Titanocene-catalysed, selective reduction of ketones in aqueous media. A safe, mild, inexpensive procedure for the synthesis of secondary alcohols via radical chemistry

Article abstract of DOI:10.1016/S0040-4039(02)02703-X

We report here a novel procedure for the reduction of ketones to secondary alcohols using catalytic quantities of commercially available Cp₂TiCl₂, inexpensive Zn dust and water as proton source. Mechanistically the reaction presumably proceeds via titanoxy radicals. In practice this reduction process has significant advantages: it shows an interesting selectivity pattern, takes place under mild conditions using safe, cheap reagents and does not require anhydrous solvents. The proton-donor activity of water under these conditions avoids the use of the frequently poisonous hydrogen-atom donors generally required to reduce free radicals. This procedure is also highly convenient for synthesising deuterium-labelled alcohols employing relatively inexpensive D₂O as deuterium source.

Full text of DOI:10.1016/S0040-4039(02)02703-X

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 1079–1082
Titanocene-catalysed, selective reduction of ketones in aqueous
media. A safe, mild, inexpensive procedure for the synthesis
of secondary alcohols via radical chemistry
Alejandro F. Barrero,a Antonio Rosales,a Juan M. Cuerva,a Andreas Gansa¨uerb and
J. Enrique Oltraa,*
aUniversidad de Granada, Departamento de Qu´ımica Orga´nica, Facultad de Ciencias, E-18071 Granada, Spain
bRheinische Friedrich-Wilhelms-Universita¨t, Kekule´-Institut fu¨r Organische Chemie und Biochemie, Gerhard-Domagk-Str. 1,
D-53121 Bonn, Germany
Received 2 November 2002; revised 22 November 2002; accepted 24 November 2002
Abstract—We report here a novel procedure for the reduction of ketones to secondary alcohols using catalytic quantities of
commercially available Cp2TiCl2, inexpensive Zn dust and water as proton source. Mechanistically the reaction presumably
proceeds via titanoxy radicals. In practice this reduction process has significant advantages: it shows an interesting selectivity
pattern, takes place under mild conditions using safe, cheap reagents and does not require anhydrous solvents. The proton-donor
activity of water under these conditions avoids the use of the frequently poisonous hydrogen-atom donors generally required to
reduce free radicals. This procedure is also highly convenient for synthesising deuterium-labelled alcohols employing relatively
inexpensive D2O as deuterium source. © 2003 Elsevier Science Ltd. All rights reserved.
The reduction of ketones to secondary alcohols is a
reaction of general interest in organic chemistry. Apart
from some chemo-biological methods,1 this reduction is
usually achieved by catalytic hydrogenation, by dissolv-
ing metals, or employing metal hydrides.2 Another
method, based upon hydrogen transfer from alcohols
catalysed by either aluminum alkoxides (the Meerwein–
Ponndorf–Verley reduction),3 or complexes containing
RuII and other metals has proved to be very fruitful,
especially for asymmetric synthesis.4–7 On the other
hand, a catalytic procedure for the chemical reduction
of ketones taking advantage of virtually free and envi-
ronmentally sound water as proton source has not until
now been developed. Although it is known that ketones
can be reduced by low-valent species of transition
metals such as samarium,8–10 chromium,11,12 and tita-
nium in aqueous media,13 the procedures reported to
date have not proved to have wide synthetic utility,
possibly because they require stoichiometric amounts of
the low-valent metals.
In 1994, RajanBabu and Nugent reported the selective
generation of free radicals from epoxides using
titanocene(III),14 and subsequently Gansa¨uer et al.
developed the catalytic version of this reaction employ-
ing 1,4-cyclohexadiene as hydrogen source.15 While
applying this method to the synthesis of natural ter-
penoids we recently found that tertiary radicals were
efficiently reduced in the presence of titanocene(III) and
Keywords: electron transfer; free radicals; homogeneous catalysis;
ketones; titanium.
* Corresponding author. Tel.: 34 958 24 80 91; fax: 34 958 24 84 37;
e-mail: joltra@ugr.es
Scheme 1. Hypothetic mechanism of the titanocene(III)-pro-
moted reduction of ketones in aqueous media.
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02703-X

1080
A. F. Barrero et al. / Tetrahedron Letters 44 (2003) 1079–1082
Table 1. Relative proportionsa (%) of compounds
obtained by titanocene(III)-promoted reduction of 5a–11a
in aqueous media
water,16 suggesting that hydrogen-atom donors such as
1,4-cyclohexadiene are not necessarily required for the
reduction of free radicals under these conditions. These
observations prompted us to attempt a novel procedure
for the reduction of ketones via ketyl radicals using
catalytic quantities of titanocene in aqueous media. We
considered that such a method would be inexpensive,
mild and safe, avoiding the use of strong bases (alkaline
metals, metal hydrides and alkoxides), dangerous
molecular hydrogen and carcinogenic 1,4-cyclohexadi-
ene. Additionally, the organometallic nature of the
catalyst might facilitate the subsequent development of
an asymmetric version using chiral organic ligands.
Substrate
Metal (equiv.)
Mn (8)
Time (h)
24
Products (%)
Our initial hypothesis (Scheme 1) was based upon the
assumption that even in the presence of water the
reaction between ketones and a trinuclear complex of
titanocene(III) such as 1,† by single electron transfer
(SET), should provide titanoxy radicals such as 2,
closely related to those proposed for the pinacol cou-
pling of aromatic and a,b-unsaturated aldehydes.20
5a
5a (B5), 5b
(\95)
6a
6a
Mn (8)
Zn (8)
48
24
6a (83), 6b (17)
6a (B5), 6b
(\95)
7a
8a
Zn (8)
Zn (8)
24
24
7a (B5), 7b
(\95)
8a (B5), 8b
(\95)
With ketones, however, the higher steric hindrance
inherent in these compounds should delay pinacol cou-
pling, and thus 2 could evolve to give the alkyl–TiIV
complex 3 by radical coupling with a second unit of
Cp2TiCl released from 1 in the vicinity of the ketyl
radical. Subsequently the organometallic derivative 3
would be hydrolysed in an irreversible step leading to 4,
which should give the corresponding secondary alcohol
after acidic quenching.
9a
10a
Zn (8)
Zn (8)
Zn (8)
Zn (8)
60
60
60
60
9a (20), 9b (80)
10a (100)
11a (18), 11b (82)
9a (10), 9b (40),
10a (50)
11a
9a+10ab
er\10a (50)
a Reductions were performed by treatment with Cp2TiCl2 (3 equiv.
per ketone) and Mn or Zn, in THF/H2O (4/1) at rt. Relative
proportions were determined on the basis of the 1H NMR spectra
of the mixtures formed in every experiment.
b Equimolar mixture.
To check our hypothesis we chose acetophenone. When
this ketone was treated with titanocene(III) generated
in situ,† in dry THF we obtained the pinacol coupling
product 2,3-diphenyl-2,3-butanediol.‡ In contrast, when
water was added to the reaction medium, acetophenone
was reduced to 1-phenylethanol as we expected. At
least 2 equiv. of Cp2TiCl2 were required to complete
this reduction, supporting the stoichiometric relation-
ships suggested in Scheme 1.
Thus we obtained alcohols 5b–9b and 11b in relative
proportions ranging from 80% to >95% (Table 1).
Throughout these experiments we made the following
observations: (i) the method seemed to be generally
valid for the reduction of aromatic ketones and cyclo-
hexanone, but failed with the acyclic ketone 10a, which
was recovered unchanged; (ii) the reaction rates
increased when Zn was used instead of Mn; (iii) the
reduction process was able to discriminate not only
between ketones and alkenes, ketones and esters, and
aromatic and aliphatic ketones, but also between cyclic
and acyclic aliphatic ketones. This noticeable selectivity
pattern is difficult to achieve by conventional reduction
methods, especially working at room temperature as we
did in all the experiments described here.
To confirm the role played by water, the experiment
was repeated using D2O instead of H2O. Thus we
obtained 1-deuterio-1-phenylethanol at a yield of 79%
(isolated product). GC–MS analysis indicated 87% deu-
terium incorporation. These results confirmed the pro-
ton-donor role of water and showed the potential value
of this procedure for the synthesis of deuterium-labelled
alcohols (D2O is the cheapest deuterated reagent
known).
We subsequently attempted the reduction of the substi-
tuted cyclohexanones 12–14 to gain more information
about the stereochemical outcome of the reaction. In all
cases the axial alcohol obtained (cis isomers 15a–17a)
predominated over the equatorial one (trans isomers
15b–17b). These results might derive from the preferred
equatorial attack by the bulky titanocene(III) upon the
cyclic titanoxy radical (Scheme 2).
Once we were confident about the role of water we
decided to explore the scope and limitations of the
method. Therefore, we assayed the reduction of
ketones 5a–11a bearing various additional functional
groups.
† Titanocene(III) can be generated in situ by stirring commercially
available Cp2TiCl2 and either Mn or Zn. Under these conditions,
trinuclear species such as 1 are obtained.17–19
‡ All products obtained were isolated and characterised by spectro-
scopic techniques.

A. F. Barrero et al. / Tetrahedron Letters 44 (2003) 1079–1082
1081
We were pleased to find that treatment of keto ester 8a
with a substoichiometric quantity (10 mol%) of com-
mercially available Cp2TiCl2,§ 2,4,6-collidine hydrochlo-
ride, Zn dust and water gave secondary alcohol 8b in
66% yield.¶
This stereochemical behaviour is similar to that
reported for bulky hydride reducing agents,21 (contrary
to that shown by titanocene borohydride),22 thus
providing an alternative method for the synthesis of
axial alcohols under non-alkaline conditions.
Additionally, titanocene-catalysed reduction of 14 gave
a mixture of 17a (25% yield) and 17b (8%), indicating
that the stereoselectivity showed by the stoichiometric
reaction was roughly retained under catalytic condi-
tions. The results obtained in the catalytic reduction of
acetophenone,§ and the keto ester 8a¶ support the con-
cept outlined in Scheme 3 and reinforce the synthetic
value of the method.
Catalysis plays a crucial role in chemical synthesis.6,23
Therefore, once we knew the main features of the
stoichiometric reduction we could focus our attention
on the development of the catalytic version, which was
the main aim of the present work. Gansa¨uer et al.
successfully used 2,4,6-collidine hydrochloride to regen-
erate Cp2TiCl2 from the alkoxy-titanium species formed
during the titanocene-catalysed opening of epoxides.15
With this idea in mind, we chose this hydrochloride to
regenerate Cp2TiCl2 from the titanium derivatives pre-
sumably formed in our process. Subsequently Cp2TiCl2
could be reduced once more to titanocene(III) by the
surplus of Zn in the medium. Thus, the catalytic cycle
depicted in Scheme 3 should be completed.
Deuterium-labelled compounds are powerful tools for
mechanistic studies of both chemical and biochemical
reactions. Nevertheless, the synthesis of deuterium-
labelled alcohols by the conventional treatment of
ketones with NaBD4 may be limited in some cases by
the relatively high cost of this reagent. In this context
we thought that
a
titanium-catalysed reduction,
employing the less expensive D2O (roughly 80-fold
cheaper than NaBD4) as deuterium source, might offer
a convenient alternative for the synthesis of labelled
alcohols. As we expected, the titanocene-catalysed
reduction of acetophenone and keto ester 8a, using
D2O instead of H2O,¶ gave 1-deuterio-1-phenylethanol
(35% yield) and the deuterated isotopomer of 8b (64%
yield), respectively. Collidine and any surplus Zn can be
easily recovered at the end of these experiments and
thus the main cost of the process derives from the
relatively inexpensive catalytic quantity of Cp2TiCl2
and the D2O consumed. These results confirm the
potential of our catalytic process for the synthesis of
some labelled alcohols in economical terms.
§ The role played by the titanium catalyst is essential and thus, when
acetophenone was treated under similar conditions but excluding
titanium, the starting ketone was recovered unchanged. When a
catalytic amount of Cp2TiCl2 was added, on the other hand,
1-phenylethanol was obtained (71% yield of isolated product).
¶ Titanocene-catalysed reduction of keto ester 8a (a model experimental
procedure): Strictly deoxygenated THF (10 mL) was added to a
mixture of commercially available Cp2TiCl2 (6.5 mg, 0.026 mmol)
and Zn dust (135 mg) under an Ar atmosphere, and the suspension
was stirred at rt until it turned lime green (after about 15 min).
Subsequently, keto ester 8a (50 mg, 0.26 mmol) in THF (5 mL),
H2O (28 mL, 1.56 mmol), and 2,4,6-collidine hydrochloride (163 mg,
1.03 mmol) were added to the green suspension, giving a deep-blue
mixture which was stirred at rt for 48 h. The suspension was then
filtered (to recover the excess of Zn) and the solvent was removed
from the filtrate, t-BuOMe (30 mL) was added to the residue and
the ethereal solution was washed with 2N HCl and brine. The
organic layer was dried over anhyd. Na2SO4 and the ether was
removed. Flash chromatography (hexane/t-BuOMe, 1/1) of the
residue afforded hydroxy ester 8b (33 mg, 66% yield), and starting
ketone 8a (13 mg, 26%). The recovered ketone was reduced once
more under similar conditions, rising the overall yield of 8b until
83%. Collidine can be also recovered at the end of the experiment
(by simple acid–base extraction),15 treated with HCl, and used
again.
Scheme 2. Stereoselective reduction of cyclohexanones pro-
moted by titanocene(III) in aqueous media.
Deuterium-labelled alcohols were obtained under the same experi-
mental conditions but using D2O instead of H2O. In these experi-
ments collidine hydrochloride was treated with D2O prior to use.
Scheme 3. Titanocene-catalysed reduction of ketones in
aqueous media.

1082
A. F. Barrero et al. / Tetrahedron Letters 44 (2003) 1079–1082
In summary, we describe here a novel procedure for the
selective reduction of ketones to secondary alcohols
catalysed by a titanocene complex in aqueous media.
Mechanistically the reaction probably proceeds via
titanoxy radicals. In practice this reduction has signifi-
cant advantages over the conventional methods: it
shows a noticeable selectivity pattern, takes place under
mild conditions using inexpensive and safe reagents and
does not require pre-dried solvents. To the best of our
knowledge this is the first procedure described for the
reduction of ketones catalysed by an organometallic
complex and employing water as proton source. The
proton-donor activity shown by water under these con-
ditions avoids the use of the frequently poisonous
hydrogen-atom donors usually required to reduce free
radicals.24 Moreover, the organometallic nature of the
catalyst is a relevant feature with regard to asymmetric
synthesis. As Professor Noyori has recently pointed
out, the virtually unlimited structural variation in the
organic ligand offers great opportunities for asymmet-
ric catalysis.23 Along these lines, we are currently pursu-
ing studies towards an enantioselective version of the
method reported here using well-established chiral
titanocene catalysts.25
5. Yamakawa, M.; Yamada, I.; Noyori, R. Angew. Chem.
2001, 113, 2900–2903.
6. Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40,
40–73.
7. Hartmann, R.; Chen, P. Angew. Chem., Int. Ed. 2001, 40,
3581–3585.
8. Girard, P.; Namy, J. L.; Kagan, H. B. J. Am. Chem. Soc.
1980, 102, 2693–2698.
9. Hasegawa, D. P.; Curran, D. P. J. Org. Chem. 1993, 58,
5008–5010.
10. Dahle´n, A.; Hilmersson, G. Tetrahedron Lett. 2002, 43,
7197–7200.
11. Patonay, T.; Hajdu, C.; Jeko, J.; Le´vai, A.; Micskei, K.;
Zucchi, C. Tetrahedron Lett. 1999, 40, 1373–1374.
12. Micskei, K.; Gyarmati, J.; Kova´cs, G.; Makleit, S.;
Simon, C.; Szabo´, Z.; Marton, J.; Hosztafi, S.; Reinke,
H.; Drexler, H. J. Eur. J. Org. Chem. 1999, 149–153.
13. For a review of low-valent titanium chemistry, see: Fu¨rst-
ner, A.; Bogdanovic, B. Angew. Chem., Int. Ed. 1996, 35,
2442–2469.
14. RajanBabu, T. V.; Nugent, W. A. J. Am. Chem. Soc.
1994, 116, 986–997.
15. Gansa¨uer, A.; Bluhm, H.; Pierobon, M. J. Am. Chem.
Soc. 1998, 120, 12849–12859.
16. Barrero, A. F.; Oltra, J. E.; Cuerva, J. M.; Rosales, A. J.
Org. Chem. 2002, 67, 2566–2571.
17. Sekutowski, D. G.; Stucky, G. D. Inorg. Chem. 1975, 14,
2192–2199.
Acknowledgements
18. Sekutowski, D.; Jungst, R.; Stucky, G. D. Inorg. Chem.
1978, 17, 1848–1855.
19. Stephan, D. W. Organometallics 1992, 11, 996–999.
20. Barden, M. C.; Schwartz, J. J. Am. Chem. Soc. 1996, 118,
5484–5485.
21. Greeves, N. In Comprehensive Organic Synthesis; Trost,
B. M.; Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 8,
pp. 1–24.
We thank our English colleague Dr. J. Trout for revis-
ing our English text, our colleague J. L. Oller for his
collaboration and the Spanish Ministerio de Educacio´n
y Cultura for the grant provided to A.R.
22. Barden, M. C.; Schwartz, J. J. Org. Chem. 1995, 60,
5963–5965.
References
23. Nobel Lecture: Noyori, R. Angew. Chem., Int. Ed. 2002,
41, 2008–2022.
1. Faber, K. Biotransformations in Organic Chemistry; 3rd
ed.; Springer: Berlin, 1997.
24. For a recent overview on free-radical chemistry, see:
Radicals in Organic Synthesis; Renaud, P., Sibi, M. P.,
Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, Basic Princi-
ples, and Vol. 2, Applications.
25. Gansa¨uer, A.; Bluhm, H. Chem. Rev. 2000, 100, 2771–
2788.
2. Larock, R. C. Comprehensive Organic Transformations;
2nd ed.; Wiley: New York, 1999.
3. Ooi, T.; Ichikawa, H.; Maruoka, K. Angew. Chem., Int.
Ed. 2001, 40, 3610–3612 and references cited therein.
4. Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc.
2000, 122, 1466–1478.