree of tilting of the Cp rings from the parallel
geometry. We found that deviations of the Cp rings
from the parallel position correlate quite well with the
critical temperature of electronic delocalization-local-
ization in mixed-valence biferrocenium salts. In this
paper, we suggest that structural micromodifications in
mixed-valence biferrocenium cations have certain im-
portant effects on the intramolecular electron-transfer
rate. To study the influence of polyalkylbiferrocenium
triiodide salts, we have prepared a series of polyalkyl-
biferrocenium triiodide salts.
In the case of mixed-valence compounds 26b,c and
27b,c, the Fe-Cp distances are 1.671(6), 1.675(2), 1.672-
(1), and 1.668(8) Å, respectively. Each value is larger
than that for the corresponding neutral compound.
Furthermore, inspection of these Fe-Cp distances in
mixed-valence cations shows that these values lie
midway between the value of 1.65 Å found for fer-
rocene31 and the value of 1.70 Å found for the ferroce-
nium ion.32 The average Fe-C bond distances in 26b,c
and 27b,c are 2.07(1), 2.068(4), 2.07(2), and 2.06(2) Å,
respectively. The average Fe-C bond distance for a
given mixed-valence cation is marginally larger than
that in the corresponding neutral biferrocene.
Furthermore, the average Fe-C distances of these
mixed-valence compounds are very similar, and the
values also lie midway between the 2.045 Å distance
observed for ferrocene31 and the 2.075 Å distance
observed for ferrocenium cations.32 Such an increase
in the Fe-C and Fe-Cp distances has been observed
when ferrocenes are oxidized to the corresponding
Resu lts a n d Discu ssion
Polyalkylbiferrocenium triiodide salts were prepared
by coupling the corresponding bromoferrocenes, using
activated Cu as a catalyst (Scheme 1). Mixed-valence
compounds 26 and 27 were prepared by oxidizing the
corresponding neutral biferrocenes with I2. The result-
ing microcrystalline samples were recrystallized from
a CH2Cl2/hexane solution. More crystalline samples,
prepared by slowly diffusing hexane into a CH2Cl2
solution containing the corresponding biferrocenium
triiodides, were used for all of the physical data mea-
surements. Before the new physical data are described,
a summary of the single-crystal X-ray structural results
obtained for 24b,c, 25b,c, 26b,c, and 27b,c are pre-
sented first. X-ray crystallographic studies of these
compounds were undertaken to elucidate the structures
which are difficult to determine with certainty using
NMR spectroscopy. It also helps in understanding the
(28) 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.
(29) 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.
(30) Dong, T.-Y.; Lee, S. H.; Chang, C. K.; Lin, K. J . J . Chem. Soc.,
Chem. Commun. 1995, 2453.
(31) Seiler, P.; Dunitz, J . D. Acta Crystallogr., Sect. B 1979, 35,
1068.
(32) Mammano, N. J .; Zalkin, A.; Landers, A.; Rheingold, A. L. Inorg.
Chem. 1977, 16, 297.
(33) Konno, M.; Sano, H. Bull. Chem. Soc. J pn. 1988, 61, 1455.

2776 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
Sch em e 1
ferrocenium cations.21,32 There is no significant differ-
ence for the average C-C bond distances in the Cp rings
in these new mixed-valence compounds. The two Cp
rings in each ferrocenyl moiety in 26b,c and 27b,c are
also nearly eclipsed, with average staggering angles of
0.4(6), 5.9(2), 16.5(9), and 3.5(5) °, respectively.
mixed-valence compounds 26b,c and 27b,c, the least-
squares planes of the Cp rings form dihedral angles of
3.9(5), 4.2(2), 4(1), and 6.6(7)°, respectively. A direct
comparison of these important structural parameters
between series of alkylbiferrocenium and polyalkylbi-
ferrocenium triiodide salts is made (Table 5). From
Table 5 and Figure 9, it can be easily seen that the
decrease of the critical temperatures for the delocaliza-
tion-localization transition (Mo¨ssbauer spectroscopy
∼107 s-1) in the series of mixed-valence biferrocenium
salts is accompanied by the increase of the tilting angle.
In our previous papers,28,29 we suggested that the
degree of tilting of the Cp rings from the parallel
geometry plays an important role in determining the
rates of electron transfer in the series of mixed-valence
biferrocenium triiodide salts. In this new series of

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2777
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 Neu tr a l Bifer r ocen es
24b
24c
25b
25c
formula
mw
cryst syst
space
group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
C32H42Fe2 C48H42Fe2 C32H42Fe2 C48H42Fe2
538.37
triclinic
P1h
730.55
monoclinic triclinic
P21/c
538.38
730.55
monoclinic
P21/n
P1h
8.175(1)
9.373(2)
10.381(2) 8.135(3)
73.07(1)
12.694(5)
17.300(6)
7.628(3)
13.342(3)
10.046(3) 6.0182(8)
10.238(3) 22.185(4)
64.80(2)
88.90(1)
65.71(1)
1.298
91.29(4)
76.20(3)
75.97(3)
1.314
100.18(1)
Fcalcd
,
1.36
1.384
g cm-3
V, Å3
688.96(2) 1786(1)
680.5(4)
1
1753.2(5)
2
Z
1
2
µ, mm-1
λ, Å
1.07
0.710 69
44.9
0.845
0.710 69
50.0
8.88
0.710 69
44.9
1.72
0.710 69
45.0
2θ limits,
deg
max, min
trans coeff
Rf
0.9991,
0.9322
0.037
1.0000,
0.8825
0.041
0.9998,
0.8611
0.044
0.9942,
0.8760
0.044
Rwf
0.038
0.050
0.051
0.043
Ta ble 2. 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 Mixed -Va len ce Bifer r ocen iu m
Tr iiod id es
F igu r e 1. ORTEP drawing for 24b, with 30% thermal
ellipsoids.
26b
26c
27b
27c
formula
mw
cryst syst
space
group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
C32H42Fe2I3 C48H42Fe2I3 C32H42Fe2I3 C48H42Fe2I3
919.08
triclinic
P1h
1111.26
triclinic
P1h
919.08
monoclinic triclinic
P21/n
1111.26
P1h
8.764(1)
9.329(3)
11.240(3)
103.08(3)
97.88(1)
102.80(2)
1.783
9.307(2)
10.626(2)
10.965(1)
76.82(1)
85.59(1)
86.29(2)
1.755
13.647(4)
10.171(2)
13.929(4)
9.517(3)
10.688(3)
10.830(4)
74.03(3)
79.81(3)
89.53(3)
1.772
116.39(2)
1.762
Fcalcd
,
g cm-3
V, Å3
855.8(4)
1
1051.5(3)
1
1732.0(9)
2
1041.4(6)
1
Z
µ, mm-1
λ, Å
7.09
5.79
6.97
6.32
0.710 69
44.9
0.710 69
45.0
0.710 69
45.0
0.710 69
44.9
2θ limits,
deg
max, min
trans coeff
Rf
0.9928,
0.8276
0.038
0.9978,
0.7107
0.034
0.9992,
0.5100
0.067
0.6294,
0.6266
0.057
Rwf
0.045
0.040
0.087
0.069
In other words, there is a correlation between the tilt
angle and the rate of electron transfer. A detailed
discussion based on density functional theoretical cal-
culations is presented in the later sections.
-
The I3 anion is at the inversion center, showing a
symmetric structure. The I-I bond distances in the
series of new mixed-valence compounds are in accord
with the accepted value of 2.92 Å reported for the free
-
I3 ion.34 The cations in this series of mixed-valence
compounds are also at the inversion centers. The site
symmetry imposed on each mixed-valence cation obvi-
ously requires that both iron ions of the cation are in
equivalent positions, and this is consistent with our
Mo¨ssbauer studies for the series of mixed-valence
compounds.
F igu r e 2. ORTEP drawing for 24c, with 30% thermal
ellipsoids.
In addition to the influence of the tilting of the Cp
rings on the electron-transfer rate, we believe that the
positioning of the triiodide anion relative to the mixed-
valence cation also plays an important role in controlling
(34) Runsink, J .; Swen-Walstra, S.; Migchelsen, T. Acta Crystallogr.,
Sect. B 1972, 28, 1331.

2778 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
clearly acts as a net electron donor. Furthermore, it is
important to find that the first or second half-wave
potential of 24b is very similar with that of 25b. The
same observation also occurs between 24c and 25c. In
other words, changing the positions of the alkyl sub-
stituents from ortho to meta does not instantaneously
change the half-wave potential. It seems that the
electronic effect of the alkyl substituent has nothing to
do with the position of substitution.
It has been demonstrated that the magnitude of the
peak-to-peak separation (∆E1/2) gives an indication of
the interaction between two Fe sites.35 Making a
comparison of the magnitude of ∆E1/2 for various poly-
alkyl-substituted biferrocenes indicates that the mag-
nitude of interaction between the two Fe sites is similar
in solution. Thus, the interaction between the two Fe
sites is insensitive to both the position of substitution
and the nature of the substituent.
F igu r e 3. ORTEP drawing for 25b, with 30% thermal
ellipsoids.
In eq 1, the abbreviations [3, 3], [2, 3], and [2, 2]
denote the dioxidized salt, the monooxidized salt, and
the neutral compound, respectively. From the value of
∆E1/2, the disproportionation equilibrium constant K (as
shown in Table 6) can be calculated. In the studies of
K
\
[2,2] + [3,3]
(1)
the intervalence transition band in the near-IR region,
quantitative calculations based on the concentration of
[2, 3] have been corrected for this equilibrium.
Electr on Tr a n sfer in th e Solid Sta te. The rates
of intramolecular electron transfer in the mixed-valence
cations 26b,c and 27b,c were estimated by variable-
temperature 57Fe Mo¨ssbauer spectroscopy (time scale
∼107 s-1). The variable-temperature 57Fe Mo¨ssbauer
spectra of these compounds are shown in Figures 10-
13. The various absorption peaks were fitted to Lorent-
zian lines. The resulting fitting parameters are col-
lected in Table 7.
F igu r e 4. ORTEP drawing for 25c, with 30% thermal
ellipsoids.
The features in all of the 80 K spectra include two
doublets, one with a quadrupole splitting (∆EQ) of ∼2
mm s-1 (Fe(II) site) and the other with ∆EQ ) ∼0.6 mm
s-1 (Fe(III) site). Both doublets have the same area, as
deduced by a least-squares fitting. 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 experiment (electron-transfer rate < ∼107
s-1 in the solid state). Increasing the sample temper-
ature in 26b,c and 27c causes the two doublets to move
together with no discernible line broadening, eventually
becoming a single average-valence doublet at temper-
atures of 295, 265, and 190 K, respectively. In the case
of 27b, the two doublets at low temperatures do not
coalesce into an average-valence doublet at 300 K
(electron-transfer rate < ∼107 s-1 at 300 K). Thus, the
micromodification of the relative positions of alkyl
substituents and the nature of alkyl substituents has a
dramatic influence on the rate of intramolecular electron
transfer. The ordering of intramolecular electron-
transfer rates is 27c > 26c > 26b > 27b.
the rate of electron transfer. The triiodide moiety in
these polyalkyl-substituted mixed-valence biferrocenium
compounds is relatively perpendicular to the fulvalenide
ligand as, found in 1,7 3-6,11,19,33 and 27a .29 With such
a similar cation-anion relationship, we can potentially
discover the influence of the tilting of the Cp rings on
the rate of electron-transfer.
Electr och em ica l Mea su r em en ts. Electrochemical
data for the neutral compounds (24b,c and 25b,c), as
well as those for some other relevant compounds, are
shown in Table 6. These binuclear biferrocenes all
undergo two successive reversible one-electron oxida-
tions to yield the mono- and then the dication. Elec-
trochemical reversibility is demonstrated by the peak-
to-peak separation between the resolved reduction and
oxidation wave maxima (Table 6) and a 1:1 relation-
ship of the cathodic and anodic peak currents (Ic/Ia
Table 6).
The effect of the alkyl substituents on the stability of
the Fe(III) state is illustrated by the shift of the half-
wave potential. In general, electron-donating groups
stabilize the ferrocenium cation, lowering the half-wave
potential, while electron-withdrawing groups have the
opposite effect. Comparison of the half-wave potentials
of 24 and 25 with those of biferrocene and 1′,1′′′-
dialkylbiferrocenes indicates that the alkyl substituent
(35) (a) Morrison, W. H., J r.; Krogsrud, S.; Hendrickson, D. N. Inorg.
Chem. 1973, 12, 1998. (b) Bunel, E. E.; Campos, P.; Ruz, J .; Valle, L.;
Chadwick, I.; Ana, M. S.; Gonzalez, G.; Manriquez, J . M. Organome-
tallics 1988, 7, 474. (c) Atzkern, H.; Huber, B.; Ko¨hler, F. H.; Mu¨ller,
G.; Mu¨ller, R. Organometallics 1991, 10, 238. (d) Chukwu, R.; Hunter,
A. D.; Santarsiero, B. D.; Bott, S. G.; Atwood, J . L.; Chassaignac, J .
Organometallics 1992, 11, 589.

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2779
F igu r e 6. ORTEP drawing for 26c, with 30% thermal
ellipsoids.
F igu r e 5. ORTEP drawing for 26b, with 30% thermal
ellipsoids.
We believe that the difference of the electron-transfer
rates in the series of biferrocenium cations is not a result
of the difference in the electronic effect of the alkyl
substituents. The Mo¨ssbauer transition temperature of
1′,1′′′-dipropylbiferrocenium triiodide (4) is 245 K. On
the other hand, the Mo¨ssbauer Tc of 26b and 27b is 295
and >300 K, respectively. Similarly, the Mo¨ssbauer Tc
of 26c (265 K) is also higher than that of 1′,1′′′-
dibenzylbiferrocenium triiodide (260 K). Hence, an
extra alkyl substituent on the Cp ring in 26b,c and 27b
does not increase the rate of electron transfer. It is clear
that the electronic effect of the alkyl is not additive by
comparing the Mo¨ssbauer Tc. Further evidence can be
gleaned from the peak-to-peak separation (∆E1/2) men-
tioned in the section discussing the electrochemical
measurements. In our previous papers, we found29 that
the deviations of the Cp rings from the parallel position
correlate quite well with the Mo¨ssbauer critical tem-
perature for electronic delocalization-localization in
mixed-valence cations of 1, 3, 6, and 27a . As shown in
Figure 9, the new series of mixed-valence compounds
26b ,c and 27c also fit into the correlation of the tilt
angle with the electron-transfer rate. Qualitatively, we
F igu r e 7. ORTEP drawing for 27b, with 30% thermal
ellipsoids.
start to interact with the ligand π orbitals as the Cp
rings tilt from the parallel position.
The IR spectra of mixed-valence compounds 26b,c
and 27b,c exhibit two C-H perpendicular bending
2
2
suggested that the metal nonbonding orbitals (dx -y , dxy)

2780 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
Ta ble 3. Selected Bon d Dista n ces (Å) a n d Bon d
An gles (d eg) for 24b,c a n d 26b,c
24b
26b
24c
26c
Distances
Fe-C(1)
2.052(4)
2.037(5)
2.044(5)
2.043(5)
2.045(5)
2.054(5)
2.046(5)
2.036(5)
2.033(5)
2.048(5)
1.470(8)
1.408(7)
1.429(6)
1.424(7)
1.412(9)
1.414(7)
1.425(7)
1.421(7)
1.426(7)
1.417(8)
1.417(8)
2.09(1)
2.07(1)
2.05(1)
2.05(1)
2.054(1)
2.09(1)
2.10(1)
2.06(1)
2.05(1)
2.06(1)
1.43(2)
1.45(1)
1.43(2)
1.43(2)
1.41(2)
1.42(2)
1.44(2)
1.44(2)
1.42(2)
1.43(2)
1.43(2)
2.912(2)
2.063(3)
2.039(3)
2.029(4)
2.039(4)
2.054(4)
2.036(3)
2.045(3)
2.042(4)
2.046(4)
2.040(4)
1.472(6)
1.419(5)
1.422(4)
1.404(6)
1.378(6)
1.429(6)
1.407(4)
1.421(5)
1.434(5)
1.399(6)
1.398(6)
2.110(4)
2.057(4)
2.049(4)
2.051(4)
2.063(4)
2.088(4)
2.096(4)
2.051(4)
2.052(4)
2.063(4)
1.438(7)
1.438(6)
1.435(6)
1.420(6)
1.406(7)
1.423(6)
1.434(6)
1.429(6)
1.421(5)
1.416(6)
1.423(5)
2.536(4)
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)
C(1)-C(1)a
C(1)-C(2)
C(1)-C(5)
C(2)-C(3)
C(3)-C(4)
C(4)-C(5)
C(6)-C(7)
C(6)-C(10)
C(7)-C(8)
C(8)-C(9)
C(9)-C(10)
I(1)-I(2)
Angles
C(2)-C(1)-C(5)
C(1)-C(2)-C(3)
C(2)-C(3)-C(4)
C(3)-C(4)-C(5)
C(1)-C(5)-C(4)
C(7)-C(6)-C(10)
C(6)-C(7)-C(8)
C(7)-C(8)-C(9)
C(8)-C(9)-C(10)
C(6)-C(10)-C(9)
I(2)-I(1)-I(2)b
107.2(4)
108.5(4)
108.0(4)
107.8(4)
108.4(4)
107.6(4)
107.3(4)
108.9(4)
107.1(4)
109.0(5)
107.6(9)
106.3(9)
109.7(9)
108(1)
108.7(9)
107.3(9)
108(1)
107.2(3)
108.5(4)
108.5(4)
109.0(4)
106.9(4)
108.2(3)
106.6(3)
108.8(4)
107.8(3)
108.5(4)
105.9(3)
108.9(4)
108.1(4)
108.3(4)
108.7(4)
107.5(3)
107.5(3)
109.0(3)
107.6(3)
108.4(3)
180.0
F igu r e 8. ORTEP drawing for 27c, with 30% thermal
ellipsoids.
109(1)
107.7(9)
108.3(1)
180.0
bands, which may be readily assigned to FeII (∼815
cm-1) and FeIII (∼845 cm-1) moieties. It is clear that
compounds 26b,c and 27b,c are all localized on the IR
time scale.
a
Symmetry equivalents: 1 - x, 1 - y, 1 - z for 24b; 1 - x, 1 -
b
y, 2 - z for 26b; and 1 - x, 1 - y, 2 - z for 26c. Symmetry
equivalents: -x, -y, -z for 26b and 26c.
Electr on Tr a n sfer in th e Solu tion Sta te. In
common with most mixed-valence compounds, the
mixed-valence compounds 26b,c and 27b,c in CH2Cl2
have an IT band at 4651 cm-1, which is not present in
the corresponding neutral compound or dioxidized ion.
A description of the width of the IT band, the extent of
electron delocalization, and the electron-transfer prop-
erties of the mixed-valence dimers has been given by
Hush.24 Hush has derived an expression for the band-
width (in cm-1) at half-maximum of the IT band of a
localized homonuclear mixed-valence dimer at 300 K as
the hopping frequency. The absorption maximum and
Ket ) νetexp(-νmax/4KBT)
(6)
νet ) (2π/p)Hab2(π/KBTνmax
)
1/2
activation parameters calculated from eqs 2-6 were
collected in Table 8, together with those for other
relevant compounds.
The experimental R2 value is the average of R2 values
for the ground and excited states. If delocalization is
small, the electronic wave functions used for the overlap
are relatively unperturbed. Thus, R2 is a direct measure
of the delocalization in the ground state. As shown in
Table 8, there is good evidence for localized valences in
26b,c and 27b,c.
1/2
∆ν1/2 ) (2310 νmax
)
(2)
where νmax is the frequency in cm-1 of the absorption
maximum. For a system with harmonic nuclear motion,
the energy of activation ∆E* is related to the energy of
the IT band maximum (νmax) by the expression:
In flu en ce of Rin g Tiltin g on Solid -Sta te Electr on
Tr a n sfer . The goal of this section is to present an
explanation for the differences of electron-transfer rates
in the series of mixed-valence cations.
∆E* ) 1/4νmax
(3)
Furthermore, the magnitude of the delocalization can
be obtained by a calculation of the delocalization pa-
rameter R2 and electronic coupling Hab from eqs 4 and
5. In eqs 4 and 5, ꢀmax is the extinction coefficient and
In the past few years, we have made considerable
progress in understanding what factors control the rate
of intramolecular electron transfer in the solid state for
mixed-valence biferrocenium salts. We found that the
factors that are potentially important in controlling the
rate of intramolecular electron transfer in a mixed-
valence biferrocenium cation include the zero-point
energy difference in the cation,12 the nature of the
counterion,11-14 and the intermolecular cation-anion
interactions.18 In fact, the most important factor in
R2 ) {(4.24 × 10-4)ꢀmax(∆ν1/2)}/{νmaxd2}
(4)
(5)
Hab ) νmax
R
d (5.1 Å) is the donor-acceptor distance.7 The rate
constant (Ket) can be estimated from eq 6, where νet is

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2781
Ta ble 4. Selected Bon d Dista n ces (Å) a n d Bon d
An gles (d eg) for 25b,c a n d 27b,c
25b
27b
25c
27c
Distances
Fe-C(1)
2.060(4)
2.036(5)
2.035(5)
2.038(5)
2.045(5)
2.047(4)
2.037(5)
2.044(5)
2.032(5)
2.040(5)
1.462(8)
1.427(7)
1.429(7)
1.422(7)
1.395(9)
1.422(7)
1.436(7)
1.417(7)
1.409(8)
1.408(8)
1.424(7)
2.10(2)
2.07(2)
2.06(2)
2.04(2)
2.06(1)
2.06(2)
2.06(2)
2.07(2)
2.09(2)
2.06(2)
1.42(3)
1.46(2)
1.46(2)
1.42(3)
1.42(3)
1.41(3)
1.46(3)
1.44(3)
1.44(3)
1.40(3)
1.39(3)
2.900(2)
2.062(8)
2.056(9)
2.028(9)
2.04(1)
2.04(1)
2.055(8)
2.025(8)
2.052(8)
2.036(9)
2.04(1)
1.48(2)
1.43(2)
1.40(2)
1.41(2)
1.40(2)
1.42(2)
1.41(1)
1.41(2)
1.41(1)
1.43(2)
1.42(1)
2.10(1)
2.06(1)
2.02(2)
2.03(2)
2.07(1)
2.05(1)
2.06(1)
2.08(1)
2.06(1)
2.04(1)
1.43(3)
1.42(2)
1.43(2)
1.38(3)
1.40(3)
1.41(2)
1.40(2)
1.43(2)
1.44(2)
1.39(2)
1.42(2)
2.920(2)
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)
C(1)-C(1)a
C(1)-C(2)
C(1)-C(5)
C(2)-C(3)
C(3)-C(4)
C(4)-C(5)
C(6)-C(7)
C(6)-C(10)
C(7)-C(8)
C(8)-C(9)
C(9)-C(10)
I(1)-I(2)
Angles
C(2)-C(1)-C(5)
C(1)-C(2)-C(3)
C(2)-C(3)-C(4)
C(3)-C(4)-C(5)
C(1)-C(5)-C(4)
C(7)-C(6)-C(10)
C(6)-C(7)-C(8)
C(7)-C(8)-C(9)
C(8)-C(9)-C(10)
C(6)-C(10)-C(9)
I(2)-I(1)-I(2)b
107.4(4)
107.7(4)
108.8(4)
108.4(4)
107.8(4)
106.1(4)
109.4(4)
107.4(4)
108.4(5)
108.7(4)
108(1)
107(2)
109(2)
110(2)
107(2)
104(2)
108(2)
108(2)
108(2)
112(2)
179.9
107.8(8)
107(1)
110(1)
107(1)
109(1)
107.2(8)
110.5(9)
105.8(8)
108.7(8)
107.7(9)
107(1)
108(2)
108(1)
109(2)
107(1)
108(1)
107(1)
109(1)
108(1)
108(1)
180.0
F igu r e 9. Plot of tilt angle vs transition temperature.
a
Symmetry equivalents: 1 - x, 1 - y, -z for 25b, 27b, 25c,
b
and 27c. Symmetry equivalents: 1 - x, -y, -z for 27b; -x, -y,
-z for 27c.
Ta ble 5. Com p a r ison of Atom ic Dista n ces (Å) a n d
An gles (d eg)
compound Fe-C
Fe-Cp stagger angle tilt angle Tc (K)a
1b
2.060(6) 1.68(4)
0.0
0.3(3) ∼365
3c
2.06(1) 1.676(5)
2.06(1) 1.677(9)
2.072(5) 1.674
2.07(1) 1.671(6)
2.068(4) 1.675(2)
2.06(1) 1.672(8)
2.07(2) 1.68(1)
2.06(2) 1.668(8)
4.8
6.6
275
245
<4.2
295
265
125
4d
1.2
6e
15.6
3.9(5)
26bf
26cf
27a g
27bf
27cf
0.4(6)
5.9(2)
3.6(1)
16.5(9)
3.5(5)
4.2(2)
5.9(7)
4(1)
>300
190
6.6(7)
F igu r e 10. Variable-temperature 57Fe Mo¨ssbauer spectra
of 26b.
a
Temperature for the localization-delocalization transition on
the Mo¨ssbauer technique. From ref 7. c From ref 33. From ref
b
d
19. e From ref 11. f This work. From ref 29.
g
The interactions between the cation and anion undoubt-
edly alter the potential energy surface of the mixed-
valence molecule and, therefore, the rate of intramo-
lecular electron transfer. Thus, the position of the anion
relative to the mixed-valence cation is rather important
in determining the rate of electron transfer. In other
words, if the charge oscillation of the I3- anion is parallel
to the electron transfer pathway, this would lead to a
stronger ability to pull the charge back and forth in the
mixed-valence biferrocenium cation. In this paper, we
would like to suggest another factor which also plays
an important role in controlling the electron-transfer
rate. We have prepared a series of polyalkylbiferroce-
nium complexes to study the influence of ring tilt
between two least-squares-fitted Cp rings on electron
controlling the intramolecular electron transfer is the
symmetry of the cation. For an asymmetric biferroce-
nium cation, the two irons are not in equivalent en-
vironments and this asymmetry results in a zero-point
energy difference for intramolecular electron transfer.
In other words, one vibronic state of the mixed-valence
cation is energetically more stable than the other state.
This explains why a delocalized electronic structure for
an asymmetic biferrocenium cation has not been ob-
served.
Recently, Hendrickson and co-workers have addressed
counterion effects on intramolecular electron-transfer
rates of biferrocenium salts.13,14 The size, shape, and
charge distribution of anions can ultimately affect the
solid-state arrangement adopted by a particular system.

2782 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
F igu r e 13. Variable-temperature 57Fe Mo¨ssbauer spectra
of 27c.
F igu r e 11. Variable-temperature 57Fe Mo¨ssbauer spectra
of 26c.
Ta ble 6. Cyclic Volta m m etr y for Va r iou s
Bifer r ocen es
a
d
compound
E1/2
,
V
∆E1/2,b V ∆,c mV Ic/Ia
K (×10-6
)
ferrocene
0.37
0.16
0.53
0.13
0.51
0.23
0.59
0.15
0.52
0.14
0.52
0.25
0.61
70
70
71
70
66
68
72
67
65
65
72
89
74
0.97
24a e
0.37
0.38
0.36
0.37
0.38
0.36
1.13
1.01
0.93
1.08
0.93
1.06
0.97
1.02
0.94
1.10
0.90
1.09
1.87
2.76
1.26
1.87
2.76
1.26
24b
24c
25a e
25b
25c
a
All half-wave potentials are referenced to the Ag+/Ag elec-
trode. b Peak separation between waves. c Peak-to-peak separation
between the resolved reduction and oxidation wave maxima.
d
Peak current ratio between the cathode and anode. e From
ref 29.
electron-transfer rate.21 Therefore, we only made a
comparison for mixed-valence compounds 1, 3-4, 6,
26b,c, 27a ,c in which the position of the triiodide anion
relative to the fulvalenide ligand is perpendicular,
rather than parallel to the fulvalenide ligand as found
in 9 and 11. Mixed-valence compounds 10-14 are also
excluded from the comparison because of the strong van
der Waals interaction between the halide substituent
in the benzyl unit and the triiodide anion.21 Under this
condition, i.e., same type of stereopacking arrangement,
there is a correlation between the tilt angle and the rate
of electron transfer in the series of biferrocenium salts
(Figure 9). The substituents in the series of biferroce-
nium cations modify the local structure of the ferrocenyl
moiety. In the qualitative view, we suggest that there
F igu r e 12. Variable-temperature 57Fe Mo¨ssbauer spectra
of 27b.
transfer and to energetically control the electron-
transfer rate.
The mixed-valence compounds 26a -c and 27a -c
serve as a very sensitive probe of the microscopic struc-
ture. We believe that the relatively minor perturba-
tions of the tilt of the Cp ring can have a pronounced
effect on the electronic structure and, therefore, the rate
of intramolecular electron transfer. As mentioned in
the section on the 57Fe Mo¨ssbauer studies, the electronic
effect of the substituents on the electron-transfer rate
was excluded. Here, we suggest that the difference in
the rates of electron transfer of 26a -c and 27a -c is a
result of the difference in the degree of tilting of the Cp
rings from a parallel geometry. As mentioned above,
the interactions between the cation and anion play an
important role in controlling the magnitude of the
2
2
is an increased (dx -y , dxy)-ring overlap as the tilt angle
increases.
The electronic ground state of ferrocene is a singlet,
4
1A1g (e2g a1g2), where the one-electron molecular orbitals
2
are predominantly d orbital in character: a1g (dz ) and
36
2
2
e2g (dx -y , dxy). As indicated by magnetic susceptibil-
(36) Duggan, D. M.; Hendrickson, D. N. Inorg. Chem. 1975, 14,
955.

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2783
Ta ble 7. 57F e Mo1ssba u er Lea st-Squ a r es-F ittin g
P a r a m eter s
δb
Γc
a
compound
T, K
∆EQ
26b
300
290
1.158
1.521
0.850
1.624
0.747
1.782
0.647
1.881
0.599
1.938
0.528
1.224
1.237
1.474
1.024
1.524
0.990
1.828
0.856
2.086
0.690
1.832
0.627
2.029
0.554
1.233
1.280
1.290
1.491
1.101
1.583
1.068
1.708
0.990
0.445
0.454
0.462
0.462
0.455
0.471
0.467
0.488
0.488
0.526
0.525
0.452
0.465
0.472
0.468
0.473
0.472
0.493
0.499
0.527
0.524
0.455
0.450
0.528
0.529
0.450
0.496
0.501
0.501
0.503
0.509
0.524
0.525
0.540
0.504, 0.426
0.454, 0.338
0.439, 0.351
0.366, 0.314
0.413, 0.332
0.293, 0.275
0.389, 0.318
0.286, 0.272
0.446, 0.364
0.284, 0.283
0.343, 0.341
0.395, 0.419
0.457, 0.478
0.393, 0.421
0.372, 0.377
0.378, 0.438
0.375, 0.400
0.403, 0.430
0.476, 0.425
0.287, 0.299
0.347, 0.336
0.267, 0.280
0.368, 0.332
0.252, 0.297
0.346, 0.325
0.433, 0.446
0.470, 0.506
0.492, 0.515
0.446, 0.480
0.407, 0.426
0.466, 0.478
0.412, 0.420
0.461, 0.465
0.438, 0.441
280
250
200
80
26c
300
270
260
250
200
80
27b
27c
300
80
300
220
200
180
150
80
a
b
Quadrupole splitting in mm s-1
.
Isomer shift referenced to
iron-foil in mm s-1
.
c Full width at half-height taken from the least-
F igu r e 14. Fragment orbitals for bending biferrocenium
triiodide from parallel geometry.
squares-fitting program. The width for the line at more positive
velocity is listed first for each doublet.
2
2
Hu¨ckel MO calculations, the e2g (dx -y , dxy) level lies
below the a1g (dz ) level in the mixed-valence biferroce-
Ta ble 8. Absor p tion Ma xim a of IT Ba n d a n d
Activa tion P a r a m eter s
2
nium cation. However, it is now generally accepted that
in mixed-valence biferrocenium cations the e2g level lies
above the a1g level, assigned on the basis of the magnetic
susceptibility and EPR results.7,40 For example, the 4.2
K EPR spectrum of biferrocenium triiodide clearly
demonstrates the large anisotropy in g values (g ) 3.58
and 1.72).7 This large anisotropy and the fast relaxation
in the cation clearly demonstrate that the e2g level must
lie above the a1g level.
a
a
∆ν1/2
∆ν1/2
Ket
Hab (10-12
R2
)
a
b
a
compound νmax ꢀmax (obsd) (calcd)
26b
26c
27b
27c
4651 1323 3044
4651 1200 3128
4651 1388 3128
4651 1139 2976
3278 0.0141 552
3278 0.0132 534
3278 0.0152 573
3278 0.0119 507
2.46
2.30
2.65
2.08
In cm-1
.
Extinction coefficient in M-1 cm-1
.
a
b
ity37 and EPR measurements,38 the electronic ground
2
In this paper, density functional theory (DFT) calcu-
lations of the energy levels of mixed-valence biferroce-
nium triiodide with tilt angles varying from 0° to 20°
have led to the different highest filled level in this
compound. The HOMO molecular level is predomi-
state of ferrocenium is a doublet, 2E2g (a1g e2g3). Lauher
and Hoffmann have derived the fragment orbitals for a
bent (Cp)2Fe unit from the parallel geometry.39 The a1g
orbital rises rapidly in energy as the Cp rings are bent
2
2
back. The e2g set also splits into orbitals of a1 (dx -y
)
2
2
nantly metal dx -y in character. Some metal dxy and
and b2 (dxy) symmetry. In our previous paper,29 we
performed extended Hu¨ckel calculations on a mixed-
valence biferrocenium cation with tilt angles varying
2
2
2
dz orbitals are also mixed into dx -y . In this case, the
2
2
2
metal dz level lies below the dx -y level. Figure 14
illustrates the fragment orbitals for the bending of
biferrocenium triiodide from a parallel geometry. As
can be seen in Figure 14, the dx -y and dxy orbitals rise
rapidly in energy as the Cp rings are bent back. As
shown in Figure 15, the LUMO-HOMO gap decreases
as the tilt angle increases. In other words, the metal
nonbonding orbitals (dx -y , dxy) start to interact with
the ligand pπ orbitals. Hence, bending back the Cp
rings leads to a larger extent of metal-ligand interac-
from 0° to 60° to see how the LUMO-HOMO gap (Ee
Ea ) correlated with the tilt angle. We found the
LUMO-HOMO gap decreases as the tilt angle in-
creases. The LUMO molecular level e1g mainly consists
of metal antibonding dxz and dyz character. The HOMO
1g-
1g
2
2
2
level, a1g, is chiefly a dz nonbonding orbital. In previous
2
2
(37) Hendrickson, D. N.; Sohn, Y. S.; Gray, H. B. Inorg. Chem. 1971,
10, 1559.
(38) (a) Horsfield, A.; Wassermann, A. J . Chem. Soc. A 1970, 3202.
Horsfield, A.; Wassermann, A. J . Chem. Soc., Dalton Trans. 1972, 187.
(b) Prins, R.; Kortbeek, A. J . Organomet. Chem. 1971, 33, C33.
(39) Lauher, J . W.; Hoffmann, R. J . Am. Chem. Soc. 1976, 98, 1729.
(40) Cowan, D. O.; Collins, R. L.; Kaufman, F. J . Phys. Chem. 1971,
75, 2025.

2784 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
matographing on silica gel (70-230 mesh), using n-hexane/
ethyl acetate (20:1) as the eluent. The first band was
unknown, and the second band was 21b (27.54% yield). The
properties of 20b are as follows. 1H NMR (CDCl3, ppm): 0.96
(t, 3H, -CH3), 1.18 (t, 3H, -CH3), 1.48 (m, 1H, -CH2-), 1.59
(m, 1H, -CH2-), 2.47 (m, 1H, -CH2-), 2.66 (m, 1H, -CH2-),
2.93 (m, 2H, -COCH2-), 4.04 (t, 2H, Cp), 4.31 (t, 2H, Cp),
4.39 (d, 2H, Cp), 4.62 (t, 1H, Cp). MS: M+ at m/ z 362, 364.
The properties of 21b are as follows. 1H NMR (CDCl3, ppm):
0.96 (t, 3H, -CH3), 1.19 (t, 3H, -CH3), 1.56 (m, 2H, -CH2-),
2.37 (m, 2H, -CH2-), 2.74 (q, 2H, -COCH2-), 4.07 (t, 2H,
Cp), 4.29 (t, 2H, Cp), 4.38 (t, 1H, Cp), 4.67 (s, 1H, Cp), 4.73 (d,
1H, Cp). MS: M+ at m/ z 362, 364.
Red u ction of 20b a n d 21b. The reduction reaction was
carried out by carefully adding, with stirring, a small portion
of AlCl3 to a mixture of the corresponding ferrocene compound
and LiAlH4 in dry ether. After 30 min, 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 evaporation of the solvent, the crude
product was chromatographed on Al2O3, eluting with hexane.
The first band was the desired compound. The yields of 22b
and 23b were approximately 95% and 67%, respectively. The
physical properties of 22b are as follows. 1H NMR (CDCl3,
ppm): 0.96 (t, 6H, -CH3), 1.49 (m, 4H, -CH2-), 2.29 (t, 4H,
-CH2-), 3.95 (t, 2H, Cp), 4.02 (s, 3H, Cp), 4.17 (t, 2H, Cp).
MS: M+ at m/ z 348, 350. The physical properties of 23b are
as follows. 1H NMR (CDCl3, ppm): 0.93 (t, 6H, -CH3), 1.51
(m, 4H, -CH2-), 2.28 (t, 4H, -CH2-), 3.96 (t, 5H, Cp), 4.18
(t, 2H, Cp). MS: M+ at m/ z 348, 350.
Ben zoyla tion of 1-Ben zyl-1′-br om ofer r ocen e. The ben-
zoylation of 19c was carried out according the same procedure
as the propionation of 1-propyl-1′-bromoferrocene. The red oily
residue was chromatographed on alumina. The first band that
eluted with n-hexane was the starting material. The second
band that eluted with n-hexane/ethyl acetate (20:1) was 20c
(13% yield), and the third band was 21c (15% yield). The
physical properties of 20c are as follows. 1H NMR (CDCl3,
ppm): 4.00 (d, 1H, -CH2-), 4.10 (t, 2H, Cp), 4.36 (t, 2H, Cp),
4.42 (t, 1H, Cp), 4.46 (t, 1H, Cp), 4.54 (d, 1H, -CH2-), 4.54
(m, 1H, Cp), 7.16 (d, 2H, -ph), 7.25 (m, 3H, -ph), 7.43 (t, 2H,
-ph), 7.53 (t, 1H, -ph), 7.79 (dd, 2H, -ph). MS: M+ at m/ z
458, 460. The physical properties of 21c are as follows. 1H
NMR (CDCl3, ppm): 3.79 (s, 2H, -CH2-), 4.10 (t, 2H, Cp),
4.34 (t, 2H, Cp), 4.50 (t, 1H, Cp), 4.88 (dt, 2H, Cp), 7.27 (m,
5H, -ph), 7.46 (t, 2H, -ph), 7.55 (t, 1H, -ph), 7.86 (d, 2H,
-ph). MS: M+ at m/ z 458, 460. Mp: 81-82 °C.
Red u ction of 20c a n d 21c. The reduction reactions of 20c
and 21c were carried out according to the same procedure as
the reduction of 20b and 21b. The physical properties of 22c
are as follows. 1H NMR (CDCl3, ppm): 3.74 (dd, 4H, -CH2-
), 4.00 (t, 2H, Cp), 4.07 (s, 3H, Cp), 4.26 (t, 2H, Cp),7.04 (m,
4H, -ph), 7.15 (m, 6H, -ph). MS: M+ at m/ z 444, 446. The
physical properties of 23c are as follows. 1H NMR (CDCl3,
ppm): 3.67 (s, 4H, -CH2-), 3.97 (t, 2H, Cp), 4.05 (s, 3H, Cp),
4.25 (t, 2H, Cp), 7.18-7.29 (m, 10H, -ph). MS: M+ at m/ z
444, 446.
F igu r e 15. HOMO-LUMO gap as function of tilt angle.
tions, which render intramolecular electron transfer
much more facile.
Exp er im en ta l Section
Gen er a l In for m a tion . All manipulations involving air-
sensitive materials were carried out by using standard Schlenk
techniques under an atmosphere of N2. Chromatography was
performed on neutral alumina (Merck, activity II). Solvents
were dried as follows: benzene, hexane, and ether were
distilled from Na/benzophenone; CH2Cl2 was distilled from
P2O5. Samples of 1-propyl-1′-bromoferrocene8 and 1-benzyl-
1′-bromoferrocene41 were prepared according to the literature
procedure.
P r op ion a tion of 1-P r op yl-1′-br om ofer r ocen e. The pro-
pionating reagent was made-up according to the Friedel-
Crafts synthesis by mixing 2.2 mL (25 mmol) of propionyl
chloride and excess AlCl3 in dry CH2Cl2 (50 mL) at 0 °C under
N2. The excess of AlCl3 was filtered off with glass wool.
The propionating reagent was added by means of a dropping
funnel over a period of about 1 h to a solution of 1-propyl-1′-
bromoferrocene (5.22 g, 17 mmol) in dry CH2Cl2 (130 mL) at
0 °C. The reaction mixture was stirred for 6 h and then poured
into an ice-water mixture. The resulting mixture was sepa-
rated after the reduction of the ferrocenium cation with
aqueous Na2S2O3. The organic layer was washed with satu-
rated aqueous NaHCO3 and H2O, and then it was dried over
MgSO4. The solvent was removed under reduced pressure.
The red oily residue was chromatographed on alumina. The
second band eluted with n-hexane/ethyl acetate (20:1) was 20b
(9% yield). The third band was a mixture of an unknown
product and 21b. This mixture could be separated by rechro-
Ullm a n Cou p lin g Rea ction of 22b,c a n d 23b,c.
A
mixture of the corresponding bromoferrocene (1 g) and acti-
vated Cu (5 g) was heated under N2 at ∼120 °C for 24 h. After
the reaction mixture was cooled to room temperature, it was
repeatedly extracted with CH2Cl2 until the CH2Cl2 extracts
appeared colorless. The solvent was evaporated and chro-
matographed. Compounds 24b,c and 25b,c were recrystal-
lized from hexane. The yields were approximately 90%. The
physical properties of 24b are as follows. 1H NMR (CDCl3,
ppm): 0.87 (t, 12H, -CH3), 1.36 (m, 8H, -CH2-), 2.03 (m, 8H,
-CH2-), 3.83 (d, 2H, Cp), 3.88 (d, 4H, Cp), 4.12 (d, 4H, Cp),
4.16 (d, 4H, Cp). MS: M+ at m/ z 538. Mp: 77.2-78 °C. The
physical properties of 24c are as follows. 1H NMR (CDCl3,
ppm): 3.46 (s, 8H, -CH2-), 3.89 (s, 6H, Cp), 4.18 (s, 4H, Cp),
4.26 (s, 4H, Cp), 6.92 (d, 8H, -ph), 7.10 (m, 12H, -ph). MS:
(41) Yamakawa, K.; Hisatome, M.; Sako, Y.; Ichida, S. J . Organomet.
Chem. 1975, 93, 219.

Mixed-Valence Biferrocenium Triiodides
Organometallics, Vol. 16, No. 13, 1997 2785
M+ at m/ z 731. Mp: 180.1-180.2 °C. The physical properties
of 25b are as follows. 1H NMR (CDCl3, ppm): 0.86 (t, 12H,
-CH3), 1.40 (m, 8H, -CH2-), 2.04 (t, 8H, -CH2-), 3.83 (s,
6H, Cp), 4.14 (t, 4H, Cp), 4.24 (t, 4H, Cp). MS: M+ at m/ z
538. Mp: 51.5-52.5 °C. The physical properties of 25c are
as follows. 1H NMR (CDCl3, ppm): 3.40 (s, 8H, -CH2-), 3.86
(s, 6H, Cp), 4.12 (s, 4H, Cp), 4.24 (s, 4H, Cp), 7.09 (d, 8H, -ph),
7.15 (d, 4H, -ph), 7.22 (d, 8H, -ph). MS: M+ at m/ z 731.
Mp: 152-153 °C.
Mixed -Va len ce Com p ou n d s 26b,c a n d 27b,c. Samples
of these mixed-valence compounds were prepared by adding
a benzene/hexane (1:1) solution containing a stoichiometric
amount of iodine to a benzene/hexane (1:1) solution of the
corresponding biferrocene at 0 °C. The resulting dark green
crystals were filtered and washed repeatedly with cold hexane.
A more crystalline sample can be prepared by slowly diffusing
hexane into a CH2Cl2 solution containing the corresponding
tics. 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.
Str u ctu r e Deter m in a tion of 24c. An orange crystal (0.5
× 0.5 × 0.3 mm) was obtained following the same procedure
as described for 24b. Data were collected to a 2θ value of 50.0°.
Absorption corrections were also made. The unit-cell dimen-
sions were obtained from 20 reflections with 9.3° < 2θ < 14.3°.
Of the 3270 unique reflections, there were 2211 with Fo
>
3.0σFo2. These data were used in the final refinement of the
structural parameters. The structure was solved and refined
as described for 24b.
Str u ctu r e Deter m in a tion of 25b. An orange crystal (0.10
× 0.15 × 0.25 mm) was obtained following the same procedure
as described for 24b. The unit-cell dimensions were obtained
from 25 reflections with 19.72° < 2θ < 26.60°. Of the 1765
unique reflections, there were 1546 with Fo > 2.0σFo2. These
data were used in the final refinement of the structural
parameters. The structure was solved and refined as described
for 24b.
biferrocenium triiodide salt. Anal. Calcd for 26b (C32H42
-
Fe2I3): C, 41.82; H, 4.61. Found: C, 41.28; H, 4.56. Anal.
Calcd for 26c (C48H42Fe2I3): C, 51.88; H, 3.81. Found: C,
51.25; H, 3.82. Anal. Calcd for 27b (C32H42Fe2I3): C, 41.82;
H, 4.61. Found: C, 41.16; H, 4.50. Anal. Calcd for 27c
(C48H42Fe2I3): C, 51.88; H, 3.81. Found: C, 51.59; H, 3.86.
P h ysica l Meth od s. 57Fe Mo¨ssbauer measurements were
made on
a constant-velocity instrument which has been
previously described.21 The absolute temperature accuracy is
estimated to be (5 K, while the relative precision is (0.5 K.
Computer fitting of the 57Fe Mo¨ssbauer data to Lorentzian
lines was carried out with a modified version of a previously
reported program.42 Velocity calibrations were made using a
99.99% pure 10 µm iron foil. Typical line widths for all three
Str u ctu r e Deter m in a tion of 25c. An orange crystal (0.08
× 0.07 × 0.15 mm) was obtained following the same procedure
as described for 24b. The unit-cell dimensions were obtained
from 25 reflections with 13.60° < 2θ < 25.90°. Of the 2294
unique reflections, there were 1118 with Fo > 2.0σFo2. These
data were used in the final refinement of the structural
parameters. The structure was solved and refined as described
for 24b.
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 Dop-
pler effects. It should be noted that the isomer shifts il-
lustrated in the figures are plotted as experimentally obtained.
Tabulated data is provided.
1H NMR spectra were run on a Varian VXR-300 spectrom-
eter. Mass spectra were obtained with a VG-BLOTECH-
QUATTRO 5022 system. The near-IR spectra were recorded
from 2500 to 900 nm in CH2Cl2 by using 1.0 cm quartz cells
with a Hitachi U-3501 spectrophotometer.
Str u ctu r e Deter m in a tion of 26b. A dark black crystal
(0.10 × 0.15 × 0.25 mm) was obtained when a layer of hexane
was allowed to slowly diffuse into a CH2Cl2 solution of 26b.
Data were collected to a 2θ value of 44.9°. The cell dimensions
were obtained from 25 reflections with 2θ in the range 19.72-
26.60°. The data were also corrected for absorption with an
empirical ψ rotation. Of the 2227 unique reflections, there
2
were 1423 with Fo > 2.0σFo2. These data were used in the
final refinement of the structural parameters. Structure
refinement was carried out in the same manner as described
for 24b. The greatest residual electron density upon comple-
tion of refinement was in the vicinity of the iodide atoms.
Close examination of difference Fourier maps in the final
stage of the structure solution revealed one disordered propyl
group in each Cp ring, labeled C15-C16 and C15′-C16′. The
site occupancy factor of the propyl substitutent was set equal
to 0.5.
Electrochemical measurements were carried out with a BAS
100W system. Cyclic voltammetry was perform with a sta-
tionary 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 of (n-C4H9)4NPF6 as the supporting elec-
trolyte. The potentials quoted in this work are relative to a
Ag/AgCl electrode at 25 °C. Under these conditions, ferro-
Str u ctu r e Deter m in a tion of 26c. A dark black crystal
(0.20 × 0.20 × 0.25 mm) was obtained following the same
procedure as described for 26b. Data were collected to a 2θ
value of 45.0°. The cell dimensions were obtained from 25
reflections with 2θ in the range 22.68-28.22°. Of the 2730
cene shows a reversible one-electron oxidation wave (E1/2
0.37 V).
)
The single-crystal X-ray determinations of compounds 24b,c,
25b,c, 26b,c, and 27b,c were carried out on an Enraf-Nonius
CAD4 diffractometer at 298 K. Absorption corrections were
made with empirical ψ rotation. The X-ray crystal data are
summarized in Tables 1 and 2. Selected bond distances and
angles are given in Tables 3 and 4. Tables of the final
positional parameters for all atoms, complete bond distances
and angles, and thermal parameters of these compounds are
given in the Supporting Information.
2
2
unique reflections, there were 2273 with Fo > 2.0σFo
.
Structure refinement was carried out in the same manner as
described for 24b.
The final structure solution revealed a badly disordered
triiodide ion. Iodine atom labeled I2 was refined with two
positions, each at 50% occupancy. Iodine atom labeled I1 was
refined with three positions, each at 33% occupancy.
Str u ctu r e Deter m in a tion of 27b. A dark black crystal
(0.08 × 0.08 × 0.15 mm) was obtained following the same
procedure as described for 26b. Data were collected to a 2θ
value of 45.0°. The cell dimensions were obtained from 25
reflections with 2θ in the range 12.48-24.50°. The data were
also corrected for absorption. Of the 2254 unique reflections,
there were 1329 with Fo2 > 2.0σFo2. Structure refinement was
carried out in the same manner as described for 24b.
Str u ctu r e Deter m in a tion of 27c. A dark black crystal
(0.05 × 0.10 × 0.10 mm) was obtained following the same
Str u ctu r e Deter m in a tion of 24b. An orange crystal (0.1
× 0.1 × 0.15 mm), which was grown by slow evaporation from
a hexane solution, was used for data collection at 298 K. Cell
dimensions were obtained from 20 reflections with 11.82° <
2θ < 30.10°. The θ-2θ scan technique was used to record the
intensities for all reflections for which 1° < 2θ < 44.9°. Of
the 1791 unique reflections, there were 1321 reflections with
2
Fo > 2.0σFo2, where σFo was estimated from counting statis-
(42) Lee, J . F.; Lee, M. D.; Tseng, P. K. Chemistry 1987, 45, 50.

2786 Organometallics, Vol. 16, No. 13, 1997
Dong et al.
procedure as described for 26b. Data were collected to a 2θ
value of 44.9°. The cell dimensions were obtained from 25
reflections with 2θ in the range 13.40-28.38°. The data were
also corrected for absorption. Of the 2715 unique reflections,
there were 1819 with Fo2 > 2.5σFo2. Structure refinement was
carried out in the same manner as described for 24b.
is comparable to 631G**. The local density approximation is
applied in the DFT calculation, where the VWN potential45 is
used. DFT computation is performed on power challenge with
the DMOL44 program.
Ack n ow led gm en t. Acknowledgments are made to
the National Science Council (Grant No. NSC86-2113-
M-110-003), Taiwan, ROC, and Department of Chem-
istry at Sun Yat-sen University. Thanks are also made
to the National High Performance Super Computer
Center for the use of their computer and software.
Molecu la r Or bita l Ca lcu la tion s. The molecular struc-
ture of biferrocenium triiodide geometry is taken from the
X-ray diffraction result. Theoretical calculations of the frag-
ment orbitals for bending of biferrocenium triiodide from a
parallel geometry to a structure with a tilt angle of up to 20°.
DFT calculations are applied to this compound. The basis set
used in DFT is double numerical plus polarization,43,44 which
Su p p or tin g In for m a tion Ava ila ble: Complete tables of
positional parameters, bond lengths and angles, and thermal
parameters for 24b,c, 25b,c, 26b,c, and 27b,c (34 pages).
Ordering information is given on any current masthead page.
(43) (a) Von Barth, U.; Hedin, L. J . Phys. C 1972, 5, 1629. (b)
Versluis, L.; Zeigler, T. J . Chem. Phys. 1988, 88, 3322.
(44) (a) Delly, B. J . Chem. Phys. 1990, 92, 508. DMol is available
commerically from BIOSYM Technologies, San Diego, CA. (b) Delly,
B. Density Functional Methods in Chemistrry; Labanowski, J ., And-
zelm, J ., Eds.; Springer-Verlag: New York, 1991; p 101. (c) Delly, B.
J . Chem. Phys. 1991, 94, 7245.
OM961032R
(45) Vosko, S. J .; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200.