Homoleptic lanthanide amides as homogeneous catalyst for the Tishchenko reaction

Article abstract of DOI:10.1002/(SICI)1521-3773(19980619)37:11<1569::AID-ANIE1569>3.0.CO;2-C

Known for about 25 years, the bis(trimethylsilyl)amides of Group 3 metals and lanthanides, M[N(SiMe₃)₂]₃, are well suited as highly efficient catalysts for the dimerization of aldehydes [Tishchenko reaction, Eq. (1)].

Full text of DOI:10.1002/(SICI)1521-3773(19980619)37:11<1569::AID-ANIE1569>3.0.CO;2-C

COMMUNICATIONS
Crystallographic Data Centre as supplementary publication no.
CCDC-100826. Copies of the data can be obtained free of charge on
application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax:
int. code ‡ (44)1223-336033; e-mail: deposit@ccdc.cam.ac.uk).
[7] M. Schelhaas, H. Waldmann, Angew. Chem. 1996, 108, 2192 ± 2219;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2056 ± 2083.
no selectivity clearly indicates that the stereogenic centers of
the substrate are unsuitable for acyclic stereocontrol.
Can one dispose of the workbench after the job is done? As
exemplified for the dihydroxylation product 18a, the benzyl
acetal can not be hydrolyzed under mild acid catalysis, but is
easily cleaved hydrogenolytically (Scheme 3). Thus one
Homoleptic Lanthanide Amides as
Homogeneous Catalysts for the Tishchenko
Reaction**
Helga Berberich and Peter W. Roesky*
The Tishchenko reaction (or Claisen ± Tishchenko reac-
tion), that is, the dimerization of aldehydes to form the
corresponding carboxylic ester [Eq. (1)], has been known for
about a century.[1] Its industrial importance is mirrored in the
Scheme 3. Cleavage of the ªworkbenchº. a) 2,2-dimethoxypropane, PPTS,
CH2Cl2, 94%; b) H2/Pd/C, EtOAc; 23 97%, 24 97%. PPTS ˆ pyridinium-
p-toluene sulfonate.
obtains the partially protected hexol derivative 23; the arene
remains bonded to one of the primary hydroxy groups as the
4-methyl benzoate protecting group. Epoxides such as 19a are
transformed into tetrahydrofuran derivatives (e.g. 24) by an
SN2 attack of the hydroxy group liberated after debenzylation.
The arene 9 that is incorporated into the macrolide can
therefore be understood as a polyfunctional protecting group
that amplifies the influence of the substrateꢁs stereogenic
centers. As such it thus represents a new type of ªstereoactive
protective groupº.[7]
great number of patents. Thus the Tishchenko ester of 3-
cyclohexenecarbaldehyde is the precursor for the formation
of epoxy resin, which is durable against environmental
influences.[2] Traditionally aluminum alkoxides[2a, 3] have been
used as homogeneous catalysts for the catalytic variations of
the Tishchenko reaction. More recently other catalysts such as
boric acid[4] and a few transition metal complexes[5] are used.
However, these alternative catalysts are either only reactive
under extreme reaction conditions (e.g. boric acid), difficult to
prepare (e.g. [(C5Me5)2LaCH(SiMe3)2]),[5a] slow (e.g.
[(C5H5)2ZrH2]),[5b] expensive (e.g. [H2Ru(PPh3)2]),[5c] or give
small yields (e.g. K2[Fe(CO)4]).[5d]
Received: December 9, 1997
Revised version: February 6, 1998 [Z11247IE]
German version: Angew. Chem. 1998, 110, 1625 ± 1628
Herein we report that the homoleptic bis(trimethylsilyl)
amides of Group 3 metals and lanthanides, M[N(SiMe3)2]3[6] 1
(M ˆ Sc, Y, Ln (lanthanide)), are high-
Keywords: diastereoselective reactions ´ dihydroxylations ´
epoxidations ´ macrocycles ´ protecting groups
ly active catalysts for the Tishchenko
reaction. Compound 1 belongs to a
class of materials that has been known
for the last 25 years. Recently, in
particular, it has proven to be a
[1] Reviews: Catalytic Asymmetric Synthesis (Ed.: I. Ojima), VCH, New
York, 1993; R. Noyori, Asymmetric Catalysis in Organic Synthesis,
Wiley, New York, 1994; new method of epoxidation: Z.-X. Wang, Y.
Tong, M. Frohn, J.-R. Zhang, Y. Shi, J. Am. Chem. Soc. 1997, 119,
11224 ± 11235.
valuable starting material in lanthanide chemistry through
the easy cleavage of the silylamide group.[7] Compound 1 can
either be prepared from a simple one-step synthesis or it can
be bought (M ˆ Y). Therefore it is even more surprising to
find that up till now there is no known use of 1 as a catalyst.
To compare the reaction rates of 1 with other catalysts the
standard reaction of benzaldehyde to benyzl benzoate was
[2] A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307 ±
1370; R. W. Hoffmann, ibid. 1989, 89, 1841 ± 1860.
[3] Rotational barriers [kJmol 1]: (E)-cyclooctene 149, (E)-cyclononene
84, (E)-cyclodecene 45; A. C. Cope, B. A. Pawson, J. Am. Chem. Soc.
1965, 87, 3649 ± 3651; A. C. Cope, K. Banholzer, H. Keller, B. A.
Pawson, J. J. Whang, H. J. S. Winkler, ibid. 1965, 87, 3644 ± 3649; G.
Binsch, J. D. Roberts, ibid. 1965, 87, 5157 ± 5162.
[4] Additions to ansa alkenes: P. K. Chowdhury, A. Prelle, D. Schomburg,
M. Thielmann, E. Winterfeldt, Liebigs Ann. Chem. 1987, 1095 ± 1099;
summary: E. Winterfeldt, Chimia 1993, 47, 39 ± 45; diastereoselective
addition to double bonds in macrocycles: W. C. Still, L. J. MacPherson,
T. Harada, J. F. Callahan, A. L. Rheingold, Tetrahedron 1984, 40, 2275 ±
2281.
[5] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, Y. Yamaguchi, Bull. Chem.
Soc. Jpn. 1979, 52, 1989 ± 1993.
[6] Crystallographic data (excluding structure factors) for the structures
reported in this paper have been deposited with the Cambridge
[*] Dr. P. W. Roesky, H. Berberich
Institut für Anorganische Chemie der Universität
Engesserstrasse, Geb. 30.45, D-76128 Karlsruhe (Germany)
Fax: (‡49)721-661921
E-mail: roesky@achibm6.chemie.uni-karlsruhe.de
[**] This work was supported by the Deutsche Forschungsgemeinschaft
and the Fonds der Chemischen Industrie (Liebig-Stipendium). Prof.
Dr. D. Fenske is also thanked for his support.
Angew. Chem. Int. Ed. 1998, 37, No. 11
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3711-1569 $ 17.50+.50/0
1569

COMMUNICATIONS
Table 2. Catalytic Tishchenko reaction with 1a as the catalyst.
chosen, whereby the reaction rate and the yield were
determined by NMR spectroscopy in C6D6 with approximate-
ly 1 mol% catalyst at 218C (Table 1). The turnover frequen-
cies (TOFs) were determined from a turnover of 50%
Reactant
Product
TOF
Yield [%][a]
Yield [%][b]
(see below).
A
comparison of La[N(SiMe3)2]3 (1a),
R ˆ phenyl
87
> 1500
> 1500
98
87
75
80
36
45[c,d]
R ˆ 3-cyclohexenyl
R ˆ cyclohexyl
R ˆ 2-furanyl
quant.
quant.
40
quant.[c]
Table 1. Catalytic Tishchenko reaction of benzaldehyde to benzyl ben-
zoate.
2
R ˆ 1-propyl
±
Catalyst
TOF
Yield [%]
T [8C]
Ref.
[a]
1a
1b
1c
87
80
63
98
98
70
88
21
21
21
RT
21
60
21
17
17
20
60
±
±
±
[a]
> 1500
90[e]
85[d]
[a]
[(C5Me5)2LaCH(SiMe3)2]
SmI2
[5a]
[a]
±
±
8
no reaction
no reaction
51
7
9
23
71
34
±
[a] Analytical approach (NMR experiment): 218C, 1 mol% catalyst in
C6D6. [b] Yield of isolated product for the preperative approach: 218C, 5 g
of reactant, 1 mol% catalyst, no solvent. [c] Reaction from 788C !RT;
trimers were formed as by-products. [d] Solvent: 50 mL of pentane/hexane
(1/1). [e] Unknown by-product.
La(OiPr)3
Al(OiPr)3
[(C5H5)2ZrH2]
[(C5H5)2HfH2]
[H2Ru(PPh3)2]
K2[Fe(CO)4]
B(OH)3
[5a]
[a]
±
[5b]
[5b]
[5c]
[5d]
[4]
250
[a] Reaction conditions (this work): 1 mol% catalyst in C6D6.
cyclohexanecarbaldehyde and 3-cyclohexanecarbaldehyde
should be particularly noted. The reactions usually happen
so quickly that only the product can be detected on NMR-
scale reactions. The catalyst in the reaction solution is also still
active after several days. Thus, a reaction that has subsided for
some time can be restarted again by the addition of new
reactant. Furthermore the suitability of 1a for the dimeriza-
tion of furfural should be emphasized, because up till now
when this reaction is carried out with aluminum or tetracar-
bonylferrate catalysts either no or very low yields are
observed. Although 1a catalyzes the dimerization of alde-
hydes, with or without an a-H atom, quickly and in high yields,
the reaction of butanal at 218C only higher coupling products
are formed.[3b, 11] If one starts the reaction at 788C then butyl
butyrate and 2-ethyl-1,3-hexanediol monobutyrate as dimeric
and trimeric products are yielded in almost the same ratio.[11a]
To present a provisional reaction mechanism, the dimeri-
Sm[N(SiMe3)2]3 (1b), and Y[N(SiMe3)2]3 (1c) shows that by
almost quantitative turnover 1a has a higher activity than the
corresponding Sm catalyst 1b, whereas the Y catalyst 1c
produces smaller yields and is less active. The dependence of
the TOF on the ionic radius of the central metal atom has
already been observed.[5a] A comparison with other easily
[8]
accessible lanthanide compounds such as SmI2 or La(Oi-
Pr)3,[5a] which are used as Tishchenko catalysts, showed that
these are inactive for the case of benzaldehyde. From the
reaction with the standard aluminum catalyst (Al(OiPr)3),[3]
under the reaction conditions described above, no quantita-
tive yield was observed. The reaction rate observed with the
use of Al(OiPr)3, which agrees with earlier mea-
sured values,[3a, 9] is roughly an order of magnitude smaller
than that when 1a is used. Although we do not know of a
more active catalyst for the Tishchenko reaction, two
1
zation of benzaldehyde was studied in more detail. H NMR
compounds are described in the literature ([(C5Me5)2LaCH-
spectroscopy and GC/MS studies show that at the beginning
of the reaction bis(trimethylsilyl)amine and benzylic acid
bis(trimethylsilyl)amide are cleaved from 1 to give a lantha-
nide alkoxide A that is, most probably, the catalytically active
species (Scheme 1). Investigations of the paramagnetic com-
pound 1b on catalytic and stoichiometric scales show that all
three amide groups are cleaved without a significant induction
[9]
(SiMe3)2][5a] and Al(OCH2Ph)3
) that could compete
with 1a. [(C5Me5)2LaCH(SiMe3)2] probably would form
the same catalytically active species as in 1a (see below),
but is, however, evidently more difficult to prepare.
Al(OCH2Ph)3 reaches almost the same TOF as 1a, but is
exclusively optimized for the dimerization of benzalde-
hyde.[9, 10]
1
time, and signals appear in the H NMR spectra that may be
The large range of applications of 1a is given in Table 2. The
reactions were carried out at 218C in C6D6 with approximately
1 mol% of catalyst in order to determine the TOF. After-
wards all the reactions were repeated on a preparative scale
(5 g of reactant) without solvent or in hydrocarbons in order
to determine the isolated yields, and to characterize the
products by 1H and 13C NMR spectroscopy as well as
elemental analysis. The work up of the reaction turned out
to be very simple: In the case of the solvent free reaction the
product can usually be easily transferred by vacuum. In the
case of the reaction with o-phthaladehyde the ester precip-
itates cleanly out of the solution. The extremely high TOFs for
attributed to SmOCH2 groups. It can be assumed that the
catalytically active species is either the same or very similar to
the compound that is formed in the Tishchenko reaction with
[(C5Me5)2LaCH(SiMe3)2].[5a] This hypothesis is supported by
the following observations: 1) with 1a similar yields are
achieved as for [(C5Me5)2LaCH(SiMe3)2], which is, inciden-
tally, significantly more difficult to prepare. 2) During the
reaction both catalysts can completely interchange the whole
1
of their original ligand shells. H NMR spectroscopic inves-
tigations show, however, that from the reaction of 1b with
stoichiometric amounts of benzylalcohol or benzylaldehyde
different compounds are formed. Kinetic investigations show
1570
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3711-1570 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1998, 37, No. 11

COMMUNICATIONS
with the catalyst solution in a tempered flask. After one day the product
was isolated by distillation.
Received: January 26, 1998 [Z11402IE]
German version: Angew. Chem. 1998, 110, 1618 ± 1620
Keywords: aldehydes ´ homogeneous catalysis ´ lanthanides
[1] a) L. Claisen, Ber. Dtsch. Chem. Ges. 1887, 20, 646 ± 650; b) W.
Tischtschenko, Chem. Zentralbl. 1906, 77, I, 1309.
[2] a) E. G. E. Hawkins, D. J. G. Long, F. W. Major, J. Chem. Soc. 1955,
1462 ± 1468, quote patent; b) F. R. Frostick, B. Phillips (Union Carbide
& Carbon Corp.), US-A 2716123, 1953 [Chem. Abstr. 1953, 50, 7852f].
[3] a) W. C. Child, H. Adkins, J. Am. Chem. Soc. 1923, 47, 789 ± 807; b) F. J.
Villani, F. F. Nord, ibid. 1947, 69, 2605 ± 2607; c) L. Lin, A. R. Day, ibid.
1952, 74, 5133 ± 5135.
[4] P. R. Stapp, J. Org. Chem. 1973, 38, 1433 ± 1434.
[5] a) S. Onozawa, T. Sakakura, M. Tanaka, M. Shiro, Tetrahedron 1996,
52, 4291 ± 4302; b) K.-I. Morita, Y. Nishiyama, Y. Ishii, Organo-
metallics 1993, 12, 3748 ± 3752; c) T. Ito, H. Horino, Y. Koshiro, A.
Yamamoto, Bull. Chem. Soc. Jpn. 1982, 55, 504 ± 512; d) M. Yama-
shita, T. Ohishi, Appl. Organomet. Chem. 1993, 7, 357 ± 361.
[6] D. C. Bradley, J. S. Ghorta, F. A. Hart, J. Chem. Soc. Dalton Trans.
1973, 1021 ± 1023.
Scheme 1. Suggested mechanism for the M[N(SiMe3)2]3-catalyzed Tish-
chenko reaction.
[7] Review: R. Anwander, Top. Curr. Chem. 1996, 179, 33 ± 112.
[8] D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990, 112, 6447 ± 6449.
[9] Y. Ogata, A. Kawasaki, Tetrahedron 1969, 25, 929 ± 935.
[10] As a catalyst for the dimerization of benzaldehyde, Al(OCH2Ph)3 is
ten times faster than Al(OiPr)3, because here the metal has already
exchanged all OiPr for OCH2Ph groups.[9] Although Al(OCH2Ph)3 is
faster than Al(OiPr)3 for this reaction, one cannot assume that this is
always the case. For our comparative studies we were not interested in
a special reaction, but rather in a wider application. Thus the literature
quoted is almost exclusively with Al(OiPr)3 as the catalyst.[3, 11a]
[11] a) G. Fouquet, F. Merger, R. Platz, Liebigs Ann. Chem. 1979, 1591 ±
1601; b) G. M. Villacorta, J. San Fillipo, Jr., J. Org. Chem. 1983, 48,
1151 ± 1154.
that 1/[reactant] is related linearly to the reaction time; that is,
a second-order reaction with respect to the reactant occurs.
Such reaction kinetics have already been established for the
aluminum-catalyzed reaction for the dimerization of benzy-
laldehyde.[9] Presumably, a molecule of the reactant coordi-
nates to A (!B), which in turn undergoes an alkoxide
transfer (!C; Scheme 1). A second molecule of the reactant
attaches itself to C followed by a hydride transfer that is
probably the rate-determining step (!D).[9]
In summary it should be emphasized that the bis(trime-
thylsilyl)amide compounds of the lanthanides represent a new
class of Tishchenko catalysts as a result of the high Lewis
acidity and the easy interchangeability of the ligand sphere.
These compounds are distinguished by a number of practical
advantages such as the ease of accessibility (one compound
can be bought), inexpensive metals, extremely high activities
(to our knowledge there are no catalysts that are more active),
and a high durability of the catalysts. These advantages lead us
to hope that 1 , which is already today a standard reagent in
organolanthanide chemistry, will find further application as a
Tishchenko catalyst.
NaSn5: An Intermetallic Compound with
Covalent a-Tin and Metallic b-Tin
Structure Motifs
Thomas F. Fässler* and Christian Kronseder
Under standard conditions tin exists in the b modification,
in which each tin atom is surrounded by six other Sn atoms in
a distorted octahedral arrangement (Figure 1a, Scheme 1,
Table 1). Below 13.28C transformation of b-tin into cubic a-
tin (diamond structure), which is thermodynamically more
Experimental Section
NMR-scale reaction: Compound 1 (0.015 mmol) was weighed under
protective gas into an NMR tube. C6D6 ( ꢀ 0.7 mL) was condensed into
the NMR tube, and the mixture was frozen at 1968C. The reactant
(1.5 mmol) was injected onto the solid mixture, and the whole sample was
frozen at 1968C. To determine the reaction kinetics the sample was
melted and mixed just before the insertion into the core of the NMR
machine (t0). The ratio between the reactant (product) and the catalyst was
exactly calculated by comparison of the integration of all CHO (CH2O)
signals with the N(SiMe3)2 signals. The latter were used as an internal
standard for the kinetic measurements.
1
stable by 2.09 kJmol , occurs.[1] In this modification, which
has a lower density, all tin atoms are four-coordinate, and the
structure can be described by invoking localized two-center,
two-electron (2c ± 2e) bonds (Scheme 1). All structurally
[*] Dr. T. F. Fässler, C. Kronseder
Laboratorium für Anorganische Chemie der
Eidgenössischen Technischen Hochschule Zürich
Universitätstrasse 6, CH-8092 Zürich (Switzerland)
Fax: (‡41)1-632-1149
Preparative-scale reaction: a) without solvent: Under protective gas the
catalyst (1 mol%) was stirred in a tempered reaction flask. The reactant
(5 g) was added directly to the catalyst. Usually an exothermic reaction was
observed. After one day the product was isolated by distillation. b) In
solution: The catalyst (1 mol%) and the reactant (5 g) were each dissolved
in pentane/hexane (1/1, 25 mL). The reactant solution was added together
E-mail: faessler@inorg.chem.ethz.ch
[**] This work was supported by the Eidgenössische Technische Hoch-
schule Zürich.
Angew. Chem. Int. Ed. 1998, 37, No. 11
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1433-7851/98/3711-1571 $ 17.50+.50/0
1571