Revelation of the nature of the reducing species in titanocene halide-promoted reductions

Article abstract of DOI:10.1021/ja0491230

The fundamental nature of Ti^III complexes generated in tetrahydrofuran by reduction of Cp₂TiCl₂ has been clarified by means of cyclic voltammetry and kinetic measurements. While the electrochemical reduction of Cp₂TiCl₂ leads to the formation of Cp₂TiCl₂^-, the use of metals such as Zn, Al, or Mn as reductants affords Cp₂TiCl and (Cp ₂TiCl)₂ in a mixture having a dimerization equilibrium constant of 3 × 10³ M^-1, independent of the metal used. Thus, we find it unlikely that the trinuclear complexes or ionic clusters known from the solid phase should be present in solution as previously suggested. The standard potentials determined for the redox couples Cp ₂TiCl₂/Cp₂TiCl₂^-, (Cp₂TiCl)₂⁺/(Cp₂TiCl)₂, Cp₂TiCl⁺/Cp₂TiCl, and Cp₂Ti ²⁺/Cp₂Ti⁺ increase in the order listed. However, the reactivity of the different Ti^III complexes is assessed as (Cp₂TiCl)₂ ? Cp₂TiCl ≈ Cp ₂Ti⁺ ? Cp₂TiCl₂^- in their reactions with benzyl chloride and benzaldehyde. None of the reactions proceed by an outer-sphere electron transfer pathway, and clearly the inner-sphere character is much higher in the case of Cp₂Ti ⁺ than for (Cp₂TiCl)₂, Cp₂TiCl, and in particular Cp₂TiCl₂^-. As to the electron acceptor, the inner-sphere character increases, going from benzyl chloride to benzaldehyde, and it is suggested that the chlorine atom in benzyl chloride and the oxygen atom in benzaldehyde may function as bridges between the reactants in the transition state.

Full text of DOI:10.1021/ja0491230

Published on Web 06/04/2004
Revelation of the Nature of the Reducing Species in
Titanocene Halide-Promoted Reductions
Rasmus Juel Enemærke, Jens Larsen, Troels Skrydstrup,* and Kim Daasbjerg*
Contribution from the Department of Chemistry, UniVersity of Aarhus,
Langelandsgade 140, 8000 Aarhus C, Denmark
Received February 17, 2004; E-mail: kdaa@chem.au.dk
Abstract: The fundamental nature of TiIII complexes generated in tetrahydrofuran by reduction of Cp2TiCl2
has been clarified by means of cyclic voltammetry and kinetic measurements. While the electrochemical
reduction of Cp2TiCl2 leads to the formation of Cp2TiCl2-, the use of metals such as Zn, Al, or Mn as
reductants affords Cp2TiCl and (Cp2TiCl)2 in a mixture having a dimerization equilibrium constant of 3 ×
103 M-1, independent of the metal used. Thus, we find it unlikely that the trinuclear complexes or ionic
clusters known from the solid phase should be present in solution as previously suggested. The standard
potentials determined for the redox couples Cp2TiCl2/Cp2TiCl2-, (Cp2TiCl)2+/(Cp2TiCl)2, Cp2TiCl+/Cp2TiCl,
and Cp2Ti2+/Cp2Ti+ increase in the order listed. However, the reactivity of the different TiIII complexes is
assessed as (Cp2TiCl)2 J Cp2TiCl ≈ Cp2Ti+ . Cp2TiCl2 in their reactions with benzyl chloride and
-
benzaldehyde. None of the reactions proceed by an outer-sphere electron transfer pathway, and clearly
the inner-sphere character is much higher in the case of Cp2Ti+ than for (Cp2TiCl)2, Cp2TiCl, and in particular
Cp2TiCl2-. As to the electron acceptor, the inner-sphere character increases, going from benzyl chloride to
benzaldehyde, and it is suggested that the chlorine atom in benzyl chloride and the oxygen atom in
benzaldehyde may function as bridges between the reactants in the transition state.
Introduction
However, in our previous study concerning the electrode
reduction of Cp2TiX2 (X ) Cl, Br, and I), we showed that
mixtures of Cp2TiX2-, Cp2TiX, and the dimer (Cp2TiX)2 are
generated, with the former being the major species only in the
case of X ) Cl.15
Low-valent metal complexes employed as single electron
transferring agents are important reagents for promoting organic
synthetic transformations.1 One of these, which has become
increasingly popular over the past few years, is bis(cyclopen-
tadienyl)titanium chloride (Cp2TiCl), promoting highly diaste-
reoselective pinacol coupling reactions,2 epoxide openings to
alkyl radicals,3 reductions of alkyl halides such as glycosyl
bromides4 and Vic-dibromides,5 Reformatsky additions,6 and
others.7-9 The generation of the trivalent titanium species in
tetrahydrofuran (THF) may be achieved by direct substitution
of two of the chlorides in TiCl3 with metalated cyclopentadiene10
or by simple reduction of Cp2TiCl2 with metals such as zinc,11
aluminum,12 and manganese.13 Cp2TiCl has likewise been
reported to be obtained after electrode reduction of Cp2TiCl2.14
The metal-based reductions have led to development of
important reaction protocols catalytic with respect to the titanium
complex, employing stoichiometric amounts of a metal re-
ductant to regenerate the catalyst after electron trans-
fer.2c-i,3q-s,u,y,gg-ii,mm,vv,4f,8,9 Such applications significantly aug-
ment the utility of this reagent for organic synthesis, while
justifying the development of more elaborate versions for
asymmetric synthesis. Despite its synthetic importance, little is
known about the mechanism of the electron transfers that it
promotes to organic substrates. Information concerning the
electronic interaction between the reductant and the substrate
in the transition state would undoubtedly be of importance for
understanding the reactivity of TiIII reductants in addition to
the development of new applications.
A complicating feature is the structure of the species produced
from the above-described reduction of Cp2TiCl2 with metals.
Previous reports have shown that trinuclear complexes (denoted
I in Chart 1) may be isolated in the solid state, depending on
the metal employed.13,16 Such complexes have therefore been
proposed to be the reducing species in solution accounting at
the same time for the high dl-selectivities observed in pinacol
couplings of aryl aldehydes in THF.17 On the other hand, ESR
and ENDOR studies of I prepared with zinc have provided
evidence for the existence of a TiIII cation in a frozen THF
(1) (a) Molander, G. A.; Harris, C. R. Chem. ReV. 1996, 96, 307. (b) Skrydstrup,
T. Angew. Chem., Int. Ed. Engl. 1997, 36, 345. (c) Molander, G. A.; Harris,
C. R. Tetrahedron 1998, 54, 3321. (d) Gansa¨uer, A. Synlett 1998, 801. (e)
Krief, A.; Laval, A.-M. Chem. ReV. 1999, 99, 745. (f) Gansa¨uer, A.; Bluhm,
H. Chem. ReV. 2000, 100, 2771. (g) Spencer, R. P.; Schwartz, J. Tetrahedron
2000, 56, 2103. (h) Gansa¨uer, A. In Radicals in Organic Synthesis; Renaud,
P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim, Germany, 2001; Vol. 2, p
2007. (i) Li, J. J. Tetrahedron 2001, 57, 1. (j) Steel, P. G. J. Chem. Soc.,
Perkin Trans. 1 2001, 2727.
(2) (a) Handa, Y.; Inanaga, J. Tetrahedron Lett. 1987, 28, 5717. (b) Barden,
M. C.; Schwartz, J. J. Am. Chem. Soc. 1996, 118, 5484. (c) Gansa¨uer, A.
J. Chem. Soc., Chem. Commun. 1997, 457. (d) Gansa¨uer, A.; Moschioni,
D.; Bauer, Eur. J. Org. Chem. 1998, 1923, 3. (e) Gansa¨uer, A.; Bauer, D.
J. Org. Chem. 1998, 63, 2070. (f) Gansa¨uer, A.; Bauer, D. Eur. J. Org.
Chem. 1998, 2673, 3. (g) Dunlap, M. S.; Nicholas, K. M. Synth. Commun.
1999, 29, 1097. (h) Halterman, R. L.; Zhu, C.; Chen, Z.; Dunlap, M. S.;
Khan, M. A.; Nicholas, K. M. Organometallics 2000, 19, 3824. (i) Dunlap,
M. S.; Nicholas, K. M. J. Organomet. Chem. 2001, 630, 125. (j) Lake, K.;
Dorrell, M.; Blackman, N.; Khan, M. A.; Nicholas, K. M. Organometallics
2003, 22, 4260.
9
10.1021/ja0491230 CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 7853-7864
7853

A R T I C L E S
Enemærke et al.
Chart 1. Proposed Structures of Reactive Trinuclear (I) and
Dimeric (II and III) Complexes in THF
matrix.18 Preliminary cyclic voltammetric studies also gave no
indications of the presence of trinuclear complexes in solution.19
Finally, a comparative study of the diastereoselectivities obtained
Figure 1. Cyclic voltammograms of 2 mM solutions of Zn-Cp2TiCl2 (s),
Mn-Cp2TiCl2 (- - -), and Al-Cp2TiCl2 (- ‚ ‚ -) recorded at a sweep
rate of 0.1 V s-1 in 0.2 M Bu4NPF6/THF.
(3) (a) Nugent, W. A.; RajanBabu, T. V. J. Am. Chem. Soc. 1988, 110, 8561.
(b) RajanBabu, T. V.; Nugent, W. A. J. Am. Chem. Soc. 1989, 111, 4525.
(c) RajanBabu, T. V.; Nugent, W. A.; Beattie, M. S. J. Am. Chem. Soc.
1990, 112, 6408. (d) Yadav, J. S.; Shekharam, T.; Gadgil, V. R. J. Chem.
Soc., Chem. Commun. 1990, 843. (e) Lowinger, T. B.; Weiler, L. Can. J.
Chem. 1990, 68, 1636. (f) Merlic, C. A.; Xu, D. J. Am. Chem. Soc. 1991,
113, 9855. (g) Yadav, J. S.; Shekharam, T.; Srinivas, D. Tetrahedron Lett.
1992, 33, 7973. (h) Merlic, C. A.; Xu, D. Tetrahedron Lett. 1993, 34, 227.
(i) RajanBabu, T. V.; Nugent, W. A. J. Am. Chem. Soc. 1994, 116, 986.
(j) Golakoti, T.; Ogino, J.; Heltzel, C. E.; Husebo, T. L.; Jensen, C. M.;
Larsen, L. K.; Petterson, G. M. L.; Moore, R. E.; Mooberry, S. L.; Corbett,
T. H.; Valeriote, F. A. J. Am. Chem. Soc. 1995, 117, 12030. (k) Clive, D.
L. J.; Magnusson, S. R. Tetrahedron Lett. 1995, 36, 15. (l) Clive, D. L. J.;
Magnusson, S. R.; Manning, H. W.; Mayhew, D. L. J. Org. Chem. 1996,
61, 2095. (m) Maiti, G.; Roy, S. C. J. Chem. Soc., Perkin Trans. 1 1996,
403. (n) Chakraborty, T. K.; Dutta, S. J. Chem. Soc., Perkin Trans. 1 1997,
1257. (o) Weigand, S.; Bru¨ckner, R. Synlett 1997, 225. (p) Yadav, J. S.;
Srinivas, D. Chem. Lett. 1997, 905. (q) Gansa¨uer, A.; Pierobon, M.; Bluhm,
H. Angew. Chem., Int. Ed. 1998, 37, 101. (r) Gansa¨uer, A.; Bluhm, H. J.
Chem. Soc., Chem. Commun. 1998, 2143. (s) Gansa¨uer, A.; Bluhm, H.;
Pierobon, M. J. Am. Chem. Soc. 1998, 120, 12849. (t) Mandal, P. K.; Maiti,
G.; Roy, S. C. J. Org. Chem. 1998, 63, 2829. (u) Gansa¨uer, A.; Lauterbach,
T.; Bluhm, H.; Noltemeyer, M. Angew. Chem., Int. Ed. 1999, 38, 2909.
(v) Jørgensen, K. B.; Suenaga, T.; Nakata, T. Tetrahedron Lett. 1999, 40,
8855. (w) Mandel, P. K.; Maita, G. Tetrahedron 1999, 55, 11395. (x)
Ferna´ndez-Mateos, A.; Mart´ın de la Nava, E.; Coca, G. P.; Silvo, A. R.;
Gonza´lez, R. R. Org. Lett. 1999, 1, 607. (y) Gansa¨uer, A.; Pierobon, M.
Synlett 2000, 1357. (z) Chakraborty, T. K.; Das, S. Chem. Lett. 2000, 80.
(aa) Rana, K. K.; Guin, C.; Roy, S. C. Tetrahedron Lett. 2000, 41, 9337.
(bb) Do¨tz, K. H.; Gomes da Silva, E. Tetrahedron 2000, 56, 8291. (cc)
Rana, K. K.; Guin, C.; Roy, S. C. Synlett 2001, 1249. (dd) Barrero, A. F.;
Cuerva, J. M.; Herrador, M. M.; Valdivia, M. V. J. Org. Chem. 2001, 66,
4074. (ee) Haruo, Y.; Hasegawa, T.; Tanaka, H.; Takahashi, T. Synlett
2001, 1935. (ff) Nakai, K.; Kamoshita, M.; Doi, T.; Yamada, H.; Takahashi,
T. Tetrahedron Lett. 2001, 42, 7855. (gg) Gansa¨uer, A.; Bluhm, H.;
Pierobon, M.; Keller, M. Organometallics 2001, 20, 914. (hh) Gansa¨uer,
A.; Pierobon, M.; Bluhm, H. Synthesis 2001, 2500. (ii) Gansa¨uer, A.;
Bluhm, H.; Lauterbach, T. AdV. Synth. Catal. 2001, 343, 785. (jj)
Chakraborty, T. K.; Tapadar, S. Tetrahedron Lett. 2001, 42, 1375. (kk)
Chakraborty, T. K.; Das, S.; Raju, T. V. J. Org. Chem. 2001, 66, 4091. (ll)
Hardouin, C.; Chevallier, F.; Rousseau, B.; Doris, E. J. Org. Chem. 2001,
66, 1046. (mm) Gansa¨uer, A.; Rinker, B. Tetrahedron 2002, 58, 7017. (nn)
Chakraborty, T. K.; Das, S. Tetrahedron Lett. 2002, 43, 2313. (oo) Parrish,
J. D.; Little R. D. Org. Lett. 2002, 4, 1439. (pp) Barrero, A. F.; Oltra, J.
E.; Cuerva, J. M.; Rosales, A. J. Org. Chem. 2002, 67, 2566. (qq) Rana,
K. K.; Guin, C.; Roy, S. C. J. Org. Chem. 2002, 67, 3242. (rr) Hardouin,
C.; Doris, E.; Rousseau, B.; Mioskowski, C. Org. Lett. 2002, 4, 1151. (ss)
Hardouin, C.; Doris, E.; Rousseau, B.; Mioskowski, C. J. Org. Chem. 2002,
67, 6571. (tt) Barrero, A. F.; Cuerva, J. M.; Alvarez-Manzaneda, E. J.;
Enrique Oltra, J.; Chahboun, R. Tetrahedron Lett. 2002, 43, 2793. (uu)
Ruano, G.; Grande, M.; Anaya, J. J. Org. Chem. 2002, 67, 8243. (vv)
Gansa¨uer, A.; Rinker, B.; Pierobon, M.; Grimme, S.; Gerenkamp, M.;
Mu¨ck-Lichtenfeld, C. Angew. Chem., Int. Ed. 2003, 42, 3687. (ww) Anaya,
J.; Ferna´ndez-Matos, A.; Grande, M.; Martia´nez, J.; Ruano, G.; Rubio-
Gonza´lez R. Tetrahedron 2003, 59, 241. (xx) Ruano, G.; Martia´nez, J.;
Grande, M.; Anaya, J. J. Org. Chem. 2003, 68, 2024. (yy) Barrero, A. F.;
Rosales, A.; Cuerva, J. M.; Oltra, J. E. Org. Lett. 2003, 5, 1935.
in the pinacol coupling of benzaldehyde with metal-reduced
Cp2TiCl2 and (Cp2TiCl)2 in the presence of trimethylsilyl
chloride allowed Dunlap and Nicholas to propose the dimer to
be the true reducing agent in THF.2i
Through a combination of kinetic and electrochemical
investigations, we disclose in this article that trinuclear com-
plexes I (or ionic clusters) are not the reacting species in THF
despite their isolation and characterization in the solid phase
and that common TiIII-based reductants are formed in all cases,
independent of the metal used. Our results substantiate the view
that (Cp2TiCl)2 and Cp2TiCl are the major reducing species in
solution. In addition, we provide characterization of the electron-
transferring abilities of TiIII reductants with the two organic
substrates, benzyl chloride and benzaldehyde, allowing further
insight into the mechanism of the pertinent reduction processes.
Results and Discussions
Cyclic Voltammetry. Cyclic voltammetry studies of the three
metal-reduced solutions of Cp2TiCl2 denoted Met-Cp2TiCl2
(Met ) Zn, Mn, and Al) display a quite interesting picture as
shown in Figure 1. All three voltammograms exhibit similar
oxidation peaks at -0.8 (designated 2/3) and -0.4 V (designated
4) vs Fc+/Fc (abbreviation for ferrocenium/ferrocene), respec-
(6) (a) Ding, Y.; Zhao, G. J. Chem. Soc., Chem. Commun. 1992, 941. (b)
Parrish, J. D.; Shelton, D. R.; Little, R. D. Org. Lett. 2003, 5, 3615.
(7) Russo, T.; Pinhas, A. R. Organometallics 1999, 18, 5344.
(8) Zhou, L.; Hirao, T. Tetrahedron 2001, 57, 6927.
(9) Barrero, A. F.; Rosales, A.; Cuerva, J. M.; Gansa¨uer, A.; Enrique Oltra, J.
Tetrahedron Lett. 2003, 44, 1079.
(10) Manzer, L. E. Inorganic Synthesis; Wiley-Interscience, 1982; Vol. XXI, p
84.
(11) Coutts, R. S. P.; Wailes, P. C.; Martin, R. L. J. Organomet. Chem. 1973,
375.
(12) Birmingham, J. M.; Fischer, A. K.; Wilkinson, G. Naturwissenschaften
1955, 42, 96.
(13) Sekutowski, D. J.; Stucky, G. D. Inorg. Chem. 1975, 14, 2192.
(14) (a) Mugnier, Y.; Moise, C.; Laviron, E. J. Organomet. Chem. 1981, 204,
61. (b) Mugnier, Y.; Moise, C.; Laviron, E. NouV. J. Chim. 1982, 6, 197.
(c) Mugnier, Y.; Fakhr, A.; Fauconet, M.; Moise, C.; Laviron, E. Acta
Chem. Scand. 1983, B27, 423. (d) Mugnier, Y.; Moise, C.; Laviron, E. J.
Organomet. Chem. 1981, 210, 69. (e) Samuel, E.; Vedel, J. Organometallics
1989, 8, 237.
(15) Enemærke, R. J.; Larsen, J.; Skrydstrup, T.; Daasbjerg, K. Organometallics
2004, 23, 1866.
(16) Sekutowski, D.; Jungst, R.; Stucky, G. D. Inorg. Chem. 1978, 17, 1848.
(17) For a discussion of trinuclear complexes in solution, see ref 1d and
references therein.
(18) Gourier, D.; Vivien, D.; Samuel, E. J. Am. Chem. Soc. 1985, 107, 7418.
(19) Enemærke, R. J.; Hjøllund, G. H.; Daasbjerg, K.; Skrydstrup, T. C. R. Acad.
Sci. 2001, 4, 435.
(4) (a) Cavallaro, C. L.; Schwartz, J. J. Org. Chem. 1995, 60, 7055. (b) Spencer,
R. P.; Schwartz, J. Tetrahedron Lett. 1996, 37, 4357. (c) Spencer, R. P.;
Schwartz, J. J. Org. Chem. 1997, 62, 4204. (d) Spencer, R. P.; Cavallaro,
C. L.; Schwartz, J. J. Org. Chem. 1999, 64, 3987. (e) Hansen, T.; Krintel,
S. L.; Daasbjerg, K.; Skrydstrup, T. Tetrahedron Lett. 1999, 40, 6087. (f)
Hansen, T.; Daasbjerg, K.; Skrydstrup, T. Tetrahedron Lett. 2000, 41, 8645.
(5) Davies, S. G.; Thomas, S. E. Synthesis 1984, 1027.
9
7854 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Reducing Species in Titanocene Halide Reductions
A R T I C L E S
Figure 2. Cyclic voltammograms of 1.0 mM Cp2TiCl2 (s), 0.4 mM
Cp2Ti+ (- - -), and 1.5 mM Cp2TiCl/(Cp2TiCl)2 (- ‚ ‚ -) recorded at a
sweep rate of 0.1 V s-1 in 0.2 M Bu4NPF6/THF. Currents are normalized
with respect to concentration to facilitate comparison.
Figure 3. Cyclic voltammograms of 2 mM Cp2TiCl2 in the presence of 0
(s), 0.5 (- - -), and 1 equiv ZnCl2 (- ‚ ‚ -) recorded at a sweep rate
of 0.1 V s-1 in 0.2 M Bu4NPF6/THF. The sweep could not be extended
past -1.3 V in the negative direction because of the deleterious adsorption
of ZnCl2.
tively, while an additional peak is observed at -1.1 V vs
Fc+/Fc (designated 1) in the case of Met ) Al. As described
below, the broad 2/3 peak actually originates from two oxidation
processes. On the reverse sweep a reduction peak (designated
5) appears at ca. -1.2 V vs Fc+/Fc, being slightly dependent
on which metal considered. The fact that the peak potentials
for the three different metal-reduced solutions are the same
indicates that the waves originate from identical species and
that trinuclear complexes are unlikely to be present in any of
these solutions. Clearly, the oxidation potential of I (see Chart
1) would be expected to differ for the different metals.
The most likely candidates as common TiIII-based species
are Cp2TiCl2-, (Cp2TiCl)2, Cp2TiCl, and Cp2Ti+. The structure
of the (Cp2TiCl)2 dimer well-characterized from the solid state11
is shown as II in Chart 1, while III presents a tentative
suggestion for its structure in THF as discussed below.
Fortunately, all of these species can be generated either
separately or as a mixture (see the Experimental Section).15 In
Figure 2, we have collected cyclic voltammograms of Cp2TiCl2,
Cp2Ti+PF6-, and the mixture Cp2TiCl/(Cp2TiCl)2 (denoted
“Cp2TiCl” in ref 15) to be able to compare their characteristics
with those of the voltammograms of Met-Cp2TiCl2 in Figure
1. In accordance with expectation, Cp2TiCl2- with an oxidation
peak potential of -1.2 V vs Fc+/Fc is found to be the easiest
one to oxidize, whereas Cp2Ti+ with a potential of -0.4 V vs
Fc+/Fc is the most difficult one. Both (Cp2TiCl)2 and Cp2TiCl
show intermediate behavior, having essentially the same oxida-
tion peak potentials of -0.8 V vs Fc+/Fc; the appearance of
the additional peak at -0.4 V vs Fc+/Fc in this case is due to
the generation of Cp2Ti+ during the cyclic voltammetric
sweep.15 The nice correspondence of these peaks with those
recorded for the metal-generated solutions strongly suggests that
the origin of peaks 1-5 can be ascribed to the electrode
processes depicted in eqs 1-4.
It should be emphasized that because cyclic voltammetry is a
dynamic technique, the species detected are not necessarily
present initially in the solution but may simply be generated
during the sweep. This also provides the explanation why the
oxidation wave pertaining to Cp2Ti+ appears in the voltammo-
gram of Cp2TiCl/(Cp2TiCl)2. Thus, one cannot a priori conclude
that all four TiIII-based molecules are present in the metal-
generated solutions. Indeed, we will show below on the basis
of an extended analysis involving both cyclic voltammetry and
kinetic measurements that the two principal TiIII species are the
dimer, (Cp2TiCl)2, in equilibrium with its monomer, Cp2TiCl,
and that the following reactions apply in the description of the
metal-based reductions of Cp2TiCl2.
Zn + 2Cp2TiCl2 f ZnCl2 + 2Cp2TiCl
Mn + 2Cp2TiCl2 f MnCl2 + 2Cp2TiCl
Al + 3Cp2TiCl2 f AlCl3 + 3Cp2TiCl
(5)
(6)
(7)
2Cp2TiCl { k } (Cp2TiCl)2
(8)
8
k-8
One of the prominent features of the Met-Cp2TiCl2 solutions
is the presence of metal chlorides and their possible involvement
in chloride transfer reactions. This explains why the 1/5 peaks
of the Cp2TiCl2/Cp2TiCl2- redox couple in the metal-generated
cases are not completely coincident with those obtained for an
authentic sample of Cp2TiCl2. As seen in Figure 3, the peaks
recorded for Cp2TiCl2 are gradually shifted upon addition of
ZnCl2 to come in closer agreement with those recorded for
Zn-Cp2TiCl2. At the same time, this is also the origin of the
variation in the relative currents of the oxidation waves for the
different Met-Cp2TiCl2 solutions. Without doubt, the chloride
abstraction proceeds with greater ease from AlCl3 than from
ZnCl2 and MnCl2 since peak 1 attributed to the oxidation of
Cp2TiCl2- can only be observed in the former case (see Figure
1). The relative importance of this peak increases upon lowering
the sweep rate. Moreover, addition of excess Bu4NCl to any of
the solutions results in the transformation of (Cp2TiCl)2,
1/5: Cp2TiCl2- - e a Cp2TiCl2
(1)
(2)
(3)
(4)
+
2: (Cp2TiCl)2 - e a (Cp2TiCl)2
3: Cp2TiCl - e a Cp2TiCl+
4: Cp2Ti+ - e a Cp2Ti2+
Cp2TiCl and Cp2Ti+ to Cp2TiCl2 as shown in Figure 4.
-
To substantiate the working hypothesis that (Cp2TiCl)2 and
Cp2TiCl should be the principal TiIII species present in the
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7855

A R T I C L E S
Enemærke et al.
Figure 4. Cyclic voltammograms of 2 mM Zn-Cp2TiCl2 in the absence
(s) and presence (‚‚‚) of 20 mM Bu4NCl recorded at a sweep rate of 0.1
V s-1 in 0.2 M Bu4NPF6/THF.
Figure 6. Cyclic voltammograms of solutions containing Zn-Cp2TiCl2
in concentrations of 0.2 (s), 0.4 (- - -), 1 (- ‚ ‚ -), and 2 mM (‚‚‚)
recorded at a sweep rate of 1 V s-1 in 0.2 M Bu4NPF6/THF. Currents are
normalized with respect to concentration.
the present case one should also take into account that the
formation of Cp2TiCl2 may occur through competing chloride
transfers to Cp2TiCl+ from the metal chlorides generated in eqs
5-7.
The evidence for the existence of the equilibrium reaction,
eq 8, involving Cp2TiCl and (Cp2TiCl)2 is provided by voltam-
mograms recorded for Zn-Cp2TiCl2 at various concentrations.
As seen in Figure 6, the broad wave appearing at -0.8 V vs
Fc+/Fc consists of two processes, with the first one becoming
more dominant as the concentration increases. Thus, the first
peak 2 is attributed to the oxidation of the dimer, eq 2, and the
second peak 3 to the oxidation of the monomer, eq 3, with the
relative height being affected by the presence of the equilibrium
reaction in eq 8. Assignment of the broad wave to the oxidation
of trinuclear complexes is ruled out by the fact that exactly the
same type of splitting is observed for Al-Cp2TiCl2 and
Mn-Cp2TiCl2 as well as the metal-free Cp2TiCl/(Cp2TiCl)2
solution analyzed previously (see Figure 2).15
Figure 5. Cyclic voltammograms of 2 mM solutions of Zn-Cp2TiCl2
recorded at sweep rates of 0.1 (s), 10 (- - -), and 50 V s-1 (- ‚ ‚ -) in
0.2 M Bu4NPF6/THF. Currents are normalized with respect to sweep rate.
Met-Cp2TiCl2 solutions, a detailed cyclic voltammetric analysis
was carried out. The first important observation is that the
relative current of the Cp2Ti+ oxidation wave 4 decreases with
increasing sweep rate as illustrated in the case of Zn-Cp2TiCl2
in Figure 5. At the same time, the number of electrons involved
Another interesting point is that eqs 3 and 9 involving the
oxidation of Cp2TiCl followed by the father-son reaction cannot
be the only pathway for the generation of Cp2Ti+ during the
cyclic voltammetric sweep. Peak 4 is also apparent at high
concentrations and low sweep rates, although the oxidation goes
almost exclusively through (Cp2TiCl)2 in eq 2 under such
conditions. This implies that Cp2Ti+ may originate from the
1
in the 2/3 wave goes from /2 toward 1. Experiments carried
out at a sweep rate of 1 kV s-1 at a 10 µm glassy carbon
ultramicroelectrode confirm this tendency as wave 4 essentially
disappears. Thus, it is concluded that Cp2Ti+ is not present in
any of the Met-Cp2TiCl2 solutions initially but that it is
generated at the electrode surface as a consequence of chemical
reactions occurring during the sweep. In a previous electro-
chemical study on metal-free TiIII-based solutions, Laviron et
al.14d suggested that Cp2Ti+ is formed in a fast father-son
reaction (eq 9) involving chloride transfer from Cp2TiCl to
Cp2TiCl+, with the latter being formed in the electrode process
of eq 3.
+
fragmentation of (Cp2TiCl)2 as well. On account of the
observations that the reduction of Cp2TiCl2 is detectable on the
reverse sweep and that the peak current ratio for the oxidation
waves 2 and 4 approaches unity at high concentrations and low
sweep rates, the fragmentation is proposed to follow the pathway
in eq 10.
k10
+
(Cp2TiCl)2
{
} Cp2Ti+ + Cp2TiCl2
(10)
k-10
Competing chloride transfers from the metal chlorides to
(Cp2TiCl)2 are found to be slower processes and thus not
considered further (see the Supporting Information).
On the basis of the above information, we propose that a
common reaction scheme, a so-called mesh scheme,20 can be
set up for the three Met-Cp2TiCl2 systems (Scheme 1).
+
Cp2TiCl + Cp2TiCl+ { k9 } Cp2Ti+ + Cp2TiCl2
(9)
k-9
Indeed, the concomitant generation of Cp2TiCl2 in eq 9 is
detectable on the reverse sweep (see Figure 1). However, in
9
7856 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Reducing Species in Titanocene Halide Reductions
A R T I C L E S
Scheme 1. Mesh Scheme Describing the Reaction Mechanism for the Electrode Reactions of Met-Cp2TiCl2 (Met ) Zn, Mn, or Al) in THF
In the Supporting Information, detailed versions of this
scheme are provided for each metal. The mesh scheme
incorporates the four relevant electrochemical processes, eqs
1-4, along with a number of chemical reactions, including eqs
8-10. Among the latter, eq 8 is a key reaction because the
equilibrium constant, K8, determines the actual composition of
the solution while the associated rate constants are essential
parameters for describing the compositional changes occurring
at the electrode surface during a chemical or electrochemical
reaction. The oxidation of both Cp2TiCl and (Cp2TiCl)2 leads
ultimately to the formation of Cp2Ti+ and Cp2TiCl2. As no peak
corresponding to the reduction of Cp2TiCl+ can be monitored
in the cyclic voltammograms, even at high sweep rates, it is
proposed that Cp2TiCl+ reacts rapidly with Cp2TiCl or in
particular metal chlorides to yield Cp2TiCl2. Likewise, the
absence at low sweep rates of a cyclic voltammetric reduction
wave pertaining to Cp2Ti2+ formed upon oxidation of Cp2Ti+
in eq 4 suggests that Cp2Ti2+ will abstract chloride from the
metal chlorides. On the other hand, the lack of peak 1 in double
sweep experiments for Met ) Zn and Mn indicates that the
chloride transfer reactions may go the other way around too,
i.e., the metal chlorides can abstract chloride effectively from
most easily oxidized. In principle, it should be possible to outrun
all chemical reactions by applying high sweep rates. However,
in the present study we found that the largest applicable value
for quantitative purposes was about 50 V s-1; at higher sweep
rates the 2 and 3 waves broadened and merged, preventing us
from measuring the equilibrium constant of eq 8 directly. Other
techniques such as chronoamperometry and sampled current
voltammetry were considered as well,21 but they did not provide
information that was not obtainable by cyclic voltammetry.
All relevant potentials, heterogeneous and homogeneous rate
constants, and equilibrium constants have been extracted by
digital simulations (using the program DigiSim 3.03)22 of
experimentally obtained voltammograms recorded at different
sweep rates of 0.1-50 V s-1 and concentrations of 0.17-2.0
mM. For these kinds of measurements, it is important that the
number of experimental observations is large to diminish the
overall uncertainty. In addition, the whole collection of volta-
mmograms recorded for each solution of Met-Cp2TiCl2 has
been simulated simultaneously to obtain the most consistent set
of results. The main emphasis in the simulation procedure has
been to reproduce the relative currents and the peak potentials
with less attention paid to the shape of the relatively broad
curves. As described in the Supporting Information, the
procedure is based partly on a subjective judgment of which
values make the better fits because of the large number of
variables present.
-
Cp2TiCl2 produced upon the electrochemical reduction of
Cp2TiCl2 in eq 1. In general, TiIV- but not TiIII-based species
thus are capable of abstracting chloride from the metal chlorides
with the Al-Cp2TiCl2 system presenting a notable exception,
as peak 1 in this case is observable.
The adjustable parameters in the simulation procedure include
standard potentials, equilibrium constants, homogeneous and
heterogeneous rate constants, symmetry factors, and diffusion
coefficients. Of these, the last two are considered to be of less
importance, and thus they are kept fixed. The values of the
standard potentials, E°, and the heterogeneous rate constants,
ks, are very important for the quality of the fits, as they
essentially determine the position of the peaks. For ks we find
relatively small values less than 0.1 cm s-1; therefore, the
From a mechanistic point of view it is crucial to determine
the exact composition of the solutions. Unfortunately, voltam-
mograms recorded at low sweep rates will be influenced strongly
by the chemical reactions preceding or following the electrode
processes. For instance, the overall oxidation of the relatively
facile Cp2TiCl/(Cp2TiCl)2 equilibrium system will mainly go
through the dimer under such conditions, as this is the species
(20) Usually the presence of at least three redox states as well as chemical states
would be required for describing such a mechanism by a mesh scheme.
Thus, in the present case one might argue that the mechanistic scheme is
(21) Amatore, C.; Jutand, A.; Khalil, F.; M’Barki, M. A.; Mottier, L. Organo-
metallics 1993, 12, 3168.
(22) Rudolph, M.; Feldberg, S. W. DigiSim, version 3.03; Bioanalytical Systems,
Inc.: West Lafayette, IN.
of the ladder type, if the (Cp2TiCl)2 species with its combination of TiIII
+
and TiIV atoms is not considered as representing a distinct redox state (see
also: Evans, D. H. Chem. ReV. 1990, 90, 739).
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7857

A R T I C L E S
Enemærke et al.
Figure 7. Recorded (-) and simulated (‚‚‚) cyclic voltammograms of 2 mM Zn-Cp2TiCl2 at sweep rates of 0.2, 1, 5, and 20 V s-1 in 0.2 M Bu4NPF6/THF.
The simulation parameters employed can be found in Table 1 and in the Supporting Information.
Figure 8. Recorded (-) and simulated (‚‚‚) cyclic voltammograms of Zn-Cp2TiCl2 in concentrations of 0.17, 0.34, 0.80, and 1.80 mM at a sweep rate of
1 V s-1 in 0.2 M Bu4NPF6/THF. The simulation parameters employed can be found in Table 1 and in the Supporting Information.
electrode processes can be considered to be quasireversible for
most of the sweep rates employed. Thus, the influence on the
peak potentials from homogeneous follow-up or preceding
reactions is not as pronounced as if the electrode processes had
been reversible. On the other hand, the homogeneous kinetics
in terms of the rate constants involved exerts a large influence
on the peak currents and the wave pattern.
The fits are acceptable as illustrated in Figures 7 and 8 for
Zn-Cp2TiCl2. In general, the oxidative sweep is described more
successfully than the reductive one, as the simulations seem to
overestimate somewhat the reduction wave of Cp2TiCl2. This
indicates that the products formed during the oxidative sweep
are not all transformed to Cp2TiCl2 as assumed in the simula-
tions, but that some of them disappear through other routes.
Indeed, it is likely as discussed elsewhere15 that a side reaction
between Cp2Ti2+ and the solvent, THF, may take place. The
latter reaction was not incorporated in the simulations as its
effect on the determination of the other parameters in any case
would be small. The simulations also showed that the father-
son reaction involving Cp2TiCl+ in eq 9 will be outrun by
chloride transfer reactions, and accordingly, k9 was set equal to
the upper limit of 5 × 103 M-1 s-1 found previously for the
Cp2TiCl/(Cp2TiCl)2 case.15 A detailed description of the exact
model parameters used and a compilation of relevant fits can
be found in the Supporting Information.
The essential data extracted from the simulation procedure
are collected in Table 1. Note that all potentials listed are
referenced against Fc+/Fc, but they can easily be converted to
9
7858 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Reducing Species in Titanocene Halide Reductions
A R T I C L E S
Table 1. Relevant Data Extracted from Cyclic Voltammograms Recorded on Met-Cp2TiCl2 Solutionsa
Zn−Cp2TiCl2
Mn−Cp2TiCl2
Al−Cp2TiCl2
Cp2TiCl/(Cp2TiCl)2b
E°/V vs Fc+/Fc c
E12°/V vs Fc+/Fc c
E3°/V vs Fc+/Fc c
E4°/V vs Fc+/Fc c
-1.20 ( 0.03
-0.81 ( 0.03
-0.75 ( 0.03
-0.43 ( 0.03
0.04
0.02
0.008
0.015
-1.20 ( 0.03
-0.82 ( 0.03
-0.77 ( 0.03
-0.43 ( 0.03
0.04
0.02
0.008
0.015
-1.24 ( 0.03
-0.82 ( 0.03
-0.76 ( 0.03
-0.43 ( 0.03
0.04
0.02
0.008
0.015
-1.27 ( 0.04
-0.81 ( 0.02
-0.75 ( 0.01
-0.41 ( 0.02
0.04
0.02
0.008
0.015
k
k
k
k
s,1/cm s-1
s,2/cm s-1
s,3/cm s-1
s,4/cm s-1
K8/ M-1
k8/ M-1 s-1
k-8/s-1
3 (2-5) × 103
2 (1-10) × 104
6.7 (2-50)
3.2 × 107
3 (2-5) × 103
1 (0.5-5) × 105
33 (10-250)
2.2 × 107
3 (2-5) × 103
5 (2-10) × 105
167 (50-1000)
3.2 × 107
3 (2-7) × 103
2 (0.5-4) × 104
6.7 (2-17)
3.2 × 107
K9/M-1d
k9/M-1s-1
k-9/s-1
<5 × 103e
<5 × 103e
< 5 × 103e
< 1.6 × 10-4e
1 (0.5-1.5) × 103g
3 (2-5) × 102g
0.3 (0.13-1)g
<5 × 103f
<1.6 × 10-4e
1 (0.01-105) × 103
3 (1-50) × 102
0.3 (10-5-30)
<2.3 × 10-4e
1 (0.01-105) × 103
3 (1-50) × 102
0.3 (10-5-30)
<1.6 × 10-4f
1 (0.01-105) ×103
1 (0.1-5) × 102
0.1 (10-5-50)
K10/M
k10/s-1
k-10/ M-1 s-1
a
The tabulated values are the ones providing the best fits, while those given in parentheses describe the intervals of tolerance, as determined from a
manual adjustment of the parameters. To minimize the number of adjustable parameters, the uncertainties on the ks values were not determined, but in
b
c
general they can be assumed to be on the order of 50%. The results for Cp2TiCl/(Cp2TiCl)2 are taken from ref 15 (denoted “Cp2TiCl” therein). Standard
d
potentials are defined as reduction potentials and can be converted to SCE by adding 0.52 V. Calculated automatically by the DigiSim program from a
thermochemical cycle. The simulations are found to be largely unaffected by the exact value of k9, and accordingly the latter is set equal to the upper limit
determined for the Cp2TiCl/(Cp2TiCl)2 system. An upper limit can be determined from simulations; see ref 15. The simulations are found to be very
e
f
g
sensitive toward changes in these parameters because of the reversible reactions occurring between aluminum chloride and the different titanium-based
species.
SCE by adding 0.52 V.23 On the basis of a simple comparison
of the potential values that exhibit a variation of as much as
0.8 V, the electron-donating abilities of the different TiIII-based
dominant one. In our previous study we showed that Cp2TiCl2
could be generated from Cp2TiCl and free Cl- with a relatively
-
large formation equilibrium constant of 104 M-1 15
Still,
.
-
-
species should be in the order Cp2TiCl2 . (Cp2TiCl)2 >
Cp2TiCl2 will not be present in the Met-Cp2TiCl2 solutions,
as chloride is firmly bound in the corresponding metal chlorides.
Only in the case of Al-Cp2TiCl2 is the chloride transfer from
AlCl3 sufficiently fast (see the Supporting Information) to ensure
production of detectable amounts of Cp2TiCl2- during a cyclic
voltammetric sweep. On the other hand, Cp2TiCl2 can easily
be made the principal TiIII species in any of the solutions, if
Bu4NCl is added. The structure of the oxidized dimer
Cp2TiCl . Cp2Ti+ from a thermodynamic point of view.
However, such a consideration does not take into account that
the reactivity of an electron donor is dependent on other factors
as well, such as self-exchange reorganization energies, electro-
static work terms, and last but not least, inner-sphere character
of the transition state. In fact, the determination of heterogeneous
rate constants in the range of 0.008-0.04 cm s-1 for all four
species would suggest that the pertinent self-exchange reorga-
nization energies are relatively large.24 Usually this would imply
large intrinsic barriers and thus kinetically slow electron-transfer
processes, unless there are favorable interactions present between
the donor and acceptor in the transition state.
-
+
(Cp2TiCl)2 is relatively fragile, endowed by a rate constant
10 of 300 s-1 for its fragmentation. As shown in our previous
study, the latter process can be further accelerated by the
k
presence of a nucleophile such as chloride.15
Kinetic Studies. With the aim of characterizing the reactivity
and electron-donating abilities of the different TiIII-based
reductants, Cp2TiCl2-, (Cp2TiCl)2, Cp2TiCl, and Cp2Ti+, kinetic
studies were carried out on their reactions with benzyl chloride
(BnCl) and benzaldehyde (PhCHO). A description of the
According to the equilibrium data, the dominant species in
all Met-Cp2TiCl2 solutions are Cp2TiCl and (Cp2TiCl)2 while
-
there will be contribution from neither Cp2Ti+ nor Cp2TiCl2
.
It should be noted that the latter is the major species generated
in the electrochemical reduction of Cp2TiCl2.15 The value of 3
× 103 M-1 found for the equilibrium constant K8 is independent
of the metal considered. Actually, it is even equal to that
obtained for a metal-free solution,15 showing that the metal
chlorides exert no influence on the position of the dimerization
equilibrium, although the associated rate constants k8 and k-8
both increase by a factor of 25 going from Met ) Zn to Al. As
to the composition of Met-Cp2TiCl2 solutions, the distribution
of TiIII species in a 0.5 mM solution, for example, will be 0.22
mM Cp2TiCl and 0.14 mM (Cp2TiCl)2, while the corresponding
numbers for a 100 mM solution will be 4 and 48 mM.
Depending on the exact reaction conditions (catalytic or
stoichiometric),1-9 either of the two species may thus be the
-
experiments involving Cp2TiCl2 and BnCl is provided else-
where.15 The measurements were performed by following the
overall absorption decay of the TiIII species and/or the equivalent
buildup of TiIV species using excess of substrate.
Benzyl Chloride. The fact that the Met-Cp2TiCl2 solutions
are mixtures of Cp2TiCl and (Cp2TiCl)2 complicates the kinetic
analysis in the sense that their reactivities have to be studied
through a variation of the equilibrium concentrations. According
to eq 8 the larger the overall concentration of Met-Cp2TiCl2
becomes, the more dominant the dimer will be. In Figure 9, we
have collected kinetic traces for the buildup of TiIV using three
different concentrations of Zn-Cp2TiCl2 in the presence of
excess BnCl. The development is distinct in the sense that the
overall kinetics becomes slower as the concentration is raised
because of the influence from the equilibrium reaction in eq 8.
The same behavior is seen for the two other Met-Cp2TiCl2
solutions.
(23) Enemærke, R. J.; Daasbjerg, K.; Skrydstrup, T. J. Chem. Soc., Chem.
Commun. 1999, 343.
(24) Enemærke, R. J.; Hertz, T.; Skrydstrup, T.; Daasbjerg, K. Chem. Eur. J.
2000, 6, 3747.
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7859

A R T I C L E S
Enemærke et al.
Analysis of the traces recorded was carried out on the basis
of the above equations on the assumption that the reactions of
eqs 12 and 14 would be so fast that their only influence on the
kinetics would be through the reaction stoichiometry. The
simulation program Gepasi 3.21 was employed,25 giving a
consistent set of data with the best fits depicted in Figure 9.
Attempts to encompass a situation in the simulations, where
the reactivity of the dimer would be negligible and the monomer
with its free coordination site the only reactive species, lowered
the quality of the fits. Furthermore, it was found that the
equilibrium constant K8 in that instance could hardly be larger
than 30 M-1, inconsistent with the value of 3 × 103 M-1
determined from the cyclic voltammetric measurements.26
In Table 2, we have collected the data extracted from the
above measurements. As seen, the reactivity of Cp2TiCl and
(Cp2TiCl)2 is largely the same with k11 ) 0.66 M-1 s-1 and k13
) 0.80 M-1 s-1, independent of the metal. A few kinetic
experiments were carried out in the presence of 0.2 M
Bu4NPF6 to have exactly the same experimental conditions as
in the cyclic voltammetric experiments. It was found that while
k11 was more than halved going to 0.30 M-1 s-1, there was
only a slight decrease in k13 to give 0.60 M-1 s-1. From an
electrostatic point of view, it would indeed be expected that
the largest influence should be seen on the smallest species.
Still, the effects are small considering that a lowering of the
rate constant by a factor of 2.5 only corresponds to an increase
in the activation energy of 0.5 kcal mol-1. This also indicates
that the charge separation in the transition state should be small.
The kinetic study of the cation Cp2Ti+ is much simpler since
it can easily be generated without any interference from the
other two TiIII reagents by adding TlPF6 to a solution of metal-
free Cp2TiCl/(Cp2TiCl)2 (see the Experimental Section). The
mechanism for the reaction between Cp2Ti+ and BnCl is
proposed to follow eqs 16-18.
Figure 9. Kinetic traces recorded for the buildup of TiIV at λ ) 535 nm
in THF for the reactions of 87 mM BnCl with 0.8 (s), 1.8 (- - -), and
4.4 mM (; ; ;) solutions of Zn-Cp2TiCl2. Absorbances are normalized
(using final absorbance ) 1) to facilitate comparison. The dotted curves
are the best fits based on simulation of eqs 8 and 11-15, providing k11
)
0.66 M-1 s-1 and k13 ) 0.80 M-1 s-1
.
The mechanistic basis for carrying out the fitting procedure
consists of eq 8 along with eqs 11-15, involving both Cp2TiCl
and (Cp2TiCl)2 as reactive species.
k11
9
Cp2TiCl + BnCl
8 [Cp2TiCl+ + Cl- + Bn•] f
Cp2TiCl2 + Bn• (11)
Cp2TiCl + Bn• f [Cp2TiCl+ + Bn-] f Cp2TiClBn (12)
k13
(Cp2TiCl)2 + BnCl
9
8 [(Cp2TiCl)2+ + Cl- + Bn•] f
Cp2TiCl + Cp2TiCl2 + Bn• (13)
(Cp2TiCl)2 + Bn• f [(Cp2TiCl)2+ + Bn-] f
Cp2TiCl + Cp2TiClBn (14)
Cp2TiClBn + BnCl f Cp2TiCl2 + Bn2 (15)
k16
9
Cp2Ti+ + BnCl
8 [Cp2Ti2+ + Cl- + Bn•] f
Cp2TiCl+ + Bn• (16)
Cp2Ti+ + Bn• f [Cp2Ti2+ + Bn-] f Cp2TiBn+ (17)
Cp2TiBn+ + BnCl f Cp2TiCl+ + Bn2
The hypothesis is that in the actual electron-transfer step between
Cp2TiCl and BnCl depicted in eq 11, the benzyl radical, Bn•, is
formed along with chloride which will be captured by Cp2TiCl+
to afford Cp2TiCl2 as revealed by previous cyclic voltammetric
studies.15 In principle, these two processes might take place
concertedly, which is the reason for showing the initial products
within brackets. Bn• is further reduced to the benzyl anion, Bn-,
in a fast second electron-transfer process (eq 12), and as this
anion is expected to be a strong ligand we suggest that
Cp2TiClBn is formed (possibly concertedly). The equivalent
reactions of (Cp2TiCl)2 are depicted in eqs 13 and 14. Formally,
the electron transfer from (Cp2TiCl)2 to BnCl in eq 13 would
produce (Cp2TiCl)2+, Cl-, and Bn• if it was not for the
immediate follow-up reaction taking place between Cl- and
(Cp2TiCl)2+. According to our previous cyclic voltammetric
studies,15 this reaction will afford Cp2TiCl and Cp2TiCl2. The
fast reduction of Bn• by (Cp2TiCl)2 shown in eq 14 leads to the
formation of Cp2TiCl and Cp2TiClBn. In the final step, eq 15,
Cp2TiClBn produced in eqs 12 and 14 reacts in a relatively
slow process with a second molecule of BnCl to afford
Cp2TiCl2 and dibenzyl, Bn2, as the final products in accordance
with experimental observations (see the Experimental Section).
(18)
In the initial electron-transfer step (eq 16), Bn• is formed along
with chloride, which according to the cyclic voltammetric
analysis will coordinate at the TiIV nucleus either concertedly
or in an immediate follow-up reaction, thereby generating
Cp2TiCl+. The benzyl radical is reduced by Cp2Ti+ in eq 17,
affording Cp2TiBn+, which upon further reaction with BnCl (eq
18) will produce the final product Bn2 along with Cp2TiCl+.
The rate constant k16 can easily be determined to be 0.44 M-1
s-1 under pseudo-first-order conditions as shown in Figure 10.
This result is also included in Table 2.
A comparison of the reactivity of the different TiIII-based
species shows quite interestingly (see Table 2) that Cp2Ti+
exhibits almost the same reactivity as (Cp2TiCl)2 and Cp2TiCl,
(25) Mendes, P. Gepasi, version 3.21; Virginia Tech: Blacksburg, VA, 1996-
1999. The program may be downloaded via the Internet at http://
www.gepasi.org.
(26) In cyclic voltammetry, Bu4NPF6 was used as the supporting electrolyte
but the presence of this salt in the kinetic experiments did not improve the
quality of the fits.
9
7860 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Reducing Species in Titanocene Halide Reductions
A R T I C L E S
Table 2. Rate Constants kobs for the Reaction between TiIII-Based Species and Benzyl Chloride or Benzaldehyde Measured in THF at 20
°C along with Estimated Electron Transfer Rate Constants, kET, and kobs/kET Ratiosa
BnCl
kET
PhCHO
kET
kobs
kobs/kET
kobs
kobs/kET
2.8 × 10-6
1.5 × 10-10
4.2 × 10-11
4.5 × 10-14
<6.8 × 103
5.3 × 109
1.6 × 1010
9.8 × 1012
<0.1
70
<2
1.5 × 10-12
1.2 × 10-28
4.7 × 10-31
<6.7 × 1010
<4.3 × 1030
-b
Cp2TiCl2
<0.019
0.80
0.66
c
(Cp2TiCl)2
Cp2TiClc
Cp2Ti+d
5.8 × 1029
0.44
a
b
All rate constants are given in units of M-1 s-1; see text for description of the methods used for the determination of kET
.
Generated electrochemically
from Cp2TiCl2; a standard potential of -1.27 V vs Fc+/Fc is employed in the calculation of kET; see ref 15. Generated as a mixture of Cp2TiCl and
(Cp2TiCl)2 in Zn-Cp2TiCl2 solutions. Generated as described in the text.
c
d
the addition of BnCl gave no such indications, nor did we see
changes in the equilibrium ratio of Cp2TiCl and (Cp2TiCl)2.
Benzaldehyde. A series of kinetic experiments was carried
out on the reaction between Zn-Cp2TiCl2 (i.e., Cp2TiCl and
(Cp2TiCl)2) and PhCHO. This reaction is of great synthetic
importance, as it presents a convenient route for accomplishing
pinacol couplings.2 The two other TiIII-based species, Cp2Ti+
and Cp2TiCl2-, were also included in this part of the study,
although the measurements pertaining to Cp2Ti+ being irrepro-
-
ducible had to be renounced, and the reactivity of Cp2TiCl2
was found to be so low that only an upper limit of 0.1 M-1 s-1
could be provided for the rate constant.
The reaction pathway in this case is assumed to consist of
an initial electron transfer from Cp2TiCl (eq 19) or (Cp2TiCl)2
(eq 20) to benzaldehyde, followed by dimerization of the thus
Figure 10. Kinetic trace recorded for the buildup of TiIV at λ ) 535 nm
formed ketyl radicals (eq 21).
in THF for the reaction of 87 mM BnCl with 0.5 mM Cp2Ti+. The dotted
curve is the best exponential fit based on eqs 16-18, providing k16 ) 0.44
M-1 s-1
.
whereas Cp2TiCl2- is a much less efficient reagent. Presumably
-
this can be explained by the fact that Cp2TiCl2 has no free
coordination site so that a binding of the acceptor to the titanium
nucleus cannot take place. This leads to an intrinsically slow
reaction possessing very little inner-sphere electron-transfer
character. On the other hand, for Cp2Ti+ and Cp2TiCl easy
replaceable solvent molecules will occupy the free ligand sites,
facilitating a coordination of benzyl chloride during the reaction
as illustrated by the structures IV and V in Chart 2. Note that
within such structures chloride can readily be transferred as a
ligand to the TiIV nucleus upon electron transfer.
In this sense, it is surprising that (Cp2TiCl)2, having no free
coordination site in the assumed symmetric structure II (known
from the solid phase, see Chart 1),11 should exhibit the same
reactivity as Cp2TiCl. Previous stereochemical investigations
have suggested that a pentavalent coordination is a possibility,2a,i
but we think that a more plausible explanation could reside in
the dimeric structure of (Cp2TiCl)2 being so weak that the same
sort of coordination as suggested for Cp2TiCl might take place
during the electron transfer (see VI, Chart 2). There is even the
possibility that the dimeric structure in THF from the very
beginning is of the half-open type depicted as III in Chart 1,
thereby leaving room for an easily accessible coordination site.
Following the electron-transfer process and the transfer of
chloride to the TiIV nucleus, the dimeric structure will be broken
completely as outlined in eq 13. It is important to note that
within this picture of concerted electron and ligand transfers
the species given in the brackets of eqs 11-17 will not exist.
Also investigated was whether the formation of any precursor
complexes would be detectable in cyclic voltammetry. However,
voltammograms of Zn-Cp2TiCl2 solutions recorded just after
Note that the TiIV species formed upon electron transfer will
be strongly coordinated at the oxygen and will not be liberated
unless other strong electrophiles such as trimethylsilyl chloride
or protons are added to the solution.1f Moreover, it is assumed
that the dimeric structure of (Cp2TiCl)2 will be broken com-
pletely as the reaction in eq 20 progresses. In previous studies
addressing the stereochemical outcome of pinacol couplings,2
the possibility has been put forth that a pentavalent coordination
involving two ketyl radicals can take place at the titanium
nucleus in either the dimer II2i or the trinuclear complex I2a,i
without the need to break bonds. Similar complexes have also
been proposed, where two of the chlorides in I have been
displaced by ketyl radicals.1f,2c,d The prominent feature of such
arrangements is that they provide straightforward explanations
for the high dl:meso ratios obtained. In contrast to this view
stands the above cyclic voltammetric analysis pointing toward
a rather fragile structure of the oxidized dimer (k10 ) 300 s-1).
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7861

A R T I C L E S
Enemærke et al.
response. The same type of measurements could not be carried
out for PhCHO because of the fast reactions occurring in eqs
19-21. From the best fits included in Figure 11 it is found that
k19 < 2 M-1 s-1 and k20 ) 70 M-1 s-1, thus showing that the
influence from eq 19, in general, will be rather small.
The rate constants obtained are all included in Table 2. While
the reactivity of (Cp2TiCl)2 and Cp2TiCl in the case of benzyl
chloride was found to be largely the same, this is certainly not
the case for benzaldehyde, where the reactivity of (Cp2TiCl)2
surpasses that of Cp2TiCl by at least a factor of 35. On this
basis the arrangement of the dimer shown as VII in Chart 2,
where one of the TiIII atoms coordinates to the carbonyl group
in benzaldehyde during the electron transfer, seems to present
a plausible structure. Simulation of the reaction schemes for
BnCl, eqs 8 and 11-15, and PhCHO, eqs 8 and 19-21 (with
k19 ) 2 M-1 s-1), using the Gepasi program shows that in the
case of a 100 mM Zn-Cp2TiCl2 solution (containing 100 mM
of the substrates) 87.5 and 99.5%, respectively, will go through
(Cp2TiCl)2. For a 0.5 mM solution, the corresponding percent-
ages are 28.5 and 85.5%. The turnover points are reached at
Figure 11. Kinetic traces recorded for the decay of TiIII at λ ) 800 nm in
THF for the reactions of 15 mM PhCHO with 1 (s), 2 (- - -), and 5
mM (; ; ;) solutions of Zn-Cp2TiCl2. The initial absorption is set at
0.0148 in all three cases to facilitate comparison. The dotted curves are the
best fits based on simulation of eqs 8 and 19-21, providing k19 < 2 M-1
s-1 and k20 ) 70 M-1 s-1
.
Zn-Cp2TiCl2 concentrations of 2.3 × 10-3 and 4.8 × 10-5
M
Chart 2. Proposed Structures for the Coordination of Benzyl
Chloride to Cp2Ti+ (IV), Cp2TiCl (V) and (Cp2TiCl)2 (VI) and of
Benzaldehyde to (Cp2TiCl)2 (VII) during the Electron Transfer
Reactions
for BnCl and PhCHO, respectively, on the assumption that the
concentrations of the substrates are 100 mM. With respect to
the product formation in the pinacol coupling reaction, it may
thus be concluded that (Cp2TiCl)2 will be the species respon-
sible, unless catalytic conditions are employed. On the other
hand, this should have no implications for the dl:meso diaste-
reoselectivities as the dimerization step in eq 21 is suggested
to be the same for the two pathways involving Cp2TiCl and
(Cp2TiCl)2. This point will be further discussed in a forthcoming
publication.27
The kinetic studies on the reactions with BnCl and PhCHO
show that the reactivity order of the four principal TiIII species
is: (Cp2TiCl)2 J Cp2TiCl ≈ Cp2Ti+ . Cp2TiCl2-. From a
thermodynamic point of view, it would be expected that
-
Cp2TiCl2 with a 0.8 V lower oxidation potential than that of
Cp2Ti+ should be the best electron donor. This indicates that at
least the electron-transfer processes involving the three former
species must be of the inner-sphere type. In this context, it is
interesting to compare the extracted rate constants for the
reaction between the different TiIII-based species and BnCl or
PhCHO, denoted generally kobs (i.e., depending on the reaction
considered it may correspond to k11, k13, k16, k19, or k20), with
those expected for electron-transfer reactions involving well-
characterized electron donors such as aromatic radical anions.
In the case of BnCl, the latter rate constants denoted kET may
The fragmentation is even faster in the presence of a nucleophile
such as chloride,15 and there would be no reasons to expect
that the ketyl radical should exert any weaker effect. Thus,
kinetically there seem to be no arguments for the view that the
dimeric structure should persist sufficiently long upon electron
transfer to influence the dimerization step in eq 21.
The traces recorded for the overall decay of the TiIII species
for varying concentrations of Zn-Cp2TiCl2 in the presence of
excess PhCHO are collected in Figure 11. This figure shows
that the kinetics is not simple because the decay consists of a
fast initial part followed by a long tail. The further analysis
was based on the assumption that eq 8 and thus the equilibrium
ratio of Cp2TiCl and (Cp2TiCl)2 would be unaffected by the
presence of PhCHO. This assumption seems reasonable on
account of the observation that at least an aliphatic aldehyde
such as propanal had no effect on the cyclic voltammetric
be deduced from the log kET vs E° plot shown in ref 24, where
A
E° denotes the standard potential of the aromatic compound.
A
Unfortunately, the same procedure for finding kET cannot be
employed in the case of PhCHO, since the electron transfer from
the aromatic radical anions to form the radical anion of
benzaldehyde involves an energetically unfavorable equilibrium
reaction. We therefore chose to calculate the required kET values
between outer-sphere electron donors with E° values as listed
in Table 1 and benzaldehyde on the basis of the equations of
Marcus and Eyring.28 The E° of benzaldehyde is -2.37 V vs
Fc+/Fc as measured by cyclic voltammetry, the reorganization
(27) Enemærke, R. J.; Larsen, J.; Hjøllund, G. H.; Skrydstrup, T.; Daasbjerg,
K. Manuscript in preparation.
(28) Eberson, L. Electron Transfer Reactions in Organic Chemistry Springer-
Verlag: Berlin, 1987.
9
7862 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Reducing Species in Titanocene Halide Reductions
A R T I C L E S
energy of the reactions is assumed to be 10 kcal mol-1, and for
the collision frequency a value of 3 × 1011 M-1 s-1 is used.28
In Table 2, the kobs/kET ratios are collected for BnCl and
PhCHO with values ranging from less than 6.8 × 103 to 4.3 ×
1030. Usually, such large numbers would be indicative of
substantial inner-sphere character of the transition state,29 but
in the present case one should exert great care in doing this
kind of assessment, since a thorough analysis should also take
into account differences in the reorganization energies, all
relevant work terms, as well as the exact association constants
for the formation of precursor complexes.24 One could easily
imagine that the latter would be of particular importance for
the reactions of PhCHO. In addition, the finding of relatively
small heterogeneous rate constants in Table 1 indicates that the
reorganization energy for the TiIII species is substantially larger
than that for aromatic radical anions. At the present stage, we
therefore prefer to develop a relative scale of uncorrected
containing 2 equiv of TlCp. While the solution was refluxed for 15
min it turned green, and a precipitation of TlCl occurred. The separation
was carried out using syringe techniques. The Zn- and Mn-Cp2TiCl2
solutions were prepared by adding THF to a flask thoroughly flushed
with argon and containing Cp2TiCl2 besides excess metal. The solutions
were stirred for 30 min until a turquoise green color had become
apparent. For preparing Al-Cp2TiCl2, activation of the aluminum foil
using Hg(NO3)2 in THF was required. The activated foil was washed
three times with THF before use. The preparation of Cp2TiPF6 was
accomplished by adding 1.1 equiv of TlPF6 dissolved in THF to a metal-
free solution of Cp2TiCl/(Cp2TiCl)2. The mixture was stirred and heated
shortly to the boiling point; blue crystals of Cp2TiPF6 along with white
crystals of TlCl then precipitated. The liquid was removed, and the
crystals were washed twice with THF. A large amount of THF was
subsequently added, and after 30 min of stirring the solution turned
pale blue. The concentration of Cp2TiPF6 in this saturated solution was
approximately 2 mM as determined from peak current measurements.31
Apparatus. Most of the electrochemical equipment was home-built,
and a description of the experimental setup is provided in ref 32. The
working electrode was a glassy carbon disk of diameter 1 mm. The
electrode surface was polished using 0.25 µm diamond paste (Struers
A/S), followed by cleaning in an ethanol bath. The counter electrode
consisted of a platinum coil melted into glass, while a silver wire in a
sintered glass containing THF/0.2 M Bu4NPF6 + 0.02 M Bu4NI served
as the reference electrode. All potentials were reported against the
ferrocenium/ferrocene (Fc+/Fc) redox couple, the potential of which
has been measured to be 0.52 V vs SCE in THF/0.2 M Bu4NPF6.23 All
handling of chemicals was performed on a vacuum line, and at no point
during the different operations was the interference of oxygen allowed.
The ohmic drop was compensated with a positive feedback system
incorporated in the potentiostat. The kinetic traces were recorded by
means of a fiber-optic spectrometer, model S1000 (dip-probe), from
Ocean Optics, using a light path length of 1 cm.
k
obs/kET ratios, which are still very useful for carrying out
comparisons, rather than paying further attention at the absolute
values.
For both BnCl and PhCHO, the kobs/kET ratios decrease in
the order Cp2Ti+, Cp2TiCl, (Cp2TiCl)2, and Cp2TiCl2-. This
would imply that there should be substantially more inner-sphere
character in the reactions of Cp2Ti+ than those of Cp2TiCl,
(Cp2TiCl)2, and in particular Cp2TiCl2-. In fact, this is in line
with expectations since the cation Cp2Ti+ compared with the
-
anion Cp2TiCl2 would be expected to have a stronger
coordination in the transition state to the developing Cl- in BnCl
and O- in PhCHO. From both a kinetic and thermodynamic
-
point of view, the features of Cp2TiCl2 (E° ) -1.27 V vs
Fc+/Fc; kobs < 0.019 M-1 s-1 and kobs/kET < 6.8 × 103 for BnCl)
are comparable with those of SmI2 (E° ) -1.41 V vs Fc+/Fc;
kobs ) 0.02 M-1 s-1 and kobs/kET ) 250 for BnCl),24 although
it is hard to carry out a direct comparison of the kobs/kET ratios
for the reasons given above. Finally, it may be concluded that
the inner-sphere character of the processes involving PhCHO
is stronger than for BnCl and that this feature is getting more
pronounced as the thermodynamic electron-donating abilities
of the TiIII complex are lowered.
Procedure. In the cyclic voltammetric experiments, 0.77 g of
Bu4NPF6 (2.0 mmol) and a small magnetic bar were added to the
electrochemical cell. The cell was closed and flushed thoroughly with
argon for 10 min. Typically, 9 mL of freshly distilled THF and 1 mL
of the appropriate standard solution containing the compound of interest
were added to the cell using a syringe, and the solution was stirred for
30 s. Special care must be taken when recording voltammograms of
the Met-Cp2TiCl2 solutions, since the metal ions present may be
reduced at low potentials and cause deleterious adsorption of metal on
the electrode surface. In general, potentials below -1.3 V vs Fc+/Fc
should be avoided. At the end of each series of experiments a small
amount of ferrocene was added, and the potential of the Fc+/Fc couple
was measured.
Experimental Section
Materials. Most chemicals were of commercial origin unless
otherwise noted. Benzyl chloride and benzaldehyde were vacuum
distilled before use. THF was distilled over sodium and benzophenone
under an atmosphere of dry nitrogen. Argon (99.99% purity) was passed
through a column of P2O5 (Sicapent). Tetrabutylammonium hexafluo-
rophosphate, Bu4NPF6, was prepared from a hot aqueous solution
containing tetrabutylammonium hydrogen sulfate and potassium hexa-
fluorophosphate. The precipitate was filtered and recrystallized from
ethyl acetate and pentane. Tetrabutylammonium chloride, Bu4NCl, was
recrystallized twice using THF. The white crystals were dried under
vacuum at 60 °C for 24 h and stored in a glovebox. All handling of
Bu4NCl had to occur in a dry atmosphere because of its hygroscopic
properties. Thallium hexafluorophosphate, TlPF6, was dried under
vacuum at 60 °C for 24 h. Thallium cyclopentandienide, TlCp, was
sublimated under vacuum at 80 °C.30 Bis(cyclopentadienyl)titanium
dichloride, Cp2TiCl2, was recrystallized from toluene. The metal-free
solution of Cp2TiCl/(Cp2TiCl)2 was synthesized as described previ-
ously.10 Typically, TiCl3 was suspended in THF and added to a solution
In the kinetic experiments, the UV-vis dip-probe was mounted
vertically in a two-necked cell containing a small magnetic bar.32 The
cell was closed and flushed with argon for 10 min before 9 mL of
freshly distilled THF was added. From a standard solution containing
the appropriate Met-Cp2TiCl2, typically 1 mL was transferred to the
cell. Benzyl chloride or benzaldehyde was added in excess, while the
solution was stirred vigorously. The decay of TiIII at the wavelength λ
) 800 nm and the concomitant buildup of TiIV at λ ) 535 nm were
followed. Analysis of the products formed in the reaction of
Zn-Cp2TiCl2 with benzyl chloride showed (after acidic quenching)
formation of both dibenzyl and toluene, the ratio of which increased
as the reaction time was prolonged. This indicates that the reaction
shown in eq 15 is relatively slow occurring in the time scale of hours.
Conclusions
Detailed cyclic voltammetric studies of Zn-, Mn-, and
Al-reduced Cp2TiCl2 solutions in THF show that the principal
(31) Bard, A. J.; Faulkner, R. Electrochemical Methods, Fundamentals and
Applications; Wiley: New York, 1980.
(32) Pedersen, S. U.; Christensen, T. B.; Thomasen, T.; Daasbjerg, K. J.
Electroanal. Chem. 1998, 454, 123.
(29) Lund, H.; Daasbjerg, K.; Lund, T.; Pedersen, S. U. Acc. Chem. Res. 1995,
28, 313.
(30) Nielson, A. J.; Rickard, C. E. F.; Smith, J. M. Inorg. Synth. 1990, 28, 315.
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7863

A R T I C L E S
Enemærke et al.
TiIII species present are a mixture of Cp2TiCl and (Cp2TiCl)2,
independent of the metal considered; the equilibrium constant
of dimerization is determined to be 3 × 103 M-1. There are no
indications of trinuclear complexes or ionic clusters involving
Cp2Ti+ as suggested previously. Nor do we see the formation
reactions of Cp2Ti+ is clearly much higher than those of
(Cp2TiCl)2, Cp2TiCl, and in particular Cp2TiCl2-. As for the
electron acceptor, the inner-sphere character of the transition
state increases going from benzyl chloride to benzaldehyde, and
it is suggested that Cl in benzyl chloride and O in benzaldehyde
could function as bridges between the reactants in the transition
state. In a forthcoming publication27 the focus will be on metal-
reduced solutions of Cp2TiBr2 and Cp2TiI2 as well as the factors
influencing the reactivities and diastereoselectivities of the
pinacol coupling reaction involving benzaldehyde.
-
of Cp2TiCl2 as found in the electrochemical reduction of
Cp2TiCl2, because chloride now becomes firmly bound by the
corresponding metal chlorides. According to the standard
potentials determined in cyclic voltammetry the electron-
donating ability of the different TiIII-based species would be
-
expected to be in the order Cp2TiCl2 . (Cp2TiCl)2
>
Cp2TiCl . Cp2Ti+. However, in reality the reactivity order is
(Cp2TiCl)2 J Cp2TiCl ≈ Cp2Ti+ . Cp2TiCl2-, as assessed in
their reactions with benzyl chloride and benzaldehyde. In
general, it is found that (Cp2TiCl)2 will be the species
responsible for the product formation, unless catalytic conditions
characterized by low concentrations of TiIII are employed. None
of the reactions studied in this article proceed by an outer-sphere
electron-transfer pathway, and the inner-sphere character of the
Acknowledgment. Torben Bo Christensen is thanked for
diligent technical assistance.
Supporting Information Available: Text giving a description
of the model parameters used in the simulations and figures
giving a compilation of relevant fits (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA0491230
9
7864 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004