Kinetics and mechanism of the proton transfer to Cp*Fe(dppe)H: Absence of a direct protonation at the metal site

Article abstract of DOI:10.1021/ja0358450

The reaction between Cp*Fe(dppe)H and a number of different proton donors (2-fluoroethanol, MFE; 2,2,2-trifluoroethanol, TFE; hexafluoro-2-propanol, HFIP; perfluoro-tert-butyl alcohol, PFTB; and trifluoroacetic acid, TFA) has been investigated spectroscopically by variable-temperature infrared, UV-visible, and NMR spectroscopy, and has been measured kinetically by the stopped-flow technique with UV-visible detection. The low-temperature IR study shows the establishment of hydrogen-bonding interactions which involve the hydride ligand as the proton accepting site. This investigation quantifies the thermodynamics of the hydrogen-bonding interaction and the basicity factor (E_j) of the hydride complex. All techniques agree in indicating an equilibration process, after the immediate hydrogen-bond formation, between the hydride complex and an intermediate dihydrogen complex, [Cp*Fe(dppe)(H₂)]⁺. The equilibrium is shifted toward the dihydrogen complex to a greater extent for the stronger alcohols and for higher alcohol/ Fe ratios. The observed equilibration rate constant is linearly dependent on the alcohol concentration, in agreement with the involvement of two alcohol molecules and the formation of a homoconjugate pair. The rate constant increases with the acidity of the proton donor (TFE < HFIP < PFTB < TFA). The rate of the subsequent irreversible isomerization leading to the classical dihydride complex, [Cp*Fe(dppe)H₂]⁺, is first order, and the rate constant does not depend on the proton donor nature. The reaction continues, if conducted in CH₂Cl₂, with a third, slower step leading to the paramagnetic [Cp*Fe(dppe)Cl]⁺ product. The kinetic data are in accord with an isomerization mechanism consisting of an intramolecular reorganization, leading in one step from the dihydrogen complex to the classical dihydride species, and disagree with the occurrence of a proton-transfer process at the metal site.

Full text of DOI:10.1021/ja0358450

Published on Web 08/07/2003
Kinetics and Mechanism of the Proton Transfer to
Cp*Fe(dppe)H: Absence of a Direct Protonation at the Metal
Site
Natalia V. Belkova,† Pavel O. Revin,† Lina M. Epstein,† Evgenii V. Vorontsov,†
Vladimir I. Bakhmutov,†,§ Elena S. Shubina,*,† Edmond Collange,‡ and
Rinaldo Poli*,‡
Contribution from the NesmeyanoV Institute of Organoelement Compounds, Russian Academy
of Sciences, VaViloV Street 28, 119991 Moscow, Russia, and Laboratoire de Synthe`se et
d’Electrosynthe`se Organome´talliques (LSEO UMR 5632), UniVersite´ de Bourgogne,
Faculte´ de Sciences “Gabriel”, 6 bouleVard Gabriel, 21000 Dijon, France
Received April 29, 2003; E-mail: poli@u-bourgogne.fr
Abstract: The reaction between Cp*Fe(dppe)H and a number of different proton donors (2-fluoroethanol,
MFE; 2,2,2-trifluoroethanol, TFE; hexafluoro-2-propanol, HFIP; perfluoro-tert-butyl alcohol, PFTB; and
trifluoroacetic acid, TFA) has been investigated spectroscopically by variable-temperature infrared, UV-
visible, and NMR spectroscopy, and has been measured kinetically by the stopped-flow technique with
UV-visible detection. The low-temperature IR study shows the establishment of hydrogen-bonding
interactions which involve the hydride ligand as the proton accepting site. This investigation quantifies the
thermodynamics of the hydrogen-bonding interaction and the basicity factor (Ej) of the hydride complex.
All techniques agree in indicating an equilibration process, after the immediate hydrogen-bond formation,
between the hydride complex and an intermediate dihydrogen complex, [Cp*Fe(dppe)(H2)]+. The equilibrium
is shifted toward the dihydrogen complex to a greater extent for the stronger alcohols and for higher alcohol/
Fe ratios. The observed equilibration rate constant is linearly dependent on the alcohol concentration, in
agreement with the involvement of two alcohol molecules and the formation of a homoconjugate pair. The
rate constant increases with the acidity of the proton donor (TFE < HFIP < PFTB < TFA). The rate of the
subsequent irreversible isomerization leading to the classical dihydride complex, [Cp*Fe(dppe)H2]+, is first
order, and the rate constant does not depend on the proton donor nature. The reaction continues, if
conducted in CH2Cl2, with a third, slower step leading to the paramagnetic [Cp*Fe(dppe)Cl]+ product. The
kinetic data are in accord with an isomerization mechanism consisting of an intramolecular reorganization,
leading in one step from the dihydrogen complex to the classical dihydride species, and disagree with the
occurrence of a proton-transfer process at the metal site.
Introduction
complex, is correspondingly faster at a hydride ligand site
relative to the metal lone pair (hydride ligands are, in general,
kinetically more basic than transition metal centers).3,4 Indeed,
dihydrogen complexes have often been detected as intermediates
along the formation of classical polyhydrides as the thermody-
namically stable final products.5-9
Proton-transfer processes to and from transition metal centers
and hydride ligand sites are key steps in many stoichiometric
and catalytic chemical and biochemical processes and have
received a great deal of attention over the last two decades.1
While protonation of a metal site affords a new hydride ligand,
proton transfer to a hydride ligand site affords a dihydrogen
ligand (nonclassical dihydride complex, or σ-H2 complex), see
Scheme 1.
Another impetus to this area has been given by the discovery
of “nonclassical dihydrogen bonding”; that is, adducts of type
I are shown to be intermediates along the proton-transfer
A fundamental issue is that of the thermodynamic versus
kinetic acidity, whereby nonclassical hydride complexes (III)
are shown to be deprotonated faster than the classical tautomers
(VI), even when the latter are stronger acids in the thermody-
namic sense.2 The reverse process, protonation of a hydride
(2) Kristja´nsdo´ttir, S. S.; Norton, J. R. In Transition Metal Hydrides; Dedieu,
A., Ed.; VCH: New York, 1992; pp 309-359.
(3) Papish, E. T.; Rix, F. C.; Spetseris, N.; Norton, J. R.; Williams, R. D. J.
Am. Chem. Soc. 2000, 122, 12235-12242.
(4) Papish, E. T.; Magee, M. P.; Norton, J. R. In Recent AdVances in Hydride
Chemistry; Peruzzini, M., Poli, R., Eds.; Elsevier: Amsterdam, 2001; pp
39-74.
(5) Parkin, G.; Bercaw, J. E. J. Chem. Soc., Chem. Commun. 1989, 255-257.
(6) Chinn, M. S.; Heinekey, D. M. J. Am. Chem. Soc. 1990, 112, 5166-5175.
(7) Jia, G.; Lough, A. J.; Morris, R. H. Organometallics 1992, 11, 161-171.
(8) Hamon, P.; Toupet, L.; Hamon, J.-R.; Lapinte, C. Organometallics 1992,
11, 1429-1431.
(9) Hamon, J. R.; Hamon, P.; Toupet, L.; Costuas, K.; Saillard, J. Y. C. R.
Acad. Sci., Ser. IIc: Chim. 2002, 5, 89-98.
† Russian Academy of Sciences.
‡ Universite´ de Bourgogne.
§ Current address: Department of Chemistry, Texas A&M University,
P.O. Box 30012, College Station, TX 77842-3012.
(1) Peruzzini, M., Poli, R., Eds. Recent AdVances in Hydride Chemistry;
Elsevier: Amsterdam, 2001.
9
11106
J. AM. CHEM. SOC. 2003, 125, 11106-11115
10.1021/ja0358450 CCC: $25.00 © 2003 American Chemical Society

Mechanism of the Proton Transfer to Cp*Fe(dppe)H
A R T I C L E S
Scheme 1
pathway to and from dihydrogen complexes. The latter may
exist as either free complexes, III, or retain the conjugate base
in a hydrogen-bonding interaction, II. Systems such as I have
also been shown to be intermediates along the pathway to the
heterolytic splitting of dihydrogen.10 Depending on the relative
strength of the proton donor and acceptor, these can be stable
entities in solution, displaying characteristic spectroscopic
signatures.11,12 The proton transfer to the metal site can
analogously be imagined as proceeding via a hydrogen-bonded
intermediate IV, leading to either free (VI) or hydrogen-bonded
(V) classical hydride product. A number of studies of adducts
such as IV are available,13 and the general implication is that
they are indeed intermediates in the proton-transfer process.
However, the interconversion between classical and nonclassical
tautomers takes place rather easily in many cases and can in
principle occur both on the free (III and VI) and on the
hydrogen-bonded (II and V) complexes, thus opening a second
possible pathway for protonation of the metal center (fast
protonation of the hydride site followed by isomerization).
that the rate of metal protonation in compound CpW(CO)2-
(PMe3)H is 9.6 times smaller than the rate of hydride protona-
tion. The rate of metal protonation was obtained indirectly from
the combination of the measured deprotonation rate of the
+
classical dihydride complex CpW(CO)2(PMe3)H2 and the
measured pKa value. It is quite possible, however, that the
deprotonation reaction occurs only after isomerization to the
nonclassical intermediate. Thus, the 9.6 factor between these
two rates could merely be related to the tautomerization
thermodynamics and does not imply the occurrence of a direct
metal protonation. More detailed information would only be
available from an independent analysis of the tautomerization
kinetics or equilibria, but, unfortunately, the nonclassical
tautomer cannot be directly observed for this system.
Previous work by other authors has addressed specific aspects
of the kinetics and mechanism of proton transfer to a hydride
complex. The work of Basallote et al. mostly concerns the
formation of dihydrogen complexes that do not evolve to the
classical isomers but rather to H2 substitution products, such as
[MH(H2)L4]+ (M ) Fe, Ru; L ) P-donor ligand).24-27 A single
exception is the formation of complexes [CpRuL2(H2)]+ (L2 )
dppm, dppe, and (PPh3)2) which further evolve to the classical
isomers, but the kinetics investigation was not extended to this
isomerization step.28 Work by Chinn and Heinekey on the pro-
tonation of a variety of CpRuH(L)(L′) and related Cp* com-
plexes,6 and the later extension to related systems by Puerta et
al.,29 on the other hand, address the mechanism of the reversible
or irreversible conversion of the dihydrogen complexes to the
corresponding dihydrides. Although the low-temperature pro-
tonation was shown to occur selectively at the hydride site for
these systems, the results of those studies cannot exclude a
competitive protonation of the metal site at ambient temperature.
To shine more light onto this basic question, we have decided
to carry out detailed investigations on the proton transfer to
complex Cp*Fe(dppe)H, prepared several years ago by Hamon
et al.8,9 A selective protonation at the hydride site with formation
Hydrogen-bonded complexes such as IV and V have been
located along the pathway of proton transfer to the metal center
when no other hydride ligands are present,14-19 but only in a
few cases have these been proven intermediates of the direct
proton transfer to the metal site.20-23 Norton has recently shown
(10) Morris, R. H. In Recent AdVances in Hydride Chemistry; Peruzzini, M.,
Poli, R., Eds.; Elsevier: Amsterdam, 2001; pp 1-38.
(11) Epstein, L. M.; Belkova, N. V.; Shubina, E. S. In Recent AdVances in
Hydride Chemistry; Peruzzini, M., Poli, R., Eds.; Elsevier: Amsterdam,
2001; pp 391-418.
(12) Epstein, L. M.; Shubina, E. S. Coord. Chem. ReV. 2002, 231, 165-181.
(13) Shubina, E. S.; Belkova, N. V.; Epstein, L. M. J. Organomet. Chem. 1997,
536-537, 17-29.
(14) Kazarian, S. G.; Hamley, P. A.; Poliakoff, M. Chem. Commun. 1992, 994-
997.
(15) Kazarian, S. G.; Hamley, P. A.; Poliakoff, M. J. Am. Chem. Soc. 1993,
115, 9069-9079.
(16) Poliakoff, M.; Howdle, S. M.; Kazarian, S. G. Angew. Chem., Int. Ed. Engl.
1995, 34, 1275-1295.
(17) Shubina, E. S.; Krylov, A. N.; Kreindlin, A. Z.; Rybinskaya, M. I.; Epstein,
L. M. J. Mol. Struct. 1993, 301, 1-5.
(18) Shubina, E. S.; Krylov, A. N.; Muratov, D. V.; Filchikov, A. A.; Epstein,
L. M. Russ. Chem. Bull. 1993, 42, 1919-1920.
(19) Epstein, L. M.; Krylov, A. N.; Shubina, E. S. J. Mol. Struct. 1994, 322,
345-352.
(24) Basallote, M. G.; Dura´n, J.; Ferna´ndez-Trujillo, M. J.; Ma´n˜ez, M. A.; de
la Torre, J. R. J. Chem. Soc., Dalton Trans. 1998, 745-750.
(25) Basallote, M. G.; Dura´n, J.; Ferna´ndez-Trujillo, M. J.; Ma´n˜ez, M. A. J.
Chem. Soc., Dalton Trans. 1998, 2205-2210.
(20) Shubina, E. S.; Krylov, A. N.; Belkova, N. V.; Epstein, L. M.; Borisov, A.
P.; Mahaev, V. D. J. Organomet. Chem. 1995, 493, 275-277.
(21) Albinati, A.; Bakhmutov, V. I.; Belkova, N. V.; Bianchini, C.; de los Rios,
I.; Epstein, L.; Gutsul, E. I.; Marvelli, L.; Peruzzini, M.; Rossi, R.; Shubina,
E. S.; Vorontsov, E. V.; Zanobini, F. Eur. J. Inorg. Chem. 2002, 1530-
1539.
(26) Basallote, M. G.; Dura´n, J.; Ferna´ndez-Trujillo, M. J.; Ma´n˜ez, M. A. Inorg.
Chem. 1999, 38, 5067-5071.
(27) Basallote, M. G.; Dura´n, J.; Ferna´ndez-Trujillo, M. J.; Ma´n˜ez, M. A. J.
Organomet. Chem. 2000, 609, 29-35.
(22) Hascall, T.; Baik, M. H.; Bridgewater, B. A.; Shin, J. H.; Churchill, D. G.;
Friesner, R. A.; Parkin, G. Chem. Commun. 2002, 2644-2645.
(23) Belkova, N. V.; Gutsul, E. I.; Epstein, L.; Shubina, E. S.; Bianchini, C.;
Zanobini, F.; Peruzzini, M., unpublished observations.
(28) Basallote, M. G.; Dura´n, J.; Ferna´ndez-Trujillo, M. J.; Ma´n˜ez, M. A.
Organometallics 2000, 19, 695-698.
(29) de los R´ıos, I.; Jime´nez-Tenorio, M.; Padilla, J.; Puerta, M. C.; Valerga, P.
Organometallics 1996, 15, 4565-4574.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 11107

A R T I C L E S
Belkova et al.
Table 1. IR (νFe-H), 1H (Hydride Resonance), and 31P NMR
of the dihydrogen complex as an observable intermediate and
the thermodynamic preference for the metal site were established
by using the strong acid HBF4. Thus, the protonation at the
hydride site and the isomerization can be analyzed independently
from the kinetic and thermodynamic points of view at the same
temperature. Our study was carried out by using the following
proton donors (HA), in order of increasing acid strength:
2-monofluoroethanol (MFE), 2,2,2-trifluoroethanol (TFE), hexa-
fluoro-2-propanol (HFIP), perfluoro-tert-butyl alcohol (PFTB),
and trifluoroacetic acid (TFA). The interaction has been
investigated by a combination of NMR, IR, and UV-visible
spectroscopy at various temperatures in the 200-290 K range.
This approach allowed us to study the overall process step by
step. The proton transfer and isomerization rates at room
temperature have been determined by both classical mixing and
rapid mixing (stopped-flow) techniques. This is the first reported
study where an independent kinetic analysis is carried out for
both the proton transfer and the isomerization steps, and the
first study of this kind where proton donors of different acid
strength are used and compared.
Data in CD2Cl2 for Cp*Fe(dppe)H and for the Products Deriving
from Its Interaction with HA
compound
ν
Fe-H/cm-1
δH (JHP/Hz)
δP
Cp*(dppe)FeH
1844 (s)
1830 (sh)
c
-17.32 (67)
b
-12.50 (broad)
-7.98 (73)
107.97
b
93.39
90.97
Cp*(dppe)FeH‚‚‚HAa
[Cp*(dppe)Fe(H2)]+
[Cp*(dppe)Fe(H)2]+
1940 (w)
a Observed for HA ) TFE. b Not significantly shifted from the values
of free Cp*Fe(dppe)H (see text). c Not observable (see text).
obtained at the low concentrations required for work in a suitable
absorbance range when using a regular 1 cm cell (ca. 5 × 10-4 M).
This phenomenon is attributed to oxidation by oxygen diffusion through
the instrument transfer lines, as confirmed by the observation of small
and irreproducible signal evolutions when shooting the same hydride
solution from both syringes. Switching to a 10-fold concentration and
to a smaller path length (1.5 mm) reduced the oxidation problem below
acceptable noise levels. Only the data that were collected within the
first 50 s were analyzed, yielding reproducible results. No oxidation
problems were evident, on the other hand, for the experiments carried
out by the more traditional long time scale monitoring. Data analyses
were carried out by using the SPECFIT31 global analysis package of
programs.
Experimental Section
All manipulations were carried out under an argon atmosphere by
standard Schlenk techniques. The Cp*Fe(dppe)H hydride was synthe-
sized according to the literature.30
Results and Discussion
(a) Interaction with HBF4. Spectral Characteristics of the
Classical and Nonclassical Protonation Products. Half-
sandwich Fe hydride complexes are relatively rare.8,9,32-37 Those
containing two or more hydride ligands display a preference
for the classical Fe(IV) isomer, the Fe(II) dihydrogen isomer
being observable only for less donating ligand environments.
For the Cp*-dppe derivative, both the dihydrogen complex,
[Cp*Fe(dppe)(H2)]+, and the classical dihydride complex,
[Cp*Fe(dppe)(H)2]+, have been described as the kinetic and
thermodynamic protonation products, respectively, resulting
from the treatment of the hydride complex with HBF4.8,9 The
faster formation of the dihydrogen complex relative to the
dihydride tautomer illustrates that proton transfer to the hydride
site is more facile than that to the metal center, as is also well
established for a number of other complexes.3,6,7
IR and UV-Visible Investigations. The IR measurements were
performed on a “Specord M82” spectrometer (IR) on 0.1-0.15 M (for
the ν(OH) measurements) or 0.02-0.025 M (for the ν(MH) measure-
ments) hydride solutions in CH2Cl2 (0.12 cm path length) using CaF2
cells. UV measurements were performed on Specord M-40 and Varian
Cary 5 spectrophotometers on 0.01 M solutions in CH2Cl2. All IR and
UV measurements were carried out by use of a home-modified cryostat
(Carl Zeiss Jena) in the 200-290 K temperature range. The cryostat
modification allows the transfer of the reagents (premixed either at
low or room temperature) directly into the cell under an inert atmosphere
and at the desired temperature. The accuracy of the temperature
adjustment was (0.5 K. This setup was used both for the variable-
temperature equilibrium studies and for the kinetics investigations at
constant temperature with UV-visible spectroscopic monitoring. An
IR study was also attempted in cyclohexane. However, the precipitation
of the cationic dihydrogen complex rendered impossible a careful
investigation in this solvent.
Because the spectral characterization of the two tautomeric
dihydride cations in solution was previously restricted to NMR,
we have reinvestigated the HBF4 protonation by IR and UV-
visible techniques, in addition to NMR. A careful reinvestigation
of the longitudinal relaxation time as a function of temperature
for the starting hydride and protonation products, however,
NMR Investigations. The NMR studies were carried out in standard
5 mm-NMR tubes containing solutions of the complexes in CD2Cl2.
1
The H and 31P NMR data were collected with a Bruker AMX 400
spectrometer operating at 400.13 and 161.98 MHz, respectively. The
conventional inversion-recovery method (180-τ-90) was used to
determine the variable-temperature longitudinal relaxation time T1. The
calculation of the relaxation times was made using the nonlinear three-
parameter fitting routine of the spectrometers. In each experiment, the
waiting period was 5 times larger than the expected relaxation time,
and 16-32 variable delays were employed. The duration of the pulses
was controlled at every temperature. Low-temperature experiments were
carried out in the 180-260 K temperature range using a TV-3000
Bruker temperature unit. The accuracy and stability of temperature were
(1 K. All mixings between the alcohols and the hydride complexes
were performed at low temperature.
1
provides useful new information. The H (hydride resonance)
and 31P NMR chemical shifts for the three species involved are
summarized in Table 1.8,9 The proton transfer is immediate and
quantitative at 200 K in CD2Cl2, as is shown by the NMR
monitoring, yielding the dihydrogen complex selectively. Sub-
sequent slow warming confirmed the previously reported
(31) Binstead, R. A.; Jung, B.; Zuberbu¨hler, A. D. Specfit/32; 2000 Spectrum
Software Associates: Chapel Hill, NC, 2000.
(32) Jetz, W.; Graham, W. A. G. Inorg. Chem. 1971, 10, 1159-1165.
(33) Paciello, R. A.; Manriquez, J. M.; Bercaw, J. E. Organometallics 1990, 9,
260-165.
Stopped-Flow Investigations. The stopped-flow kinetic runs were
carried out at 25 °C with a Hitech SF-61-DX2 apparatus coupled to a
Hitech diode-array UV-visible spectrophotometer. Given the extreme
air-sensitivity of the hydride compound, unacceptable results were
(34) Hamon, P.; Hamon, J.-R.; Lapinte, C. J. Chem. Soc., Chem. Commun. 1992,
1602-1603.
(35) Jime´nez-Tenorio, M.; Puerta, M. C.; Valerga, P. Organometallics 1994,
13, 3330-3337.
(36) Tilset, M.; Fjeldahl, I.; Hamon, J. R.; Hamon, P.; Toupet, L.; Saillard, J.
Y.; Costuas, K.; Haynes, A. J. Am. Chem. Soc. 2001, 123, 9984-10000.
(37) Jia, G.; Ng, W. S.; Yao, J.; Lau, C.-P.; Chen, Y. Organometallics 1996,
15, 5039-5045.
(30) Roger, C.; Hamon, P.; Toupet, L.; Rabaaˆ, H.; Saillard, J.-Y.; Hamon, J.-
R.; Lapinte, C. Organometallics 1991, 10, 1045-1054.
9
11108 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Mechanism of the Proton Transfer to Cp*Fe(dppe)H
A R T I C L E S
Table 2. Longitudinal Relaxation Times at 400 MHz in CD2Cl2 for
Cp*Fe(dppe)H and for the Products Deriving from Its Interaction
with TFE (TFE/Fe ) 3)
T /ms
1
T/K
FeH
FeH + TFE
Fe(H )+
2
190
200
220
225
230
235
240
245
250
255
260
25
12
3.9
24
15
11
349.7
328.2
311.4
280
241.4
223.4
200
9.6
0.67
10.2
182.8
173
conversion to the classical dihydride complex at temperatures
above 250 K. The classical dihydride product is stable in
CD2Cl2 at room temperature for a few minutes. Longer mon-
itoring (few hours) revealed, however, a previously undetected
further evolution with formation of a paramagnetic product,
which is shown to correspond to the previously described
complex [Cp*Fe(dppe)Cl]+.30 No equivalent decomposition
occurs in THF. This decomposition did not prevent us from
studying the details of the proton-transfer process in CH2Cl2,
which occurs on a much faster time scale.
Figure 1. Spectral changes (IR, above; UV-visible, below) observed for
the protonation of Cp*Fe(dppe)H by HBF4 in CH2Cl2. (a) Before addition
of HBF4 (T ) 200 K); (b) after addition of HBF4 (T ) 200 K); (c) T ) 290
K (after a few minutes); (d) T ) 290 K, after several hours. [Fe] ) 0.03 M
(IR); 0.01 M (UV-visible); HBF4/Fe ) 1.
The relaxation times as a function of temperature are shown
in Table 2. It can be noted that complex Cp*Fe(dppe)H does
not show a T1 minimum. Rather, T1 keeps decreasing as the
temperature increases. Fitting the T1 data for the dihydrogen
complex [Cp*Fe(dppe)(H2)]+, on the other hand, gives a T1min
time of 9.7 ms at 230 K, in good agreement with the minimum
of 7 ms obtained by Hamon et al. at 223 K and 300 MHz.9
This behavior is not expected because the two complexes have
practically the same inertia moment and should therefore exhibit
a T1min at approximately the same temperature when working
at the same field strength. This discrepancy can only be
rationalized by the presence of small amounts of the paramag-
netic complex [Cp*Fe(dppe)H]+ in the neutral precursor. This
complex has been shown to be stable, even at room tempera-
ture,8 which is a relatively rare occurrence for paramagnetic
hydrides.38 The two complexes must be involved in a degenerate
exchange with a rate that increases with temperature, and
additional observations made in the presence of fluorinated
alcohols (vide infra) confirm this view. It is clear that the
concentration of the cationic paramagnetic complex is very small
and therefore does not affect the resonance line width, whereas
it affects the relaxation times. Its presence does not affect, on
the other hand, the T1 time of the dihydrogen complex, [Cp*Fe-
(dppe)(H2)]+, because a hypothetical exchange between these
two complexes would be much slower than the degenerate
exchange with Cp*Fe(dppe)H.
features in the 2000-1750 cm-1 region which correspond to
overtones of aryl group vibrations, see spectrum a. The HBF4
addition at 200 K causes the essentially complete disappearance
of this band, see spectrum b. Unfortunately, the IR spectra
(measured in 3000-1500 cm-1 region) do not reveal any
absorbance for ν(H-H) and ν(MH2) vibration modes. These
bands are known to be very weak and often hidden under strong
CH vibrations or overtones.39 This spectrum remains essentially
unchanged upon raising the temperature, until it reaches ca. 250
K. At higher temperatures, conversion to the classical dihydride
complex occurs, yielding eventually spectrum c. A weak band
at 1940 cm-1 can clearly be attributed to the cationic dihydride
complex, while the higher frequency shoulders of this band are
probably overtones of the aryl group vibrations. Such a
significant high-frequency shift (∆νMH ) +100 cm-1) of the
ν(MH) band is typical for transition metal protonation yielding
a cationic classical product.20,40,41
Yet another experiment was carried out with UV-visible
monitoring. The resulting spectral changes are shown in the
lower part of Figure 1. The starting hydride complex has a
relatively strong and broad metal-ligand charge-transfer band42
with a maximum around 388 nm (ꢀ ) 2370 L mol-1 cm-1),
spectrum a. The dihydrogen complex, spectrum b, has a much
weaker and featureless absorption (ꢀ ) 640 L mol-1 cm-1 at
388 nm). Evolution to the classical dihydride complex is
accompanied by a minor change of the spectrum, which is more
In a subsequent experiment carried out in CH2Cl2, IR
monitoring in the Fe-H stretching region provided the spectral
changes shown in the upper part of Figure 1. The starting
hydride complex is characterized by a relatively strong ν(MH)
band at 1844 cm-1 (ꢀ ) 8.81 and 6.32 L mol-1 cm-1 at 200
and 290 K, respectively) in CH2Cl2 solution, which is displaced
relative to the solid state (1869 cm-1),8 plus a few weaker
(39) 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.
(40) Girling, R. B.; Grebenik, P.; Perutz, R. N. Inorg. Chem. 1986, 25, 31-36.
(41) Abdur-Rashid, K.; Fong, T. P.; Greaves, B.; Gusev, D. G.; Hinman, J. G.;
Landau, S. E.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2000, 122,
9155-9171.
(38) Poli, R. In Recent AdVances in Hydride Chemistry; Poli, R., Peruzzini,
M., Eds.; Elsevier Science: Amsterdam, 2001; pp 139-188.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 11109

A R T I C L E S
Belkova et al.
Figure 3. Temperature dependence of ln K for the interaction between
Cp*Fe(dppe)H and TFE in CH2Cl2. [Fe] ) 0.13 M; [TFE] ) 0.01 M.
Figure 2. IR spectra in the ν(RFO-H) stretching region showing the
interaction between Cp*Fe(dppe)H and MFE or TFE in CH2Cl2. Solid
line: [Fe] ) 0.15 M; [MFE] ) 0.015 M. Dashed line: [Fe] ) 0.13 M;
[TFE] ) 0.01 M.
enthalpy values obtained by the two different methods agree
rather well with each other, confirming once again the ap-
plicability of the correlation given by eq 1. The equilibrium
constant for the hydrogen-bond formation is K ) 10 (298 K)
and 870 (200 K). It can reasonably be predicted that the more
acidic alcohols HFIP and PFTB provide higher equilibrium
constants for hydrogen-bond formation, leading to the prediction
that most of the hydride complex is in the hydrogen-bonded
form in the presence of excess alcohol, especially at low
temperatures.
Table 3. Parameters of the Dihydrogen-Bonding Interaction
between Cp*Fe(dppe)H and MFE or TFE in CH2Cl2
νOH(free)
cm-1
νOH(bonded)
cm-1
∆ν
cm-1
∆H°a
kcal mol-1
ROH
P
i
Eb
j
MFE
TFE
0.74
0.89
3608
3594
3362
3240
246
354
-4.6
-5.9
1.35
1.38
a ∆H° parameter calculated by eq 1, mean error (0.4 kcal mol-1 b Ej
.
parameter calculated by eq 2, ∆H11° ) -4.6 kcal mol-1 for CH2Cl2.16
18∆ν
-∆H° )
(1)
∆ν + 720
accurately determined by the analysis of the stopped-flow kinetic
data (vide infra). The dihydride spectrum is given by trace c (ꢀ
) 1093 L mol-1 cm-1 at 388 nm). However, the UV-visible
monitoring shows that this product is unstable in dichlo-
romethane at room temperature and decomposes over several
hours to afford trace d, which is characterized by a weak d-d
transition with maximum at 466 nm (ꢀ ) 587 L mol-1 cm-1).42
No analogous evolution occurs in THF. This behavior is fully
consistent with the results of the NMR investigation described
above. The spectral characteristics of the nonclassical and
classical hydrides and for the final [Cp*Fe(dppe)Cl]+ decom-
position product now allow us to study the mechanism of the
protonation process in more detail.
(b) Interaction with MFE and TFE. Determination of
Hydrogen-Bonding Enthalpy and Basicity Factor Ej of the
Proton Accepting Site. The interaction of MFE and TFE (on
a short time scale, vide infra) with excess Cp*Fe(dppe)H leads
to hydrogen-bond formation without any complication from
proton-transfer processes.43 This interaction could conveniently
be investigated according to well-established protocols.11 The
spectral features in the ν(RFO-H) stretching region (Figure 2)
show the expected stronger interaction with the more fluorinated
alcohol, as indicated by the greater ∆ν (νOH(free)-νOH(bonded)).
The resulting interaction enthalpies were obtained by Iogansen’s
empirical correlation (eq 1)44 and are reported in Table 3. They
show that the hydrogen bonds between complex Cp*Fe(dppe)H
and MFE or TFE have medium strength (4-6 kcal/mol). The
enthalpy for the interaction with TFE was also determined by
van’t Hoff’s method, which required investigations at variable
temperature and furnishes also the interaction entropy. The
equilibrium shifts toward the hydrogen-bonded adduct upon
cooling from 290 to 250 K, see Figure 3, yielding ∆H° ) -5.4
( 0.3 kcal mol-1 and ∆S° ) -13.6 ( 0.9 cal mol-1 K-1. The
Use of the empirical relationship in eq 245 led to the
calculation of the Ej basicity factors, see Table 3. As expected,
the values obtained for Ej by use of the two different alcohols
are essentially identical. This places the Cp*Fe(dppe)H complex
among the most basic hydrides so far investigated, for ex-
ample, [ReH3(η4-N(CH2CH2PPh2)3)] (Ej ) 1.45)21 and {(MeC-
(CH2PPh2)3}Ru(CO)H2 (Ej ) 1.39).46 Note, however, that
hydrogen bonding is established with the metal atom in the first
case and with the hydride ligand in the second case. Therefore,
our next step was the establishment of the hydrogen-bonding
site for the title Fe hydride complex.
∆H°
∆H11°Pi
Ej )
(2)
(c) Establishment of the Hydrogen-Bonding Site. It is now
well established that hydrogen bonding to a hydride ligand site,
also termed “nonclassical or dihydrogen bonding”, causes a low-
frequency shift of the M-H band,11 whereas the involvement
of the metal lone pair(s) as proton acceptor(s) for hydrogen
bonds shifts other M-X stretching vibrations (notably the M-H
band) to a higher frequency range.12,20,21 Therefore, we have
studied the interaction between the hydride complex and
different proton donors by IR spectroscopy in the ν(MH) region.
The alcohol addition was first carried out and studied at low
temperatures. All alcohols induce the development of a low-
frequency shift of the ν(MH) band (1836-1828 cm-1), which
indicates the formation of the dihydrogen-bonded complex
Cp*(dppe)FeH‚‚‚HORF. The extrapolation of the variable-
temperature data for hydrogen bonding with TFE (Figure 3)
gives a formation constant of 870 at 200 K. Thus, essentially
all of the hydride complex is in the hydrogen-bonded state under
these conditions. Consequently, the ν(MH) band at 1830 cm-1
(42) Lever, A. B. P. Inorganic Electronic Spectroscopy, 2nd ed.; Elsevier:
Amsterdam, 1984.
(43) An NMR investigation also confirmed the absence of any proton transfer
at room temperature in the presence of 12 equiv of MFE.
(44) Iogansen, A. V. Hydrogen Bond; Nauka: Moscow, 1981; p 134.
(45) Iogansen, A. V. Theor. Exp. Khim. 1971, 7, 312-317.
(46) Bakhmutov, V. I.; Bakhmutova, E. V.; Belkova, N. V.; Bianchini, C.;
Epstein, L. M.; Masi, D.; Peruzzini, M.; Shubina, E. S.; Vorontsov, E. V.;
Zanobini, F. Can. J. Chem. 2001, 79, 479-489.
9
11110 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Mechanism of the Proton Transfer to Cp*Fe(dppe)H
A R T I C L E S
Table 4. Band Widths at Half Height for [Cp*Fe(dppe)(H2)]+ and
Hydrogen-Bonded Cp*Fe(dppe)H Adducts in CD2Cl2 as a Function
of Temperaturea
w1/2/Hz
T/K
FeH‚‚‚TFE
FeH‚‚‚HFIP
FeH‚‚‚PFTB
180
200
220
230
240
250
260
65
142
230
320
480
triplet
170
230
290
440
190
270
350
510
760
Figure 4. IR study of the interaction between Cp*Fe(dppe)H (0.038 M)
and TFE in CH2Cl2 at 200 K. (a) Without TFE. (b) With TFE (3 equiv).
1100
1280
in Figure 4 (trace b) can be unambiguously attributed to the
Fe-H stretching vibration in the dihydrogen-bonded complex.
The absence of a high-frequency shoulder shows that no
significant amount of hydrogen bond at the metal center, that
is, Cp*(dppe)HFe‚‚‚HORF, is present, see Figure 4.47
a CFeH° ) 0.05-0.07 M. Identical values were observed for different
alcohol/Fe ratios.
be assigned to a hypothetical Cp*Fe(dppe)(ORF) product, which
may result from dihydrogen evolution. However, a longer
NMR monitoring revealed the slow formation (few hours at
room temperature) of the same paramagnetic product, [Cp*Fe-
(dppe)Cl]+, which is observed by using HBF4.
The initial proton transfer to yield the intermediate dihydrogen
complex is more shifted toward the right, at constant alcohol/
Fe ratio, as the alcohol acidity increases (TFE < HFIP < PFTB)
and, for the same alcohol, as the alcohol/Fe ratio increases. For
instance, the FeH/Fe(H2)+ ratio is approximately 45:55 when
using PFTB/Fe ) 1, 75:25 when using HFIP/Fe ) 3, and
40:60 at TFE/Fe ) 10 (220 K).
In conclusion, the IR study shows the establishment of
hydrogen bonding involving the hydride ligand as the first step
of the proton transfer leading to the dihydrogen complex. In
comparison with previous literature data,12 the rather high Ej
value suggests that this hydride can be protonated by fluorinated
alcohols such as HFIP and PFTB in CH2Cl2. Indeed, the
ν(MH) band intensities observed in the presence of these alco-
hols dramatically decrease with a temperature decrease, espe-
cially when PFTB is used. This contrasts with the intensity in-
crease in the absence of alcohol, see section a. This change is
reversible in the low-temperature range and signals a reversible
proton-transfer process leading to the dihydrogen complex.
(d) Interaction with PFTB, HFIP, and TFE: NMR
The starting hydride resonance (mixture of starting hydride
in fast equilibrium with the hydrogen-bonded adducts) broadens
significantly as the temperature increases (see Table 4), whereas
the resonance of the dihydrogen complex maintains approxi-
mately the same width (ca. 25 Hz) independent of the temper-
ature, nature of the alcohol, and alcohol/Fe ratio. This phenom-
enon cannot result from a rapid exchange between the free and
hydrogen-bonded hydride complexes, nor from an exchange
between the hydrogen-bonded complex and the proton-transfer
product. In the first case, the peak should sharpen upon
increasing the temperature because coalescence is already
achieved, whereas in the second case the broadening phenom-
enon should be equally observable on both resonances, contrary
to the observation. The most likely explanation for the observed
phenomenon is the self-exchange process with trace amounts
of the one-electron oxidation product, the paramagnetic [Cp*Fe-
(dppe)H]+ ion, whose presence is also responsible for the
unusual temperature dependence of the T1 time for the starting
hydride complex (vide supra). Further indication that this is the
correct explanation is the measurement of unrealistically small
relaxation times (T1) for the broad hydride resonance in the
presence of TFE (see Table 2). The longitudinal relaxation time
has been frequently used as a criterion for establishing the
presence of dihydrogen-bonding interactions. The T1 value
slightly decreases relative to that of the starting hydride complex,
because of the additional dipolar relaxation induced by the
hydrogen-bonded proton. The presence of dihydrogen bonding
for the Cp*Fe(dppe)H complex, however, could not possibly
give such low values for T1 (e.g., 0.67 ms at 240 K, see Table
2), even lower than for the dihydrogen complex product where
the H-H distance is certainly shorter. This value can only result
from the fast relaxation induced by the unpaired electron in the
oxidized 17-electron complex, with which the starting hydride
complex is in fast degenerate exchange. Shaking the NMR
solutions with Zn/Cu alloy inside the NMR tube did not,
1
Investigations. H NMR investigations of CD2Cl2 solutions
containing Cp*Fe(dppe)H and the three fluorinated alcohols
provided only qualitative information on the nature of the species
involved and on the chemical equilibria. In all cases, the alcohol
addition to the hydride solution at low temperature yielded two
hydride resonances, a broadened triplet at δ ca. -17.3 and a
broad resonance at δ ca. -12.5. These chemical shifts are
essentially identical to those of free Cp*Fe(dppe)H and
Cp*Fe(dppe)(H2)+BF4-, respectively, and are not significantly
affected by the temperature, alcohol nature, and alcohol/Fe ratio.
The dihydrogen complex, Cp*Fe(dppe)(H2)+, is probably
present in solution as a free ion, III, rather than hydrogen bonded
to the alkoxide ion, II (Scheme 1), as will be further argued
later. The -17.3 resonance is attributed to the rapidly exchang-
ing equilibrium mixture of Cp*Fe(dppe)H and Cp*(dppe)-
FeH‚‚‚PFTB (whose presence is proven by the IR study, vide
supra). The 31P{1H} spectrum correspondingly gives two signals
at δ ca. 93.8 and at δ ca. 108. These are also essentially
unshifted with respect to the resonances of Cp*Fe(dppe)-
(H2)+BF4 and free Cp*Fe(dppe)H, respectively.
-
The nature of the alcohol and the alcohol/Fe ratio strongly
affect the proton-transfer equilibrium position with the nonclas-
sical intermediate and the temperature at which further evolution
to the classical product starts to occur (ca. 250 K for PFTB,
270 K for HFIP, and room temperature for TFE). There was
no evidence for the formation of products other than the classical
dihydride complex on a short time scale (few minutes) at room
temperature. In particular, there were no resonances that could
(47) The occasional observation of a high-frequency shoulder is attributed to
variable amounts of the 17-electron complex [Cp*Fe(dppe)H]+,8 resulting
from adventitious oxidation of the extremely air-sensitive hydride complex.
Indeed, a synthesized sample of this complex yields an Fe-H stretching
vibration at 1868 cm-1 in CH2Cl2.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 11111

A R T I C L E S
Belkova et al.
unfortunately, eliminate this problem. This occurrence unfor-
tunately renders a more thorough NMR characterization of
hydrogen bonding for the Cp*Fe(dppe)H hydride complex
impossible.
The reason the starting hydride band is broader and T1 is
shorter in the presence of alcohol may be the result of the partial
conversion to the protonation product, increasing the [Cp*Fe-
(dppe)H]+/Cp*Fe(dppe)H ratio. Additional oxidation upon
introduction of the alcohol cannot be excluded. In any event,
the extent of this oxidation process remains small. If a significant
amount of hydride complex had been oxidized, the resonance
would not only broaden but also shift to an extent that would
depend on temperature. In fact, this is not the case, the chemical
shift being essentially temperature independent as in the alcohol-
free solution.
Figure 5. UV-visible study of the interaction between Cp*Fe(dppe)H and
HFIP (5 equiv) in CH2Cl2. [Fe] ) 0.02 M. (a) Before the HFIP addition.
(b) T ) 270 K. (c) T ) 200 K. The other intermediate spectra correspond
to 265-225 K (5 K steps).
(e) Interaction with PFTB, HFIP, and TFE: IR Investiga-
tions. Low-temperature IR spectroscopic studies also provide
qualitative indications on the interaction between Cp*Fe(dppe)H
and fluorinated alcohols, leading to proton transfer. The essential
findings from the NMR study were confirmed by the IR
investigation: (i) the proton-transfer equilibrium between the
hydrogen-bonded system and the dihydrogen complex is shifted
toward the protonation product to a greater extent for stronger
alcohols and for higher alcohol/Fe ratios; (ii) conversion of the
overall mixture to the final classical dihydride product started
to occur at a higher temperature for the weaker alcohols.
Figure 6. Temperature dependence of ln K for the interaction between
Cp*Fe(dppe)H and HFIP in CH2Cl2.
followed by transformation to the dihydride isomer at temper-
atures greater than 250 K and finally to [Cp*Fe(dppe)Cl]+ over
several hours at room temperature. With HFIP, the proton-
transfer equilibrium could be investigated in a wider temperature
range and for several HFIP/Fe ratios. In the 200-270 K
temperature range, the UV-visible spectral changes are per-
fectly reversible and indicate, in agreement with the IR evidence,
a temperature dependence of this equilibrium, see Figure 5. The
UV-visible properties of the dihydrogen-bonded complexes are
essentially indistinguishable from those of the free hydride
complex, as confirmed by the fast kinetics studies, vide infra.
This is reasonable, because the hydrogen-bonding interactions
should perturb the electronic structure of the metal center only
in a minimal way. The least absorbing species in this equilibrium
solution, however, is the dihydrogen complex, see Figure 1.
Therefore, the overall changes shown by Figure 5 are consistent
with the presence of a greater relative amount of the dihydrogen
complex at lower temperatures.
With PFTB, the complete disappearance of the starting ma-
terial occurs when using a 3-fold excess at 200 K in CH2Cl2,
whereas an equilibrium situation is obtained when the alcohol
is used stoichiometrically. As in the NMR experiment, the
isomerization started to occur upon raising the temperature
above 250 K. For the HFIP experiment, larger excess amounts
of alcohol are required to consume a substantial amount of the
hydride precursor at low temperature. For TFE, only ca. 50%
of dihydrogen complex was formed when using a 15-fold excess
of TFE at 200 K (cf., 98% for HFIP under otherwise identical
conditions). In addition, the proton transfer is much slower (>30
min for TFE vs <5 min for HFIP at 200 K). No quantitative
kinetics information, however, can be obtained using this
technique. In all three cases, the reversibility of the IR spectral
changes with temperature (below the temperature at which
isomerization to the final classical dihydride product starts to
occur) qualitatively indicates that the proton-transfer equilibrium
shifts toward the dihydrogen complex when temperature de-
creases, but no quantitative thermodynamic information can be
derived.
The reversibility of this process enabled us to obtain the
equilibrium constant for eq 3 assuming that all hydride is in
the dihydrogen-bonded form at 200-270 K (vide supra, section
b) and that the equilibrium involves a second alcohol molecule
(vide infra, section g). The van’t Hoff plot (Figure 6) gives the
enthalpy (∆H° ) -6.6 ( 0.9 kcal mol-1) and the entropy (∆S°
) -17.2 ( 1.4 eu) of the process shown in eq 3.
(f) Interaction with PFTB, HFIP, and TFE: Preliminary
UV-Visible Investigations. Finally, the Cp*Fe(dppe)H-HA
interaction was investigated by UV-visible spectroscopy in
CH2Cl2, because this technique is more adept to yielding
quantitative results. The study confirmed the two-step pathway
already established by the IR and NMR techniques, followed
by the slow decomposition to the paramagnetic [Cp*Fe(dppe)-
Cl]+. The evolution of the UV-visible spectrum is qualitatively
identical with respect to the corresponding protonation by HBF4
(see Figure 1, bottom part). This evolution appeared best suited
to detailed kinetic investigations of the proton-transfer process,
which will be described below in section g.
Cp*Fe(dppe)H‚‚‚HA + HA a
[Cp*Fe(dppe)(H2)]+[AHA]- (3)
For TFE, the slow equilibration rates made it impractical to
determine the accurate equilibrium position at low temperatures.
In addition, the smaller proton-transfer equilibrium constant
required the use of greater alcohol concentrations, with conse-
quent crystallization of the excess alcohol at low temperatures
in dichloromethane.
The interaction of Cp*Fe(dppe)H with PFTB (1:5 ratio) at
200 K immediately gives the dihydrogen complex quantitatively,
9
11112 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Mechanism of the Proton Transfer to Cp*Fe(dppe)H
A R T I C L E S
Scheme 2
Scheme 3
For the PFTB process, the first measurable step was complete
within a fraction of a second and could be accurately measured
only by stopped-flow kinetics. The second, slower step was also
accessible from the stopped-flow data. The two processes occur
on quite different time scales, and the first process is essentially
quantitative even for the lowest [HA]/[FeH] ratio, rendering the
kinetic analysis straightforward. These two processes correspond
to the proton-transfer equilibrium yielding the dihydrogen
intermediate and to the isomerization process, respectively. This
is confirmed by comparison of the extrapolated spectra of
intermediate and product (as obtained from a global Specfit
analysis) with those obtained from the HBF4 protonation (see
Figure 1). The first-order dependence of the proton-transfer rate
constant (k3obs) on the alcohol concentration is shown in Figure
7a. The intercept is zero within experimental error (0.1 ( 0.7
s-1), confirming the assumption of irreversibility for this step.
The subsequent rate constants, k4obs and k5obs, on the other hand,
do not show any dependence on the alcohol concentration. All
rate constants are collected in Table 5.
(g) Kinetics Investigations of the Proton Transfer from
TFE, HFIP, PFTB, and TFA. The collective spectroscopic
studies of the hydride-HA interactions described in the previous
section suggest Scheme 2, in line with previous knowledge in
this field as outlined in the Introduction. In principle, each
different hydrogen-bonded complex can be seen as an inter-
mediate leading to a different proton-transfer product. The
dihydrogen bond (K1) would lead to the nonclassical product
(K3 ) k3/k-3), while the hydrogen bond to the metal center (K2)
would lead to the classical product (k4′). However, the classical
dihydride complex may also be obtained directly from the
nonclassical tautomer (k4). The second hypothesis appears
consistent with the observation by Hamon et al. that the
isomerization also takes place for the isolated dihydrogen
complex in the solid state.9 The results of a more detailed kinetic
investigation by UV-visible monitoring allow us to throw more
light on this dichotomy.
The reactions with the three fluorinated alcohols TFE, HFIP,
and PFTB were carried out in CH2Cl2 at 298 K under pseudo-
first-order conditions with HA/Fe ratios in the 30-200 range.
The entire kinetics investigation required the use of both
stopped-flow and classical mixing and monitoring methods. In
all cases, the spectrum recorded immediately after mixing (ca.
1 ms) is indistinguishable from that of the pure hydride in the
absence of alcohol,48 proving that the hydrogen bonding does
not significantly perturb the electronic structure of the metal
center. The establishment of hydrogen bonds is a diffusion-
controlled step.49 The investigations yielded the rate constants
for three separate processes, as indicated in Scheme 3 [Fe )
Cp*Fe(dppe)]. Although the initial equilibrium may include,
in principle, the hydrogen-bonded adducts involving the hydride
site and those at the metal site, [FeH‚‚‚HA] and [HFe‚‚‚HA],
the latter species was not found in the present case.
For the HFIP run, the first step required a few seconds and
partially overlapped with the second kinetics. In addition, the
above determined thermodynamic data for this proton-transfer
step (extrapolated equilibrium constant of 7.97 at 298 K) suggest
that this may be equilibrated, even for the relatively high HFIP/
Fe ratios employed in this kinetic study. However, a good fit
was possible for the model A f B f C, yielding individual
observed rate constants for each step. The first step (proton
transfer) has a first-order dependence on HFIP, and the intercept
is close to the experimental error (k-3obs ) 0.21 ( 0.13 s-1),
see Figure 7b. Thus, this step is almost irreversible under the
experimental conditions used for this kinetic experiment. Even
though it is affected by a large uncertainty, the estimated k3obs
/
k-3obs ratio is consistent with the estimated proton-transfer
equilibrium constant from the van’t Hoff analysis (vide supra).
The values of k4obs and k5obs are independent from the alcohol
concentration. Furthermore, and most notably, they are also
independent of the alcohol nature, the values obtained for PFTB
and HFIP being very close to each other, see Table 5.
Contrary to the PFTB and HFIP reactions, the proton transfer
from TFE led to a single measurable kinetics for the transforma-
tion of FeH to the classical dihydride product. The reason for
this result is that the second step is faster than the first one.
This was too slow for the stopped-flow time scale, but it could
be accessed by regular mixing and monitoring in an airtight
cuvette. A complication in this case was the contamination by
the final transformation to the paramagnetic chloro complex.
However, the SPECFIT global analysis afforded reasonably
(48) Comparative shootings of the hydride solution against the alcohol and
against neat CH2Cl2 show an identical first spectrum (ca. 1 ms).
(49) Scheiner, S. Hydrogen bonding: A Theoretical PerspectiVe; Oxford
University Press: New York, 1997.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 11113

A R T I C L E S
Belkova et al.
knowledge that they are diffusion-controlled, barrierless reac-
tions.49 The subsequent step is proton transfer to the hydride
site, whose rate and equilibrium constant increase in the order
TFE < HFIP < PFTB < TFA, as the acidity of the proton donor
increases. This rate has a first-order dependence on the alcohol
concentration (established experimentally for the three alcohols).
If one neglects the establishment of the hydrogen bonds, the
first-order dependence on the alcohol concentration would point
to a rather simple bimolecular elementary process, whereby one
alcohol molecule transfers its proton to the hydride complex in
a direct, single-step process. This would also explain the rate
constant dependence on the proton donor acidity. However, the
measured thermodynamic parameters for hydrogen bonding with
TFE (Figure 3) allow us to estimate a 6:1 FeH‚‚‚HX/FeH ratio
under the conditions used for the stopped-flow kinetics study.
When HFIP and PFTB are used, the constant for dihydrogen-
bond formation will be even higher. Therefore, the [FeH‚‚‚HA]
complex predominates in the mixture which transforms further
into nonclassical protonation product.
The starting point for the first kinetics process is not the pure
hydride precursor but rather the equilibrium mixture with the
hydrogen-bonded adducts, Scheme 3. A straightforward ma-
nipulation of the rate expression leads to eq 4, which can be
simplified as shown at high alcohol concentrations (quantitative
formation of the hydrogen-bonded adducts). Therefore, the
observed first-order dependence on HA implies the intervention
of a second molecule in the rate-determining step, as indicated
in Scheme 4. If a second alcohol molecule were not involved
in the proton-transfer step, the k3obs expression would lead to
an essentially alcohol-independent observed rate constant. This
observation suggests that a direct proton transfer leading to a
salt which contains a free alkoxide anion does not occur. The
involvement of a second alcohol molecule presumably renders
the proton-transfer process thermodynamically more favorable
and faster via the establishment of a hydrogen-bonding interac-
tion and the formation of an alcohol/alkoxide homoconjugate
pair (see Scheme 4). This homoconjugate pair formation has
been noted in other proton-transfer processes to hydride
complexes3,50 and is comforted by theoretical calculations.51
Because of this homoconjugate pair formation, we consider it
unlikely that the alkoxide ion also remains hydrogen bonded to
the dihydrogen ligand in the dichloromethane solvent used. The
product, therefore, is probably present in solution in the form
of free ions.
Figure 7. Pseudo-first-order rate constants for the first step of the
proton transfer from HA to Cp*Fe(dppe)H [k3obs; HA ) PFTB (a), HFIP
(b), TFE (c)].
Table 5. Observed Rate Constants for the Reaction between
Cp*Fe(dppe)H and HAa
HA
k
3obs[HA]-1/s-1 M-1
k
4obs/s-1
k
5obs/s-1
TFA
b
(6.4 ( 0.2) × 10-2
PFTB 156 ( 5
HFIP
TFE
(7.6 ( 0.2) × 10-2 (2.3 ( 0.3) × 10-4
(6.8 ( 0.2) × 10-2 (2.1 ( 0.3) × 10-4
5.4 ( 0.9
(1.5 ( 0.3) × 10-3
c
(2.2 ( 0.6) × 10-4
a See Scheme 3. b Too fast to be measured by stopped-flow methods.
c Not measurable because faster than k3obs
.
precise values for both k3obs and k5obs. The value of k3obs was
linearly dependent on the alcohol concentration, with a zero
intercept within experimental error as for the previous cases,
(1.6 ( 3) × 10-4 s-1. The value of k5obs does not show any
significant dependence on the alcohol concentration and is very
close to those observed with the other alcohols (see Table 5).
To further widen the acidity range of the proton donor, we
also carried out a kinetic study of the protonation with trifluoro-
acetic acid (TFA). With this acid, the first step was too fast to
measure by the stopped-flow technique, the first recorded spec-
trum after ca. 1 ms corresponding already to that of the inter-
mediate dihydrogen complex (i.e., spectrum b in Figure 1).
Further transformation to the classical tautomer (spectrum c in
Figure 1) followed clean first-order kinetics, the k4obs being once
again independent of the acid concentration and close in value
to those observed for the fluorinated alcohols (see Table 5).
In summary, proton donors instantaneously establish hydrogen
bonding at the hydride site. This is in agreement with the general
k3K1[HA]2
k3K1
k3obs
)
≈
[HA]
K1 + K2
1 + (K1 + K2)[HA]
The question remains as to the intimate mechanism of the
tautomerization process, as indicated by the two possible
pathways a and b in Scheme 2. Both pathways would lead to a
zero-order dependence on the proton donor concentration (see
Supporting Information), as was experimentally observed.
Therefore, they cannot be distinguished on this basis. However,
the independence of the rate constant on the nature of the proton
donor strongly points toward an internal reorganization mech-
(50) Belkova, N. V.; Ionidis, A. V.; Epstein, L. M.; Shubina, E. S.; Gruende-
mann, S.; Golubev, N. S.; Limbach, H. H. Eur. J. Inorg. Chem. 2001,
1753-1761.
(51) Belkova, N. V.; Besora, M.; Epstein, L. M.; Lledo´s, A.; Maseras, F.;
Shubina, E. S. J. Am. Chem. Soc. 2003, 125, 7715-7725.
9
11114 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Mechanism of the Proton Transfer to Cp*Fe(dppe)H
A R T I C L E S
Scheme 4
anism (pathway b in Scheme 2). The alternative possibility of
reversible deprotonation followed by protonation at the metal
center would certainly lead to a kinetics dependent on the nature
of the acid.
studies, therefore, do not necessarily prove that a competitive
metal protonation would not occur at ambient temperature.
Conclusion
It is interesting to compare our results with those reported
for the related ruthenium systems (Ring)RuH(L)(L′) (Ring )
Cp, Cp*; L,L′ ) tertiary phosphine ligands).6,29 For these, the
occurrence of an intramolecular rearrangement mechanism
leading from the dihydrogen complex to the dihydride isomer
was proposed, in part, on the basis of the observation that the
rate constant is invariant with the acid concentration. However,
we have shown that this rate should not depend on the acid
concentration even for a deprotonation/reprotonation mechanism
(see Supporting Information). The observation by Chinn and
Heinkey6 that the isomerization is faster than H/D scrambling
in the presence of external acids is much stronger evidence in
favor of an intramolecular mechanism. Our observations, based
on the comparison of isomerization rates in the presence of acids
of different strength, are complementary and lead to the same
conclusion for the related Fe system. In addition, however, they
show the absence of a kinetically viable direct protonation of
the metal site under the same experimental conditions in which
the isomerization process occurs. For the previous studies carried
out by Heinekey and Puerta on the hydridoruthenium complexes,
as well as for the previous study by Hamon et al. on Cp*FeH-
(dppe), the proton-transfer process is carried out at very low
temperature, where the direct protonation of the metal site may
be disfavored by large negative activation entropies. Those
The direct and independent observation, at the same temper-
ature, of the kinetics for the faster protonation of the hydride
ligand site in complex Cp*FeH(dppe) and for the isomerization
of the resulting Cp*Fe(H2)(dppe)+ intermediate to the thermo-
dynamically preferred classical dihydride product has allowed
us to obtain strong evidence in favor of a direct isomerization
process, not involving the reversible deprotonation of the hydride
site and protonation of the metal site. The unambiguous presence
of a direct proton-transfer pathway to the metal site, for any
complex which contains hydride ligands, remains to be estab-
lished. Further investigations will be necessary to probe the
generality or limitations of the results reported here.
Acknowledgment. This work was supported by the European
Commission’s RTN Program (HPRN-CT-2002-00176), by
INTAS (00-00179), by RFBR (No. 02-03-32194, 03-03-6283),
and by NATO (PST.CLG.978453).
Supporting Information Available: Derivation of equations
for the observed rate constant of the tautomerization process
under the two different mechanistic hypotheses (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA0358450
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 11115