An experimental and computational study of 1,1′-ferrocene diamines

Article abstract of DOI:10.1016/j.poly.2012.08.063

A systematic experimental and computational study of six 1,1′-ferrocene and the corresponding ferrocenium diamines is presented. The influence of alkyl, aryl, and silyl substituents on the nitrogen atoms was studied from a geometric and an electronic point of view.

Full text of DOI:10.1016/j.poly.2012.08.063

Polyhedron 52 (2013) 377–388
Contents lists available at SciVerse ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
An experimental and computational study of 1,10-ferrocene diamines
⇑
´
Selma Duhovic, Paula L. Diaconescu
Department of Chemistry and Biochemistry, University of California, Los Angeles, CA 90095, United States
a r t i c l e i n f o
a b s t r a c t
A systematic experimental and computational study of six 1,10-ferrocene and the corresponding ferroce-
nium diamines is presented. The influence of alkyl, aryl, and silyl substituents on the nitrogen atoms was
studied from a geometric and an electronic point of view.
Article history:
Available online 1 September 2012
Dedicated to Alfred Werner on the 100th
Anniversary of his Nobel Prize in Chemistry
in 1913.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Ferrocene
Substituted amines
CV
DFT
1. Introduction
glovebox. Solvents were purified using a two-column solid-state
purification system by the method of Grubbs and co-workers [23]
Its unintentional discovery in 1951 [1,2] has led to the rebirth
of organometallic chemistry and perhaps no other compound is
more synonymous with that field than ferrocene. Because of its
versatility, ferrocene has intrigued chemists for over 60 years. Its
derivatives have found applications in fields ranging from petro-
chemistry [3–5] to medicine [6–9]. The use of ferrocene derivatives
in our group stems from our interest in supporting ligands with
electronic and steric properties capable of influencing bonding
and enhancing reactivity of transition metals, lanthanides, and
actinides [10–20]. More specifically, our group works with chelat-
ing 1,10-diamidoferrocene ligands, which, as a result of their redox
activity, are capable of accommodating both electronic and steric
changes of metals in various oxidation states [21,22]. In order to
examine the interaction between iron and those metal centers,
we first set out to study systematically how the modification of
nitrogen substituents changes both structural and functional as-
pects of 1,10-ferrocene diamines. Here we report our experimental
and theoretical findings.
and transferred to the glovebox without exposure to air. NMR sol-
vents were obtained from Cambridge Isotope Laboratories, de-
gassed, and stored over activated molecular sieves prior to use. 1H
and 13C NMR spectra were recorded on a Bruker300 spectrometer
(supported by the NSF Grant CHE-9974928) at room temperature
in C6D6. 1H and 13C chemical shifts are reported with respect to
C6D6 residual peak, 7.16 ppm and 128.5 ppm, respectively. Cyclic
voltammetry measurements were conducted on a CH Instruments
CHI630D potentiostat using a 2 mm platinum disk as the working
electrode, 3 mm glassy carbon disk as the counter electrode, and
0.25 mm silver wire as the pseudo-reference electrode. UV–Vis
spectra were recorded on a Varian Carey 5000 spectrophotometer
from 230 to 1600 nm using matched 1 cm quartz cells; all spectra
were obtained using a solvent reference blank in a cuvette with
an air-free Teflon adapter. IR spectra were recorded on a JASCO
FT-IR-420 spectrophotometer from 4000 to 400 cmꢀ1 using a sealed
liquid cell from International Crystals Laboratory with a 1 mm path
length and KBr windows. CHN analysis was performed on an Exeter
Analytical, Inc. CE-440 Elemental Analyzer. Variable temperature
magnetic susceptibility measurements were conducted on
a
Quantum Design MPMS XL superconducting quantum interference
device (SQUID) magnetometer.
2. Experimental methods
All amines, which are easily oxidized in air, were synthesized
following previously published procedures and their purity was
determined by comparison of 1H NMR shifts. Shafir et al. reported
the synthesis of fc[NH2]2 (fc = 1,10-ferrocenylene, NNH) [24],
fc[NHSiMe3]2 (NNTMS) [25], and fc[NHMes]2 (NNMES) [26], while
2.1. General considerations
All experiments were performed under a dry nitrogen atmo-
sphere using standard Schlenk techniques or an MBraun inert-gas
Siemeling, et al. reported the synthesis of fc[NHCH2(t-Bu)]2 (NNNP
)
⇑
Corresponding author.
E-mail address: pld@chem.ucla.edu (P.L. Diaconescu).
[27]. Amines fc[NHSi(t-Bu)Me2]2 (NNTBS) [10] and fc[NHXyl]2
0277-5387/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.poly.2012.08.063

´
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
378
(NNXYL) [18] were previously prepared in our group. Dimethyl-
phenylsilyl chloride was purchased from Alfa Aesar and used as re-
ceived. Tetrabutylammonium hexafluorophosphate (TBAPF6) was
purchased from Sigma–Aldrich and recrystallized from THF before
use.
tetrahydrofuran, dichloromethane, and chloroform but insoluble
in hexanes, pentane, toluene, and benzene. Note: The fine powder
becomes a thick sludge upon exposure to diethyl ether and is hard
to handle. Elemental Anal. Calc. for C35H40FeI3N2 (925 g/mol): C,
45.43; H, 4.36; N, 3.03. Found: C, 45.10; H, 4.27; N, 3.18%.
2.2. Synthesis of fc[NHSiPhMe2]2 (NNDMP
)
3. Experimental results
In a 20-mL scintillation vial charged with a stir bar, NNH
(496 mg, 2.29 mmol) was dissolved in dichloromethane and cooled
to ꢀ190 °C for 30 min. At room temperature, triethylamine
(533 mg, 5.28 mmol) and dimethylphenylsilyl chloride (809 mg,
4.7 mmol) were slowly added to the above solution. The mixture
was stirred for 24 h. Dichloromethane was removed under reduced
pressure to yield an orange oil, which was then dissolved in n-pen-
tane and filtered through alumina. The solution was concentrated
and stored at ꢀ35 °C. Large, dark-red clusters of crystalline needles
formed after 12 h (337 mg, 30% yield). 1H NMR (300 MHz, 25 °C,
C6D6) d (ppm): 7.61 (m, 4H, C6H5), 7.22 (m, 6H, C6H5), 3.71 (s,
8H, Cp–CH), 2.16 (s, 2H, NH), 0.35 (s, 12H, Si(CH3)2). 13C NMR
(500 MHz, 25 °C, C6D6) d (ppm): 139.6 (s, 2C, Si-C6H5), 134.5 (s,
6C, meta/para-C6H5), 130.1 (s, 4C, ortho-C6H5), 65.1 (s, 6C, HN–
C5H4), 60.7 (s, 4C, HN–C5H4), ꢀ0.78 (s, 4C, Si–CH3). Elemental Anal.
Calc. for C26H42N2Si2Fe (485 g/mol): C, 64.45; H, 6.66; N, 5.78.
Found: C, 64.41; H, 6.65; N, 5.55%.
None of the published syntheses of 1,10-ferrocene diamines
were supplemented with comprehensive characterization. Only
1H and 13C NMR spectra, along with elemental analyses, were pro-
vided. Cyclic voltammetry and IR spectroscopy studies of NNH and
NNTBS were also reported [10,24]. A single-crystal X-ray structure
accompanied the former. Therefore, a systematic analysis of their
spectroscopic and electrochemical features relative to ferrocene
was deemed necessary. This report focuses on the 1,10-ferrocene
diamines most commonly used in our group (Chart 1). The discus-
sion begins with experimental results, which are followed by in-
sights from computational analyses.
3.1. Cyclic voltammetry
Because the oxidation state of iron affects the reactivity of metal
complexes supported by disubstituted ferrocene ligands, under-
standing the basic electrochemical properties of 1,10-ferrocene dia-
mines will ultimately enable us to enhance that reactivity through
rational design. One of the most useful characteristics of ferrocene-
based ligands is their reversible oxidation. Cyclic voltammetry
experiments were conducted in THF solutions of tetrabutylammo-
nium hexafluorophosphate (TBAPF6) using a platinum disk as the
working electrode.
2.3. Oxidation of NNTMS, NNTBS, NNDMP, NNMES
2.3.1. [NNTMS][I]
In a 20-mL scintillation vial, a hexanes solution of I2 (3.5 mL,
21 mM, 0.08 mmol) was added to a cold n-pentane solution of
NNTMS (63 mg, 0.17 mmol). The reaction mixture was stirred at
room temperature. A fluffy olive green solid precipitated immedi-
ately but adhered to the vial wall after 15 min, leaving a clear solu-
tion that was then decanted. The product was washed with 10 mL
of diethyl ether and dried under reduced pressure. [NNTMS][I] is
soluble in tetrahydrofuran, dichloromethane, and chloroform but
insoluble in hexanes, n-pentane, diethyl ether, toluene, and ben-
zene. Elemental Anal. Calc. for C16H38N2Si2FeI (487 g/mol): C,
39.43; H, 5.79; N, 5.75. Found: C, 39.16; H, 5.23; N, 5.64%.
All 1,10-ferrocene diamines exhibit one quasi-reversible redox
event and are more easily oxidized than the parent ferrocene
(Fig. 1). This decrease in oxidation potential is expected because
the electron-donating ability of the amino substituents can stabi-
lize the positive charge on iron upon oxidation. Taking only elec-
tronic factors into account, the oxidation potential of NNTBS
should have the greatest negative shift because tert-butyl is a bet-
ter electron donor than either a methyl or phenyl substituent. Sim-
ilarly, the oxidation potential of NNMES should have the smallest
shift because mesityl is less electron donating than all three silyl
substituents. Although the observed order of redox potential devi-
ates from the expected trend based solely on electronic factors, the
difference between the redox potentials is insignificant.
2.3.2. [NNDMP][I]
In a 20-mL scintillation vial, a hexanes solution of I2 (6 mL,
12 mM, 0.07 mmol) was added to a cold n-pentane solution of
NNDMP (45 mg, 0.132 mmol). The reaction mixture was stirred at
room temperature. A bright green solid precipitated immediately
but adhered to the vial wall after 15 min, leaving a clear solution
that was then decanted. The solid was washed with diethyl ether
and dried under reduced pressure. The crude powder was then dis-
solved in approximately 5 mL of THF and crashed out of solution
with addition of an equal volume of hexanes. The precipitate was
collected on a fine-porosity frit and washed with a mixture of
hexanes and diethyl ether. [NNDMP][I] is soluble in tetrahydrofu-
ran, dichloromethane, and chloroform but insoluble in hexanes,
n-pentane, diethyl ether, toluene, and benzene. Elemental Anal.
Calc. for C26H42N2Si2FeI (621 g/mol): C, 50.43; H, 5.29; N, 4.42.
Found: C, 51.07; H, 5.27; N, 4.58%.
The calculated value of the oxidation (see Supporting Informa-
tion for details) potential of ferrocene (0.89 V versus SCE) is in good
agreement with the experimental data (0.80 V [28]). While the
same method correctly predicts lower oxidation potentials for
1,10-ferrocene diamines relative to the parent compound, the
magnitude of the difference as well as the relative order within
the series are not consistent with experimental data (Table 1). For
example, cyclic voltammetry measurements reveal that the
oxidation potential follows the order fc > NNTBS > NNDMP
>
R =
NNH
H
SiMe3
SiMe2 Bu
SiMe2Ph
R
NH
NNTMS
NNTBS
NNDMP
NNNP
2.3.3. [NNMES][I3]
t
In a 20-mL scintillation vial, a hexanes solution of I2 (5 mL,
21 mM, 0.08 mmol) was added to a cold n-pentane solution of
NNMES (56 mg, 0.124 mmol). A yellow-green solid precipitated
immediately and the reaction mixture was stirred at room temper-
ature for 10 min. The precipitate was allowed to settle and after the
supernatant was decanted, it was washed with 10 mL of toluene
and dried under reduced pressure. [NNMES][I3] is soluble in
Fe
t
CH2 Bu
NH
R
NNXYL
3,5-Me2C6H3
2,4,6-Me3C6H2 NNMES
Chart 1. 1,10-Ferrocene diamines.

´
379
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
the orbitals of relatively pure metal character and thus should be
essentially insensitive to the substitution of the cyclopentadienyl
rings [29,33]. The intensity of these forbidden transitions, however,
is greater than that observed for the parent ferrocene (Fig. 2).
Ferrocenium exhibits several bands in the UV and visible re-
gions. For example, a band centered at 617 nm (e )
= 450 Mꢀ1 cmꢀ1
is sensitive to ligand substitution and has therefore been attributed
to ligand-to-metal charge transfer. More specifically, it corre-
sponds to a symmetry-allowed excitation from a ligand-based e1u
to the iron-based e2g (HOMO) orbital [31,34]. Similarly, bands cen-
tered at 283 nm (
e
e
= 9700 Mꢀ1 cmꢀ1), 250 nm (
e -
= 15900 Mꢀ1
cmꢀ1) and 198 nm (
= 14200 Mꢀ1 cmꢀ1) have been assigned to a
single excitation from the ligand-based 1e2u orbital to the iron-
based 2e1g orbital [31]. Finally, four weak bands between 565 nm
and 380 nm
spin-allowed d–d transitions [31].
(e ) have been attributed to
= 150–350 Mꢀ1 cmꢀ1
Fig. 1. Cyclic voltammetry of 1,10-ferrocene diamines in THF with TBAPF6 as the
Absorption spectra of amine-substituted ferrocenium deriva-
tives exhibit one broad band in the visible-NIR region. Based on
the assignment of the parent ferrocenium, it is likely that the band
in the visible-near IR region correspond to a combination of charge
transfer from substituted Cp ligands to Fe3+ and spin-allowed d–d
transitions. The appearance of ligand-to-metal charge transfer
bands at lower energies than in ferrocenium is consistent with
the electrochemical measurements described previously. Two
strong bands in the UV–Vis region are also apparent in the spectra
of 1,10-ferrocenium diamines (Fig. 2). More specifically, these
supporting electrolyte at 50 mV/s scan rate.
Table 1
Experimental and calculated (in THF, PCM method) oxidation potentials of 1,10-
ferrocene diamines.
Oxidation potential
Experimental
Calculated
(V vs. SCE)
(V vs. Fc0/+
)
(V vs. SCE)
NNH
ꢀ0.60a
ꢀ0.69
ꢀ0.70
ꢀ0.72
ꢀ0.73
–
–
bands are centered at 451 nm (
= 1584 Mꢀ1 cmꢀ1) in NNTMS+, 445 nm (
383 nm (
= 4070 Mꢀ1 cmꢀ1) in NNTBS+, 446 nm (
cmꢀ1) and 379 nm ( = 3471 Mꢀ1 cmꢀ1) in NNMES+, and 447 nm
= 1444 Mꢀ1 cmꢀ1) and 387 nm ( = 3195 Mꢀ1 cmꢀ1) in NNDMP+
e
= 945 Mꢀ1 cmꢀ1) and 379 nm
= 3270 Mꢀ1 cmꢀ1) and
= 2673 Mꢀ1
NNTBS
NNDMP
NNMES
NNTMS
ꢀ1.57
ꢀ1.58
ꢀ1.60
ꢀ1.61
ꢀ1.47
ꢀ1.43
ꢀ1.43
ꢀ1.46
(e
e
e
e
-
e
a
From Ref. [24].
(e
e
.
Based on the intensity of these absorption bands and on the fact
that electrochemical data showed a negligible difference between
the electron-donating abilities of the substituted amines, these
bands are assigned to ligand-to-metal charge transfer.
NNMES > NNTMS
,
while the calculations show fc > NNDMP
>
NNMES > NNTMS > NNTBS. The trend predicted by calculations fol-
lows that based on electronic effects. Since both the calculated
and experimental differences between all amines are small, it is
possible that the changes are a consequence of solvent and/or steric
effects.
3.3. Vibrational spectroscopy
In order to gain further insight into the electronic structures, we
studied the infrared spectra of each 1,10-ferrocene diamine in both
neutral and oxidized states, then used computational methods to
facilitate the interpretation of the major vibrational modes. The
calculations were performed for structures in the gas phase, and
in the case of oxidized species, the counter anions were excluded.
Consequently, the calculated bands are systematically shifted from
those observed experimentally. Both the extent of shifting and the
band intensities were dependent on the basis set. For example, as
3.2. Electronic absorption spectroscopy
The absorption spectrum of ferrocene contains two bands in the
UV–Vis region (325 nm and 440 nm), both corresponding to spin-
allowed d–d transitions [29–32]. As shown in Fig. 2, the absorption
spectra of all 1,10-ferrocene diamines exhibit one band centered
between 447 and 452 nm. This red shift is very small and expected
because the electronic transition that occurs at this energy involves
Fig. 2. UV–Vis spectra of neutral (left) and oxidized (right) 1,10-ferrocene diamines.

´
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
380
shown in Fig. 3 for ferrocene, inclusion of polarization functions on
all atoms in the computational method (6-31G⁄) yielded a better
agreement with the experimental spectrum.
The assignment of the fundamental vibrational modes for ferro-
cene is well established [35–37]. Based on our calculations, ferro-
cene shows four strong bands at low energy. The experimental
band at 1106 cmꢀ1 corresponds to both symmetric and asymmet-
ric C–C stretching, while the band at 1004 cmꢀ1 corresponds to C–
H rocking. The lowest energy band at 478 cmꢀ1 is due to ring-metal
stretching, while the band centered at 494 cmꢀ1 corresponds to
symmetric and asymmetric ring-metal-ring tilting [36]. It has been
shown that the perpendicular bending of C–H bonds is affected by
the electron density on iron, and is thus the most diagnostic of its
oxidation state [38]. The frequency of this vibration is centered at
815 cmꢀ1 in ferrocene (Fig. 3; in our hands, the corresponding
vibration was observed at 822 cmꢀ1) and upon oxidation shifts to
851 cmꢀ1. Similar effects have also been observed for polyferro-
cene species [38–42]. Appending the primary amine functionality
to both rings shifts this band to lower frequencies in NNH
Fig. 4. Experimental IR spectra of ferrocene and NNH in CH2Cl2.
(Fig. 4), resulting in a much broader band centered at 801 cmꢀ1
.
Unsurprisingly, our calculations reveal that the band arising
from perpendicular C–H bending shifts to higher energy upon oxi-
dation of the iron center (Fig. 7). For example, symmetric bending
is centered at 851 in NNTMS+ (825 cmꢀ1 in NNTMS) and 847 cmꢀ1 in
NNTBS+ (824 cmꢀ1 in NNTBS), respectively. This band is weak rela-
tive to those arising from Cp C–N stretching at ꢁ1500 cmꢀ1 and
methyl C–H wagging at ꢁ930 cmꢀ1. According to our calculations,
the intense bands around 800 cmꢀ1 (or 900 cmꢀ1 from 6-31G⁄)
found in the spectra of the silyl amines correspond to various
vibrational modes involving methyl groups on the silyl substitu-
ents. As shown in Fig. 7, only silyl derivatives exhibit a strong band
in this region, thus supporting this interpretation.
This width is due to separate vibrations of the individual C–H com-
ponents on the cyclopentadienyl rings. For example, the perpen-
dicular C–H bending of the two fragments closest to the nitrogen
is located at 847 cmꢀ1, while the same mode involving those far-
thest from nitrogen is located at 800 cmꢀ1
.
The spectral features are similar for all 1,10-ferrocene diamines
(Fig. 5) and were interpreted based on the results of DFT calcula-
tions. The N–H and C–H stretches are apparent between
ꢁ3000 cmꢀ1 and 3500 cmꢀ1 (Fig. 5). The strongest band, observed
at approximately 1500 cmꢀ1, corresponds to Cp C–N stretching,
while the weaker bands at 1200–1300 cmꢀ1 are due to wagging
of Cp rings relative to the iron center. At approximately
1100 cmꢀ1, a single band is present only in NNDMP and NNMES
and involves stretching of the phenyl rings. The set of bands in
its vicinity (1000–1100 cmꢀ1) represent asymmetric twisting of
the Cp rings. All 1,10-ferrocene diamines have three strong bands
between 830 cmꢀ1 and 940 cmꢀ1, which are due to various stretch-
ing and bending modes involving the carbon and hydrogen atoms
of the amine substituents. The bands of those bearing silylamino
substituents, however, appear to be sharper and better defined
than those of NNMES. Reflecting the greater electron density on iron
due to donating amine substituents, the perpendicular C–H bend-
ing was calculated to occur at lower energy for all 1,10-ferrocene
diamines (750–800 cmꢀ1). Moreover, the corresponding bands
are predicted to be inherently weak, and, consequently, are not ob-
served experimentally (Fig. 6). Finally, the ring-metal-ring tilting
seems to be insensitive to substitution, and appears between
4. Computational treatment
4.1. Methods
The computational study of the title compounds was conducted
with GAUSSIAN09 [43], using the unrestricted hybrid functional
B3LYP. Moreover, atoms were modeled with an all-electron 6-
31G⁄ basis set and, unless noted otherwise, all species discussed
in this report correspond to minima with no imaginary frequen-
cies. For the sake of comparison, we also employed an all-electron
LANL2DZ [44,45] basis set. In order to ascertain the accuracy of the
computational methods used, we performed the same calculations
on ferrocene/ferrocenium using different methods and basis sets,
and compared our results to both experimental [46,47] and theo-
retical [48–51] data reported previously. As shown in Table 2,
400 cmꢀ1 and 500 cmꢀ1
.
Fig. 3. Comparison of experimental and theoretical IR spectrum of ferrocene.

´
381
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
Fig. 5. IR spectra of neutral 1,10-ferrocene diamines in dichloromethane.
Fig. 6. IR spectrum of NNTMS in the region 2000–400 cmꢀ1 obtained experimentally and computationally.
results obtained using the unrestricted B3LYP functional (uB3LYP)
[52,53] and 6-31G⁄ [54–60] basis set were in best agreement with
the experimental data. Because no structural parameters exist for
1,10-ferrocene diamines, we built upon the crystallographic coordi-
nates of NNH and optimized the geometry of each 1,10-ferrocene
diamine using a fragment-based algorithm [61]. No thermal
parameters were specified, so all calculated temperature-depen-
dent values assume 298 K.
What follows is a discussion of structural features and frontier
molecular orbitals. Each section begins with a discussion of the
neutral species, followed by a comparison with the oxidized deriv-
atives. Furthermore, to put our results into context, we preface

´
382
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
Fig. 7. IR spectra of oxidized 1,10-ferrocene diamines in dichloromethane.
Table 2
Comparison of the distance between Fe and the center of Cp rings obtained
experimentally and with different computational methods.
Theory
uB3LYP
Basis
Fe–Cp (Å)
LANL2DZ
cc-pVTZ
6-31G⁄
6-31G⁄
6-31G⁄
1.73
1.69
1.65
1.62
1.65
1.66a
uBP86
uBLYP
Experimental
a
From Refs. [47,62].
each discussion with an overview of the parent compounds (i.e.
ferrocene and/or NNH).
Fig. 8. Comparison of calculated and experimental distances between iron and Cp
4.2. Structural features
rings in ferrocene and ferrocenium derivatives. aFrom Ref. [65].
The structural parameters of the optimized geometry of ferro-
cene are dependent on the basis set (Table 2). For example, using
LANL2DZ, the distance between iron and the center of each Cp ring
is 1.73 Å and 1.76 Å in ferrocene and ferrocenium, respectively. The
values decrease to 1.65 Å and 1.68 Å when the 6-31G⁄ basis set is
used. As mentioned above, the latter is in better agreement with
the experimental data [46,48,63,64]. As shown in Fig. 8, the change
in Fe–Cp distance upon oxidation of the iron center is nonetheless
insignificant (0.029 Å in LANL2DZ and 0.023 Å in 6-31G⁄).
A more significant difference lies in the conformation of the Cp
rings relative to each other. Previously published studies have
established that the Cp rings are staggered in the solid state [66–
69] but eclipsed in the gas phase [46,62,70]. Similarly, others have
calculated the rotational barrier to be 0.9–1.1 kcal/mol and from
crystallographic data, concluded that up to 60% of ferrocene mole-
cules in a single crystal feature Cp rings in an eclipsed conforma-
tion [47]. Theoretical studies have also predicted the greater
stability of the eclipsed isomer [48]. Consistent with these studies,
our calculations show that in the gas phase, ferrocene has D5h sym-
metry. However, the staggered isomer is predicted to be essentially
isoenergetic (+0.11 kcal/mol). Such a low barrier to rotation is con-
sistent with previous studies. While both basis sets predict ferro-
cene to be eclipsed, 6-31G⁄ shows that the rings are staggered
when iron is in the higher oxidation state. More specifically, con-
vergence to an eclipsed isomer of ferrocenium using 6-31G⁄ is

´
383
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
E-1,1’
E-1,2’
E-1,3’
Synperiplanar
Synclinal Eclipsed
Anticlinal
S-1,2’
Anticlinal Staggered
S-1,1’
Synclinal
S-1,3’
Antiperiplanar
Fig. 9. Possible conformations of symmetrically disubstituted ferrocene systems.
first-order saddle point, which corresponds to a transitional struc-
ture for the rotation of the Cp rings.
Unsurprisingly, the optimized geometry of each 1,10-ferrocene
diamine derivative was dependent on the conformation of the ini-
tial structure. The most stable structures (Fig. 11), which all further
discussion will be based upon, did not depend on the basis set,
however. It is evident that the orientation of the amino substitu-
ents changes to either a 1,20 (NNTMS, NNTBS, NNDMP, NNNP) or 1,30
(NNXYL and NNMES) eclipsed conformation, as a consequence of ste-
ric and electronic interactions. Moreover, in the latter class, both
amine hydrogen atoms point outward, while in the silyl and alkyl
derivatives, only one H atom points inward.
Fig. 10. Experimental (solid) and calculated (opaque) structure of E-1,10-NNH.
Geometry optimizations of the respective oxidized 1,10-ferro-
cene diamines were conducted by using the optimized Cartesian
coordinates of the neutral counterparts at two different conforma-
tions but changing the charge and multiplicity values in the input
file. The most stable conformations are shown in Fig. 12. Interest-
ingly, all ferrocenium diamines maintained the structural motif of
accompanied by one imaginary frequency centered at approxi-
mately ꢀ5 cmꢀ1, suggesting that the structure corresponds to a
transitional state for the rotation of the Cp rings. The barrier to
rotation is less than 1 kcal/mol. These results are consistent with
previous theoretical and experimental reports of various ferroce-
nium salts [65,71–73].
Having established the validity of both methods, we performed
geometry optimizations and frequency analysis for six 1,10-
ferrocene diamines bearing silyl, alkyl, and aryl substituents, as
well as the parent 1,10-ferrocene diamine. For the sake of clarity,
the following discussion will use structural motifs and nomencla-
ture (Fig. 9) in line with the established categorization of 1,10-
disubstituted ferrocene systems [63,64].
the corresponding ferrocene diamine, with the exception of NNNP
,
which gave the 1,30 conformer. In addition, the degree of staggering
of the Cp rings increased in the oxidized diamines and varies from
small in NNTMS+ to medium in NNTBS+ and NNDMP+ to large in
NNNP+, NNXYL+, and NNMES+
.
As shown in Table 3, the distance between the Cp carbon and
nitrogen atoms decreases for all 1,10-ferrocene diamines, while
the distance between iron and the ring’s center (as well as each
carbon atom) increases upon removal of an electron. For example,
the distance between iron and each Cp ring changes by approxi-
mately 0.07 Å (Table 3). This is significantly greater than the in-
Structural features of NNH (Fig. 10), calculated using both basis
sets, are in good agreement with those measured experimentally. It
has previously been shown that two different conformations of
NNH exist in the same crystallographic unit cell [24]. For E-1,10-
NNH, the distance between iron and the center of the Cp ring
was overestimated by less than 4% using LANL2DZ (Fe–Cp dis-
tance = 1.74 Å versus 1.66 Å). When polarization functions were
added to all atoms, the structure was modeled more accurately
(Fe–Cp distance = 1.658 Å). Vibrational analysis of E-1,10-NNH
using LANL2DZ and 6-31G⁄, however, yielded one imaginary
frequency centered at ꢀ23 cmꢀ1 and ꢀ33 cmꢀ1, respectively. This
implies that both methods model the eclipsed geometry as a
crease in ferrocene (
D = 0.02 Å) and is a reflection of weaker
metal–ligand bonding brought about upon oxidation [74]. It has
been suggested that this increase in metal–ligand distance is due
to removal of an electron from the bonding e2 orbital or the simul-
taneous electronic rearrangement [46]. While increasing the oxida-
tion state decreases the metal–ligand distance for main group
metal ions (due to the decrease in ionic size as well as the increase
in electron affinity of the metal center), the same effect is not al-
ways true for transition metals [75].

´
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
384
Fig. 11. Conformation of neutral 1,10-ferrocene diamines at minimum potential energy: E-1,20-NNTMS, E-1,20-NNTBS, E-1,20-NNDMP, E-1,20-NNNP, E-1,30-NNXYL, and E-1,30-
NNMES
.
Fig. 12. The most stable conformations of oxidized ferrocene diamines (at the minimum potential energy surface): E-1,20-NNTMS+, S-1,30-NNTBS+, E/S-1,20-NNDMP+, S-1,30-
NNNP+, S-1,30-NNXYL+, and S-1,30-NNMES+
.

´
385
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
Table 3
A more interesting structural change brought on by oxidation is
the distortion of the cyclopentadienyl rings. That is, the planar con-
formation of both rings becomes disrupted upon oxidation
(Fig. 13). This deviation is more pronounced in alkyl- and aryl- than
in silyl-substituted 1,10-ferrocene diamines. Loss of ring aromatic-
ity and disruption of conjugation is expected to weaken the overlap
between Fe3+ and ring orbitals [76]. Because the iron-ring interac-
tion provides the greatest stabilization within the ferrocene back-
bone [31], such a structural perturbation may be important not
only for the oxidized 1,10-ferrocene diamines, but also for the metal
centers that they support.
Change in structural parameters observed upon oxidation of discussed compounds.
6-31G⁄
Fe–Cp (Å)
Fe2+
CCp–N (Å)
Fe2+
Fe3+
D
Fe3+
D
NNMES
NNTMS
NNTBS
NNXYL
NNDMP
NNNP
fc
1.661
1.661
1.661
1.661
1.662
1.665
1.654
1.734
1.731
1.730
1.730
1.730
1.733
1.677
0.073
0.070
0.069
0.069
0.068
0.068
0.023
1.413
1.399
1.400
1.400
1.400
1.400
–
1.353
1.357
1.358
1.355
1.357
1.345
–
ꢀ0.060
ꢀ0.041
ꢀ0.042
ꢀ0.045
ꢀ0.043
ꢀ0.055
–
4.3. Electronic structure of neutral and oxidized 1,1-ferrocene
diamines
Both LANL2DZ and 6-31G⁄ basis sets paint a qualitatively simi-
lar picture of the electronic structure of ferrocene (Fig. 14). The
ground state configuration of eclipsed ferrocene (D5h) was calcu-
lated to possess highest occupied orbitals of a01 and e02 symmetry.
The energy difference between these two orbitals is small and their
relative ordering in the staggered conformation has been debated
[29–32,51,77,78]. Studies favoring a1 > e2 were based on a stag-
gered conformation. Indeed, one report describing the substitution
effects on the electronic structure of ferrocene has shown that the
energy order of a1 and e2 orbitals is dependent on the relative con-
formation of the Cp rings [79]. Our calculations of the eclipsed iso-
mer show that e02 > a01 (Fig. 14).
As shown in Figs. 15 and 16, the HOMO remains iron-based in
all except NNXYL and NNMES species, whose highest occupied
molecular orbitals show a mixing of iron and nitrogen atomic orbi-
tals. Furthermore, the energy of the HOMO in all derivatives in-
creases relative to ferrocene (Fig. 17). The increase in energy of
the highest occupied molecular orbital is greatest in derivatives
Fig. 13. Distortion of the cyclopentadienyl ring as defined by the out-of-place
displacement of the amino carbon.
Fig. 14. MO diagrams of ferrocene (D5h) and its derivatives (C1); energy levels showing electron pairs represent the highest occupied MO.

´
386
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
Fig. 15. Atomic contributions to highest occupied (left) and lowest unoccupied (right) molecular orbitals.
Fig. 16. HOMO (left) and LUMO (right) of 1,10-ferrocene diamines.
bearing silyl substituents. This is expected because silyl groups are
stronger electron donors than aryl substituents.
5. Conclusions
As shown in Fig. 17, with the exception of NNDMP, NNXYL, and
NNMES, the energy of the lowest unoccupied molecular orbital also
increases relative to ferrocene. These exceptions, however, are con-
sistent with the nature of the LUMO on 1,10-ferrocene diamines
bearing aromatic groups (Figs. 15 and 16), which is localized lar-
gely on the phenyl rings, with predominant contributions originat-
We have presented a systematic experimental and computa-
tional treatment of several 1,10-ferrocene and ferrocenium dia-
mines bearing alkyl, aryl, and silyl substituents. Electrochemical
measurements reveal that the amine groups on Cp rings donate
electron density to the iron center, allowing it to undergo oxidation
more easily. The extent of this change in oxidation potential,
however, is not greatly dependent on the various substituents
ing from the phenyl
p orbitals.

´
387
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
Fig. 17. Energy of highest occupied and lowest unoccupied molecular orbitals relative to those of ferrocene, which are defined as 0 kcal/mol.
appended to the amine group. Similarly, electronic spectra of 1,10-
ferrocenium diamines reveal that the energy of charge transfer
from Cp to Fe(III) does not depend on the nitrogen substituents.
Accordingly, DFT calculations also show that the distance between
iron and the Cp rings, is nearly identical within the series. Finally,
computational treatment of 1,10-ferrocene diamines shows that
aromatic substituents on the nitrogen atoms impose the greatest
influence on the electronic structure of the neutral species not only
by lowering the energy difference between frontier orbitals but
also by changing the nature of the lowest occupied molecular orbi-
tal from metal to ligand character.
´
[15] S. Duhovic, M.J. Monreal, P.L. Diaconescu, Inorg. Chem. 49 (15) (2010) 7165.
[16] W. Huang, S.I. Khan, P.L. Diaconescu, J. Am. Chem. Soc. 133 (27) (2011) 10410.
[17] E.M. Broderick, N. Guo, C.S. Vogel, C. Xu, J. Sutter, J.T. Miller, K. Meyer, P.
Mehrkhodavandi, P.L. Diaconescu, J. Am. Chem. Soc. 133 (24) (2011) 9278.
[18] J.A. Lee, B.N. Williams, K.R. Ogilby, K.L. Miller, P.L. Diaconescu, J. Organomet.
Chem. 696 (25) (2011) 4090.
[19] M.J. Monreal, P.L. Diaconescu, Organometallics 27 (8) (2008) 1702.
[20] B.N. Williams, W. Huang, K.L. Miller, P.L. Diaconescu, Inorg. Chem. 49 (24)
(2010) 11493.
[21] A. Shafir, J. Arnold, J. Am. Chem. Soc. 123 (37) (2001) 9212.
[22] D. Astruc, Acc. Chem. Res. 19 (12) (1986) 377.
[23] A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J. Timmers,
Organometallics 15 (5) (1996) 1518.
[24] A. Shafir, M.P. Power, G.D. Whitener, J. Arnold, Organometallics 19 (19) (2000)
3978.
[25] A. Shafir, M.P. Power, G.D. Whitener, J. Arnold, Organometallics 20 (7) (2001)
1365.
Acknowledgments
[26] A. Shafir, J. Arnold, Inorg. Chim. Acta 345 (2003) 216.
[27] U. Siemeling, C. Farber, M. Leibold, C. Bruhn, P. Mucke, R.F. Winter, B. Sarkar, M.
von Hopffgarten, G. Frenking, Eur. J. Inorg. Chem. 2009 (31) (2009) 4607.
[28] N.G. Connelly, W.E. Geiger, Chem. Rev. 96 (2) (1996) 877.
[29] D.R. Scott, R.S. Becker, J. Organomet. Chem. 4 (5) (1965) 409.
[30] A.T. Armstrong, F. Smith, E. Elder, S.P. McGlynn, J. Chem. Phys. 46 (11) (1967)
4321.
[31] H.B. Gray, Y.S. Sohn, N. Hendrickson, J. Am. Chem. Soc. 93 (15) (1971) 3603.
[32] M.-M. Rohmer, A. Veillard, M.H. Wood, Chem. Phys. Lett. 29 (3) (1974) 466.
[33] V. Weinmayr, J. Am. Chem. Soc. 77 (11) (1955) 3012.
[34] R. Prins, J. Chem. Soc., Chem. Commun. 5 (1970) 280.
[35] E.R. Lippincott, R.D. Nelson, Spectrochim. Acta 10 (3) (1958) 307.
[36] E.R. Lippincott, R.D. Nelson, J. Chem. Phys. 21 (7) (1953) 1307.
[37] E.R. Lippincott, R.D. Nelson, J. Am. Chem. Soc. 77 (19) (1955) 4990.
[38] J.A. Kramer, D.N. Hendrickson, Inorg. Chem. 19 (11) (1980) 3330.
[39] T.Y. Dong, D.N. Hendrickson, C.G. Pierpont, M.F. Moore, J. Am. Chem. Soc. 108
(5) (1986) 963.
[40] T.Y. Dong, T.L. Hsu, J. Organomet. Chem. 367 (3) (1989) 313.
[41] J. Jiao, G.J. Long, L. Rebbouh, F. Grandjean, A.M. Beatty, T.P. Fehlner, J. Am.
Chem. Soc. 127 (50) (2005) 17819.
We acknowledge Kathy I. On for synthesizing NNDMP. This work
was supported by the U.S. Department of Energy, Office of Basic
Energy Sciences, Heavy Element Chemistry program under Grant
No. ER15984, by the National Science Foundation under a CAREER
award (Grant CHE-0847735), and the Sloan Foundation.
Appendix A. Supplementary data
Experimental (cyclic voltammograms, UV-Vis and IR spectra)
and computational (XYZ coordinates, frequency analyses, and re-
dox potentials) details can be found in the supporting information
file. Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.poly.2012.
08.063.
[42] S.K. Pal, K. Alagesan, A.G. Samuelson, J. Pebler, J. Organomet. Chem. 575 (1)
(1999) 108.
References
[43] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro,
M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, T. Keith, R.
Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken,
C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R.
Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski,
G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B.
Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision B.01,
Gaussian, Inc., Wallingford, CT, 2009.
[1] T.J. Kealy, P.L. Pauson, Nature 168 (4285) (1951) 1039.
[2] S.A. Miller, J.A. Tebboth, J.F. Tremaine, J. Chem. Soc. (1952) 632.
[3] K.E. Ritrievi, J.P. Longwell, A.F. Sarofim, Combust. Flame 70 (1) (1987) 17.
[4] T. Hirasawa, C.-J. Sung, Z. Yang, A. Joshi, H. Wang, Combust. Flame 139 (4)
(2004) 288.
[5] A. Braun, F.E. Huggins, K.E. Kelly, B.S. Mun, S.N. Ehrlich, G.P. Huffman, Carbon
44 (14) (2006) 2904.
[6] S. Top, B. Dauer, J. Vaissermann, G. Jaouen, J. Organomet. Chem. 541 (1–2)
(1997) 355.
[7] S. Top, A. Vessieres, C. Cabestaing, I. Laios, G. Leclercq, C. Provot, G. Jaouen, J.
Organomet. Chem. 637–639 (2001) 500.
[8] L. Delhaes, H. Abessolo, C. Biot, L. Berry, P. Delcourt, L. Maciejewski, J. Brocard,
D. Camus, D. Dive, Parasitol. Res. 87 (3) (2001) 239.
[9] S. Top, A. Vessieres, G. Leclercq, J. Quivy, J. Tang, J. Vaissermann, M. Huché, G.
Jaouen, Chem. Eur. J. 9 (21) (2003) 5223.
[44] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1) (1985) 299.
[45] P.J. Hay, T.H. Dunning, Modern Theoretical Chemistry, vol. 3, Plenum, New
York, 1976.
[46] A. Haaland, Acc. Chem. Res. 12 (11) (1979) 415.
[10] M.J. Monreal, C.T. Carver, P.L. Diaconescu, Inorg. Chem. 46 (18) (2007) 7226.
[11] C.T. Carver, M.J. Monreal, P.L. Diaconescu, Organometallics 27 (3) (2008) 363.
[12] E.M. Broderick, P.L. Diaconescu, Inorg. Chem. 48 (11) (2009) 4701.
[13] C.T. Carver, D. Benitez, K.L. Miller, B.N. Williams, E. Tkatchouk, W.A. Goddard,
P.L. Diaconescu, J. Am. Chem. Soc. 131 (29) (2009) 10269.
[14] P.L. Diaconescu, Comments Inorg. Chem. 31 (5–6) (2010) 196.
[47] A. Haaland, J.E. Nilsson, Acta Chem. Scand. 22 (8) (1968) 2653.
[48] M.J. Mayor-Lopez, J. Weber, Chem. Phys. Lett. 281 (1–3) (1997) 226.
[49] W. Klopper, H.P. Lüthi, Chem. Phys. Lett. 262 (5) (1996) 546.
[50] H. Koch, P. Jorgensen, T. Helgaker, J. Chem. Phys. 104 (23) (1996) 9528.
[51] K. Pierloot, B.J. Persson, B.O. Roos, J. Chem. Phys. 99 (11) (1995) 3465.
[52] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (2) (1988) 785.

´
S. Duhovic, P.L. Diaconescu / Polyhedron 52 (2013) 377–388
388
[53] A.D. Becke, J. Chem. Phys. 98 (7) (1993) 5648.
[67] P.F. Eiland, R. Pepinsky, J. Am. Chem. Soc. 74 (19) (1952) 4971.
[68] J.D. Dunitz, L.E. Orgel, Nature 171 (4342) (1953) 121.
[69] J.D. Dunitz, L.E. Orgel, A. Rich, Acta Crystallogr. 9 (4) (1956) 373.
[70] R.K. Bohn, A. Haaland, J. Organomet. Chem. 5 (5) (1966) 470.
[71] R.F. Kirchner, G.H. Loew, U.T. Mueller-Westerhoff, Theor. Chim. Acta 41 (1)
(1976) 1.
[72] I. Chàvez, A. Alvarez-Carena, E. Molins, A. Roig, W. Maniukiewicz, A. Arancibia,
V. Arancibia, H. Brand, J. Manuel Manriquez, J. Organomet. Chem. 601 (1)
(2000) 126.
[54] R. Ditchfield, W.J. Hehre, J.A. Pople, J. Chem. Phys. 54 (2) (1971) 724.
[55] M.M. Francl, W.J. Pietro, W.J. Hehre, J.S. Binkley, M.S. Gordon, D.J. DeFrees, J.A.
Pople, J. Chem. Phys. 77 (7) (1982) 3654.
[56] P.C. Hariharan, J.A. Pople, Mol. Phys. 27 (1) (1974) 209.
[57] P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (3) (1973) 213.
[58] W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (5) (1972) 2257.
[59] V.A. Rassolov, J.A. Pople, M.A. Ratner, T.L. Windus, J. Chem. Phys. 109 (4) (1998)
1223.
[60] V.A. Rassolov, M.A. Ratner, J.A. Pople, P.C. Redfern, L.A. Curtiss, J. Comput.
Chem. 22 (9) (2001) 976.
[73] P.S. Bagus, U.I. Walgren, J. Almlof, J. Chem. Phys. 64 (6) (1976) 2324.
[74] P. Carty, K.C. Clare, J.R. Creighton, E. Metcalfe, E.S. Raper, H.M. Dawes, Inorg.
Chim. Acta 112 (2) (1986) 113.
[61] D.C. Young, Computational Chemistry:
A Practical Guide for Applying
Techniques to Real World Problems, Wiley, 2001.
[62] A. Haaland, Top. Curr. Chem. 53 (1975) 1.
[75] M. Swart, Inorg. Chim. Acta 360 (1) (2007) 179.
[76] A.G. Nagy, I. Dezsi, M. Hillman, J. Organomet. Chem. 117 (1) (1976) 55.
[77] D.C. Driscoll, P.A. Dowben, N.M. Boag, M. Grade, S. Barfuss, J. Chem. Phys. 85 (9)
(1986) 4802.
[78] C.J. Ballhausen, J.P. Dahl, O. Buchardt, G.E. Olsen, C. Pedersen, J. Toft, Acta
Chem. Scand. 15 (1961) 1333.
[63] S.I. Kirin, H.-B. Kraatz, N. Metzler-Nolte, Chem. Soc. Rev. 35 (4) (2006) 348.
[64] K.-S. Gan, T.S.A. Hor, 1,10-Bis(diphenylphosphino)ferrocene: coordination
chemistry, organic synthesis, and catalysis, in: T.H. Antonio Togni (Ed.),
Ferrocenes: Homogeneous Catalysis, Organic Synthesis, Materials Science,
VCH Publishers, New York, 1995, p. 3.
[79] G. Zhang, H. Zhang, M. Sun, Y. Liu, X. Pang, X. Yu, B. Liu, Z. Li, J. Comput. Chem.
28 (14) (2007) 2260.
[65] R. Martinez, A. Tiripicchio, Acta Crystallogr., Sect. C 46 (2) (1990) 202.
[66] W. Pfab, E.O. Fischer, Z. Anorg. Allg. Chem. 274 (6) (1953) 316.