Dihydrogen to dihydride isomerization mechanism in [(C5Me 5)FeH2(Ph2PCH2CH2PPh 2)]+ through the experimental and theoretical analysis of kinetic isotope effects

Article abstract of DOI:10.1021/ic061428n

The isomerization of complex [Cp*Fe(dppe)(η²-H ₂)]⁺, generated in situ by low-temperature protonation of Cp*Fe(dppe)H with either HBF₄ or CF₃COOH, to the dihydride tautomer trans-[Cp*Fe(dppe)(H)₂]⁺ is irreversible and follows first-order kinetics in the -10 to +15°C range with ΔH^? = 21.6 ± 0.8 kcal mol^-1 and ΔS^? = 5 ± 3 eu. The isomerization rate constant is essentially independent of the nature and quantity of a strong acid. Density functional theory (DFT) calculations on various models, including the complete system at both the quantum mechanics/ molecular mechanics (QM/MM) and full QM levels, probe the relative importance of steric and electronic effects for the relative stability of the nonclassical and classical isomers and identify two likely isomerization mechanisms: a direct pathway involving simultaneous H-H bond breaking and cis-trans isomerization and a via Cp pathway involving agostic C₅Me₅H intermediates. Both pathways are characterized by activation energies in close correspondence with the experimental value (21.3 and 22.2 kcal mol^-1, respectively). Further kinetic studies were carried out for the Cp*Fe(dppe)H + CF ₃COOD and Cp*Fe(dppe)D + CF₃COOD systems at 273 K. The [Cp*Fe(dppe)(η²-HD)]⁺ complex establishes a very rapid isotope redistribution equilibrium with the η²-H ₂ and η²-D₂ analogues. The equilibrium constant value (K = 3.3 ± 0.3) indicates a significant equilibrium isotope effect. Simulation of the rate data provides access to the individual isomerization rate constants k_HH, k_HD, and k_DD for the three isotopomers, yielding kinetic isotope effects: k _HH/k_HD = 1.24 ± 0.01 and k_HD/k _DD = 1.58 ± 0.01 (and, consequently, k_HH/k _DD = 1.96 ± 0.02). The analysis of the DFT-calculated frequencies, using the [Cp*Fe(dhpe)H₂]⁺ model system, for the [Cp*Fe(dhpe)(η²-XY)]⁺ isotopomers as well as transition states for the direct (TS _dir) and via Cp (TS_rot) pathways (X = H, D) allowed the computation of the expected isotope effects. A comparison with the experiment strongly suggests that the mechanism occurs via the direct pathway for the present system, although the small difference in the calculated energy barriers suggests that the via Cp pathway may be preferred in other cases.

Full text of DOI:10.1021/ic061428n

Inorg. Chem. 2006, 45, 10248−10262
Dihydrogen to Dihydride Isomerization Mechanism in
+
[(C5Me5)FeH2(Ph2PCH2CH2PPh2)] through the Experimental and
Theoretical Analysis of Kinetic Isotope Effects
|
Miguel Baya,†,‡,§ Olivier Maresca,‡ Rinaldo Poli,*,† Yannick Coppel,† Feliu Maseras,‡, Agust´ı Lledo´s,*,‡
Natalia V. Belkova, Pavel A. Dub, Lina M. Epstein, and Elena S. Shubina*,
Laboratoire de Chimie de Coordination, UPR CNRS 8241 lie´e par conVention a` l’UniVersite´
Paul Sabatier et a` l’Institut National Polytechnique de Toulouse, 205 Route de Narbonne,
31077 Toulouse Cedex, France, Departament de Qu´ımica, Edifici cn, UniVersitat Auto`noma de
Barcelona, 08193 Bellaterra, Catalonia, Spain, and NesmeyanoV Institute of Organoelement
Compounds, Russian Academy of Sciences, VaViloV Street 28, 119991 Moscow, Russia
Received July 29, 2006
The isomerization of complex [Cp*Fe(dppe)(η
2-H2)]+, generated in situ by low-temperature protonation of
Cp*Fe(dppe)H with either HBF4 or CF3COOH, to the dihydride tautomer trans-[Cp*Fe(dppe)(H)2]+ is irreversible
1
and follows first-order kinetics in the
−10 to
+15
°
C range with
∆
Hq
)
21.6
±
0.8 kcal mol- and
∆
Sq
)
5
±
3
eu. The isomerization rate constant is essentially independent of the nature and quantity of a strong acid. Density
functional theory (DFT) calculations on various models, including the complete system at both the quantum mechanics/
molecular mechanics (QM/MM) and full QM levels, probe the relative importance of steric and electronic effects for
the relative stability of the nonclassical and classical isomers and identify two likely isomerization mechanisms: a
“direct” pathway involving simultaneous H−H bond breaking and cis−trans isomerization and a “via Cp” pathway
involving agostic C5Me5H intermediates. Both pathways are characterized by activation energies in close
correspondence with the experimental value (21.3 and 22.2 kcal mol-1, respectively). Further kinetic studies were
carried out for the Cp*Fe(dppe)H
[Cp*Fe(dppe)(
2-HD)]+ complex establishes a very rapid isotope redistribution equilibrium with the
analogues. The equilibrium constant value (K 3.3 0.3) indicates a significant equilibrium isotope effect. Simulation
of the rate data provides access to the individual isomerization rate constants kHH, kHD, and kDD for the three
isotopomers, yielding kinetic isotope effects: kHH/kHD 1.24 0.01 and kHD/kDD 1.58 0.01 (and, consequently,
HH/kDD 1.96
0.02). The analysis of the DFT-calculated frequencies, using the [Cp*Fe(dhpe)H2]+ model system,
for the [Cp*Fe(dhpe)(
2-XY)]+ isotopomers as well as transition states for the “direct” (TSdir) and “via Cp” (TSrot
pathways (X H, D) allowed the computation of the expected isotope effects. A comparison with the experiment
+
CF3COOD and Cp*Fe(dppe)D
+
CF3COOD systems at 273 K. The
η
η
2-H2 and 2-D2
η
)
±
)
±
)
±
k
)
±
η
)
)
strongly suggests that the mechanism occurs via the “direct” pathway for the present system, although the small
difference in the calculated energy barriers suggests that the “via Cp” pathway may be preferred in other cases.
Introduction
initial stages of their interaction with dihydrogen.1 Subse-
quently, depending on the properties of the metal and ligands,
an oxidative addition process may take place, leading to a
dihydride, or the dihydrogen complex may remain as the
stable final product. An equilibrium between the two isomeric
forms may also be observed. Dihydrogen complexes are also
sometimes obtained by protonating hydride precursors under
kinetically controlled conditions, preceding in some cases
the rearrangement to the thermodynamic classical dihydride
product.2-7 For the group 8 half-sandwich systems such as
It is nowadays well established that most 16-electron LnM
fragments form a dihydrogen complex LnM(η2-H2) in the
* To whom correspondence should be addressed. E-mail: poli@
lcc-toulouse.fr (R.P.), agusti@klingon.uab.es (A.L.), shu@ineos.ac.ru (E.S.S.).
Fax: +33-561553003 (R.P.).
† UPR CNRS 8241 lie´e par convention a` l’Universite´ Paul Sabatier et a`
l’Institut National Polytechnique de Toulouse.
‡ Universitat Auto`noma de Barcelona.
§ Current address: Departamento de Qu´ımica Inorga´nica, Instituto de
Ciencia de Materiales de Arago´n, Universidad de Zaragoza-CSIC, 50009
Zaragoza, Spain.
| Current address: Institute of Chemical Research of Catalonia (ICIQ),
43007 Tarragona, Catalonia, Spain.
(1) Kubas, G. J. Metal Dihydrogen and σ-Bond Complexes; Kluwer
Russian Academy of Sciences.
Academic/Plenum Press: New York, 2001.
10248 Inorganic Chemistry, Vol. 45, No. 25, 2006
10.1021/ic061428n CCC: $33.50
© 2006 American Chemical Society
Published on Web 11/15/2006

Analysis of Kinetic Isotope Effects
Scheme 1
affording an intermediate cyclopentadiene complex, followed
by ring rotation and transfer back to the metal center, has
also been contemplated but discarded on the basis of indirect
bond energy arguments.3 A theoretical study of this isomer-
ization process has not been tackled yet, and no reports on
the transition states or a comparison of energy barriers for
different mechanisms can be found.
We have recently reported experimental investigations,
including a stopped-flow kinetic study, of the hydrogen
bonding and proton transfer to Cp*Fe(dppe)H with a variety
of proton donors (HA) of different acid strength.10,11 This
study has highlighted the presence of a hydrogen-bonded
intermediate involving the hydride site (and not the metal
site), Cp*Fe(dppe)H‚‚‚HA, the reversibility of the proton
transfer step (see Scheme 2), and the need of a second proton
donor molecule to trigger the proton transfer process. An
accompanying theoretical investigation has mostly focused
on the first step, i.e., proton transfer to the hydride ligand.11
Concerning the subsequent isomerization process, an intra-
molecular mechanism is suggested by an identical rate when
using proton donors of different strength at the same
temperature. The alternative reversible deprotonation, fol-
lowed by a slower protonation at the metal site, is incon-
sistent with this experimental observation.10 Our attempts
to carry out an Eyring analysis of this isomerization step
were complicated by the reversibility of the preceding proton-
transfer step, which resulted in strong coupling between two
key rate constants in the mathematical fitting procedure, with
the first rate constant leading from the dihydrogen intermedi-
ate to the final dihydride product (k2, irreversible step) and
the second one leading from the same intermediate back to
the starting monohydride complex (k-1; see the Supporting
Information of ref 11).
[CpRuLL′H2]+ (Cp ) cyclopentadienyl; L ) tertiary phos-
phine)3,4 and iron analogues,6,7 this interconversion implies
an important stereochemical change. In these systems, two
different mutual arrangements of two ligands are possible:
cisoid and transoid. Because there are no reports of the
coexistence of dihydrogen and cis-dihydride isomers, the
interconversion must entail passing from the three-legged-
piano-stool dihydrogen complex to a transoid square-based
four-legged-piano-stool dihydride (Scheme 1).8 Thus, the
cleavage of the H-H bond is accompanied by a cis-trans
rearrangement, which involves the migration of one H atom
from one side of the molecule to the opposite one. The
isomerization of the dihydrogen complex [Cp*Fe(dppe)(η2-
H2)]+ is particularly interesting because the chelate nature
of the diphosphine ligand may impose an extra obstacle to
the H migration to yield the observed trans structure for the
dihydride product.
The isomerization kinetics have been studied by NMR
spectroscopy for the cationic ruthenium dihydrogen com-
plexes.3,9 Reported activation enthalpies range from 16 to
21 kcal mol-1, with small and negative activation entropies.
The highest values have been obtained for bidentate phos-
phines with sterically demanding ligands.9 The small ∆Sq
supports an intramolecular mechanism in which the dihy-
drogen ligand has lost rotational freedom in the transition
state.3 A dissociative mechanism has been proposed when
bulky and good electron-donor ligands are present9 but was
discarded for a simpler system on the basis of kinetic
evidence (no retardation effect was observed in the presence
of excess ligand).3 The possibility of an intermolecular
deprotonation followed by reprotonation of the neutral
hydride intermediate was ruled out for the ruthenium system
on the basis of the absence of an isotope scrambling for a
selectively generated (HD) system.3,9 Finally, an alternative
intramolecular mechanism proceeding by transfer of an H
atom from the dihydrogen ligand to the Cp ligand, thereby
In this report, we present a new experimental study that
has led to the determination of the accurate rate constant
and activation parameters for the isomerization process of
the [Cp*Fe(dppe)H2]+ system when the counterion is either
-
BF4 or CF3COO-. These investigations include the deter-
mination of the kinetic isotope effects (KIEs) kHH/kHD and
k
HD/kDD, obtained from the comparative isomerization rates
of complexes [Cp*Fe(dppe)(H2)]+ (kHH), [Cp*Fe(dppe)(HD)]+
(kHD), and [Cp*Fe(dppe)(D2)]+ (kDD). To the best of our
knowledge, the experimental determination of the isotope
effects for the transformation of a nonclassical [(H2)/(HD)/
(D2)] to a classical [(H)2/(H)(D)/(D)2] dihydride complex is
unprecedented. This is accompanied by a theoretical inves-
tigation whose purpose is 2-fold. First, it seeks to determine
the subtle factors tuning the relative stability between the
nonclassical dihydrogen and the classical dihydride isomers
of the [(C5R5)Fe(R′2PCH2CH2PR′2)H2]+ complexes (R ) H,
(2) Parkin, G.; Bercaw, J. E. J. Chem. Soc., Chem. Commun. 1989, 255-
257.
(3) Chinn, M. S.; Heinekey, D. M. J. Am. Chem. Soc. 1990, 112, 5166-
5175.
(4) Jia, G.; Lough, A. J.; Morris, R. H. Organometallics 1992, 11, 161-
171.
(5) Papish, E. T.; Rix, F. C.; Spetseris, N.; Norton, J. R.; Williams, R. D.
J. Am. Chem. Soc. 2000, 122, 12235-12242.
(6) Hamon, J. R.; Hamon, P.; Toupet, L.; Costuas, K.; Saillard, J. Y. C.
R. Chim. 2002, 5, 89-98.
(7) Roger, C.; Hamon, P.; Toupet, L.; Rabaaˆ, H.; Saillard, J.-Y.;
Hamon, J.-R.; Lapinte, C. Organometallics 1991, 10, 1045-
1054.
i
Me; R′ ) H, Ph, Pr), separating electronic and steric
contributions. Second, it explores the intimate details of the
possible intramolecular isomerization mechanisms. In par-
ticular, discrimination between two pathways showing very
(8) Ontko, A. C.; Houlis, R. J. F.; Schnabel, C.; Roddick, D. M.; Fong,
T. P.; Lough, A. J.; Morris, R. H. Organometallics 1998, 17, 5467-
5476.
(9) De Los R´ıos, I.; Jime´nez-Tenorio, M.; Padilla, J.; Puerta, M. C.;
Valerga, P. Organometallics 1996, 15, 4565-4574.
(10) Belkova, N. V.; Revin, P. O.; Epstein, L. M.; Vorontsov, E. V.;
Bakhmutov, V. I.; Shubina, E. S.; Collange, E.; Poli, R. J. Am. Chem
Soc. 2003, 125, 11106-11115.
(11) Belkova, N. V.; et al. Chem.sEur. J. 2005, 11, 873-888.
Inorganic Chemistry, Vol. 45, No. 25, 2006 10249

Baya et al.
Scheme 2
similar computed activation barriers (both close to the
experimental value) will be possible through the analysis of
the KIEs. The computational work involves pure quantum
mechanics (QM) and mixed QM/molecular mechanics (MM)
methods.
Computational Details
QM calculations were performed with the Gaussian 98
package12 at the density functional theory (DFT) B3PW91
level.13-15 Furthermore, coupled-cluster16,17 single-point cal-
culations on the DFT-optimized structures, using both single
and double substitutions18-21 and including triple excitations
noniteratively,22 were performed to obtain more reliable
energies. Core electrons of the Fe and P atoms were
described using the effective-core pseudopotentials of Hay-
Wadt,23,24 and valence electrons were described with the
standard LANL2DZ basis set associated with the ECP.12 In
the case of the P atoms, a set of d-type functions was added.25
C and H atoms nonbonded to the metal were described with
a 6-31G basis set.26 The two H atoms bonded to the Fe atom
were described with a 6-31G(d,p) basis set of functions.27
Experimental Part
All manipulations were carried out under an argon atmosphere
by standard Schlenk techniques. Complexes Cp*Fe(dppe)H and
Cp*Fe(dppe)D were synthesized according to the literature.7 NMR
characterization of the deuteride complex in C6D6: 2H NMR δ
-16.9 (t, JPD ) 10.4 Hz); 31P{1H} NMR δ 111.5 (t, JPD ) 10.3
1
Hz). In addition, the H NMR spectrum shows contamination by
the hydride complex Cp*Fe(dppe)H (25%, by integration of the
hydride resonance against the Cp* resonance). The acids HBF4‚OEt2,
CF3COOH, and CF3COOD were purchased from Aldrich and used
as received.
Mixed QM/MM calculations were performed with the
IMOMM method28 with a program built from modified
versions of two standard programs: Gaussian 9812 for the
QM part and mm3(92)29 for the MM calculations. The
IMOMM approach will be used to point out the effect of
The NMR studies were carried out in standard 5-mm NMR tubes
1
containing solutions of the complexes in CD2Cl2. The H and 31P
NMR data were collected with a Bruker AMX 400 spectrometer
operating at 400.13 and 161.98 MHz, respectively. The low-
temperature kinetic experiments with 31P{1H,2H} NMR moni-
toring were carried out on a Bruker AV500 spectrometer equipped
with a 5-mm triple-resonance inverse probe with dedicated 31P
(12) Frisch, M. J.; et al. Gaussian 98; Gaussian, Inc.: Pittsburgh, PA, 1998.
(13) Perdew, J. P. In Electronic Structure of Solids; Ziesche, P., Eschrig,
H., Eds.; Akademie Verlag: Berlin, 1991; Vol. p∧11.
(14) Perdew, J. P. Phys. ReV. B: Condens. Matter Mater. Phys. 1986, 33,
8822-8824.
1
channel operating at 500.33 MHz for H NMR, 202.54 MHz for
31P NMR , and 76.80 MHz for 2H NMR. For the purpose of
quantitative concentration measurements for the kinetic runs, the
31P NMR spectra were recorded with inverse-gated 1H and 2H
(15) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(16) Pople, J. A.; Krishnan, R.; Schlegel, H. B.; Binkley, J. S. Int. J.
Quantum Chem. 1978, 14, 545-560.
1
decoupling. For 31P and H NMR, a typical pulse width corre-
(17) Bartlett, R.; Purvis, G. Int. J. Quantum Chem. 1978, 14, 561-81.
(18) Cizek, J. AdV. Chem. Phys. 1969, 14, 35.
sponding to a 30° flip angle and 30-s relaxation delays was used
to obtain reliable integration data. In the case of overlapping
resonances, the quantitative evaluation was realized by peak
deconvolution (see the Supporting Information). All chemical shifts
for 1H and 2H NMR are relative to tetramethylsilane using a residual
peak of the solvent as a secondary standard. 31P NMR chemical
shifts were referenced to an external 85% H3PO4 sample. The
temperature was regulated with a Bruker TV-3000 unit. Temper-
ature calibration was determined using a methanol chemical shift
thermometer. The temperature accuracy and stability was (1 K.
All mixings between the acid and hydride complexes were
performed at -80 °C.
(19) Purvis, G. D.; Bartlett, R. J. J. Chem. Phys. 1982, 76, 1910-1918.
(20) Scuseria, G. E.; Janssen, C. L.; Schaefer, H. F. J. Chem. Phys. 1988,
89, 7382-7387.
(21) Scuseria, G. E.; Schaefer, H. F. J. Chem. Phys. 1989, 90, 3700-3703.
(22) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. J. Chem. Phys. 1987,
87, 5968-5975.
(23) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284-298.
(24) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.
(25) Ho¨llwarth, A.; et al. Chem. Phys. Lett. 1993, 208, 237-240.
(26) Hehre, W.; Ditchfie, R.; Pople, J. J. Chem. Phys. 1972, 56, 2257-
2261.
(27) Harihara, P.; Pople, J. Theor. Chim. Acta 1973, 28, 213-222.
(28) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170-1179.
(29) Allinger, N. L. mm3(92); Quantum Chemistry Program Exchange:
Bloomington, IN, 1992.
10250 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
Table 1. Results of the Isomerization Kinetics of [Cp*Fe(dppe)(η2-X2)]+ to [Cp*Fe(dppe)(X)2]+ (X ) H, D)'
a
a
a
run
complex
acid
HBF4
HBF4
HBF4
HBF4
HBF4
CF3COOH
CF3COOD
CF3COOD
[HA/Fe]
T (°C)
104kHH (s-1
)
104kHD (s-1
)
104kDD (s-1
)
1
2
3
4
5
6
7
8
Cp*Fe(dppe)H
Cp*Fe(dppe)H
Cp*Fe(dppe)H
Cp*Fe(dppe)H
Cp*Fe(dppe)H
Cp*Fe(dppe)H
Cp*Fe(dppe)Da
Cp*Fe(dppe)H
1
1
1
1
3
1
1
1
-10
0
5
15
0
0.70 ( 0.09
3.21 ( 0.07
5.92 ( 0.15
28.3 ( 0.7
3.53 ( 0.22
3.10 ( 0.12
0
0
0
1.73 ( 0.12
1.56 ( 0.01
3.07 ( 0.02
2.48 ( 0.01
a The standard deviations reported for runs 1-6 were calculated keeping in mind an estimated 10% error in the NMR integration. For run 7, see the
Supporting Information. b The starting material was contaminated by ca. 25% of CpFe(dppe)H.
the substituents located both on the Cp group and on the
phosphine ligand. The substituents are methyl, phenyl, and
isopropyl groups treated in the MM part through the MM3
force field. The same basis set as that in the pure quantum
computations was used at the same level of theory for the
QM part. In the IMOMM computations, all of the geo-
metrical parameters were optimized except for the distances
between the atoms linking the QM and MM parts. The KIEs
were calculated as described in the Results and Discussion
section using harmonic vibrational frequencies obtained from
frequency calculations. The frequencies of each isotopically
substituted system were obtained at the fixed geometry
previously optimized for the corresponding unsubstituted
isotopomer. These calculations were carried out with Gauss-
ian 03.30
using CF3COOH or HBF4 in greater than a 2-fold excess.
Thus, the system is not amenable to a clean kinetic study in
dichloromethane. A change of solvent has been envisaged,
but no ideal system has so far been found: tetrahydrofuran
reduces considerably the thermodynamic drive to proton
transfer (because of the free acid stabilization by hydrogen
bonding with the solvent molecules) and polymerizes under
strongly acidic conditions; toluene cannot keep the produced
salts in solution at sufficient concentrations.
A convenient solution to this problem was found by
recurring to the NMR technique. The advantage in this case
is that the relative amounts of dihydrogen and dihydride
complexes can be measured directly and independently, even
in the presence of any amount of the [Cp*Fe(dppe)Cl]+
decomposition product (a paramagnetic complex). Therefore,
the rate at which the intermediate dihydrogen complex decays
is directly accessible with accuracy. In addition, working at
low temperatures (thereby slowing down the isomerization
process to a suitable time scale for NMR monitoring) reduces
the impact of the acid-catalyzed decomposition in dichloro-
methane. The sum of the integrated intensities for the
dihydrogen and dihydride complexes, monitored against that
of an internal standard, provides a direct gauge of the system
stability.
Solvent effects were taken into account by means of
polarized continuum model calculations31,32 using standard
options.30 Individual solvation cavities were added on the H
atoms directly bonded to the Fe atom. The free energies of
solvation were computed in dichloromethane (ꢀ ) 8.93) at
the geometries optimized in the gas phase.
Results and Discussion
1. Isomerization Kinetics of [Cp*Fe(dppe)(η2-X2)]+ (X
) H, D). As outlined in the Introduction, the accurate
determination of the isomerization rate constant for the
process shown in Scheme 2 at different temperatures was
thwarted by the reversibility of the proton-transfer step when
a mildly acidic proton donor (i.e., (CF3)nCH3-nOH with n
) 1-3) was used. In an attempt to circumvent this problem,
we have carried out stopped-flow studies using stronger acids
(CF3COOH and HBF4), to force the proton-transfer step to
quantitative conversion and then measure the subsequent
isomerization under clean pseudo-first-order conditions.
Unfortunately, these experimental conditions introduce an-
other complication. The final dihydride product is unstable
in dichloromethane in the presence of an excess of strong
acids, leading to the oxidized chloride complex [Cp*Fe-
(dppe)Cl]+, as described previously.10 This phenomenon is
not observed when using the (CF3)nCH3-nOH proton donors,
even in large excess amounts, whereas it is observed when
After generation of the intermediate dihydrogen complex
quantitatively in situ by low-temperature (-80 °C) addition
of the appropriate amount of a strong acid (see Table 1 for
details), monitoring was carried out at four different tem-
peratures (-10, 0, +5, and +15 °C, runs 1-4, respectively).
Both the decay of the dihydrogen complex resonance
(examples are shown in Figure 1, top) and the growth of the
1
final dihydride complex resonance were monitored by H
NMR spectroscopy. The sum of the integral, after normaliza-
tion relative to the solvent peak that was used as an internal
standard, was relatively constant throughout the reaction,
showing limited decomposition (<10% when using 1 equiv
of a strong acid; a maximum 17% decomposition during 153
min was found for run 5, where a greater excess of acid
was used). However, because the acid-catalyzed decomposi-
tion occurs only for the classical product and because the
isomerization is irreversible,10 the decay rate of the non-
classical reagent is an accurate measure of the isomerization
rate.
(30) Frisch, M. J.; et al. Gaussian 03, revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(31) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027-2094.
(32) Amovilli, C.; Barone, V.; Cammi, R.; Cance`s, E.; Cossi, M.; Menucci,
B.; Pomelli, C. S.; Tomasi, J. AdV. Quantum Chem. 1998, 32, 227-
261.
The nonclassical resonance decay gives an excellent fit
to the first-order rate law at each temperature (examples are
shown in Figure 1, above), yielding the rate constants kHH
Inorganic Chemistry, Vol. 45, No. 25, 2006 10251

Baya et al.
Figure 2. Eyring analysis of the isomerization rate constant kHH
.
2. Role of the Electronic and Steric Effects in the
Relative Stabilities of the Dihydrogen and Dihydride
Complexes. Before analysis of the possible rearrangement
pathways in detail from the computational point of view, it
is important to gain a full appreciation of the model effect
on the relative energies. As a general trend in computational
chemistry, models are used to represent the real molecular
systems. This is mainly done in order to deal with affordable
computation time and system resources, but this procedure
also allows one to introduce the complexity of the real
systems in a gradual way, analyzing at each step the changes
suffered by the system. This approach will be used to separate
electronic and steric contributions to the relative stabilities
of the dihydrogen and dihydride forms. Usually, the system
is simplified by replacing substituents with H atoms. If we
apply this general procedure to the [(C5R5)Fe(R′2PCH2CH2-
PR′2)H2]+ complexes, we can define four models, as depicted
in Chart 1, together with the real complexes. The smallest
model is the [CpFe(dhpe)H2]+ complex 1 (dhpe ) PH2CH2-
CH2PH2), where both the methyl groups of the Cp* ligand
and the phenyl or isopropyl substituents of the phosphine
ligand are replaced by H atoms. In the second model [Cp*Fe-
(dhpe)H2]+ (2), only the phosphine substituents are replaced
by H atoms, keeping the whole Cp*. Two additional models
can be generated by introducing the real phosphine substit-
uents while keeping the model Cp ligand: [CpFe(dppe)H2]+
(3) for R′ ) Ph and [CpFe(dippe)H2]+ (5) for R′ ) iPr. The
four models can be related to the two experimentally reported
complexes [Cp*Fe(dppe)H2]+ (4)6,33 and [Cp*Fe(dippe)H2]+
(6).34 The dihydride form of all of the species considered in
the calculations and the numbering scheme are presented in
Chart 1. The cisoid (dihydrogen) and transoid (dihydride)
isomers will be labeled c and t, respectively.
Figure 1. 1H NMR monitoring of the [Cp*Fe(dppe)(η2-H2)]+ resonance
-
decay (BF4 salt in CD2Cl2).
reported in Table 1. The three different kinetic runs carried
out at 0 °C (runs 2, 5, and 6) show that the isomerization
rate does not significantly depend on the acid concentration
(compare runs 2 and 5) or on its nature (compare runs 2 and
6). The latter observation confirms our previous findings of
identical isomerization rates using different (CF3)nCH3-nOH
proton donors.10 A comparison of runs 6 and 7, on the other
hand, shows a significant KIE. Monitoring of run 7 could
not be carried out conveniently by 1H NMR because the Cp*
and dppe peaks of the reactant and product are not
significantly spread apart. The quantitative determination was
based on the integration of the inverse-gated doubly de-
coupled 31P{1H,2H} NMR resonance (see the Experimental
Part). It should be remarked that the ca. 25% of the
Cp*FeH(dppe) isotopomer in the deuteride starting material
may slightly affect the resulting rate constant value, given
that the 1H-containing nonclassical impurity, [Cp*Fe(η2-HD)-
(dppe)]+, isomerizes faster (vide infra). Therefore, the value
of kDD reported in Table 1 should be considered as an upper
estimate of the isomerization rate constant for complex
[Cp*Fe(dppe)(η2-D2)]+ (a more accurate value for kDD is
available from an alternative approach, vide infra). The
Eyring analysis of the rate constants kHH of runs 1-4 (Figure
2) yields the activation parameters ∆Hq ) 21.6 ( 0.8 kcal
mol-1 and ∆Sq ) 5 ( 3 eu. These values are similar to those
previously reported for related ruthenium derivatives.3,9 The
enthalpy value resulting from this analysis is useful for a
comparison with the computed energies for the transition
states of different potential rearrangement pathways (vide
infra).
The optimized structures of the isomers for the model
complex 1 are depicted in Figure 3. The structural features
of the dihydrogen and dihydride isomers for the systems 1-6
are analyzed in detail in the Supporting Information. The
relative energies of the dihydrides 1t-6t with respect to the
corresponding dihydrogen isomers 1c-6c are given in Table
2. Our computed values for 1 and 4 (+6.6 and -3.9 kcal
(33) Hamon, P.; Toupet, L.; Hamon, J.-R.; Lapinte, C. Organometallics
1992, 11, 1429-1431.
(34) Jime´nez-Tenorio, M.; Puerta, M. C.; Valerga, P. Organometallics 1994,
13, 3330-3337.
10252 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
Chart 1
Figure 3. Optimized geometries of the dihydrogen (left) and dihydride
(right) isomers of 1.
Table 2. Relative QM Energies (in kcal mol-1; QM/MM Values in
Parentheses) of the Dihydride Complexes, with Respect to the
Corresponding Dihydrogen Isomer
complex
∆E
[CpFe(dhpe)H2]+, 1
[Cp*Fe(dhpe)H2]+, 2
[CpFe(dppe)H2]+, 3
[Cp*Fe(dppe)H2]+, 4
[CpFe(dippe)H2]+, 5
[Cp*Fe(dippe)H2]+, 6
6.6
4.9 (5.9)a
1.4 (5.7)b
-3.9 (0.1)b
-3.6 (0.9)c
-8.6 (-6.2)c
a The methyl groups of the Cp* ligand are in the MM part. b The phenyl
substituents of the dppe ligand are in the MM part. c The isopropyl
substituents of the dippe ligand are in the MM part.
of the protonation of the Cp*Fe(dppe)H complex,6,10,11 we
expect that it must be more stable than the related non-
classical dihydrogen complex. However, the nonclassical
dihydrogen complex is the most stable isomer for all of the
models except when the computation is carried out on the
real complex [Cp*Fe(dppe)(H)2]+. It is clear that in our case
model systems cannot account for the experimental trends
on the relative stabilities of the two isomers.
The substituents tune the relative stabilities of the dihy-
drogen and dihydride forms, and thus the ability of the metal
fragment to break the H-H bond, by a combination of
electronic and steric effects. We will apply a simple analysis
already used on other systems to separate and quantify both
contributions on the relative energies of the two isomers.36-38
To this aim, we have carried out additional QM/MM
computations. The approach relies on two basic assump-
tions: (1) the full QM calculation on the whole system
describes correctly the experimental behavior and thus
includes both electronic and steric contributions; (2) the QM/
MM calculation describes only the steric effects of the groups
included in the MM part, leaving out their electronic effects.
From the small model 1, the substituents have been added
in order to build up the real complexes, but in these QM/
MM calculations, they have been introduced at the MM level.
By this scheme, only the steric contributions of the substitu-
ent are responsible for the changes in the relative stabilities
of the two isomers.
mol-1, respectively) compare well with those previously
reported using a different functional (+4.3 and -4.4 kcal
mol-1, respectively).6 To check the accuracy of our meth-
odology, we have recalculated the energy of the dihydrogen/
dihydride couple of the smallest system, 1, with the highly
correlated CCSD(T) method. There is a reasonable agreement
between both methods because the dihydride lies 9.4 kcal
mol-1 above the dihydrogen complex at the CCSD(T) level.
Moreover, the experimental trend on the stabilities is well
reproduced by our calculations. The dihydrogen complex was
found to be more stable than the dihydride only for the Cp/
dppe set of ligands.6,10,33-35 We have also checked whether
solvation affects the dihydrogen/dihydride relative stabilities
by calculating the energy difference between the dihydrogen
and dihydride forms of 2 in dichloromethane. The gas-phase
energy difference (4.9 kcal mol-1 in favor of the dihydrogen
complex) is only slightly decreased to 4.5 kcal mol-1 in
dichloromethane. Therefore, solvation plays a very minor
role in this equilibrium. On the contrary, the equilibrium is
strongly influenced by the Cp and phosphine substituents.
The introduction of the methyl substituents in the Cp ring
favors the dihydride, reversing the stabilities. The preference
for the dihydride isomer is considerably enhanced by the
presence of the isopropyl substituents in the phosphine.
Because the classical dihydride is the thermodynamic product
The relative energy between the two isomers computed
with model 1 is the reference. No substituents are present,
(36) Barea, G.; Maseras, F.; Jean, Y.; Lledo´s, A. Inorg. Chem. 1996, 35,
6401-6405.
(37) Bustelo, E.; Carbo, J.; Lledo´s, A.; Mereiter, K.; Puerta, M.; Valerga,
P. J. Am. Chem. Soc. 2003, 125, 3311-3321.
(38) Maresca, O.; Maseras, F.; Lledo´s, A. New J. Chem. 2004, 28, 625-
630.
(35) Jia, G.; Ng, W. S.; Yao, J.; Lau, C.-P.; Chen, Y. Organometallics
1996, 15, 5039-5045.
Inorganic Chemistry, Vol. 45, No. 25, 2006 10253

Baya et al.
the extra stabilization of the dihydride isomer. The branched
isopropyl groups have a more important steric hindrance than
the planar phenyl groups, rendering the dihydride the favored
isomer, even with the Cp ligand (1 vs 5).
At a first glance, it could seem surprising that the
coordination number increase associated with the dihydrogen
f dihydride interconverson could be favored on steric
grounds. However, the destabilizing steric interactions are
not developed between the piano-stool legs but rather
between the phosphine substituents and the C5 ring. They
are signaled by the notable increase in the [C5 ring-
(centroid)]-Fe-[P2(centroid)] (R) angle on going from the
dihydrogen (R of about 150°) to the dihydride structures (R
of about 180°); see the Supporting Information. In a thorough
study of the η2-dihydrogen complexes of ruthenium, it was
already pointed out that the interaction of phosphine sub-
stituents with the Cp or Cp* ring may be significant in
determining the position of the dihydrogen f dihydride
equilibrium.3 Our study fully confirms this hypothesis and
quantifies the magnitude of such an interaction.
The electronic effects also stabilize the dihydride isomer.
The largest effect is found with the phenyl groups. The role
of the phenyl substituents is particularly important in the
preferential stabilization of the isomer with the highest
oxidation state. The omission of the phenyl substituents in
the model complexes is the main reason for the poor results
they give for the dihydrogen/dihydride relative energies.
3. Computational Study of the Dihydrogen f Dihy-
dride Isomerization. We now turn to the analysis of the
isomerization mechanism. As mentioned in the Introduction,
several mechanisms have been suggested for the isomeriza-
tion of the related dihydrogen ruthenium derivatives:3,9,39,40
(1) a direct intramolecular rearrangement through a trigonal-
bipyramidal transition state; (2) a dissociative process with
the partial or total phosphine ligand decoordination; (3) a
relay of the migrating hydride via the Cp ligand; (4) a
deprotonation of the cis isomer followed by protonation of
the metal at the trans position. The first three possibilities
are intramolecular processes, while the latest involves a
proton-donor molecule. Because we have accumulated
evidence, including the present study, that the isomerization
rate of 4 is independent of the nature of the proton donor,6,10,11
only intramolecular mechanisms have been explored. In
addition, the new experimental results shown in this paper
set an experimental value to calibrate the highest enthalpy
point along the computed reaction coordinate.
Figure 4. Substituent steric and electronic effects on the stabilization of
the classical dihydride isomer for the complexes drawn in Chart 1.
neither on the C5 ring nor on the phosphine ligand, and the
dihydrogen is 6.6 kcal mol-1 more stable than the dihydride.
This value is reduced to 4.9 kcal mol-1 when performing a
full QM calculation of the [Cp*Fe(dhpe)H2] system 2,
whereas a QM/MM calculation of the same system with the
methyl substituents of the Cp* considered only at the MM
level gives the dihydrogen 5.9 kcal mol-1 more stable than
the dihydride. The total contribution of the methyls is thus
1.7 kcal mol-1, of which 0.7 kcal mol-1 is attributed to the
steric effect. Afterward, the electronic effects of the methyl
groups (1.0 kcal mol-1) can be estimated by subtracting
(1.7-0.7 kcal mol-1) the steric effect from the total contribu-
tion. Of course, the electronic effect could also be evaluated
directly as the difference (5.9-4.9 kcal mol-1) between the
relative QM and QM/MM energies reported in Table 2, with
an identical result. The same scheme can be applied to the
phenyl and isopropyl substituents. The results are sum-
marized in Figure 4.
From 1 to 2, both the electronic and steric effects are quite
small and are of the same order of magnitude. However, the
effect of the phenyl groups (from 1 to 3) is more pronounced.
The dihydride isomer is stabilized about 6.0 kcal mol-1, but
this energy is not enough to reverse the thermodynamic
preference for the dihydrogen. The extra stabilization is
mainly due to electronic effects, with the steric effects not
being relevant. When we compare the relative energies of
the real systems containing the bulky phosphines and the
Cp* ligand, a stronger influence of the Cp* methyl groups
can be appreciated: 5.3 and 5.0 kcal mol-1 for the dppe and
dippe phosphines, respectively. In both cases, the substitution
of Cp by Cp* favors the dihydride complex. Assuming that
the electronic effect of the methyls will be similar in the
three cases, stabilizing the dihydride by about 1.0 kcal mol-1,
it is evident that steric factors of the ring substituents work
in favor of the dihydride in the real complexes. Phenyl and
isopropyl substituents have a major influence in the ther-
modynamic preference for the dihydride isomer. Changing
the H model substituents with phenyls in the Cp*-containing
complexes (2 vs 4) favors the dihydride by 8.8 kcal mol-1,
with a similar contribution of electronic and steric effects.
The influence of the isopropyls (2 vs 6) in favoring the
dihydride is even higher (13.5 kcal mol-1), although in this
case, the weight of the steric effects is the main cause for
To keep the computational demands affordable, the study
of the reaction mechanisms was performed at a full QM level
only for the 1 and 2 model complexes and at the QM/MM
level with the IMOMM method for the real 4 complex (with
phenyl groups at the MM level). Additional single-point
energy computations at the full QM level on the real 4
complex were performed on the geometry of the intermedi-
ates and transition states located along the QM/MM potential
energy surface (QM/IMOMM).
(39) Cayuela, E.; Jalon, F. A.; Manzano, B. R.; Espino, G.; Weissensteiner,
W.; Mereiter, K. J. Am. Chem. Soc. 2004, 126, 7049-7062.
(40) Ryan, O. B.; Tilset, M. J. Am. Chem. Soc. 1991, 113, 9554-9561.
10254 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
Figure 5. Direct (right) and “via Cp” (left) pathways for the cis-dihydrogen f trans-dihydride isomerization. The optimized geometries are shown for the
2 system (the Cp* and dhpe H atoms have been omitted for the sake of clarity), whereas the energies (in kcal mol-1) correspond to the QM calculations on
the QM/MM optimized 4 system.
Table 3. Main Geometrical Parameters (Distances in Angstroms and
Direct Intramolecular Rearrangement. As discussed
Angles in Degrees) of the Transition States for the Two Reported
Dihydrogen f Dihydride Interconversion Mechanisms of 2a
above, in these piano-stool compounds, the cis-dihydrogen
f trans-dihydride isomerization implies two processes:
oxidative addition of dihydrogen and structural rearrangement
of the piano-stool legs. These two processes may proceed
in one or two steps. The simple cleavage of the H-H bond
in the dihydrogen structure leads to a cis-dihydride. Although
there are no reports of cis-dihydride [(C5R5)FeH2(PP)]+
complexes, either isolated or spectroscopically detected, we
considered also the possibility of a transient cis-dihydride
intermediate. Starting from the dihydrogen complex and
taking as a reaction coordinate the H-H distance, we
calculated the energy curve for the H-H breaking. The
energy was continuously increasing, and no stationary point
corresponding to a cis-dihydride intermediate could be
located. Thus, the reaction appears to be taking place in a
single step.
2c
TSdir
TSrot
2t
Fe-H
1.576
1.580
0.869
2.220
2.208
2.121
2.159
144.8
85.2
32.0
1.482
1.490
2.282
2.149
2.203
2.115
2.202
1.483
1.644
2.274
2.165
2.183
2.104
2.238
1.482
1.482
2.698
2.166
2.167
2.105
2.135
Hf-Hm
Fe-P
Fe-C(Cp)range
Rb
158.9
162.3
85.1
179.3
P1-Fe-P2
Hf-Fe-Hm
Cpc*-Fe-Hf-Hm
84.0
100.3
100.7
151.8
153.6
119.0
91.5
131.0
179.7
178.8
134.0
134.0
c
36.9
c
Cpc*-Fe-P1-P2
115.8
126.0
128.0
c
Cpc*-Fe-P1
c
Cpc*-Fe-P2
a Values for the initial dihydrogen and the final dihydride complexes
are also included. b R ) angle [C5 ring(centroid)]-Fe-[P2(centroid)]. c Cpc*
) Cp* centroid.
Then, we explored the potential energy surface for the
intramolecular rearrangement taking the H-Fe-H angle as
the reaction coordinate. While the two hydride positions are
symmetry-equivalent in the starting dihydrogen complex,
they become inequivalent in the transition state. They will
be identified as the “fixed” (Hf) and “migrating” (Hm)
positions, respectively. The Hf-Fe-Hm angle must undergo
a large change during the process: from 32° in the dihy-
drogen complex to 131° in the dihydride complex. Starting
from the maximum of this monodimensional energy profile,
we were able to locate the transition state for the direct
intramolecular rearrangement (TSdir). This presents a single
imaginary frequency (298i cm-1), with the opening of the
Hf-Fe-Hm angle as the main component of its associated
eigenvector. The mechanism of the process is illustrated in
Figure 5 (right-hand side), together with the structure of the
TSdir for the 2 system. The main geometric parameters of
the TSdir for the Cp*-dhpe system are given in Table 3,
while the full list of calculated frequencies as given as the
Supporting Information. Similar structural values, not re-
ported for the sake of brevity, are obtained for the Cp-dhpe
and Cp*-dppe systems.
The transition state has a dihydride nature. The H-H bond
appears totally broken, as reflected by the long H-H
separation, whereas the two Fe-H bonds are already formed.
Consequently, the oxidative addition process is already
completed at the transition state level; the structure can be
Inorganic Chemistry, Vol. 45, No. 25, 2006 10255

Baya et al.
Table 4. Relative Energies (∆E, kcal mol-1) of the Species Involved in the Two Mechanisms
dihydrogen
TSCpcis
ICpcis
TSRot
TSdir
ICptrans
TSCptrans
12.5
dihydride
[CpFe(dhpe)H2]+
[Cp*Fe(dhpe)H2]+
[Cp*Fe(dppe)H2]+a
[Cp*Fe(dppe)H2]+b
0.0
0.0
0.0
0.0
30.1
28.9
24.5
21.3
6.6
4.9
0.1
15.3
14.9
12.1
11.2
30.0
23.9
22.2
11.5
5.9
3.4
-3.9
a QM/MM calculation with the phenyl groups in the MM part. b Full QM calculation at the QM/MM-optimized geometries.
The transition state TSdir lies 30.1 and 28.9 kcal mol-1
above the dihydrogen for 1 and 2, respectively (Table 4). In
these model systems, the effect of the Cp methyl substituents
is to decrease slightly the potential energy barrier. We have
also computed the enthalpy and free energy of activation,
namely, ∆Hq ) 27.1 kcal mol-1 and ∆Gq ) 30.7 kcal mol-1
for 2 at T ) 298.15 K. The small, almost zero, activation
entropy (∆Sq ) -3.9 eu) is not inconsistent with the small
value observed in this work for the iron complex (vide supra)
and with the experimentally reported ∆Sq (-3.3 ( 0.7 eu)
in related ruthenium complexes.3 As was already found for
the dihydrogen-dihydride equilibrium, the energy barrier of
the rearrangement is only slightly affected by solvent effects.
The computed barrier in dichloromethane for the 2 system
is 28.6 kcal mol-1, very close to the gas-phase value (28.9
kcal mol-1).
As reported for the relative stabilities of the dihydrogen
and dihydride complexes, the effects of the substituents
(electronic and steric) are also major on the transition-state
energy. The stabilization found for the Cp* transition state
is similar to those found for the minima, in agreement with
the dihydride nature of the transition state. The same trend
could by inferred for the phenyl effects, and this is indeed
the case. The potential energy barrier for the real system 4,
with the phenyl groups described at the QM/MM level, is
computed as 24.5 kcal mol-1, and full QM energy computa-
tions at the fixed QM/MM minima and transition state further
lower the energy barrier to 21.3 kcal mol-1. Thus, the effect
of the phenyl substituents is to considerably decrease the
isomerization barrier. The calculated barrier is very close to
the experimental value of 21.6 ( 0.8 kcal mol-1 (see above)
and is also close to the barrier of the related isomerization
of the [Cp*Ru(dippe)(H2)]+ complex.9
Chart 2
described as an iron(IV) dihydride. Nevertheless, the R angle
is closer to its initial value in the dihydrogen complex than
to the 180° value in the final dihydride. The phosphine ligand
is also experiencing a significant repositioning. The dihedral
angle formed by the C5 ring(centroid), the Fe atom, and the
two P atoms takes a midway value (about 150°) between
the 115° in the dihydrogen and the 180° in the trans-
dihydride. The movement of the phosphine ligand can also
be appreciated from the two Cpc*-Fe-P angles. These are
close to each other in both the starting dihydrogen and final
dihydride complexes, and the values do not change much
on going from one to the other (126° and 128° for the
dihydrogen complex and 134° and 134° for the dihydride
complex). However, one angle is opened to 150°, whereas
the other one is closed to 120°, in the transition state. This
illustrates the swing of the entire diphosphine ligand toward
one side of the molecule, which is necessary to allow passage
of one hydride ligand from the side, to move from the front
to the back of the metal center.
Given that both the transition state and the final trans-
dihydride are four-legged-piano-stool complexes, they can
be analyzed in relation to the mechanisms invoked for the
CpMo(CO)2LR cis-trans interconversion.3,41 Within this
model, the two trigonal-bipyramidal-like transition states
presented in Chart 2 can be envisaged. The geometric
parameters of the transition states are in closer agreement
with a structure of type B, describing a movement in which
one of the phosphine arms has moved down to allow the
migration of the hydride ligand toward its final position. The
generation of a four-legged-piano-stool product with a
transoid disposition of the E and H ligands resulting from
the oxidative addition of a E-H bond by a (C5R5)ML2
fragment is a relatively common result, not well explained
yet. A transition state closely related to TSdir has been
reported for the cis-trans isomerization of a hydridoaryl
complex of rhenium.42 It could also be at work for the
oxidative addition of an alkyne C-H bond to yield a trans-
alkynyl hydride.37
Phosphine Dissociation Mechanism. We have also
considered a dissociative isomerization mechanism, in which
one of the P atoms in the [Cp*Fe(dhpe)(H2)]+ complex
dissociates from the metal, yielding an unsaturated interme-
diate where the diphosphine acts as a monodentate ligand.
The oxidative addition would subsequently take place in this
intermediate. However, the nonclassical dihydrogen complex
with one phosphine arm fully decoordinated is 30.0 kcal
mol-1 above the saturated dihydrogen complex. This high
Fe-P binding energy value led us to discard this mechanism
for the systems under study.
Isomerization through the Cp Ring. Some examples of
hydrogen migration involving the metal center and the Cp
ring can be found in the literature for iron compounds.
Examples are provided by [η4-C5H5(exo-D)]Fe(CO)3,43
(41) Faller, J. W.; Anderson, A. S. J. Am. Chem. Soc. 1970, 92, 5852-
5860.
(42) Clot, E.; Oelckers, B.; Klahn, A. H.; Eisenstein, O.; Perutz, R. N. J.
Chem. Soc., Dalton Trans. 2003, 4065-4074.
(43) Whitesides, T. H.; Shelly, J. J. Organomet. Chem. 1975, 92, 215-
226.
10256 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
Scheme 3
for the simpler Cp model complex, but this search was
unsuccessful. All of our attempts to obtain such species
reverted to the dyhdrogen complex. Evidently, the simpler
Cp ligand is not basic enough for this system to be protonated
by the dihydrogen. We must point out that, for those CpFe
complexes where the hydrogen migration from the metal
center to the Cp ring has been experimentally observed, this
process occurs under more drastic conditions. In terms of
the currently examined mechanism, replacing Cp by Cp*
should facilitate the migration, making it possible under
milder conditions.
The relative energies of the intermediates and transition
states for the “via Cp” mechanism for the 2 and 4 systems
are given in Table 4. In the model system, the ICpcis
intermediate lies 14.9 kcal mol-1 above the dihydrogen and
the transition state that connects both is 0.4 kcal mol-1 above
the intermediate. The trans-Cp*-protonated complex is 3.4
kcal mol-1 more stable than the cis-protonated derivative.
The transition state for the rotation of the protonated ring is
found to be 30.0 kcal mol-1 above the initial dihydrogen
complex in the gas phase and 30.8 kcal mol-1 in dichloro-
methane. The energy barrier for the Cp* rotation is 15.1 kcal
mol-1. The presence of the strong agostic bond, which is
preserved along the rotation, is the main reason for this high
value for a Cp* rotational process. Therefore, the rate-
determining step of the “via Cp” mechanism is the rotation
of the protonated Cp*. The computed activation enthalpy
and free energy are ∆Hq ) 29.2 kcal mol-1 and ∆Gq ) 30.7
kcal mol-1. This intramolecular mechanism also implies a
small negative activation entropy (∆Sq ) -5.0 eu). In the
gas phase, the energy barriers of the “direct” and “via Cp”
mechanisms for the dihydrogen f trans-dihydride intercon-
version in the model 2 system are very similar (about 30
kcal mol-1). Dichloromethane slightly favors the direct
mechanism, increasing the difference between both barriers
from 1.1 to 2.2 kcal mol-1. Next, we examine how the
presence of the phenyl substituents in the real system affects
this result.
For the 4 complex, we have not been able to find the
transition states TSCpcis and TSCptrans because of the flatness
of the potential energy surface around the intermediates. The
main effect of the phenyl groups is to stabilize all of the
species involved in the mechanism with respect the dihy-
drogen complex (see Table 4). Given that the phenyl groups
are described at the MM level, the explanation for this
behavior should be in the steric effects. On the basis of the
results presented above on the relative stabilities of the
dihydrogen and dihydride forms, the main reason for this
behavior could be attributed to a destabilization of the
dihydrogen complex by the steric interaction between phenyl
and methyl groups. Very close values of the energy barriers
for the “direct” and “via Cp” mechanisms are found also
for the real system at the QM/MM level: 24.5 kcal mol-1
(“direct”) and 23.9 kcal mol-1 (“via Cp”).
Fe(CO)3(η4-C6H7Ph),44 and (η5-C5H5)FeH(triphos) (triphos
) Ph2PCH2CH2PPh-CH2CH2PPh2).45 These isomerizations
take place under more drastic conditions relative to the
dihydrogen f dihydride interconversion examined here. In
addition, this mechanism was briefly contemplated for the
analogous ruthenium system by Chinn and Heinekey3 but
considered unlikely on the basis of bond strength consider-
ations. During the thorough exploration of the potential
energy surface for the dihydrogen f dihydride rearrangement
in the [Cp*Fe(dhpe)(H2)]+ system, we discovered the exist-
ence of low-lying intermediates with a protonated C5 ring.
This finding prompted us to fully explore the “via Cp”
mechanism.
This mechanism appears as a multistep pathway in which
the Hm atom migrates from the cis to the trans position via
a successive series of H jumps from and to the metal, together
with a rotation of the protonated C5 ring. The mechanism of
the process is illustrated in Figure 5 (left-hand side), along
with the structure of all intermediates for the 2 system. The
most relevant geometrical parameters of the TSrot (one
imaginary frequency with 87i cm-1) for the Cp*-dhpe
system are given in Table 3, while the full list of calculated
frequencies are given as Supporting Information. The two
intermediates ICpcis and ICptrans are formally unsaturated, 16-
electron dieneiron(II) complexes, [(η4-C5Me5H)Fe(dppe)H]+.
However, they are stabilized by an agnostic interaction
established by the newly created σ C-H bond and the Fe
center, which keeps the C5 ring essentially planar. The C-H
and Fe-H distances (1.248 and 1.661 Å in ICpcis; 1.216 and
1.757 Å in ICptrans) confirm the presence of this interaction.
The H-H bond is completely broken in ICpcis (H-H ) 2.005
Å). In the rotation transition state TSrot, the proton is midway
between the two positions. The H-H distance is almost
constant along the rotation. The TSrot also presents a σ C-H
agostic interaction with the Fe center (C-H) 1.179 Å;
Fe-H ) 1.483 Å). The transition states TSCpcis and TSCptrans
have geometries very near those of the corresponding
intermediates, ICpcis and ICptrans, respectively. The optimized
geometries of all stationary points are available as Supporting
Information.
This first step can be viewed as an intramolecular
heterolytic splitting of the H-H bond of the dihydrogen
ligand. In cationic complexes, the η2-H2 ligand can be acidic
enough to protonate a basic cis ligand; see Scheme 3,46
whereas the basic properties of a Cp* ligand are well-known.
We have searched the related ring-protonated intermediate
To introduce the electronic effects of the phenyl substit-
uents, we have performed full QM computations at the fixed
QM/MM structures of the optimized intermediates and
transition state of the 4 system. The relative energies of all
(44) Whitesides, T. H.; Neilan, J. P. J. Am. Chem. Soc. 1973, 95, 5811-
5813.
(45) Davies, S. G.; Felkin, H.; Watts, O. J. Chem. Soc., Chem. Commun.
1980, 159-160.
(46) Kubas, G. J. AdV. Inorg. Chem. 2004, 56, 127-177.
Inorganic Chemistry, Vol. 45, No. 25, 2006 10257

Baya et al.
Scheme 4
Figure 6. Example of a 31P NMR spectrum obtained during the kinetic
run of the isomerization experiment starting from Cp*FeH(dppe) +
CF3COOD (1:1) at 0 °C in CD2Cl2.
process do not scramble the hydride ligands. The latter
statement was verified by an independent experiment, where
a sample of [Cp*Fe(dppe)(H)2]+ was generated and then
treated with an equivalent amount of CF3COOD. No
deuterium incorporation occurred over 2 h at 273 K (the
isomerization process is extensive over this time scale, as
can be seen in Figure 1).
of the intermediates and TSrot slightly decrease relative to
the dihydrogen (Table 4). At this level, the energy barrier
for the “via Cp” pathway remains again close to that
computed for the “direct” pathway and to the experimentally
determined values. Thus, the possibility of a dihydrogen f
dihydride interconversion via hydrogen migration to the Cp*
ring and involving an heterolytic H-H bond breaking
appears as a feasible and competing pathway, together with
the expected direct oxidative addition mechanism in these
cationic iron half-sandwich complexes. Therefore, the com-
putational work carried out so far does not allow a clear-cut
distinction of the two alternative mechanisms. For this reason,
we have decided to carry out a more detailed analysis of the
KIE.
4. Isomerization Kinetics of [Cp*Fe(dppe)(η2-HD)]+.
The isomerization kinetics starting from the hydride complex
Cp*Fe(dppe)H and the deuterated trifluoroacetic acid,
CF3COOD, in a 1:1 ratio provided additional information
on the KIE. We carried out this experiment hoping to be
able to measure the isomerization rate constant of the pure
η2-HD complex (kHD) and compare this with the results of
the rate constants given in Table 1 for the η2-H2 (kHH) and
η2-D2 (kDD) complexes. As it turns out, the outcome of this
experiment was more complex and interesting than expected
because the nonclassical species obtained by low-temperature
protonation establishes a rapid equilibrium with the mixture
of the other two isotopomers. This means that, although the
proton transfer to the hydride site is effectively quantitative
when using a strong acid such as CF3COOD, the process is
still subject to rapid reversibility, generating both CF3COOH
and Cp*Fe(dppe)D, which further lead to the other two
nonclassical complexes as shown in Scheme 4. There-
fore, subsequent isomerization yields all three classical
products.
[Cp*Fe(dppe)(η2-H2)]+ + [Cp*Fe(dppe)(η2-D2)]+ a
2[Cp*Fe(dppe)(η2-HD)]+ (1)
The information on the individual concentrations of the
three nonclassical and the three classical species from a
kinetic run carried out at 0 °C, starting with approximately
equimolar amounts of Cp*FeH(dppe) and CF3COOD, was
obtained by integration of the inverse-gated doubly decoupled
2
31P{1H, H} NMR resonances. An example of a recorded
spectrum is shown in Figure 6. The double decoupling from
proton and deuterium is essential to generating sufficiently
sharp resonances and allowing a satisfactory deconvolution
of the individual components, especially for the nonclassical
isotopomers. The inverse-gated procedure, together with long
relaxation times and the use of a small pulse angle (see details
in the Experimental Section), allowed one to remove or at
least minimize the distortion of the integrated intensities
caused by the nuclear Overhauser effect and by an incom-
plete magnetization recovery. Visual inspection of Figure 6
shows that the classical structure experiences a much greater
isotope shift than the nonclassical one. This is quite expected
because the isotope effect in the former is transmitted from
the hydride ligands to the detected 31P nucleus through
stronger Fe-H covalent bonds.
The evolution of the signal intensities is shown in Figure
7. The presence of a KIE is already quite evident from a
qualitative analysis of the experimental data. In fact, the
concentration of the η2-H2 complex is always below that of
the η2-D2 complex, whereas the classical dihydride product
always has a greater concentration than the dideuteride
product. Fitting the calculated decays of the three nonclassical
species and the calculated growth of the three classical
species to the experimental data gave the solid lines indicated
However, useful kinetics information could still be ob-
tained from this experiment because (i) the equilibration
between the three nonclassical isotopomers according to the
equilibrium shown in eq 1 is much more rapid than the
isomerization rate, (ii) the isomerization is irreversible, and
(iii) the classical dihydride products of the isomerization
10258 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
From the results in Table 1 (run 8), we can derive kHH
/
kHD ) 1.24 ( 0.01 and kHD/kDD ) 1.58 ( 0.01. Isotope
effects have been previously reported for the oxidative
addition of dihydrogen to transition metals but only for
reactions between a metal complex and free dihydrogen, to
the best of our knowledge. Furthermore, they only relate to
the addition of dihydrogen versus dideuterium. Examples (kH/
kD in parentheses) are the additions to Fe(CO)4 (1.1 ( 0.1
at 24 °C),47 to Ir(PPh3)2(CO)Cl (1.09 at 25 °C),48,49 to
[Rh(PPh3)2Cl] (1.5 at 23 °C),50 to [Cr(CO)5(C6H12)] (1.9 at
23 °C),51 and to [W(PMe3)3I2] (1.2 ( 0.2 at 60 °C).52 While
the oxidative addition process may take place in a single
step, a local energy minimum for an intermediate dihydrogen
complex may also occur. In the latter case, when the slow
step is the intramolecular rearrangement of the dihydrogen
ligand, the observed isotope effect results from the combina-
tion of an EIE for the dihydrogen coordination step and a
KIE for the rearrangement. The dihydrogen coordination
equilibrium constant has been measured for a number of
systems and is always characterized by an inverse EIE.
Examples (with the KH/KD in parentheses) are [Ir(PCy3)2-
HCl2] (0.50 at -13 °C),53 [Ir(PtBu2Me)2H2Cl] (0.37 at -13
°C),54 [Os(PiPr3)2(CO)(Cl)H] (0.35 at 85°C),55 [Cr(CO)3-
(PCy3)2] (0.65 ( 0.15 at 22 °C),56 and [W(PMe3)4I2] (0.63
( 0.05 at 60 °C).52 Thus, a greater normal KIE for the
isomerization process would be necessary to yield an
observed normal KIE for the oxidative addition. Our
determined kHH/kDD value is 1.96 ( 0.02 at 273 K, i.e.,
comparable to the highest reported value for an oxidative
addition process, namely, 1.9 for [Cr(CO)5(C6H12)].51 The
present report appears to be the first direct investigation of
the KIE for the rearrangement process of a dihydrogen
intermediate to the final dihydride product and the first
complete analysis encompassing the H2, HD, and D2 isoto-
pomers. Note, however, that the rearrangement does not
consist of a simple H-H bond breaking completed by the
formation of the two M-H bonds with little change in the
rest of the coordination sphere because it occurs in the
majority of the other cases. A cis to trans rearrangement also
takes place. Thus, the current results should not be extrapo-
lated in general.
Figure 7. Evolution of the concentrations of (a) nonclassical and (b)
classical complexes during the isomerization process following the reaction
of Cp*Fe(dppe)H with CF3COOD at 0 °C in CD2Cl2: diamonds, dihydride;
squares, monohydride monodeuteride; triangles, dideuteride. The solid lines
are the best fit on the basis of the kinetic model, as outlined in the text and
in the Supporting Information.
in Figure 7 for the optimized rate constant values reported
in Table 1 (run 8). The details of the data analysis, which
required the numerical integration of the coupled differential
equations and a global least-squares fitting procedure, are
available as Supporting Information. Note that the optimized
value for kHH is identical within experimental error to that
independently (and more directly) obtained from the
CF3COOH protonation (Table 1, run 6). In addition, the
optimized value of kDD is slightly lower than that obtained
directly from the CF3COOD protonation of the deuteride
sample (Table 1, run 7), which must be considered an upper
estimate, as discussed above.
The data analysis (see the Supporting Information) used
the assumptions that the mixture of the three nonclassical
isotopomers is instantaneously equilibrated by virtue of eq
1, with an equilibrium constant taken as the average
calculated from the experimental data (K ) 3.3 ( 0.3). Note
that this value is significantly different from that expected
for a statistical distribution without an isotope effect (K )
4). When the fitting procedure was repeated by manually
altering this K value to 4 or to other values, the obtained
residual was greater, confirming the significance of this
deviation. Such observations prove that, besides the KIE
affecting the dihydrogen-dihydride isomerization, there is
an additional equilibrium isotope effect (EIE) on the isotope
redistribution equilibrium among the three nonclassical
species. The reason for this EIE will be analyzed in the next
section.
We can conclude that the isomerization process from
[Cp*Fe(dppe)(η2-H2)]+ to 4 has a small isotope effect and
(47) Wang, W.; Narducci, A. A.; House, P. G.; Weitz, E. J. Am. Chem.
Soc. 1996, 118, 8654-8657.
(48) Zhou, P.; Vitale, A. A.; San Filippo, J., Jr.; Saunders, W. H., Jr. J.
Am. Chem. Soc. 1985, 107, 8049-8054.
(49) Abu-Hasanayn, F.; Goldman, A. S.; Krogh-Jespersen, K. J. Phys.
Chem. 1993, 97, 5890-5896.
(50) Wink, D. A.; Ford, P. C. J. Am. Chem. Soc. 1987, 109, 436-442.
(51) Church, S. P.; Grevels, F.-W.; Hermann, H.; Schaffner, K. J. Chem.
Soc., Chem. Commun. 1985, 30-32.
(52) Hascall, T.; Rabinovich, D.; Murphy, V. J.; Beachy, M. D.; Friesner,
R. A.; Parkin, G. J. Am. Chem. Soc. 1999, 121, 11402-11417.
(53) Gusev, D. G.; Bakhmutov, V. I.; Grushin, V. V.; Vol’pin, M. E. Inorg.
Chim. Acta 1990, 177, 115-120.
(54) Hauger, B. E.; Gusev, D.; Caulton, K. G. J. Am. Chem. Soc. 1994,
116, 208-214.
(55) Bakhmutov, V. I.; Bertra´n, J.; Esteruelas, M. A.; Lledo´s, A.; Maseras,
F.; Modrego, J.; Oro, L. A.; Sola, E. Chem.sEur. J. 1996, 2, 815-
825.
(56) Bender, B. R.; Kubas, G. J.; Jones, L. H.; Swanson, B. I.; Eckert, J.;
Capps, K. B.; Hoff, C. D. J. Am. Chem. Soc. 1997, 119, 9179-9190.
Inorganic Chemistry, Vol. 45, No. 25, 2006 10259

Baya et al.
Table 5. Frequency Analysis of the EIE for the Isotope Redistribution
Equilibrium in Complex [Cp*Fe(dhpe)(η2-XY)]+ (2-XY) (X, Y ) H, D)
at 273 K
that this is not symmetric relative to the two hydride ligands.
This is not surprising because the two hydride ligands are
inequivalent in the rearrangement process (one migrates to
the new position, and the second one remains on the starting
site). Both KIEs and the EIE will now be analyzed from a
theoretical point of view, in terms of the two mechanistic
pathways that were found energetically compatible with the
measured activation enthalpy.
ν/cm-1
2-H2 2-HD 2-D2
assignment
MMI EXC ZPE
K
2762.7 2423.5 1956.3 ν(H-H)
1810.3 1561.0 1283.9 νas(Fe-H)
1.09 0.95 0.71 2.94
1.05 0.98 0.93 3.84
0.95 1.03 1.22 4.78
1096.9 903.1 785.6
νsym(Fe-H)
680.4 651.5 512.9 δ(Fe-H2)out-of-plane 1.22 1.00 0.75 3.65
585.5 533.2 534.6 δ(Fe-H2)in-plane
468.1 370.6 359.4 τ(Fe-H2)
total
0.91 1.00 1.15 4.19
0.82 1.00 1.26 4.10
0.97 0.96 0.88 3.30
5. Theoretical Analysis of the EIE. On the basis of the
energy-optimized structure of the nonclassical model com-
plex [Cp*Fe(dhpe)(η2-H2)]+, i.e., system 2, frequency cal-
culations were run for the HD and D2 analogues. For the
sake of clarity, the three isotopomers have been named 2-H2
(nondeuterated), 2-HD (monodeuterated), and 2-D2 (dideu-
terated). The EIE can be obtained directly from the calculated
∆G of the isotope redistribution equilibrium (see thermo-
chemical data in the Supporting Information). This is
obtained from a statistical mechanics treatment that uses
unscaled normal-mode frequencies. The resulting value of
the equilibrium constant for eq 1 is 3.81. The slight
disagreement with the experiment is expected, given the
various approximations used by the calculation.
Table 6. Calculated Activation Free Energies in kcal mol-1 [and
Relative ∆∆Gq Values with Respect to ∆Gq(HH) in Parentheses] for the
“Direct” and “via Cp” Mechanisms at 273 K (Model 2 System)
mechanism
∆Gq(HH)
∆Gq(HmDf)
∆Gq(DmHf)
∆Gq(DD)
direct
29.26 (0.00) 29.41 (0.15) 29.54 (0.28)
29.70 (0.44)
via Cp
31.27 (0.00) 31.41 (0.14) 31.15 (-0.12) 31.31 (0.04)
the experimental value. As Table 5 shows, the EIE is
dominated by the ZPE term, and the most significant
contribution to the reduction of K from the statistical value
of 4 is given by the ν(H-H) vibration. However, the
contribution of the other isotope-sensitive modes cannot be
neglected.
From the knowledge of the individual frequencies, the
equilibrium constant may also be calculated through the
classical Bigeleisen-Mayer treatment, where the constant
is given by the expression in eq 2. For the definition of the
6. KIE on the Dihydrogen-Dihydride Isomerization.
Analogous frequency calculations were run on the isoto-
pomers of the two transition states corresponding to the two
mechanisms found compatible with the experimental activa-
tion barrier. We must also consider that the symmetry of
the dihydrogen ligand is broken on going from the initial
nonclassical system to the transition state. While the two
hydride positions are symmetry-equivalent in the starting
dihydrogen complex, they become inequivalent in the
transition state. They will be identified as the “fixed” (Hf/
Df) and “migrating” (Hm/Dm) positions, respectively (see
Figure 5). This implies that two pathways for each mecha-
nism must be considered in the case of the monodeuterated
species, depending on whether the migrating atom is H or
D.
As for the EIE analysis, the KIE can be derived directly
from the calculated free energies, in this case the activation
free energies. These are reported in Table 6, and the
corresponding ZPE and thermal enthalpy values are included
in the Supporting Information. Clearly, the migration of
hydrogen is favored over the migration of deuterium in the
mixed 2-HD species for the “direct” mechanism, whereas
the reverse is true for the “via Cp” mechanism. The resulting
calculated isotope effects are given in Table 7. The ∆∆Gq
values for the two different transition states of the 2-HD
systems (HmDf and DmHf) are small in each mechanism
(-0.117 kcal mol-1 for TSdir and 0.264 kcal mol-1 for TSrot),
allowing the calculation of k(HmDf)/k(DmHf) ) 1.241 for the
“direct” and 0.615 for the “via Cp” mechanism. Thus, a
considerable fraction of the HD sample will isomerize by
the less favorable pathway (HmDf/DmHf ) 55.4:44.6 for the
“direct” and 38.1:61.9 for the “via Cp” mechanism). The
observed KIE must take this into account. Thus, the
calculated effective rate of isomerization for the 2-HD
K ) 4(MMI)(EXC)(ZPE)
(2)
MMI (mass moment of inertia), EXC (excitations of vibra-
tional energy levels), and ZPE (zero-point vibrational energy)
terms, see the Supporting Information and the literature.57,58
The factor of 4 takes into account the doubled concentration
of the HD species in the equilibrium expression. The
advantage of the Bigeleisen-Mayer treatment is to pinpoint
the isotope-sensitive modes that are mostly responsible for
the observed effect. Related analyses have been previously
reported for the secondary isotope effects in the coordination
of C2H4 vs C2D4 to the Os2(CO)8 moiety59 and for the primary
isotope effects in the H2/D2, CH4/CH3D, and CH4/CD4
oxidative additions to trans-Ir(PR3)2(CO)X,60 in the H2/D2
coordination to W(CO)3(PCy3)2,56 and in the H2/D2/T2
coordination and oxidative addition to W(CO)5.61 The
isotopic redistribution between H2, HD, and D2 complexes
has never been analyzed to the best of our knowledge.
This type of analysis was carried out for our system on
the basis of the selected, most isotope-sensitive frequencies,
reported in Table 5 (no scaling factors have been used). Only
these normal modes are sensitive to the isotope nature and
contribute significantly to the EIE. A full list of vibrational
frequencies is provided in the Supporting Information section.
The calculated K value is 3.30 at 273 K, in agreement with
(57) Bigeleisen, J.; Mayer, M. G. J. Chem. Phys. 1947, 15, 261-267.
(58) Melander, L. Isotope effects on reaction rates; The Ronal Press Co.:
New York, 1960.
(59) Bender, B. R. J. Am. Chem. Soc. 1995, 117, 11239-11246.
(60) Abu-Hasanayn, F.; Krogh-Jespersen, K.; Goldman, A. S. J. Am. Chem.
Soc. 1993, 115, 8019-8023.
(61) Janak, K.; Parkin, G. Organometallics 2003, 22, 4378-4380.
10260 Inorganic Chemistry, Vol. 45, No. 25, 2006

Analysis of Kinetic Isotope Effects
Table 7. Frequency Analysis of the KIE for the Irreversible
Isomerization of [Cp*Fe(dhpe)(η2-AB)]+ to [Cp*Fe(dhpe)(A)(B)]+ (A,
B ) H, D) at 273 K
Table 8. Isotope-Sensitive Frequencies (cm-1) for the Transition States
of the “Direct” (TSdir) and “via Cp” (TSrot) Mechanisms
(a) “Direct” Mechanism
b
c
X/Ya
VP*
EXC*
ZPE*
kX/kY
kX/kY
TSdir-H2 TSdir-HmDf TSdir-DmHf TSdir-D2
assignment
(a) “Direct” Mechanism
-298.3
507.5
536.5
587.0
591.7
654.5
667.5
677.0
778.7
1025.3
1921.0
2027.7
-296.8
601.2
523.6
593.8
568.5
466.1
545.4
668.4
671.4
1024.9
1369.7
2027.5
-227.2
507.1
536.3
586.8
589.0
651.2
662.6
675.6
783.1
719.3
1921.3
1448.9
-226.6
598.8
522.6
593.7
567.3
465.1
544.3
664.7
668.9
721.3
1369.0
1449.4
ν*
HH/HmDf
HH/DmHf
HmDf/DD
DmHf/DD
0.938
0.753
0.772
0.961
1.050
1.126
1.093
1.019
1.313
1.380
1.521
1.447
1.304
1.619
1.693
1.364
1.299
1.536
1.680
1.421
(b) “Via Cp” Mechanism
δ(Fe-Hf)in-plane
HH/HmDf
HH/HmDf
HmDf/DD
HmDf/DD
1.018
0.997
1.018
1.038
1.027
1.100
1.068
0.998
1.150
0.742
0.824
1.277
1.282
0.788
0.826
1.344
1.215
0.816
0.897
1.336
δ(Fe-Hm)out-of-plane
ν(Fe-Hf)
ν(Fe-Hm)
a The m and f labels refer to the “moving” and “fixed” atoms involved
in the rearrangement. b From the ∆Gq analysis (see the text). c From eq 3
and the selected frequencies in Table 8.
(b) “Via Cp” Mechanism
TSrot-H2 TSrot-HmDf TSrot-DmHf TSrot-D2
species will be kHD ) 0.554k(HmDf) + 0.446k(DmHf) for the
“direct” pathway and kHD ) 0.381k(HmDf) + 0.619k(DmHf)
for the “via Cp” pathway. This leads to the calculation of
the following values for the KIE:
-86.8
364.8
394.7
440.1
469.1
531.2
537.0
548.6
637.3
686.8
768.6
772.9
801.6
819.1
823.7
846.6
929.9
983.2
1059.8
1083.3
1289.7
1408.1
1924.1
1952.3
-85.9
352.1
394.3
411.6
448.5
530.6
544.0
542.0
529.4
686.7
756.2
776.7
802.3
603.3
822.4
834.0
929.9
982.9
1059.7
1083.2
1289.7
1408.3
1926.1
1390.2
-86.6
364.3
382.1
439.3
467.6
519.6
535.4
546.1
637.1
681.4
768.3
772.8
834.4
813.6
821.3
847.3
699.9
1121.7
998.1
1075.1
906.9
1353.3
1386.2
1950.4
-85.7
351.8
381.7
411.5
447.8
516.8
543.2
539.6
529.1
681.4
755.9
776.6
834.7
602.6
817.5
832.6
700.0
1121.7
997.7
1075.0
906.6
1353.3
1381.3
1396.4
ν*
(a) direct
(b) via Cp
k
k
HH/kHD ) 1.428
HD/kDD ) 1.547
kHH/kHD ) 0.924
kHD/kDD ) 1.147
δ(Fe-Hf)
The computed values are in much better agreement with the
experimental ones (1.24 ( 0.01 and 1.58 ( 0.01, respec-
tively) for the “direct” mechanism. This result must, there-
fore, be considered in strong support of the occurrence of
the “direct” mechanism and against the “via Cp” mechanism.
As for the EIE analysis, the KIEs can also be calculated
by an extension of the Bigeleisen-Mayer method,58,62 with
the relevant relationship becoming that of eq 3. The
δ(Fe-Hm-C)
ν
sym(Fe-Hm;Fe-C)
νas(Fe-Hm;Fe-C)
sym(Fe-Hm;C-Hm)
ν(Fe-Hf)
KIE ) kX/kY ) (VP*)(EXC*)(ZPE*)
(3)
ν
expressions of the VP*, EXC*, and ZPE* terms are based
on the starting materials and transition states rather than on
the starting materials and products. They are again available
in the Supporting Information and in the literature.
around TSdir. On the other hand, the “via Cp” mechanism
involves a transition state TSrot connecting two agostic
intermediates and having itself a significant agostic interac-
tion. The C-Hm and Fe-Hm bonds undergo little changes
around TSrot, resulting in essentially isotope-independent
imaginary frequencies.
Table 8 contains only those frequencies that are sensitive
to the isotope nature and contribute significantly to the KIE.
The full list of frequencies is available in the Supporting
Information. The frequency analysis in this case is not as
straightforward as that for the EIE, especially for the “via
Cp” mechanism, because the Fe-X/Y and X-Y stretching
and bending motions are heavily mixed with other molecular
motions in several normal modes. Only those normal modes
having a major component from the Fe-X/Y bonds have
been explicitly assigned in Table 8. The calculation of the
various terms of eq 3 leads to the values collected in Table
7. It can be seen that the ZPE* term contributes the most
but the VP* term also gives an important contribution,
especially for the “direct” mechanism when comparing
different moving atoms (Hm/Dm). This is mostly due to the
ν*X/ν*Y term contained in VP* (ratio of the imaginary
frequencies; see the Supporting Information) because large
changes at the level of the Fe-Hm and Hf-Hm bonds occur
The corresponding KIE values are again listed in Table
7. We can grossly attribute this difference to the variation
of the chemical environment of the migrating atom, Hm/Dm,
on going from the starting dihydrogen complex and the
transition state. For the “direct” mechanism, the strong H-H
and two weak Fe-H bonds (in-phase and out-of-phase
stretching vibrations at 1096.9 and 1810.3 cm-1 for the 2-H2
species, respectively) are transformed into two stronger
Fe-H bonds, yielding stretching vibrations at 1921.0 (for
Hf) and 2027.7 (for Hm) cm-1. However, the loss of the strong
H-H bond in the starting material (νHH ) 2762.7 cm-1)
dominates, yielding a normal KIE. The fact that KIE(HH/
HD) is greater when the migrating atom is D (1.536, primary
isotope effect) and lower when it is H (1.299, secondary
isotope effect) is not solely related to the bond stretching
vibrations because ν(Fe-Xm) > ν(Fe-Xf) for both X ) H
and D. Thus, the bending modes play a dominant role. The
(62) Bigeleisen, J. J. Chem. Phys. 1949, 17, 675-678.
Inorganic Chemistry, Vol. 45, No. 25, 2006 10261

Baya et al.
Chart 3
by protonation of Cp*Fe(dppe)H with CF3COOD results in
a rapid isotope redistribution equilibrium, yielding complexes
[Cp*Fe(dppe)(η2-H2)]+ and [Cp*Fe(dppe)(η2-D2)]+; each of
the three isotopomers isomerizes to the corresponding
classical product with its own rate constant (kHH, kHD, and
k
DD). The resulting KIEs at 273 K are kHH/kHD ) 1.24 (
0.01 and kHD/kDD ) 1.58 ( 0.01 (and, consequently, kHH
/
kDD ) 1.96 ( 0.02). This type of isotope effect analysis, for
a rearrangement process of a dihydrogen complex to the
isomeric dihydride and encompassing the three H2, HD, and
D2 isotopomers, is unprecedented to the best of our knowl-
edge. The isotope redistribution EIE and the KIEs were also
derived from the calculated normal-mode frequencies of the
dihydrogen isotopomers and the transition states. The
computed values are in excellent agreement with the
experimental values, but only for the “direct” isomerization
mechanism, providing strong support in favor of this
rearrangement pathway and against the “via Cp” pathway
for this system. However, given the closeness of the
calculated activation barriers, it seems reasonable to think
that subtle modifications in the metallic system could favor
a mechanism of the “via Cp” type.
same argument rationalizes why KIE(HD/DD) is greater
when the migrating atom is H. For the “via Cp” mechanism,
on the other hand, the two hydride ligands yield a relatively
strong Fe-Hf bond (its stretching vibration at 1952.3 cm-1
in TSrot is similar to that calculated in TSdir) and an agostic
Fe‚‚‚Hm-C moiety. The strength of the C-H bond is reduced
from that of a regular C-H bond, but the combination of
the C-Hm and Fe‚‚‚Hm vibrations, which are mixed with an
Fe-C component, yields three relatively high-frequency
normal modes, as shown in Chart 3. The overall effect is a
strengthening of the bonding environment for the Hm atom
and a weakening for the Hf atom, leading to an inverse
primary KIE (migrating D) and a normal secondary KIE
(migrating H).
Acknowledgment. We thank the European Commission
through the HYDROCHEM program (Contract No. HPRN-
CT-2002-00176) for support of this work. Additional bilateral
(PICS, France-Russia) and national support from the CNRS
(France), from the RFBR (Grants 05-03-22001 and 05-03-
32415) and the Division of Chemistry and Material Sciences
of RAS (Russia), and from MEC (Project CTQ2005-09000-
CO2-01; Spain) is also gratefully acknowledged. A.L. thanks
the Generalitat de Catalunya for a Distincio´ per a la Promocio
de la Recerca Universita`ria. N.V.B. thanks the Russian
Science Support Foundation for an individual grant. M.B.
thanks the Spanish Ministerio de Educacio´n y Ciencia for a
postdoctoral fellowship.
Conclusions
We have reported here a detailed investigation of the
irreversible isomerization process of the dihydrogen inter-
mediate [Cp*Fe(dppe)(η2-H2)]+, obtained by low-temperature
protonation of Cp*Fe(dppe)H, to the dihydride complex
trans-[Cp*Fe(dppe)(H)2]+. The computational tool illustrates
the need to include the steric and electronic effects of the
Cp and diphosphine substituents to quantitatively reproduce
the relative stabilities of nonclassical and classical isomers
and the activation barrier of the isomerization process.
However, the calculations also indicate that two different
pathways are likely candidates for the isomerization mech-
anism: a “direct” pathway (∆Eq ) 21.3 kcal mol-1) and a
“via Cp” pathway (∆Eq ) 22.2 kcal mol-1), never suggested
previously, which involves Cp*H intermediates (cf. 21.6 (
0.8 kcal mol-1 from the experiment). Sorting between these
pathways was possible through the investigation of isotope
effects. Generation of the [Cp*Fe(dppe)(η2-HD)]+ complex
Supporting Information Available: Details of the kinetic
analysis for the Cp*Fe(dppe)H + CF3COOD reaction, optimized
geometries (Cartesian coordinates) for the calculated species, and
calculated normal-mode frequencies used for the EIE and KIE
analyses. This material is available free of charge via the Internet
at http://pubs.acs.org.
IC061428N
10262 Inorganic Chemistry, Vol. 45, No. 25, 2006