Dehydrogenative Silylation of Amines and Hydrosilylation of Imines Catalyzed by Ytterbium-Imine Complexes

Article abstract of DOI:10.1021/jo982154r

Dehydrogenative silylation of primary and secondary amines with triphenylsilane was catalyzed by ytterbium-imine complexes, [Yb(η²-Ph₂CNAr)(hmpa)_n], to give aminosilanes in good yields. In the reaction with diphenyl- and p

Full text of DOI:10.1021/jo982154r

J . Org. Chem. 1999, 64, 3891-3895
3891
Deh yd r ogen a tive Silyla tion of Am in es a n d Hyd r osilyla tion of
Im in es Ca ta lyzed by Ytter biu m -Im in e Com p lexes
Ken Takaki,* Tohru Kamata, Yoshimi Miura, Tetsuya Shishido, and Katsuomi Takehira
Department of Applied Chemistry, Faculty of Engineering, Hiroshima University,
Kagamiyama, Higashi-Hiroshima 739-8527, J apan
Received October 27, 1998
Dehydrogenative silylation of primary and secondary amines with triphenylsilane was catalyzed
by ytterbium-imine complexes, [Yb(η2-Ph2CNAr)(hmpa)n], to give aminosilanes in good yields. In
the reaction with diphenyl- and phenylsilanes, diaminosilanes were formed as major products.
Whereas n- and sec-alkylamines were readily silylated, tert-alkylamines and aromatic amines
exhibited lower reactivities. Moreover, hydrosilylation of imines has been achieved by using
phenylsilane and the imine complexes (Ar ) Ph, C6H4F-4), giving rise to mono- and diaminosilanes.
The two reactions were in agreement as regards the product selectivities and yields.
In tr od u ction
complexes 1 were found to catalyze irreversible isomer-
ization7 and dehydrogenative silylation of terminal al-
kynes.8 For the synthetic application of the imine com-
plexes 1, we studied their catalyst activity on the hy-
drosilylation of imines and found that the reaction took
place efficiently.
Synthetic reactions based on lanthanide reagents have
been extensively explored in the last two decades because
of their unique properties.1 Particularly, olefin transfor-
mations catalyzed by organolathanides such as hydro-
genation2 and hydrosilylation3 have attracted much
attention, because they have been proven to be compa-
rable or, in some cases, superior to those catalyzed by
late transition metals. However, the catalytic reaction
of carbonyl and imino compounds instead of olefins has
been rarely investigated. Recently, Marks et al. reported
the first example of lanthanide-catalyzed imine hydro-
genation, in which acyclic imines were reduced under
more vigorous conditions than for olefin hydrogenation.4
It has also been demonstrated that stoichiometric reac-
tion of imine with lanthanocene hydrides proceeds via
C-H activation rather than imine insertion to the Ln-H
bond.4,5
As in olefin hydrosilylation with organolathanides,3
imine hydrosilylation would involve also two key steps:
insertion of imine to the Ln-H bond and silanolysis of
the resulting Ln-N bond (eq 1). The second step is very
rare9 and seems to be difficult, compared to the reaction
of olefin. Thus, we studied at first a catalytic dehydro-
genative silylation of amines with the imine complexes
1 to confirm the second step.10 Next, our effort was
addressed to generation of the Ln-H species from the
complexes 1 for the first step and then to imine hydrosi-
lylation. We describe herein these results.
Catalysts used for these olefin transformations have
been limited to trivalent lanthanocene complexes, which
are rather difficult for handling and large-scale synthesis.
Previously, we reported that divalent Yb-imine com-
plexes [Yb(η2-Ph2CNAr)(hmpa)n] (1) were easily prepared
from Yb metal and diarylimines in THF and HMPA.6 The
Resu lts a n d Discu ssion
As a model reaction, dehydrogenative silylation of
amylamine (2a ) was tested with various hydrosilanes and
imine complexes 1 (Table 1). When triphenylsilane and
amylamine (1:1) were successively added to a solution of
1a (3 mol %) in THF, generated in situ from Yb metal
and diphenymethylideneaniline in the presence of 4 equiv
of HMPA,11 N-triphenylsilylamylamine (3a ) was pro-
duced in 61% yield, along with vigorous hydrogen evolu-
* Phone: 81-824-24-7738. Fax: 81-824-22-7191. E-mail: ktakaki@ipc.
hiroshima-u.ac.jp.
(1) (a) Edelmann, F. T. In Topics in Current Chemistry 179;
Herrmann, W. A., Ed.; Springer-Verlag: Berlin, 1996; pp 247-276.
(b) Imamoto, T. Lanthanides in Organic Synthesis; Academic Press:
London, 1994. (c) Molander, G. A. Chem. Rev. 1992, 92, 29-68. (d)
Kagan, H. B.; Namy, J . L. Tetrahedron 1986, 42, 6573-6614.
(2) (a) J eske, G.; Lauke, H.; Mauermann, H.; Schumann, H.; Marks,
T. J . J . Am. Chem. Soc. 1985, 107, 8111-8118. (b) Molander, G. A.;
Hoberg, J . O. J . Org. Chem. 1992, 57, 3266-3268. (c) Conticello, V.
P.; Brard, L.; Giardello, M. A.; Tsuji, Y.; Sabat, M.; Stern, C. L.; Marks,
T. J . J . Am. Chem. Soc. 1992, 114, 2761-2762. (d) Giardello, M. A.;
Conticello, V. P., Brard, L.; Gagne, M. R.; Marks, T. J . J . Am. Chem.
Soc. 1994, 116, 10241-10254. (e) Haar, C. M.; Stern, C. L.; Marks, T.
J . Organometallics 1996, 15, 1765-1784.
(6) Makioka, Y.; Taniguchi, Y.; Fujiwara, Y.; Takaki, K.; Hou, Z.;
Wakatsuki, Y. Organometallics 1996, 15, 5476-5478.
(7) Makioka, Y.; Saiki, A.; Takaki, K.; Taniguchi, Y.; Kitamura, T.;
Fujiwara, Y. Chem. Lett. 1997, 27-28.
(8) Takaki, K.; Kurioka, M.; Kamata, T.; Takehira, K.; Makioka, Y.;
Fujiwara, Y. J . Org. Chem. 1998, 63, 9265-9269.
(9) Opposite transposition (Ln-N + Si-H f Ln-Si + N-H) was
proposed to account for the role of phenylsilane accelerator in lantha-
nocene-catalyzed imine hydrogenation by Marks et al. Simultaneously,
they documented the formation of a small amount of hydrosilylation
product (<10%); see: ref 4.
(10) Dehydrogenative silylation of amines catalyzed by lithium
amides has been well-known; see: Kraus, C. A.; Nelson, W. K. J . Am.
Chem. Soc. 1934, 56, 195-202.
(3) (a) Sakakura, T.; Lautenschlager, H.-J .; Tanaka, M. J . Chem.
Soc., Chem. Commun. 1991, 40-41. (b) Molander, G. A.; J ulius, M. J .
Org. Chem. 1992, 57, 6347-6351. (c) Molander, G. A.; Nichols, P. J .
J . Am. Chem. Soc. 1995, 117, 4415-4416. (d) Molander, G. A.; Retsch,
W. H. Organometallics 1995, 14, 4570-4575. (e) Fu, P.-F.; Brard, L.;
Li, Y.; Marks, T. J . J . Am. Chem. Soc. 1995, 117, 7157-7168.
(4) Obora, Y.; Ohta, T.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc.
1997, 119, 3745-3755.
(5) Radu, N. S.; Buchwald, S. L.; Scott, B.; Burns, C. J . Organome-
tallics 1996, 15, 3913-3915.
10.1021/jo982154r CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/06/1999

3892 J . Org. Chem., Vol. 64, No. 11, 1999
Takaki et al.
Ta ble 1. Deh yd r ogen a tive Silyla tion of Am yla m in ea
tion (run 1). Catalyst activity of 1 was dependent on the
substituents on the imine nitrogen. Thus, the complexes
1 with electron-withdrawing groups generally gave better
yields of 3a than those with electron-donating ones (runs
1 and 2 vs 4-6). However, 2,6-xylyl catalyst 1c showed
the highest activity among the imine complexes tested,
suggesting that a steric factor can also play an important
role in the reaction (run 3). With respect to the reactivity
of hydrosilanes, those with a phenyl substituent gave the
aminosilanes 3 in over 90% yields (runs 3, 7, and 8),
whereas trialkyl and trialkoxysilanes decreased the yield
to about 20% (runs 9 and 10).
Dehydrogenative silylation of various primary and
secondary amines with Ph3SiH was studied by using the
imine complex 1c (Table 2). Of the primary amines, n-
and sec-alkylamines 2a and 2b were readily converted
to 3a and 3f in 93% and 97% yields, respectively (runs 1
and 2). Although reaction of tert-alkylamine 2c was very
slow (run 3), the product 3g was formed in 81% yield by
increasing the catalyst amount (10 mol %) and reaction
time (run 4). Secondary amines 2e-g were similarly
silylated in good yields (runs 6-8). In contrast, aniline
was converted to the aminosilane 3h in lower yield (56%)
than alkylamines (run 5), and no reaction took place with
N-methylaniline (run 9).
Dehydrogenative silylation of amines could be per-
formed with diphenylsilane and phenylsilane, giving rise
to aminosilanes and diaminosilanes (eqs 2 and 3). In the
Ta ble 2. Deh yd r ogen a tive Silyla tion of Va r iou s Am in es
w ith P h 3SiHa
reaction of amylamine (2a ) with diphenylsilane, when the
ratio of 2a to Ph2SiH2 was unity, aminosilane 4a and
diaminosilane 5a were obtained in 18% and 35% yields,
respectively. Use of 2 equiv of 2a resulted in the selective
formation of 5a . N-Benzylamylamine (2i) was silylated
with phenylsilane (3 equiv) in the presence of the imine
complex 1a to afford aminosilane 7b in 80% yield,
together with a minor diaminosilane 8b.
The dehydrogenative silylation of amine would be
explained as follows (Scheme 1). The imine complex 1
reacts with amine to yield bisamide complex A. Silanoly-
sis of the Yb-N bond of A affords the aminosilane and
hydride complex B.9 Then, bisamide A is regenerated by
the reaction of B with amine, along with hydrogen
evolution. Of the two amide moieties of A, only one amide
derived from the amine would be silylated; the other from
the imine complex 1 remains unchanged, which is
deduced from the fact that N-methylaniline was not
silylated at all (run 9, Table 2).12 Moreover, if another
bisamide, [(R′2N)2Yb], is formed by the double silanolysis
of A or by further reaction of A with R′2NH and works
as a real intermediate, all imine complexes 1 bearing
different imine ligands should exhibit similar catalyst
activity.
After it was confirmed that silanolysis of lanthanide
amide, i.e., the second step of imine hydrosilylation (eq
1), took place readily, we tried to generate a hydride
species from the imine complex 1 for the first step.
(11) Although the accurate number of the coordinated HMPA is
unknown, equivalents of this additive used for generation of 1 are
tentatively shown as a coordination number. The product 3a was
formed in very low yields (<5%) in the absence of HMPA.
(12) No aminosilanes derived from the imine ligands were detected
in any reactions.

Yb-Imine-Catalyzed Silylation
J . Org. Chem., Vol. 64, No. 11, 1999 3893
a
Ta ble 3. Hyd r osilyla tion of Im in es w ith P h SiH3
Sch em e 1
Sch em e 2
Although 1 can be regarded as a sort of lanthanide carbyl
complex, no reaction occurred under atmospheric hydro-
gen. Accordingly, we expected that hydride complex B
generated in the dehydrogenative silylation of secondary
amines could be utilized to initiate the reaction as shown
in Scheme 2. Stoichiometric reaction of triphenylsilane
with secondary amines by using 1a or 1c gave aminosi-
lanes quantitatively, suggesting a generation of B. In
fact, this reaction mixture reduced adamantanone to
2-adamantanol in good yields. However, addition of a
4-fold excess of hydrosilane and aldimines to the mixture
did not produce any hydrosilylation or reduction prod-
ucts.
Next, we intended to generate the hydride species
directly from the imine complexes 1 and hydrosilane. For-
tunately, it was found that phenylsilane was oligomerized
with hydrogen evolution in the presence of 1.13 It has
been reported that dehydrogenative oligomerization of
the silane catalyzed by lanthanocene complexes proceeds
through a hydride and silyl species.14 Therefore, similar
intermediates could be generated in the present system,
and if the origomerization is suppressed, the hydride
species can be available for the imine hydrosilylation.
Reaction of benzylidenemethylamine (6a ) with an
equimolar amount of phenylsilane in the presence of 1a
(5 mol %) gave aminosilane 7a and diaminosilane 8a in
7% and 65% yields based on 6a (7% and 32% yields based
on the silane), respectively, along with a very small
amount of the oligomer (eq 4). Of course, no reaction took
the most active catalyst in the dehydrogenative silylation
of amines, did not promote any reaction. In addition,
triphenylsilane and diphenysilane could not be used for
the present hydrosilylation, probably because a hydride
species was not generated in the reaction mixture.15 The
amount of HMPA in the complex 1a affected the product
selectivity and yields.11 When 2 equiv of the additive was
used (n ) 2), 7a and 8a were formed in 28% and 20%
yields, respectively, whereas the reaction with 1a ′ (n )
6) gave the products in 5% and 75% yields, respectively.
Hydrosilylation of aldimines and ketimines with phen-
ylsilane was investigated by using the imine complex 1a ′
(Table 3). Aldimines 6a -f were converted to aminosilanes
7 and diaminosilanes 8 in good yields (runs 1-6). The
diaminosilanes 8 were selectively obtained from N-
methylimines (runs 1 and 4), whereas the ratio of 7 to 8
was increased in the reaction of N-pentylimines (runs 2,
5, and 6). Ketimines 6g-i were less reactive than
aldimines (runs 7-9). Although hydrosilylation of ketimine
6g produced aminosilane 7g in 56% yield as the sole
product, little or no product was obtained in the reaction
of 6h and 6i.
For comparison of the hydrosilylation of imines with
the dehydrogenative silylation of amines, amines corre-
sponding to the imines listed in Table 3 were treated with
(13) For example, treatment of PhSiH3 with 1b′, [Yb(η2-Ph2CNC6H4F-
4)(hmpa)6], (10 mol %, rt, 17 h) produced polyphenylsilanes with
vigorous hydrogen evolution, from which oligomer (MW ) 2966, MW
/
MN ) 1.1) was isolated in 29% yield by aqueous workup, followed by
short-column chromatography on silica gel using hexanes-ethyl
acetate (1/1).
place in the absence of 1a (THF-HMPA, rt ∼ refluxing
temperature). Substitution of 1a with 1b caused an
increase of the oligomer, though the products 7a and 8a
were obtained in similar yields. The imine complex 1c,
(14) Forsyth, C. M.; Nolan, S. P.; Marks, T. J . Organometallics 1991,
10, 2543-2545.
(15) Oligomerization and hydrogen evolution were not observed on
treatment of the two silanes with 1.

3894 J . Org. Chem., Vol. 64, No. 11, 1999
Takaki et al.
Ta ble 4. Deh yd r ogen a tive Silyla tion of Am in es w ith
of bisamide A. In the latter, insertion of imine to hydride
species B, which is generated in situ by the reaction of
the imine complex 1 with phenylsilane, affords bisamide
A. Subsequent silanolysis of one Yb-N bond of A with
phenylsilane yields aminosilane and B, as described in
Scheme 1. The aminosilanes, particularly those derived
from the less crowded imine, would be more reactive than
phenylsilane in the second step, probably because metal
coordination at their nitrogen atom facilitates this pro-
cess. Therefore, diaminosilanes could be obtained as the
major product in the most reactions. Decreased reactivity
of ketimines would be attributed to the lower efficiency
in the silanolysis step, as substituents on the amide
moieties are secondary hydrocarbons. In fact, dehydro-
genative silylation of 2n -p gave yields of the products
lower than those of the amines corresponding to aldi-
mines (runs 7-9 vs 1-6, Table 4). In Scheme 3, insertion
of imine to the Ln-H bond may compete with a meta-
lation reaction, which follows by the reaction with
another molecule of imine to yield coupling products.4,5
However, this metalation process could be negligible in
the present system, because the hydrosilylation pro-
ceeded effectively, at least for aldimines, and neither
C-silylated imine nor a coupling product was detected.
In summary, catalytic dehydrogenative silylation of
amines and hydrosilylation of imines have been achieved
by using Yb-imine complexes. The two reactions proceed
through common intermediates and thus exhibit similar
product selectivities and yields. Although the present
reactions include restrictions on the hydrosilanes and
substrates, they provide a potentially useful method for
the preparation of aminosilanes and diaminosilanes
because readily available nonlanthanocene catalysts can
be utilized under mild conditions.
a
P h SiH3
Sch em e 3
Exp er im en ta l Section s
Gen er a l. 1H and 13C NMR spectra were recorded at 270
and 67 MHz, respectively. IR spectra were taken on a FT-IR
spectrophotometer. Mass spectra were obtained at 70 eV on a
GC-MS apparatus. Microanalyses were performed at our
analytical laboratory. All reactions were carried out under
argon. THF was distilled from sodium/benzophenone ketyl
immediately prior to use. HMPA was distilled from CaH2 and
stored over molecular sieves. Ytterbium metal (40 mesh) was
washed with anhydrous hexane under argon and dried in
vacuo. Imines 6 were prepared from the corresponding amines
and aldehydes or ketones by the conventional procedure.
Amines 2i and 2k -p were prepared by the reduction of imines
6b and 6d -i with LiAlH4. All other materials were com-
mercially available and were used after distillation.
Gen er a l P r oced u r e for Deh yd r ogen a tive Silyla tion of
Am in es. Ytterbium metal (12.9 mg, 0.075 mmol) and diphen-
ylmethylidene-2,6-xylylamine (21.4 mg, 0.075 mmol) were
placed in a Schlenk tube. HMPA (53 µL, 0.3 mmol) and THF
(1 mL) were added to the mixture. After addition of MeI (0.2
µL) to activate the metal, the mixture was stirred for 4 h at
room temperature to yield a homogeneous reddish-black
solution of the Yb-imine complex 1c. Hydrosilane (2.5 mmol)
and amine (2.5 mmol) were successively added to the solution,
and the mixture was stirred at room temperature for the
appropriate time as indicated in Tables 1 and 2. The reaction
mixture was directly passed through a short alumina column
or subjected to Kugelrohr distillation without an aqueous
workup. In the reaction with PhSiH3 (Table 4), a solution of
the silane and amine in THF (1 mL) was added to the imine
complex 1a ′ to avoid oligomerization of the silane.
equimolar amounts of phenylsilanes in the presence of
the imine complex 1a ′ (5 mol %), and these results are
summarized in Table 4. As can be seen in Tables 3 and
4, the two reactions are in good agreement as regards
product selectivities and yields, except for runs 2 and 6.
These results indicate that the two reactions proceed
through the same intermediates.
A possible catalytic cycle for the imine hydrosilylation
is outlined in Scheme 3. The cycle is essentially similar
to Scheme 1 proposed for the dehydrogenative silylation
of amines, in that they both include bisamide A and
hydride B as common intermediates. The only difference
between Schemes 1 and 3 is the process for the formation
Gen er a l P r oced u r e for Hyd r osilyla tion of Im in es. The
imine complex 1a ′ was prepared from Yb metal (34.6 mg, 0.2
mmol), diphenylmethylideneaniline (51.5 mg, 0.2 mmol), and

Yb-Imine-Catalyzed Silylation
J . Org. Chem., Vol. 64, No. 11, 1999 3895
HMPA (209 µL, 1.2 mmol) in THF (1 mL) as above. A solution
of phenylsilane (216 mg, 2 mmol) and imine (2 mmol) in THF
(1 mL) was added to the mixture, and stirring was continued
for 20 h at room temperature with monitoring by GC. The
products were isolated as amines on hydrolysis, as described
bellow.
by GC-MS of the mixture; yield was determined by GC using
the relative intensity of 3a .
Reaction mixtures of the hydrosilylation and dehydrogena-
tive silylation with PhSiH3 (Tables 3 and 4) were very labile
on workup similar to the above. The two reactions gave the
same products 7 and 8 monitored by GC analysis, wherein 7
was gradually changed to 8 as the reaction proceeded. On
hydrolysis of the mixture, 7 and 8 disappeared immediately,
and the corresponding amine was formed, which was identified
with an authentic sample by GC. Product yields were deter-
mined on the basis of the ratio of GC peak areas of the starting
substrates 6 or 2 and products 7 and 8, assuming that all
relative intensities were equal. This calculation agreed fairly
well with a quantity of amine 2 obtained by the hydrolysis of
the mixture of the imine hydrosilylation.
Id en tifica tion of P r od u cts a n d Deter m in a tion of Th eir
Yield s. Aminosilanes 3, 4, and 7 and diaminosilanes 5 and 8,
products of the dehydrogenative silylation of amines and
hydrosilylation of imines, are quite unstable and readily
hydrolyze to the corresponding amines and silanol or siloxane.
Thus, their complete purification through aqueous workup
could not be done. However, of the products, aminotriphenyl-
silanes 3a and 3f-k were relatively stable and could be
directly chromatographed on neutral alumina (ca. 70 mm ×
15 mm i.d.) with anhydrous hexanes-ethyl acetate (20/1)
eluent to give almost pure samples, which were identified by
1H and 13C NMR. Their GC yield was determined by using an
internal standard such as eicosane. Aminosilanes 3b-d were
isolated by Kugelrohr distillation directly from the reaction
mixture, which contained a small amount of impurities such
Su p p or tin g In for m a tion Ava ila ble: Characterization
data (1H and 13C NMR and MS spectra) for products 3 and
5a . This material is available free of charge via the Internet
at http://pubs.acs.org.
1
as HMPA and siloxanes, judged by H and 13C NMR spectra.
The product 3e was highly unstable and thus detected only
J O982154R