Understanding the exceptional hydrogen-atom donor characteristics of water in TiIII-mediated free-radical chemistry

Article abstract of DOI:10.1021/ja105670h

In recent years solid evidence of HAT reactions involving water as hydrogen atom source have been presented. In this work we demonstrate that the efficiency of titanocene(III) aqua complexes as an unique class of HAT reagents is based on two key features: (a) excellent binding capabilities of water toward titanocene(III) complexes and (b) a low activation energy for the HAT step. The theory has predictive capabilities fitting well with the experimental results and may aid to find more examples of this remarkable radical reaction.

Full text of DOI:10.1021/ja105670h

Published on Web 08/19/2010
Understanding the Exceptional Hydrogen-Atom Donor
Characteristics of Water in TiIII-Mediated Free-Radical
Chemistry
Miguel Paradas,† Araceli G. Campan˜a,‡ Tania Jime´nez,† Rafael Robles,†
J. Enrique Oltra,† Elena Bun˜uel,‡ Jose´ Justicia,† Diego J. Ca´rdenas,*,‡ and
Juan M. Cuerva*,†
Department of Organic Chemistry, Faculty of Sciences, UniVersity of Granada,
Campus FuentenueVa s/n, E-18071 Granada, Spain, and Department of Organic Chemistry,
Faculty of Sciences, UniVersidad Auto´noma de Madrid, Cantoblanco E-28049, Madrid, Spain
Received June 28, 2010; E-mail: diego.cardenas@uam.es; jmcuerva@ugr.es
Abstract: In recent years solid evidence of HAT reactions involving water as hydrogen atom source have
been presented. In this work we demonstrate that the efficiency of titanocene(III) aqua complexes as an
unique class of HAT reagents is based on two key features: (a) excellent binding capabilities of water
toward titanocene(III) complexes and (b) a low activation energy for the HAT step. The theory has predictive
capabilities fitting well with the experimental results and may aid to find more examples of this remarkable
radical reaction.
( 0.07 kcal mol-1)4 would preclude any potential HAT reaction
to carbon-centered radicals. Despite this general assumption,
Introduction
Reduction of carbon-based radicals is an essential and widely
spread process in organic chemistry (Scheme 1).1 It lies at the
heart of many fundamental reactions in organic synthesis, such
as the Barton-McCombie deoxygenation reaction, reductive
radical cyclizations, low valent metal-mediated carbonyl reduc-
tions, etc. Two main mechanisms have been proposed to carry
out this essential step: (a) a direct hydrogen atom transfer (HAT)
from common donors (1,4-cyclohexadiene (1,4-CHD), thiols,
hypophosphorus acid and its salts, HSnR3 and silicon-based
derivatives) to the carbon-centered radical2 or (b) a stepwise
electron and proton transfer (ET/PT).3 The synthetic methods
based on the first process are limited by the availability of
suitable hydrogen atom donors, which are usually unstable,
toxic, expensive, and/or foul smelling, thus seriously restricting
the application for large scale preparations.
we have recently described that, in fact, water becomes an
excellent hydrogen atom donor in the presence of bis(cyclo-
pentadienyl)titanium(III) chloride (Cp2TiCl)5 toward aliphatic
carbon radicals.6-8 To explain the experimental results, we
proposed that the H-OH bond is weakened upon coordination
to titanocene(III), acting the corresponding aqua-complex 1 as
an efficient hydrogen atom donor (Scheme 2). The reaction
energy for the process was computed at DFT level, showing
(4) Ruscic, B.; Wagner, A. F.; Harding, L. B.; Asher, R. L.; Feller, D.;
Dixon, D. A.; Peterson, K. A.; Song, Y.; Qian, X.; Ng, C.-Y.; Liu, J.;
Chen, W.; Schwenke, D. W. J. Phys. Chem. A 2002, 106, 2727–2747.
(5) The single-electron-transfer reagent bis(cyclopentadienyl)titanium(III)
chloride can be generated in situ by stirring commercial Cp2TiCl2 with
Mn dust in THF, where it exists as a mixture of the mononuclear
Cp2TiCl(THF) and the dinuclear (Cp2TiCl)2 species; see: (a) Rajan-
Babu, T. V.; Nugent, W. A. J. Am. Chem. Soc. 1994, 116, 986–997.
(b) Enemærke, R. J.; Larsen, J.; Skrydstrup, T.; Daasbjerg, K. J. Am.
Chem. Soc. 2004, 126, 7853–7864. (c) Half-open dimers are also active
compound for binding Lewis-bases and can be considered as the real
precursor of monomeric Cp2TiCl(THF) and others Cp2TiCl(L) related
species: Gansa¨uer, A.; Barchuk, A.; Keller, F.; Schmitt, M.; Grimme,
S.; Gerenkamp, M.; Mu¨ck-Lichtenfeld, C.; Daasbjerg, K.; Svith, H.
J. Am. Chem. Soc. 2007, 129, 1359–1371.
In this context, water would be a remarkable, safe, and cheap
HAT reagent. Nevertheless, it is generally believed that the high
bond dissociation energy (BDE) of the H-OH bond (117.59
(6) (a) Cuerva, J. M.; Campan˜a, A. G.; Justicia, J.; Rosales, A.; Oller-
Lo´pez, J. L.; Robles, R.; Ca´rdenas, D. J.; Bun˜uel, E.; Oltra, J. E.
Angew. Chem., Int. Ed. 2006, 45, 5522–5526. (b) Jime´nez, T.;
Campan˜a, A. G.; Bazdi, B.; Paradas, M.; Arra´ez-Roma´n, D.; Segura-
Carretero, A.; Ferna´ndez-Gutie´rrez, A.; Oltra, J. E.; Robles, R.; Justicia,
J.; Cuerva, J. M. Eur. J. Org. Chem. 2010, 4288–4295. (c) This
observation was also extended to the reduction of ketyl radicals:
Paradas, M.; Campan˜a, A. G.; Marcos, M. L.; Justicia, J.; Haidour,
A.; Robles, R.; Ca´rdenas, D. J.; Oltra, J. E.; Cuerva, J. M. Dalton
Trans. 2010, DOI: 10.1039/c001689f.
† University of Granada.
‡ Universidad Auto´noma de Madrid.
(1) (a) Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions; VCH: Weinheim, Germany, 1996; p 4. (b) Curran, D. P.
In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I.,
Semmelhack, M. E., Eds.; Pergamon: Oxford, 1991; Vol. 4, pp 715-
777.
(2) HAT mechanism is a subfamily of the more general proton-coupled
electron transfer (PCET) mechanism: (a) Mayer, J. M. Annu. ReV.
Phys. Chem. 2004, 55, 363–390. (b) Huynh, M. H. V.; Meyer, T. J.
Chem. ReV. 2007, 107, 5004–5064.
(7) For other precedent of water acting as HAT reagent: Spiegel, D. A.;
Wiberg, K. B.; Schacherer, L. N.; Medeiros, M. R.; Wood, J. L. J. Am.
Chem. Soc. 2005, 127, 12513–12515.
(3) Organometallic compounds, derived from a radical heterocoupling
between a carbon radical and metallic species, can be involved in the
initial reduction step, leading the global ET/PT after a subsequent
protonolysis.
(8) Other remarkable, safe and cheap HAT reagent, such as H2, has been
described: Gansa¨uer, A.; Fan, C.-A.; Keller, F. J. Am. Chem. Soc.
2008, 130, 6916–6917.
9
12748 J. AM. CHEM. SOC. 2010, 132, 12748–12756
10.1021/ja105670h  2010 American Chemical Society

Exceptional H-Atom Donor Characteristics of H2O
A R T I C L E S
Scheme 1. Commonly Accepted Mechanisms for Carbon-Centered
allyl or benzyl radicals as well as alkenyl and aryl radicals might
also undergo the same process. Additionally, related reduction
reactions of carbon-centered radicals using methanol activated
by B-alkylcatecholboranes have been also reported.13 Taking
into account our mechanistic hypothesis and this result, other
compounds with hydrogen-heteroatom bonds, able to coordinate
to Cp2TiCl, might also be potential radical reducing reagents.
Bearing this idea in mind, we present in this paper an in depth
study of the exceptional hydrogen atom donor behavior of water
toward different carbon-centered radicals in the presence of
titanocene(III) complexes compared to other potential hydrogen
atom donors containing hydrogen-heteroatom bonds.
Radical Reduction
Scheme 2. Mechanism of Carbon-Centered Radical Reduction by
Aqua-Complex 1 and Calculated Reaction Energy for the
Homolytic O-H Bond Dissociation
Results and Discussion
Based on our mechanistic hypothesis, two key factors seem
to control the success of the reduction step: (i) the coordination
of the potential hydrogen atom donor to the titanocene(III)
complex and (ii) the efficiency of the HAT process. Interestingly,
the influence of both factors is accessible from both experimental
and theoretical point of views.
Study of the Coordination Capabilities of Potential
Hydrogen Atom Donors toward (Cp2TiCl)2. (Cp2TiCl)2 (2) is a
highly air-sensitive complex, which can be prepared in pure
form. The coordination capabilities of different potential
hydrogen atom donors toward this complex were evaluated
experimentally using UV-vis spectroscopy.14 In our present
study, we selected ligands with H-O bonds, such as water,
methanol, and phenol, and with H-N bonds, such as primary
(N-octylamine) and secondary (N,N-dibutylamine) amines. Am-
monia gas was excluded because it is difficult to handle properly
and reproducibly in the glovebox. We initially chose tetrahy-
drofurane (THF) as solvent since it is the usual one in Cp2TiCl
mediated reactions.
Addition of water and primary or secondary amines (octy-
lamine and dibuthylamine, respectively) to 10 mM solutions of
(Cp2TiCl)2 in THF promoted marked changes in the corre-
sponding spectra (Figure 1-3). In the case of water, the
disappearance of the absorption band at 456 nm, characteristic
of complex 2, is observed upon addition of relatively low
amounts of additive (8 equiv of water).15 A similar trend is
observed using both amines (1 equiv).
On the other hand, the addition of methanol or phenol only
promoted slight variations in the UV-vis spectra in the same
conditions (Figures 4a and 5a). In the case of methanol, up to
17 equiv of additive were necessary to promote similar changes
to that observed with water or amines. With phenol, we only
obtained minor changes in the UV-vis spectra. Competitive
experiments also confirmed that water has better coordination
ability toward Ti(III) than methanol and phenol. Addition of 8
equiv of water to 1:8 mixtures of (Cp2TiCl)2/methanol and
(Cp2TiCl)2/phenol yielded the same UV-vis spectrum as
complex 2 in the presence of same amounts of water (Figures
4b and 5b).
that the homolytic O-H bond dissociation is extraordinarily
favored by coordination of H2O to Cp2TiIIICl. Reaction energy
decreases from a calculated value of 108.1 kcal mol-1 for H2O
to 49.4 kcal mol-1 for 1. This fact was strongly supported by
kinetic measurements. In accord with our proposal, Newcomb
et al. have determined the rate constant for the HAT reaction
of Cp2TiCl-H2O to secondary alkyl radicals using indirect kinetic
methods (k25 ) 1.0 × 105 M-1 s-1).9 Although with the actual
experimental data a direct comparison with widely used Bu3SnH
is not possible, it is worth noting that the rate constant is only
1 order of magnitude lower (k ) 1.4 × 106 M-1 s-1).10
This remarkable behavior of water further facilitated the
development of an efficient synthesis of alcohols and ꢀ-deu-
terated alcohols from epoxides with anti-Markovnikov
regiochemistry,6b the control of the ending step in titanocene-
catalyzed radical cyclizations in the straightforward synthesis
of complex polycyclic terpenoids,11 and the development of a
new hydrogenation reaction of alkenes and alkynes based on
an unprecedented HAT reaction from water to common
hydrogenation catalysts (Pd/C, Rh/C, Wilkinson catalyst, etc.),
avoiding the use of potentially dangerous hydrogen gas.12 All
these results might also have important consequences on
fundamental mechanistic considerations of organic reactions in
aqueous media. Moreover, from a practical point of view, water
can be used as a safe and inexpensive HAT reagent instead of
actual ones.
Nevertheless, the general scope of this uncommon radical
reduction is unknown despite of its interest. Simple alkyl radicals
seem to be suitable substrates for this transformation, but related
(9) Jin, J.; Newcomb, M. J. Org. Chem. 2008, 73, 7901–7905.
(10) Newcomb, M. Tetrahedron 1993, 49, 1151–1176.
A similar UV-vis study of methanol and phenol in benzene
was carried out to avoid any competitive binding process with
the solvent (Figures 6 and 7). Water had to be excluded from
the study due to its immiscibility with benzene. The behavior
(11) (a) Barrero, A. F.; Oltra, J. E.; Cuerva, J. M.; Rosales, A. J. Org.
Chem. 2002, 67, 2566–2571. (b) Barrero, A. F.; Rosales, A.; Cuerva,
J. M.; Oltra, J. E. Org. Lett. 2003, 5, 1935–1938. (c) Justicia, J.;
Rosales, A.; Bun˜uel, E.; Oller-Lo´pez, J. L.; Valdivia, M.; Haidour,
A.; Oltra, J. E.; Barrero, A. F.; Ca´rdenas, D. J.; Cuerva, J. M.
Chem.sEur. J. 2004, 10, 1778–1788. (d) Justicia, J.; Oller-Lo´pez,
J. L.; Campan˜a, A. G.; Oltra, J. E.; Cuerva, J. M.; Bun˜uel, E.; Ca´rdenas,
D. J. J. Am. Chem. Soc. 2005, 127, 14911–14921. (e) Justicia, J.; Oltra,
J. E.; Cuerva, J. M. J. Org. Chem. 2005, 70, 8265–8270.
(12) Campan˜a, A. G.; Este´vez, R. E.; Fuentes, N.; Robles, R.; Cuerva, J. M.;
Bun˜uel, E.; Ca´rdenas, D. J.; Oltra, J. E. Org. Lett. 2007, 9, 2195–
2198.
(13) Pozzi, D.; Scanlan, E. M.; Renaud, P. J. Am. Chem. Soc. 2005, 127,
14204–14205.
(14) For a previous spectroscopic study (mainly UV-vis and 17O NMR)
of titanocene(III) aqua complexes see reference.8
(15) Taking into account that (Cp2TiCl)2 is a dimer, 1 equiv of water means
a molar ratio (Cp2TiCl)2: H2O of 1:2.
9
J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010 12749

A R T I C L E S
Paradas et al.
Figure 1. UV-vis spectra of 2 (green) in THF (0.01 M) with increasing
amounts of H2O: 0.5 equiv (blue), 1 equiv (red), 1.5 equiv (yellow), 2 equiv
(purple), 2.5 equiv (orange), 3 equiv (dark green), 4 equiv (dark blue), 6
equiv (dark red), 9 equiv (dark yellow), and 13 equiv (light green).
Figure 4. (a) UV-vis spectra of 2 (green) in THF (0.01 M) with increasing
amounts of MeOH: 0.5 equiv (blue), 1 equiv (red), 1.5 equiv (yellow), 2
equiv (purple), 2.5 equiv (orange), 3 equiv (dark green), 4 equiv (dark blue),
6 equiv (dark red), 10 equiv (dark yellow), 15 equiv (light green), 21 (light
blue), 28 (light red), 45 (light purple), and 55 (black). (b) Competitive
experiments: 2 (green), 2 + 13 equiv of phenol (blue), and 2 + 13 equiv
of phenol and water (red).
Figure 2. UV-vis spectra of 2 (green) in THF (0.01 M) with increasing
amounts of N-octylamine: 0.05 equiv (blue), 0.1 equiv (red), 0.15 equiv
(yellow), 0.2 equiv (purple), 0.3 equiv (orange), 0.4 equiv (dark green),
0.5 equiv (dark blue), 0.7 equiv (dark red), 0.9 equiv (dark yellow), 1.2
equiv (light green), 1.5 (light blue), 2 (light red), and 3 (light purple).
Figure 3. UV-vis spectra of 2 (green) in THF (0.01 M) with increasing
amounts of di-N-butylamine: 0.1 equiv (blue), 0.9 equiv (red), and 2 equiv
(orange).
of both additives in benzene and THF was very similar. This
fact suggests that, at least, THF is not a competitive ligand when
methanol is used as additive. Moreover, the independence of
the behavior of both additives from the solvent used, strongly
supports that the titanocene(III) complex has similar structure
in both solutions at those concentrations. A dimeric structure,
(Cp2TiCl)2, seems to be plausible in benzene solutions of
complex 2, according to the low coordination ability of benzene.
Therefore, the dimeric structure is retained in THF solutions
even in the presence of a large excess of a potential ligand, as
THF molecules.16 Under this assumption, we could obtain the
Figure 5. (a) UV-vis spectra of 2 (green) in THF (0.01 M) with increasing
amounts of PhOH: 0.5 equiv (blue), 1 equiv (red), 2 equiv (yellow), 4 equiv
(purple), 8 equiv (orange), 14 equiv (dark green), 22 equiv (dark blue), 32
equiv (dark red), 47 equiv (dark yellow), 77 equiv (light green), 122 (light
blue), 182 (light red), and 262 (light orange). (b) Competitive experiments:
2 (green), 2 + 13 equiv of phenol (blue), and 2 + 13 equiv of phenol and
water (red).
corresponding binding constants from the titration curves.17
Consequently with the UV-vis spectra, the highest binding
constant in THF corresponds to N-octylamine (Koctylamine ) 1412
M-1). Water and methanol have binding constant values of 101.6
and 2.5 M-1, respectively. For phenol we could only determine
(16) The corresponding UV-vis study of a 0.01 M solution of complex 2
in benzene in the presence of increasing amounts of THF was also
carried out. We did not observe significant changes in the UV-vis
spectrum even in the presence of high amounts of THF. See Supporting
Information for details.
(17) We could not obtain a titration curve for N, N-dibutyl amine.
9
12750 J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010

Exceptional H-Atom Donor Characteristics of H2O
A R T I C L E S
Table 1. Binding Energies of Potential Hydrogen Atom Donors to
Cp2TiCl19
ligand (L)
∆(E+ZPE), kcal mol-1
H2O
D2O
-12.3
-12.8
-10.4
-6.4
Cp2TiCl
MeOH
PhOH
NH3
MeNH2
Me2NH
THF
Figure 6. UV-vis spectra of 2 (green) in benzene with increasing amounts
of MeOH: 0.5 equiv (blue), 1 equiv (red), 1.5 equiv (yellow), 2 equiv
(purple), 2.5 equiv (orange), 3 equiv (dark green), 4 equiv (dark blue), 6
equiv (dark red), 10 equiv (dark yellow), 15 equiv (light green), 21 (light
blue), 28 (light red), 45 (light purple), and 55 (black).
-4.6
-13.6
-12.1
-8.7
-4.6
These (relative) values of binding energies could explain the
profile of the titration experiments. That is, dimeric (Cp2TiCl)2
is stable in the presence of methanol and phenol, and low
concentrations of the corresponding Cp2TiCl(HXR) active
hydrogen atom donor species is expected in these cases.
Consequently, the rates of HAT for methanol and phenol will
be masked by this unfavorable coordination step. In the case of
phenol, the situation is even less favorable owing to an efficient
competence of THF molecules toward the binding sites.
According to the calculated values, water and primary amines
are able to yield the monomeric structures, which explains the
marked UV-vis changes. Nevertheless, the interpretation of the
UV-vis spectra of mixtures of titanocene(III) and amines might
be not so simple, since subsequent deprotonation steps of the
initial amine complexes could yield the corresponding amido
complexes.18 Noteworthy, deuterium oxide has a slightly better
coordination ability than H2O and this fact could have some
influence when water-based or D2O-based experiments are
compared. In this sense, free titanocene(III) concentration would
be lower in experiments using D2O, and thus, slightly slower
reaction rates are expected.
Thermodynamic Profile of the HAT Process. The second key
parameter to consider is the intrinsic HAT capabilities of the
studied additives after coordination to titanocene(III). We have
previously suggested that the unprecedented ability of water as
hydrogen atom donor is due to a decrease of the O-H BDE
favored by coordination to Cp2TiIIICl. We wondered whether
the above-mentioned closely related systems, including amines,
alcohols, or phenols, would show the same trend. Theoretical
calculations gave us the (relative) BDE values for the free
species HXR and for the corresponding Cp2TiCl(HXR) com-
plexes, which let us estimate the H-X bond weakening. The
results are summarized in Table 2.
Figure 7. UV-vis spectra of 2 (green) in benzene with increasing amounts
of PhOH: 22 equiv (blue), 32 equiv (red), 47 equiv (yellow), 77 equiv
(purple), 122 equiv (orange), 182 equiv (dark green), and 262 equiv (dark
blue).
that the binding constant is in the range of 10-2 M-1. We could
also determine the binding constant value for methanol in
benzene, which increases up to 6.7 M-1
.
Binding Energies. The experimental results suggest that only
water, amines, and, to a lesser extent, methanol are able to
dissociate the dimeric (Cp2TiCl)2. To clarify this key point we
performed a computational study using density functional theory
(DFT). The (relative) binding energies of water and the others
potential hydrogen atom donors to Cp2TiCl were calculated at
the B3LYP/6-31G(d) level. In this theoretical study, we have
included water, methanol, phenol, and methyl- and dimethy-
lamine as simplified structures of the primary and secondary
amines. Ammonia and THF have been also included for
comparison. The binding energy of the dimeric (Cp2TiCl)2 was
also calculated because it is another essential parameter for the
understanding of the nature of the hydrogen atom donor species
in solution.
We were pleased to find that theoretical results fit with the
experimental ones (see Table 1). We can clearly observe that
ammonia and methylamine have binding energies, -13.6 and
-12.1 kcal mol-1, respectively, similar to water (-12.3 kcal
mol-1). Methanol (-6.4 kcal mol-1) and phenol (-4.6 kcal
mol-1) possesses a coordination capability close to THF (-4.6
kcal mol-1), while a secondary amine, such as dimethylamine,
presents intermediate values (-8.7 kcal mol-1). Interestingly,
the binding energy of the titanocene(III) dimer is -10.4 kcal
mol-1, which is between the calculated energies for water and
methanol or phenol.
As can be seen, water is the species showing the highest bond
weakening by coordination with Cp2TiCl (∆(BDE) ) 58.6 kcal
(19) This study has been restricted to Cp2TiCl. Nevertheless, it has been
described that changes in the Cp substitution in titanocene(III)-
complexes have significant variations in their chemical reactivity and
consequently coordination capabilities can be also affected: Duthaler,
R. O.; Hafner, A. Chem. ReV. 1992, 92, 807–832. In fact, calculated
binding energies for water and (tBuCp)2TiCl and Cp2*TiCl are -10.63
and -8.30 kcal mol-1, respectively.
(18) For a recent reference see: Paniagua, C.; Mosquera, M. E. G.; Jacobsen,
H.; Jimenez, G.; Cuenca, T. Organometallics 2009, 28, 6975–6980.
9
J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010 12751

A R T I C L E S
Paradas et al.
Table 2. BDE Values of the X-H Bond in Cp2TiCl(HXR)
Complexes
Scheme 3. Mechanistic Studies of Titanocene(III)-Mediated
Reduction of Epoxide 3
HXR
BDE, kcal mol-1 Cp2TiCl(HXR) BDE, kcal mol-1 ∆(BDE), kcal mol-1
H2O
D2O
108.1
110.3
93.5
74.3
100.4
92.4
H2O
D2O
49.4
51.4
42.6
39.6
65.0
63.7
67.6
58.6
58.8
50.9
34.7
35.4
28.7
19.1
MeOH
PhOH
NH3
MeNH2
Me2NH
MeOH
PhOH
NH3
MeNH2
Me2NH
aqua complexes actually takes place (Scheme 3). Under
anhydrous standard conditions (Cp2TiCl2/Mn), the main product
is the allylic alcohol 5 (58%) and not the expected deoxygenated
product derived from a potential Ti(IV) organometallic inter-
mediate.22 However, when we treated epoxide 3 in the presence
of D2O (10 equiv) under the standard conditions, reduced
derivative 4 (69%, 83% deuterium incorporation) was obtained.
This result indicated that the water-mediated reduction of the
transient tertiary radical was faster than the potential radical
trapping by a second Cp2TiCl species. Additionally, when we
repeated the experiment in the presence of an equimolar mixture
of D2O and 1,4-CHD (10 equiv each), we obtained 4 (78%)
with a 83% deuterium incorporation, revealing that the transfer
of deuterium from D2O was even faster than that of hydrogen
from 1,4-CHD.23 This kinetics ruled out the possibility of 4
being formed by the hydrolysis of an alkyl-TiIV complex.
We explored the reactivity of epoxide 3 (1 mmol) using
pregenerated 2 (1 mmol)24 and different potential hydrogen atom
donors (5 mmol) in THF. The amount of additive was selected
to ensure the presence of unbound titanocene(III) in the reaction
media, which is essential for the epoxide radical opening. The
results are summarized in Table 4.
86.7
Table 3. ∆H for the Reduction of t-Butyl Radical by Cp2TiCl(HXR)
Complexes
HXR
∆H, kcal mol-1
H2O
D2O
-42.1
-42.2
-48.9
-54.9
-26.5
-27.8
-23.2
MeOH
PhOH
NH3
MeNH2
Me2NH
mol-1). Of the other two oxygenated systems studied, only
methanol has a similar behavior (∆(BDE) ) 50.9 kcal mol-1).
Phenol, ammonia, and primary and secondary amines showed
a significantly lower bond weakening (∆(BDE) ) 19.1-35.4
kcal mol-1).
Although the reactions are not quantitative, the total yields
of isolated compounds are similar and consistent. Therefore,
conclusions derived from the relative relationships of 4 and 5
can be assumed. First of all, the results clearly show that a direct
relationship between the acidity of the hydrogen donor and the
efficiency of the reduction process does not exist, although it
should be expected if a protonolysis of an alkyltitanium
intermediate had been involved.25 As we can see, water is the
Nevertheless, the feasibility of the HAT reaction is related
to not only the bond weakening but also the exotermicity of
the global reaction. In that case, the creation of the new H-C
bond has to be taken into account. Consequently, the endoter-
micity of the H-X bond dissociation in the titanocene(III)
complex can be surpassed more easily with low values of BDE.
By this reasoning methanol (BDE of Cp2TiCl(MeOH) ) 42.6
kcal mol-1) and phenol (BDE of Cp2TiCl(PhOH) ) 39.6 kcal
mol-1) would be the best HAT reagents and not water. In this
sense, we have computed the enthalpy change of a model HAT
reaction. For simplicity, we selected the reduction of t-butyl
radical using the corresponding Ti(III)-complexes. In all cases,
the reaction was exotermic, and as we expected based on the
values in Table 3, phenol and methanol showed the higher
exotermicities.
All of these theoretical results suggest that water, methanol,
and phenol should be good HAT reagents in the reduction of
tertiary radicals, with phenol or methanol being the best.
Nevertheless, the coordination capabilities described before must
be also taken into account in real experiments.
Influence of the H-Atom Donor on the Reduction of
Epoxides Mediated by Titanocene(III). Fortunately, these theo-
retical predictions could be tested experimentally. We selected
epoxide 3 as the model substrate20,21 and carried out a set of
experiments to guarantee that a HAT reaction from titanium(III)
(21) The benzoate group was introduced to avoid losses by evaporation,
guaranteeing the homogeneity of the results and as a UV marker for
an easy detection by analytical techniques.
(22) Although it has been described that bis(cyclopentadienyl)titanium(III)
chloride, generated in situ by stirring commercial Cp2TiCl2 with Mn
dust in anhydrous THF, exists as an equilibrium mixture of the
monomer Cp2TiCl and the dinuclear species (Cp2TiCl)2, we postulate
that monomeric Cp2TiCl is the active species in the mixed dispro-
portionation process: Justicia, J.; Jime´nez, T.; Morcillo, S. P.; Cuerva,
J. M.; Oltra, J. E. Tetrahedron 2009, 65, 10837–10841.
(23) A control experiment using 1,4-CHD (10 equiv each) as HAT reagent
gave a mixture of 4 (40%) and 5 (32%). Although the calculated kinetic
values of t-butyl radical and 1,4-CHD (∆H ) -21.6 kcal mol-1
,
Ea ) 6.9 kcal mol-1) suggest that should be better than water, it can
not coordinate with Cp2TiCl. In this situation the concentration of free
Cp2TiCl is high enough to promote secondary reactions such as
deoxygenation processes. Pre-association processes between aqua
complex and the tertiary radical cannot be ruled out either: Hammerum,
S. J. Am. Chem. Soc. 2009, 131, 8627–8635. See also: Miyazaki, S.;
Kojima, T.; Mayer, J. M.; Fukuzumi, S. J. Am. Chem. Soc. 2009, 131,
11615–11624.
(24) In the usual experimental conditions, interferences owing to the
presence of both Mn dust (usually required for the in situ generation
of titanocene(III) from titanocene(IV) precursors) and stoichiometric
amounts of MnCl2 (formed upon oxidation of the metallic co-reductor)
have to be taken into account. Therefore, for comparison, we carried
out the same experiments in the presence of Mn dust or MnCl2
obtaining similar results.
(25) In such a case, according to the pKa values, a higher amount of reduced
product should be expected in the reaction with phenol compared with
water. Nevertheless, the actual yields follow the opposite trend (17%
and 84%, respectively).
(20) In our preliminary communication (ref 6a), we selected 6,7-epoxyneryl
acetate as model substrate for tertiary radical generation and study. It
presented some unique characteristics which helped us to distinguish
between different mechanisms. Nevertheless, we obtained mixtures
of cyclized and uncyclized products complicating the comparison
between experiments.
9
12752 J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010

Exceptional H-Atom Donor Characteristics of H2O
A R T I C L E S
Table 4. Yields of Compounds 4 and 5 Using Different Hydrogen
Atom Donors and Experimental Conditions in THF
Table 5. Yields of Compounds 4 and 5 Using Different
Experimental Conditions and Hydrogen Atom Donors in Benzene
entry
R-X-H (D)
yield (%)
4:5
D inc. (%)
1
2
3
4
5
H2O
D2O
MeOH
MeOD
PhOD
77
86
84
75
60
100:0
78:22
100:0
60:40
40:60
92
73
72
entry
R-X-H (D)
yield (%)
4:5
D inc. (%)
1
2
3
4
5
6
7
H2O
D2O
MeOH
MeOD
PhOD
Bu2NH
CH3(CH2)7NH2
84
87
83
80
85
71
81
100:0
90:10
55:45
33:66
20:80
0:100
20:80
84
Table 6. Activation Energies for the Reduction of the t-Butyl
Radical by Cp2TiCl(HXR) Complexes
58
47
a
a We used a 1:2 CH3(CH2)7NH2:2 molar ratio to ensure the presence
of active 2. When we used a 10:1 CH3(CH2)7NH2:2 molar ratio, only
starting material was recovered.
HXR
∆Ea, kcal mol-1
H2O
D2O
8.8
9.8
best hydrogen atom donor among the tested compounds, and
its behavior is significantly different compared with closely
related MeOH and PhOH. This apparent contradiction with the
theoretical findings can be easily explained by taking into
account that the concentrations of the corresponding titanoce-
ne(III) complexes of methanol and phenol in THF must be very
low, compared to free titanocene (Cp2TiCl)2, as we have
previously determined. These experimental results suggest that
the remarkable behavior of water as HAT reagent in Ti(III)
chemistry is due to not only the weakening of the H-OH bond
but also the high affinity of water for Ti(III) complexes.
Deuterium oxide gave similar global yields although low
amounts of alkene 5 could also be isolated. This is a relevant
result, since it suggests that the hydrogen atom transfer is the
slowest step in the overall reaction mechanism. In fact, we could
determine an estimated isotopic effect of 4.7, which is consistent
with our proposal.
The situation was different when deuterated phenol was used
as the additive. The low deuteration incorporation in compound
4 showed that the reduction product cannot derive exclusively
from a HAT from a titanocene(III):phenol complex. This is also
consistent with the absence of such species as we previously
determined by UV-vis and theoretical calculations. Under these
conditions, a HAT reaction from the solvent could take place.26
A direct disproportionation reaction between the corresponding
tertiary radicals cannot be excluded either.
Both binding energy to Cp2TiCl and H-N bond dissociation
energy suggest that amines could be good candidates. Therefore,
we carried out the reaction with model primary (octyl amine)
and secondary (dibutyl amine) amines. In the first case, the
starting epoxide was recovered unalterated. The higher coor-
dination abilities of nitrogen ligands compared to oxygen ones
led octylamine and (Cp2TiCl)2 to form a stable complex, and
consequently, the amounts of unbounded (Cp2TiCl)2 are so low
that the initial radical epoxide opening cannot occur. In order
to observe the behavior of octyl amine as HAT reagent, we
changed the reaction conditions by lowering the amount of
amine below the (Cp2TiCl)2 concentration. In these new
conditions, the reaction could be carried out, but only low yields
of reduction product 4 were obtained. Dibutyl amine could be
used in the standard proportion, but as in the case of octylamine,
allylic alcohol 5 was the main product. These experiments rule
MeOH
PhOH
NH3
MeNH2
Me2NH
7.9
4.0
14.7
12.4
15.3
out that amines can be used as efficient hydrogen atom donors
in this reaction, despite their excellent coordination capabilities.
The above shown binding constant of MeOH to titanocene(III)
in benzene suggests that the amount of the active species,
Cp2TiCl(MeOH), in that solvent should be higher. Therefore,
an increase of saturated alcohol 4 is expected. Although pure
benzene cannot be used directly as solvent in titanocene(III)-
mediated transformations, the experiments using pregenerated
(Cp2TiCl)2 could be carried out. We were pleased to find that,
in benzene, alcohol 4 is the sole product when methanol is used
as an additive. MeOD showed the same profile confirming that
the hydrogen atom also comes from methanol in benzene
solutions. When phenol-d6 was used as an additive in benzene,
small amounts of deuterated 4 could be isolated. This result
can be explained taking into account that a direct HAT process
from uncoordinated phenol can also take place as previously
reported.27 Consequently with the absence of THF, the deute-
rium incorporations are higher in benzene. It is worth noting
that incomplete deuteration can be correlated with an increase
of alkene 5, which can derive from a direct disproportionation
process (Table 5).
Theoretical studies can be used again to understand the
experimental results. The corresponding activation energies (Ea)
for triplet transition state (TS) were calculated (Table 6). A
correlation between the exotermicity and the Ea of all of the
reactions exists, as it is predicted by the Evans-Polanyi
equation. The higher exotermicities correlate with the minor
activation energies. Therefore, phenol showed an activation
energy of only 4.02 kcal mol-1, whereas the HAT process
involving dimethylamine has an activation energy of 15.33 kcal
mol-1. The activation energy for water is placed in the
intermediate value of 8.82 kcal mol-1. We also calculated the
key thermodynamic and kinetic parameters of the HAT reaction
from THF to t-butyl radical since it cannot be excluded in
principle (∆H ) -2.1 kcal mol-1, Ea ) 11.9 kcal mol-1). These
values suggest that HAT processes from THF cannot compete
(26) Newcomb et al. have reported that under similar conditions significant
HAT reactions from THF (2-4 × 103 M-1 s-1) are operative: see
ref 9.
(27) Franchi, P.; Lucarini, M.; Pedulli, G. F.; Valgimigli, L.; Lunelli, B.
J. Am. Chem. Soc. 1999, 121, 507–514.
9
J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010 12753

A R T I C L E S
Paradas et al.
Table 7. Enthalpies and Activation Energies for the Reduction of
Different Carbon-Centered Radicals by Cp2TiClOH2
∆H, kcal mol-1
∆Ea, kcal mol-1
entry
R
Figure 8. Calculated transition state for the H-atom transfer from 1 to
t-butyl radical (DFT level).
1
2
3
4
5
6
7
8
9
10
11
12
13
CH3-CH2
-48.5
-51.6
-45.0
-42.1
-43.5
-34.2
-58.3
-59.6
-37.2
-37.1
-36.9
-37.0
-36.6
6.3
5.5
7.2
8.8
8.4
16.1
0.4
3.0
12.4
12.2
7.5
12.1
13.9
CH2OH-CH2
CH3-CH-CH3
(CH3)3C
efficiently with water, methanol or phenol Cp2TiCl complexes,
even in the presence of high molar relationship.
CH2OH(CH3)2C
CH2CHCH2
CH2dCH
C6H5
C6H5CH2
If we compare the Ea values for the reduction of the t-butyl
radical with deuterium oxide or water, we can also conclude
that this transformation has a high isotopic effect, estimated at
5.08 based on the calculated values at 298 K (assuming the
same frequency factor). The only experimental values of isotopic
effect for this transformation are 3.35,6a 4.4,9 and 4.7 obtained
in the reduction of primary, secondary, and tertiary radicals (this
work), respectively, which are in good agreement with the
calculated one.
p-F-C6H5
p-OMe-C6H5CH2
p-OAc-C6H5CH2
p-CO2Me-C6H5CH2
and therefore should be considered. In this case, the Ea values
are slightly lower than the corresponding to non polar radicals
(entries 2 and 5). A decrease of 0.8 and 0.4 kcal mol-1 is
observed in primary and tertiary radicals, respectively. The
highest calculated Ea was obtained in the HAT to allyl radical
(16.1 kcal mol-1). Benzyl radicals were also studied, showing
values of Ea in between those calculated for alkyl and allyl ones
(12.1 kcal mol-1). In this case, substitution in para position might
affect the Ea because of an expected π-interaction between the
π-type radical and the π-aromatic.29 Nevertheless, we could not
find a simple relationship between the exotermicity of the
reaction or any other parameter and the Ea. All of Ea values are
in the range of 12.1-13.9 kcal mol-1 with the exception of
p-OMe substituted benzyl radical which calculated Ea is 7.5 kcal
mol-1. This abnormally low value suggests that the TS is
affected by important polar effects or a change of mechanism
is taking place. Again, inspection of the SOMO’s of the
corresponding TS’s suggests a HAT mechanism rather than a
proton-coupled electron transfer. If our mechanistic hypothesis
based on HAT processes from titanocene aqua complex is
correct, the experimental results should fit with the above-
mentioned theoretical predictions. Low Ea must be correlated
with efficient HAT processes at room temperature. However,
experiments are complicated in this case by the fact that side
reactions can become the main reaction with some substrates.
While in the study of tertiary radicals the interference of side
reactions could be avoided, the study of secondary and primary
radicals is more complex. As an example, primary radicals can
be efficiently captured by a molecule of Cp2TiCl and a
subsequent hydrolysis the alkyl-Ti(IV) intermediate can mask
a real HAT mechanism. In our preliminary communication,6a
we could demonstrate unequivocally that titanocene(III)-aqua
complex is able to promote a HAT reaction from coordinated
water to primary alkyl radicals generated by a cyclization of
caryophyllene oxides. Noteworthy, this result fits the predicted
Ea calculated for the HAT transfer to primary radicals.
HAT reactions are considered to be a subfamily of proton-
coupled electron transfer (PCET) reactions. Although PCET
processes avoid high energy intermediates such as carbanions
derived from a stepwise electron transfer, in this case, a careful
inspection SOMOs of the TSs suggests a HAT mechanism rather
than a proton-coupled electron transfer. The electron density
between O and H and the absence of connection between the
metal complex and the carbon radical imply that the proton and
the electron are being transferred involving the same set of
atomic orbitals.28
Reduction of Other Carbon-Centered Radicals. From the
above exposed results, we now know that two key factors are
at the heart of the success of water as hydrogen atom donor in
titanocene(III) chemistry: a strong coordination between water
and the titanocene(III) complex and a low activation energy
(estimated at less than 8.8 kcal mol-1 based on the calculated
values). Nevertheless, little is known about the generality of
this reaction since a systematic study of the nature of the carbon-
based radical has not been performed to date. For that reason,
we paid attention to the reactivity of 1 toward different carbon-
centered radicals to study the scope of the reaction. Bearing in
mind the solid concordance between the experimental and
theoretical results until this moment, we carried out theoretical
calculations over a range of HAT reactions from aqua complex
1 and different carbon-centered radicals (Figure 8). We have
used hydroxyradicals instead of the corresponding titanoxy
species in the modelization of the process for computational
convenience. The results are summarized in Table 7.
The theoretical values show that an increase of the stabiliza-
tion of the radical results in an increase of the activation energy.
The reduction of phenyl and vinyl radicals (entries 7 and 8) is
predicted to take place in an essentially barrierless pathway
(0.4-3.0 kcal mol-1). The activation energy of simple alkyl-
type radicals is also low (6.3-8.8 kcal mol-1, entries 1, 3, and
4) compatible with fast reaction rates. Although we have
considered mainly non polar radicals, that is, with no heteroatom
placed near the carbon-centered radical, in radical reductive
opening of epoxides an oxygen atom is present in the R-position
(29) In the literature, there are some cases where the activation energy in
HAT processes can be correlated with Hammett σ parameters. In that
case the activation energies, the spin densities and the HOMO levels
of the TS correlated well with the σ parameters: Wang, Y.; Kumar,
D.; Yang, C.; Han, K.; Shaik, S. J. Phys. Chem. B 2007, 111, 7700–
7710.
(28) For a similar analysis of the transition state, see: DiLabio, G. A.;
Johnson, E. R. J. Am. Chem. Soc. 2007, 129, 6199–6203.
9
12754 J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010

Exceptional H-Atom Donor Characteristics of H2O
A R T I C L E S
Scheme 4. Reduction of Epoxide 6
Scheme 6. Dimerization of Farnesyl Chloride (12) in the Presence
of Water
Scheme 5. TiIII-Promoted Cyclization of Epoxyalkyne 7 in the
Presence of H2O or D2O
Scheme 7. Dimerization of Benzylic Bromide (14) in the Presence
of Water
with (Cp2TiCl)2 under anhydrous conditions until all starting
material was consumed (5 h, TLC analysis). Then, D2O was
added and compound 9 was obtained (66%) without any
deuterium incorporation, which would have occurred in the
deuterolysis of a hypothetical organometallic (or carbanion-type)
intermediate.
Based on the calculated Ea values, the reduction of allylic-
type radicals should be cumbersome at room temperature. Such
radicals can be generated by (Cp2TiCl)2 from allylic halides.
As model substrate we selected farnesyl chloride 12. Conse-
quently with the Ea values, we could not observe significant
amounts of the corresponding reduction product even in the
presence of high concentrations of water (20 equiv). Homodimer
13 was the only product (71%; Scheme 6).33 This is in
agreement with the reported rate constant for the dimerization
of allyl radicals (4.0-8.3 × 109 M-1 s-1).34
The reduction of secondary radicals, with a predicted Ea value
of 7.2 kcal mol-1, may also occur by a HAT mechanism. When
simple epoxide 6 was used as model substrate, a similar kinetic
reasoning to that for compound 3 could not be performed.
Although alcohol 7 is the only isolable product in the presence
of water (Scheme 4), 1,4-CHD is unable to promote such
reduction and competitive experiments between D2O and 1,4-
CHD were not conclusive. Nevertheless, the isotopic effect of
this reduction was 4.4, which is similar to described for this
HAT process.
Related to this result, in 2008, Newcomb et al. determined
experimentally the Ea of the HAT reaction from Cp2TiClOH2
to a secondary radical. The obtained value (5.5 kcal mol-1) is
in excellent agreement with the calculated one (7.2 kcal mol-1
within the error expected for calculations.
)
The predicted activation energies for the reduction of benzylic
radicals are also compatible with a fast reaction. Nevertheless,
the rate constant for the dimerization of benzyl radical is
described as being around 109 M-1 s-1.35 Therefore, the expected
products in the reduction of the corresponding benzylic bromides
by (Cp2TiCl)2 is the corresponding homodimer, even in the
presence of an excess of water. In fact, when we carried out
the reaction with functionalized benzyl bromide 14 in the
presence of 20 equiv of water we exclusively obtained dimer
15 (90%; Scheme 7).36,37
Interestingly, products derived from the homodimerization
of benzylic radicals have been described in titanocene(III)
chemistry when styrene oxides are used as starting materials.38
As an example, the reductive opening of styrene oxide (16)
mediated by (Cp2TiCl)2 in anhydrous conditions yields the
corresponding dimer 1739 (48%; Scheme 8). The isolation of
the dimer guarantees that organometallic intermediates are not
involved in those reaction conditions. When the reaction was
HAT processes to phenyl and vinyl radicals are usually very
fast reactions with very low Ea, as theoretical calculations
predict. In this situation HAT from THF becomes an important
reaction, which can mask the HAT from the acuacomplex 1 in
THF. In fact, calculated Ea for the HAT from THF to these
radicals ranges from 2.0 kcal mol-1 for vinyl radical to
practically barrierless (0.1 kcal mol-1) for phenyl radical.30 The
generation of such radicals using (Cp2TiCl)2 (2) is not a simple
task. Nevertheless, vinyl radicals can be obtained after a
titanocene(III)-mediated cyclization of epoxi-alkynes in THF.31
To preclude any potential HAT from THF,32 we decided to carry
out that cyclization reaction in benzene using (Cp2TiCl)2. The
cyclization of model epoxi-alkyne 8 in the presence of water
(10 equiv) in benzene gave the cyclization product 9 albeit in
moderate yield (37%, Scheme 5).
When D2O was used instead of H2O, we obtained isotopomers
10 and 11 with 92% deuterium incorporation (Scheme 5),
confirming that the hydrogen atom came from water. The above
observations strongly suggested a reaction mechanism via
unprecedented H-atom transfer from water to a sp2 carbon-
centered radical, mediated by Cp2TiIII(OH2)Cl. Additionally, a
potential alternative mechanism via formation and subsequent
hydrolysis of an organometallic vinyl-TiIV intermediate was
ruled out by the following experiments. Epoxide 8 was treated
(33) Barrero, A. F.; Herrador, M. M.; Qu´ılez del Moral, J. F.; Arteaga, P.;
Arteaga, J. F.; Die´guez, H. R.; Sa´nchez, E. M. J. Org. Chem. 2007,
72, 2988–2995.
(34) Throssel, J. J. Int. J. Chem. Kinetics 1972, 4, 273–276.
(35) k ) 4.6 × 109 M-1s-1: Claridge, R. F. C.; Fischer, H. J. Phys. Chem.
1983, 87, 1960–1967.
(36) The functionalization was selected to obtain reduction products which
would not be volatile.
(37) Barrero, A. F.; Herrador, M. M.; Qu´ılez del Moral, J. F.; Arteaga, P.;
Akssira, M.; El Hanbali, F.; Arteaga, J. F.; Die´guez, H. R.; Sa´nchez,
E. M. J. Org. Chem. 2007, 72, 2251–2254.
(30) Cp2TiClOD2 showed similar profile to vinyl radical: Ea ) 1.51 kcal
mol-1, ∆H ) 58.1 kcal mol-1
.
(31) Gansa¨uer, A.; Pierobon, M.; Bluhm, H. Angew. Chem., Int. Ed. 2002,
41, 3206–3208.
(38) (a) Gansa¨uer, A.; Ndene, N.; Lauterbach, T.; Justicia, J.; Winkler, I.;
Mu¨ck-Lichtenfeld, C.; Grimme, S. Tetrahedron 2008, 64, 11839–
11845. (b) Ferna´ndez-Mateos, A.; Encinas Madrazo, S.; Herrero
Teijo´n, P.; Rubio Gonza´lez, R. J. Org. Chem. 2009, 74, 3913–3918.
(39) Compound 17 was isolated as a 1:2 mixture of meso:dl stereoisomers.
(32) When the reaction is carried out in THF, we only obtained the
nondeuterated products in the presence of D2O, which suggests that
the HAT from THF is the main reaction.
9
J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010 12755

A R T I C L E S
Paradas et al.
Scheme 8. Dimerization and Reduction of Stylbene Oxide (16)
centered radicals of diverse nature. The success of this
transformation is based on two key features: (a) an excellent
binding capabilities of water toward titanocene(III) complexes
and (b) a low activation energy for the HAT step. Therefore,
the observed reactivity can be explained in the framework of
an unprecedented HAT reaction involving water. The theory
has predictive capabilities fitting well with the experimental
results, and may aid to find more examples of this remarkable
radical reaction. We are currently working in obtaining a
complete theoretical foundation of this phenomenon and in the
extension of these studies to other metals such as Sm, Cr, or
Co.
Table 8. Enthalpies, Activation Energies and Yields for the
Reduction of Different Styrene-Derived Radicals by Cp2TiClOH2
epoxide
R
∆H, kcal mol-1
∆Ea, kcal mol-1
alcohol
yield, %
Acknowledgment. We thank to the Spanish Ministry of
Sciences and Innovation (MICINN; Projects CTQ2008-06790 and
CTQ2007-60494/BQU) and the Regional Government of Andaluc´ıa
(Project P06-FQM-01726 and P09-FQM-04571) for financial sup-
port. T.J. thanks MICINN for her FPU fellowship. M.P. thanks
financial support from the Regional Government of Andaluc´ıa. J.J.
thanks the University of Granada for his “Reincorporacio´n de
Doctores” contract. We also thank the Centro de Computacio´n
Cient´ıfica-UAM for computation time.
16
19
20
H
F
OAc
-36.1
-36.1
-39.9
13.9
13.7
18.2
18
21
22
43
52
48
repeated in the presence of water we could obtain alcohol 18
(43%) which can not derive from the protonolysis of an
organotitanium. In this situation the HAT process could be
operative. This study was extended to para substituted styrene
oxides 19-20 and apparently the reduction process takes place
even when substrates with a high Ea are used (Table 8).
Supporting Information Available: General experimental
procedures and computational details. UV-vis and 13C NMR
spectra of all new compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
Conclusions
We have demonstrated that our initial assumption that
titanocene(III) aqua complexes are a unique class of HAT
reagents is correct. They are able to reduce efficiently carbon-
JA105670H
9
12756 J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010