Large scale synthesis of N-benzyl-4-formylpiperidine through partial reduction of esters using aluminum hydride reagents modified with pyrrolidine

Article abstract of DOI:10.1016/S0040-4020(01)00143-0

The modification of sodium bis(2-methoxyethoxy)aluminum hydride (SMEAH) with pyrrolidine provided a highly selective reducing agent to transform N-benzyl-4-ethoxycarbonylpiperidine into N-benzyl-4-formylpiperidine 1 under mild conditions. However, this simple modification led to a significant amount of N-benzyl-4-(pyrrolidin-1-ylmethyl)piperidine 4 due to overreduction of an intermediate. Our extensive research revealed that an alkaline base such as potassium tert-butoxide could suppress the formation of the by-product to give the desired aldehyde, enabling us to establish a viable synthetic process for a key intermediate of donepezil hydrochloride. The potential applications of this reagent are also described.

Full text of DOI:10.1016/S0040-4020(01)00143-0

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 2701±2710
Large scale synthesis of N-benzyl-4-formylpiperidine through
partial reduction of esters using aluminum hydride reagents
modi®ed with pyrrolidine
Taichi Abe,p Toyokazu Haga, Shigeto Negi, Yukio Morita, Keizou Takayanagi and
Kimio Hamamura
Process Research Laboratories, Eisai Co. Ltd., 22-Sunayama, Hasaki-machi, Kashima-gun, Ibaraki 314-0255, Japan
Received 6 November 2000; accepted 29 January 2001
AbstractÐThe modi®cation of sodium bis(2-methoxyethoxy)aluminum hydride (SMEAH) with pyrrolidine provided a highly selective
reducing agent to transform N-benzyl-4-ethoxycarbonylpiperidine into N-benzyl-4-formylpiperidine 1 under mild conditions. However, this
simple modi®cation led to a signi®cant amount of N-benzyl-4-(pyrrolidin-1-ylmethyl)piperidine 4 due to overreduction of an intermediate.
Our extensive research revealed that an alkaline base such as potassium tert-butoxide could suppress the formation of the by-product to give
the desired aldehyde, enabling us to establish a viable synthetic process for a key intermediate of donepezil hydrochloride. The potential
applications of this reagent are also described. q 2001 Elsevier Science Ltd. All rights reserved.
Donepezil hydrochloride (Ariceptw)1±3 is a selective
inhibitor of acetylcholinesterase (AChE) and is the ®rst
promising agent with this mode of action for the treatment
of mild to moderate dementia of the Alzheimer's type. This
new drug was approved ®rst in the US in 1997 and later in
67 countries,4 and has provided an effective remedy for this
central nervous system dysfunction illness (Scheme 1).
Donepezil hydrochloride is synthesized through the Aldol
condensation of 5,6-dimethoxy-1-indanone and N-benzyl-4-
formylpiperidine (1), ensuing double bond reduction and
hydrochlorination.5 The aldehyde 1 was originally manu-
factured from N-benzylpiperidone via the Wittig reaction
followed by acid hydrolysis of the enol ether, as shown in
Scheme 2. However, the use of expensive (methoxymethyl)-
triphenylphosphonium chloride raised the production cost,
prompting us to seek a more economical process for 1.
O
MeO
Our literature survey rendered us with three other methods6±8
for the synthesis of 1. Two of them were dif®cult to imple-
ment at the production scale because they require costly
reagents and tedious procedures,6,7 but one of the methods
appeared to be quite promising. This method started with
less expensive ethyl isonipecotate (INPE) which was
N
.
HCl
MeO
Aricept (Donepezil Hydrochloride)
Scheme 1.
OMe
CHO
O
MeOCH2P+Ph3Cl-
HCl
KOt-Bu
N
N
N
Ph
1
Ph
Ph
Keywords: N-benzyl-4-formylpiperidine; partial ester reduction; sodium bis(2-methoxyethoxy)aluminum hydride.
Corresponding author. Tel.: 1479-46-4971; fax: 1479-46-4959; e-mail: t2-abe@hhc.eisai.co.jp
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00143-0

2702
T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
COOEt
CHO
COOEt
reducing
PhCH2Cl
base
reagent
N
N
N
H
Ph
Ph
1
INPE
2
Scheme 3.
converted to N-benzylester (2), followed by partial reduc-
tion of the ester group with diisobutylaluminum hydride
(DIBAL-H) to give the aldehyde 1 in 92% yield (Scheme
3).8 However, the method required harsh cryogenic condi-
tions, such as 2788C, which represented a problem for large
scale production. Thus, we set out to ®nd a more economical
and milder reducing agent than DIBAL-H.
Following extensive screening of secondary amines, we
found that SMEAH and pyrrolidine were the best combina-
tion, although a signi®cant amount of by-products were
formed. Our attempts to suppress the formation of the by-
products led to a viable process for 1, a key intermediate of
donepezil hydrochloride. We wish to report here a new
reducing agent named Red-ALP, effect of additional base
on the reaction, the established synthetic process and the
applicability of Red-ALP reduction to other substrates.
There have been only two reports on partial ester reduction
other than DIBAL-H reduction. In the ®rst method, sodium
bis(2-methoxyethoxy)aluminum hydride (SMEAH) modi-
®ed with N-methylpiperazine or morpholine was reported
to transform methyl benzoate to benzaldehyde under mild
conditions and in excellent yields.9 Secondly, the modi®ca-
tion of aluminum hydride with N-methylpiperazine was
described as an expedient reductant for aliphatic and
aromatic esters, carboxylic acids and amides to yield their
corresponding aldehydes.10±12 Of the two methods, the
modi®ed SMEAH reagent seemed to satisfy our needs,
because it is commercially available, easy to handle even
at the production scale, and the reaction temperature
reported was mild (e.g. ice-cooling).
1. Results and discussion
INPE was alkylated with benzyl chloride at the N-position
in the presence of NaHCO3 to give the ethyl ester 2 in 93%
yield. The ester 2 was then reduced with SMEAH modi®ed
with N-methylpiperazine or morpholine in THF at 108C
(entries 1 and 2, Table 1). As this reaction showed poor
selectivity, several other secondary amines were examined.
We noticed that when the reaction mixture was quenched
with NaOH solution, the aldehyde 1 easily formed enamines
with secondary amines in the extract. Thus, to determine
Table 1. Reduction of 2 with SMEAH modi®ed with various secondary amines in THF
Entry
Amine
Product ratioa,b (%)
Selectivity
Aldehyde 1
Alcohol 3
Others
4
Aldehyde/alcohol
1
66
25
2.6
HN
NMe
O
2
3
4
5
58
69
63
71
35
21
28
3
6
9
1.7
3.3
HN
HN
HN
HN
8
2.3
26
23.7
Refer to Section 2 for reaction conditions.
The crude product was obtained after the treatment described in Section 2.
Product ratio was determined by GC.
a
b
CHO
CHO
N
N
NaOH aq
HCl aq
+
+
+
N
Cl-
H
N
N
H
H
N
Ph
Ph
Ph
1
1
pH 8.5
Scheme 4.

T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
2703
N
O
N
COOEt
CHO
CH2OH
+
+
+
N
N
N
N
N
Ph
N
Ph
Ph
Ph
Ph
O
O
O
O
Na+
Al-
2
1
3
4
5
H
Red-ALP
Scheme 5.
the pyrrolidinylmethyl by-product 4 increased to 30%. We
also noticed that the yield of 4 varied from 30 to 3% from lot
to lot of SMEAH. Therefore, we suspected that a key factor
was present in the reagent itself.
Table 2. Effects of additional base on the reduction of 2
Entry
Additive
Product ratioa (%)
1
2
3
4
5
As shown in Scheme 6, SMEAH is manufactured from
NaH, Al and 2-methoxyethanol, and has been reported to
take a number of forms through equilibriums.13 The equi-
librium shown in the right hand appeared to be less
plausible, but seemed to be worthy of consideration. We
hypothesized that an excess or shortage of one of the
components in the equilibrium may contribute to the
differences in yield of by-product 4.
1
2
3
4
5
6
±
65.6
95.9
93.5
95.2
91.8
70.0
1.8
0.9
±
0.2
4.7
2.3
1.6
1.5
3.1
3.2
2.0
2.4
30.3
±
±
±
±
25.0
0.3
1.7
1.5
1.2
1.5
0.3
NaOC2H4OCH3
NaOC(CH3)3
KOC(CH3)3
PhONa
Et3N
All reactions were carried out at 208C and additive/SMEAHˆ0.14.
a
Product ratio was determined by GC.
more precisely the selectivity, the products were treated
with aqueous hydrochloric acid solution to hydrolyze
the enamine, and then were re-extracted with organic
solvent at pH 8.5, taking advantage of the difference in
pKa between 1 (pKaù10) and secondary amines (pKaù12)
(Scheme 4).
To test this hypothesis, NaOCH2CH2OCH3, one of the
possible species in the reagent, was added to the reducing
agent. This resulted in complete suppression of the forma-
tion of 4 (entry 2). Commercially available alkaline bases
such as sodium tert-butoxide and potassium tert-butoxide
(KTB) were examined and found to have the same effect
(entries 3 and 4). Sodium phenoxide, a weaker base, showed
also a similar effect (entry 5), whereas triethylamine could
not suppress the formation of the by-product 4 (entry 6).
Among the alkaline bases, KTB was chosen as the most
suitable additive considering its availability, cost and
solubility in organic solvents like THF.
The reduction of SMEAH modi®ed with piperidine slightly
improved the product ratio to 3.3 while a non-cyclic base,
such as diethylamine, gave poor selectivity (2.3) (entries 3
and 4). Among the secondary amines, the ®ve-membered
base, pyrrolidine (entry 5) gave remarkable selectivity
(23.7), although a large amount of other by-products were
formed. This combination was named Red-ALP.
1.2. Optimization of the reaction conditions
The structure of the major and minor by-products was
identi®ed as pyrrolidinylmethyl product (4) and amide (5),
The reaction conditions such as reagent ratios, reaction
temperature, and solvent were optimized. First, the amount
of KTB was investigated. As shown in Table 3, the molar
ratio of KTB against SMEAH was varied from 1 to 28%
molar (entries 1±4). The formation of the by-product 4 at
1% molar was 3.5%, and from 7 to 14% molar was almost
suppressed. On the other hand, when 28% molar KTB was
used, the reaction did not reach completion. Seven percent
molar was concluded to be the most suitable ratio.
1
respectively, by H NMR and mass spectroscopic analysis.
A detailed scheme of Red-ALP reduction for the ester 2 is
shown in Scheme 5.
1.1. Minimization of the pyrrolidinylmethyl product
Considering commercial scale production, the reaction
solvent was changed from THF to t-butyl methyl ether
(MTBE) (Table 2, entry 1). Under these conditions the
aldehyde/alcohol selectivity slightly improved, although
Secondly, the addition sequence was examined. When Red-
ALP was added to a mixture of 2 and KTB in MTBE, the
H
O
O
O
O
O
H
Na+
Al-
ONa
+
Al
H
+
+
NaH
Al
O
O
OH
O
H
SMEAH
Scheme 6.

2704
T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
Table 3. Optimization of reduction conditions
Entry
Pyrrolidine/SMEAH
SMEAH/2
Temp. (8C)
KOC(CH3)3/SMEAH (mol%)
Product ratioa (%)
1
2
3
4
5
1
2
3
4
1.05
1.05
1.05
1.05
1.05
1.05
1.15
1.25
1.50
1.25
1.25
1.15
1.05
1.05
1.4
1.4
1.4
1.4
1.4
1.4
1.7
2.0
2.8
2.0
2.0
1.5
1.4
1.4
20
20
20
20
20
10
20
20
20
10
0
1
7
92.5
94.4
90.0
72.5
77.0
52.3
96.3
95.0
95.4
97.7
97.2
97.5
88.5
95.5
±
±
2.7
4.2
3.5
±
0.2
±
±
17.2
±
±
±
0.4
±
±
1.2
1.6
1.8
2.3
1.3
1.3
1.3
2.9
3.1
1.3
1.3
1.4
1.7
1.9
14
28
14
14
7
7
7
7
7
1.5
21.5
2.3
0.5
0.8
0.6
±
±
±
±
0.6
0.2
3.3
3.3
14.0
6.5
1.0
1.4
1.2
1.0
1.1
0.9
3.4
0.9
5b
6c
7
8
9
10
11
12d
13e
14f
10
10
10
7
7
7
3.3
0.5
a
Product ratio was determined by GC.
KTB was added to a solution of 2 in MTBE.
Addition order was reversed.
Total amount of the reaction was reduced by 1/3 (to 3.6 volume of 2 from 11.4 volume).
Toluene was used as solvent.
THF was used as solvent.
b
c
d
e
f
production of 4 was suppressed to zero while the aldehyde/
alcohol selectivity decreased to 5.5 (entry 5). When the
order of addition was reversed, a higher ratio of 4 was
produced (entry 6). These results indicated that KTB
interacts with SMEAH to prevent the formation of the
by-product.
mechanism for Red-ALP reduction in Scheme 7. As
Mukaiyama10 described for the mechanism of aluminum
hydride-N-methylpiperazine reduction system, the reaction
between Red-ALP and the ester 2 could proceed in two
different ways to provide intermediates (9) and (10). In
route A, Red-ALP would simply act as a Lewis acid toward
the ester 2 to form the adduct 9. The hydrolysis of the
intermediate 9 would give the amide 5, and the reduction
of 9 with intra- or intermolecular hydride would give rise to
another intermediate (11), which could be hydrolyzed into
the aldehyde 1. In route B, Red-ALP would act as a reducing
agent by passing its hydride to the ester 2 to form the inter-
mediate 10. The hydrolysis of 10 would give the aldehyde 1
and the alcohol 3. The enamine (6) could be formed from 1
and pyrrolidine as already mentioned. This scheme suggests
three candidates which could yield 4: the amide 5, the
enamine 6, and the intermediate 11. To identify the real
intermediate, the candidates were reduced with Red-ALP.
The reduction of the amide 5 with Red-ALP gave the amine
4 and 1, however, since only trace amounts of the amide 5
were detected in the Red-ALP reaction, we concluded that
the amide 5 was less likely to be the intermediate. The
reduction of the enamine 6 with Red-ALP did not give 4,
thus eliminating this as a possible route. Although it was
dif®cult to obtain direct evidence, the overreduction of 11
appeared to be the most likely mechanism for the formation
of 4.
Next, we investigated the ratio between pyrrolidine and
SMEAH (entries 2, 7±9). As the ratio of pyrrolidine was
increased, alcohol 3 slightly declined, amide 5 slightly
increased, and more SMEAH was needed to complete the
reaction. From the practical viewpoint, the use of an excess
of pyrrolidine was thought to be more advantageous because
it gave a clear solution while Red-ALP at the ratio of 1.05
gave a cloudy suspension. We thus concluded that 1.15±
1.25 was the most adequate ratio.
To optimize the reaction temperature, the reaction was
carried out at 20, 10, 08C (entries 8, 10 and 11). At 208C,
slightly more alcohol (3) was produced than under the other
conditions. The most adequate temperature was determined
to be 108C. Further efforts to minimize the amount of reac-
tion solvent led us to the optimal reaction conditions (entry
12).
Concerning the reaction solvent, toluene did not deter the
formation of 4 and seemed to increase the formation of
alcohol 3, as KTB was insoluble (entry 13). THF gave
results similar to those of MTBE, however, a larger amount
of solvent like MTBE was necessary to extract the products
(entry 14). We could not ®nd any merits for THF, and
concluded that MTBE was the best solvent for the reaction.
In order to con®rm this hypothesis, the aldehyde 1 was
reacted with Red-ALP. Scheme 8 illustrates a possible
scheme for the reaction. The reaction between 1 and Red-
ALP would probably give two intermediates, 11 and 12.
From the intermediate 11, the aldehyde 1 and by-product
4 could be obtained. On the other hand, intermediate 12
would give the alcohol 3. GC analysis of the reaction
mixture showed a 48:26:25 mixture of 1, 3, 4 as expected
(Table 4, entry 1). When additional 0.5 equiv. of Red-ALP
was added to the reaction mixture of entry 1, the amount of
alcohol 3 did not change and the amount of 4 increased,
Overall, we optimized the reaction conditions and manufac-
tured the key intermediate aldehyde 1 up to a 250 kg scale.
1.3. Effect of alkaline bases on Red-ALP reduction
To elucidate the role of KTB, we draw the most plausible

T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
2705
COOEt
N
Ph
2
N
O
O
O
O
Na+
Al-
Red-ALP
H
N
H
N
H
N
N
(RO)2Al-O
OEt
Ph
(RO)2Al-O
OEt
Ph
Route A
Route B
Na+
Na+
R: CH3OCH2CH2
10
9
1) H-
H-
H2O
H2O
2) H2O
O
H
OH
Ph
H
N
O
N
(RO)2Al-O
H
H2O
Na+
N
N
N
N
Ph
Ph
Ph
1
3
5
11
H
N
H-
N
N
N
N
Ph
Ph
4
6
Scheme 7.
whereas 1 decreased (entry 2). This indicated that the extra
amount of Red-ALP reduced the intermediate 11
exclusively to give 4. Thus, the overreduction of 11 by
Red-ALP is the more plausible route for the formation of
the pyrrolidinylmethyl by-product 4.
However, the above mechanism is not enough to explain the
effect of alkaline bases. Considering the components of
SMEAH, we postulated that two kinds of active species
might be present in SMEAH. One is aluminum trialkoxide,
which could be formed if sodium is de®cient in SMEAH and

2706
T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
O
H
N
Ph
1
N
O
O
O
O
Na+
Al-
Red-ALP
H
Route A
Route B
N
(RO)2AlO
Na+
H
N
H
(RO)2Al-O
Na+
N
N
H
H
Ph
Ph
11
12
H-
H2O
H2O
O
H
OH
Ph
N
N
N
N
Ph
Ph
1
4
3
Scheme 8.
would work as a Lewis acid in the reaction. The other is
trivalent aluminum hydride as already shown in Scheme 6
(Table 5).
Table 4. Reduction of 2 with Red-ALP
Entry
Red-ALP (equiv.)
Product ratioa (%)
1
3
4
To verify our thesis, we investigated the effect of several Al
compounds which were analogous species present in
SMEAH. When aluminum tri(isopropoxide), a Lewis acid-
type compound, was added to the Red-ALP reduction,
although the reaction did not reach completion, 47.8% of
by-product 4 was produced (entry 2). The addition of both
1
2
1.0
Entry 110.5
48
29
26
24
25
47
a
Product ratio was determined by GC.
Table 5. Effects of additional aluminum compounds on the Red-ALP reduction of 2
Entry
Time (h)
Additive (additive/SMEAH)
Product ratioa (%)
1
2
3
4
5
1
2
3
4
5
6
3
2
±
Al(O-iPr)3 (0.14)
KOC(CH3)3 (0.07)
Entry 31Al(O-iPr)3 (0.21)
Entry 31Red-ALP (0.7)
Entry 31DIBAL-H (0.07)
65.6
37.7
96.5
96.5
58.2
67.9
1.8
10.8
±
±
±
1.6
2.5
2.2
2.3
6.7
6.7
30.3
47.8
±
±
33.8
24.6
0.3
1.9
1.3
1.3
1.4
1.7
1
3
1
±
a
Product ratio was determined by GC.

T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
2707
aluminum tri(isopropoxide) and KTB (entry 4) caused
complete suppression of by-product, similar to the result
obtained without aluminum tri(isopropoxide) (entry 3).
Thus, a Lewis acid type compound affected the Red-ALP
reduction, but did not promote the formation of 4. When
further Red-ALP (1.0 equiv.) was added to the reaction of
entry 3, the amount of 4 rose from 0 to 33.8%, indicating the
presence of an active species (entry 5). The addition of
0.1 equiv. of DIBAL-H, a trivalent aluminum hydride, to
the reaction mixture of entry 3 increased the content of 4
from 0 to 24.6% within 1 h, which revealed that the active
species was a trivalent aluminum hydride (entry 6).
KTB or other alkaline bases as deterrents of overreduction
could be attributed to the transformation of the trivalent
aluminum hydride to quadrivalent aluminum hydride,
which might be too weak to reduce the intermediate 11.
Lastly, we explored the scope of this new reduction system.
Table 6 shows a summary of the results obtained. In general,
benzoic acid esters could be transformed to their corre-
sponding aldehydes in excellent yields, regardless of the
electron withdrawing or donating substituents. The Red-
ALP KTB system gave a similar yield to that of the
SMEAH-N-methylpiperazine system9 (entry 2), indicating
that benzoic acid esters can be reduced to aldehydes with
either reducing agent effectively. For aliphatic carboxylic
acid esters, the Red-ALP KTB system also proved to be
effective. The low yield of 2-ethylbutylaldehyde was
attributed to its high solubility in water which made it
dif®cult to extract ef®ciently (entry 11). On the contrary,
the Red-ALP KTB system was not effective for cinnnamic
acid esters due to the formation of a complex mixture
(entry 9).
The putative mechanism of formation of the by-product is
outlined in Scheme 9. The chain reaction could start with a
small amount of trivalent aluminum hydride compound
present in the reagent that is active enough to reduce the
intermediate 11 to give 4. After the reaction, the reagent
could lose its hydride but maintain its trivalent form. The
trivalent aluminum hydride could be regenerated through
equilibrium to reduce another molecule 11. The effect of
H
N
(RO)2Al-O
H
Na+
O
N
O
Al
H-
N
H
Ph
11
N
H-
N
O
Al
O
O
O
Na+
Al-
OR
O
H
N
R: CH3OCH2CH2
Ph
4
N
RO
11
H-
Al- OR
Al
O
Na+
H
OR
H-
N
OR
O
O
Al-
Al
Al- OR
O
O
O
Na+
4
Na+
OR
O
H
H-
OR
O
N
O
Na+
Al-
Al- OR
RO
Al
OR
OR
Na+
H
the same reactions contin e
Scheme 9.

2708
T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
Table 6. Reduction of other substrates with Red-ALP in the presence of KTB
Entry
Product
yield (%)
93b
Substrate
CHO
1
COOMe
97b
a
2
97b
COOMe
COOMe
CHO
CHO
3
4
100b
MeO
MeO
Cl
Cl
98b
85b
39b
5
6
COOMe
COOMe
CHO
CHO
NC
NC
CHO
COOMe
7
8
97c
82c
70c
COOMe
CHO
d
9
COOMe
COOMe
CHO
10
11
58c
CHO
a
d
N-Methylpiperazine was used as the base. bDetermined by HPLC. cDetermined by GC. In the absence of KTB. Reaction conditions and procedures are
described in Section 2.
In summary, we established a viable synthetic process for
the manufacture of the key intermediate aldehyde 1 for
donepezil hydrochloride using a new selective reducing
agent, SMEAH modi®ed with pyrrolidine. Furthermore,
we proposed the mechanism of overreduction to give the
pyrrolidinylmethyl by-product 4, and revealed the
versatility of the new reducing system to other substrates.
We believe that our ®ndings could provide a new useful
method for the transformation of ester to aldehyde and
promote further understanding of the complex reducing
agent, SMEAH.
column length: 30 m, ID: 0.25 mm, initial temperature:
1508C, ®nal temperature.: 2508C, temperature rate: 58C/
min, initial time: 10 min, ®nal time: 5 min, injection
temperature: 2508C, detector temperature: 3008C, injection
volume: 2 mL, total ¯ow: ca 60 mL/min, split rate: 1/80,
HEWLETT PACKARD 5890 SERIES II). All esters,
aldehydes, and alcohols used as substrates, products and
authentic samples in Table 6 were obtained from Tokyo
Chemical Industry.
2.1.1. N-Benzyl-4-ethoxycarbonylpiperidine (2). In a
200 L reactor was placed INPE (165 kg, 1051 mol) and
NaHCO3 (89 kg, 1059 mol) in 50% EtOH/H2O (284 L).
To the mixture was added dropwise benzyl chloride
(133 kg, 1051 mol) over 1 h at 408C. After stirring for 2 h
at 808C, n-hexane (295 L) and H2O (295 L) were added, and
the organic layer was separated, washed with H2O (295 L)
and concentrated in vacuo to give 241.6 kg of 2 in 93%
yield. This product was used for the next step without
2. Experimental
2.1. General
EI-MS and FAB-MS were obtained with a JEOL JMS
1
HX100. H and 13C NMR spectra were measured on a
JEOL JNM-a600 using tetramethylsilane as an internal
standard. GC conditions were as follows (column: DB-1,
1
puri®cation. Yellow oil; H NMR (400 MHz, CDCl3) d
1.24 (t, Jˆ7 Hz, 3H), 1.70±1.90 (m, 4H), 1.95±2.05 (m,

T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
2709
2H), 2.12±2.32 (m, 1H), 2.80±2.90 (m, 2H), 3.48 (s, 2H),
4.12 (q, Jˆ7 Hz, 2H), 7.20±7.40 (m, 5H). 13C NMR
(100 MHz, CDCl3) d 14.19, 28.28, 41.23, 52.91, 60.22,
63.24, 126.94, 128.16, 129.05, 138.41, 175.25. Anal.
Calcd for C15H21NO2: C, 72.88; H, 8.43; N 5.65. Found:
C, 72.84; H, 8.56; N, 5.66.
(100 MHz, CDCl3) d 23.38, 31.11, 35.35, 53.75, 54.61,
63.13, 63.54, 126.82, 128.07, 129.23, 138.52. Mass 258
(M1H)1.
2.2.4. N-Benzyl-4-(pyrrolidin-1-ylcarbonyl)piperidine
1
(5).14 White wax; H NMR (400 MHz, CDCl3) d 1.70±
2.10 (m, 8H), 2.25±2.40 (m, 1H), 2.85±3.05 (m, 2H),
3.35±3.60 (m, 6H), 3.45 (s, 2H), 7.20±7.45 (m, 5H).
2.2. Screening of secondary amines
2.2.5. N-Benzyl-4-(pyrrolidin-1-ylmethylene)piperidine
(6). Colorless oil as a mixture of 1 and 6; 1H NMR
(400 MHz, CDCl3) d 1.60±1.80 (m, 8H), 2.35±2.45 (m,
4H), 2.70±2.95 (m, 2H), 2.95±3.02 (m, 2H), 3.43±3.44
(m, 2H), 5.67 (s, 1H), 7.20±7.40 (m, 5H).
Typical procedure. A solution of N-methylpiperazine
(2.00 g, 20 mmol) in THF (5 mL) was added dropwise to
a 66% toluene solution of SMEAH (20 mmol) in THF
(20 mL) around 2208C over 5 min, and the mixture was
stirred at room temperature for 1 h. The prepared reducing
agent was added dropwise to a solution of 2 (2.48 g,
10 mmol) in THF (10 mL) with ice-cooling over 1 h.
After 2 h, the reaction mixture was quenched with 2N
NaOH, and extracted with MTBE (100 mL). The organic
layer was separated, extracted with 2N HCl solution
(20 mL), and the pH of the mixture was adjusted to ca 8.5
with K2CO3 solution. The organic layer was subjected to GC
analysis.
2.2.6. 1-Benzyl-4-formylpiperidine (1). Reduction of (2) in
the presence of KTB. In a 500 L reactor was placed a 66%
toluene solution of SMEAH (221 kg, 723 mol) in MTBE
(242 L) under a nitrogen atmosphere. To the solution was
added dropwise a solution of pyrrolidine (60 kg, 838 mol) in
MTBE (64 L) at 2208C to 258C, and the resulting mixture
was stirred overnight at around 208C. Then KTB (5.8 kg,
52.1 mol) in THF (14.8 L) was added in one portion and
stirred for 1 h, to give a clear solution of Red-ALP-KTB.
2.2.1. N-Benzyl-4-formylpiperidine (1). Reduction of (2)
without KTB. In a 200 mL reactor was placed a 66% toluene
solution of SMEAH (31.0 g, 101.9 mmol) in MTBE
(109 mL) under a nitrogen atmosphere. To the solution
was added dropwise a solution of pyrrolidine (7.6 g,
107.0 mmol) in MTBE (25 mL) at 220 and 258C, and
stirred overnight at 208C to prepare the Red-ALP solution.
In a 1030 L reactor was placed 2 (120 kg, 485 mol) in
MTBE (130 L). To the solution was added dropwise the
Red-ALP-KTB agent below 158C, followed by stirring for
1 h. The mixture was quenched with 4N NaOH solution
(572 L) at 158C. The organic layer was separated, washed
with 4N NaOH solution (146 L) and H2O (580 L). To the
organic layer was added a solution of conc. HCl (118 L) in
H2O (448 L) below ca. 158C and was stirred for 0.5 h. Then
the pH of the mixture was adjusted with 25% NaOH solu-
tion (117 L) to 8.5±8.7. The organic layer was separated,
washed with H2O (428 L) and concentrated in vacuo to give
crude 94.5 kg of 1 in 95% yield. This product was puri®ed
by reduced pressure distillation (bp 130±1348C/1 mmHg) to
In a 300 mL four necked-¯ask was placed 2 (18.0 g,
72.8 mmol) in MTBE (72 mL). To this solution was added
dropwise the prepared Red-ALP solution at 208C, followed
by stirring for 2 h. The reaction mixture was quenched with
4N NaOH solution (86 mL) at 158C. The aqueous layer was
separated, washed with 4N NaOH solution (11 mL) and
H2O (86 mL). To the organic layer was added a solution
of conc. HCl (18 mL) in H2O (68 mL) pre-cooled in an
ice-bath at ca. 158C and stirred for 0.5 h. Then, the pH of
the mixture was adjusted with 8N NaOH (16 mL) solution
to 8.5. The organic layer was separated, washed with H2O
(61 mL) twice and concentrated in vacuo to give crude
14.1 g of 1 in 95% yield. The mixture ratio was determined
by GC. 1/2/3/4/5ˆ65.6:1.8:1.6:30.3:0.3.
1
give 1 as a colorless oil. H NMR (400 MHz, CDCl3) d
1.62±1.74 (m, 2H), 1.83±1.92 (m, 2H), 2.06±2.14 (m,
2H), 2.18±2.26 (m, 1H), 2.77±2.85 (m, 2H), 3.49 (s, 2H),
7.20±7.35 (m, 5H), 9.64 (s, 1H). 13C NMR (100 MHz,
CDCl3) d 25.43, 47.98, 52.45, 63.21, 127.00, 128.17,
129.03, 138.24, 203.99. Anal. Calcd for C13H17NO: C,
76.66; H, 8.40; N, 6.88. Found: C, 76.81; H, 8.43; N, 6.89.
2.3. Reduction of other substrates
Retention time of each compound: 1ˆ14 min, 2ˆ20 min,
3ˆ16 min, 4ˆ25 min, 5ˆ32 min, 6ˆ26 min.
Typical procedure. A solution of pyrrolidine (1.1 g,
15.4 mmol) in MTBE (3.3 mL) was added dropwise to a
66% toluene solution of SMEAH (14.7 mmol) in MTBE
(8.9 mL) around 2208C over 20 min and stirred at room
temperature for 1 h. A solution of KTB (16.5 mg,
1.47 mmol) in THF (0.7 mL) was added to the mixture.
The above reducing agent was added dropwise to a solution
of methyl 4-methylbenzoate (1.1 g, 7.3 mmol) in MTBE
(4.0 mmol) around 108C over 1 h and stirred for 2 h. The
reaction mixture was quenched with 2N HCl (80 mL) and
the organic layer was separated. The content of 4-methyl-
benzaldehyde was determined by HPLC using authentic
material.
2.2.2. N-Benzyl-4-hydroxymethylpiperidine (3). Yellow
oil; H NMR (400 MHz, CDCl3) d 1.21±1.35 (m, 2H),
1
1.42±1.56 (m, 1H), 1.69 (d, Jˆ12 Hz, 2H), 1.90±2.03 (m,
2H), 2.90 (d, Jˆ12 Hz, 2H), 3.46 (d, Jˆ7 Hz, 2H), 3.49 (s,
2H), 7.20±7.36 (m, 5H). 13C NMR (100 MHz, CDCl3) d
28.76, 38.53, 53.39, 63.44, 67.79, 126.89, 128.09, 129.21,
138.34. Anal. Calcd for C13H19NO: C, 75.81; H, 9.31; N,
6.83. Found: C, 76.06; H, 9.33; N, 6.82. Mass 206 (M1H)1.
2.2.3. N-Benzyl-4-(pyrrolidin-1-ylmethyl)piperidine (4).
Yellow oil; H NMR (400 MHz, CDCl3) d 1.20±1.35 (m,
1
4H), 1.40±1.55 (m, 1H), 1.70±1.85 (m, 4H), 1.90±2.03 (m,
2H), 2.29 (d, Jˆ7.3 Hz, 2H), 2.42±2.48 (m, 4H), 2.88 (d,
Jˆ11.7 Hz, 2H), 3.49 (s, 2H), 7.20±7.36 (m, 5H). 13C NMR
For entries 1±7, the reactions were carried out in a similar

2710
T. Abe et al. / Tetrahedron 57 (2001) 2701±2710
5. Sugimoto, H.; Tsuchiya, Y.; Higurashi, K.; Karibe, N.; Iimura,
Y.; Sasaki, A.; Yamanishi, Y.; Ogura, H.; Araki, S. EP 296560
A2; Chem Abstr. 1989, 110, 173102.
manner to that described above and the yields were calcu-
lated by HPLC using authentic samples.
6. Hermans, B.; Van D, P. Chem Abstr. 1969, 70, 57585.
7. Miwa, K.; Aoyama, T.; Shioiri, T. Synlett 1994, 2, 109.
8. Chen, Y. L. EP 441517 A2; Chem Abstr. 1991, 115, 232222.
9. Kanazawa, R.; Tokoroyama, T. Synthesis 1976, 526±527.
10. Muraki, M.; Mukaiyama, T. Chem. Lett. 1975, 875±878.
11. Muraki, M.; Mukaiyama, T. Chem. Lett. 1974, 1447±1450.
12. Muraki, M.; Mukaiyama, T. Chem. Lett. 1975, 215±218.
13. Cerny, Z. J. Organomet. Chem. 1996, 516, 115±122.
14. The authentic sample was prepared by the following route.
For entries 8±11, the yields were calculated by GC using
authentic samples.
HPLC conditions; L-column (4.6 mm£250 mm), Detection
254 nm, ¯ow rateˆ1.0 mL/min, mobile phase; for entries
1,2,3,5,7; CH3CN/H2O/70%HClO4ˆ400:600:1, and for
entries 4,6; CH3OH/H2Oˆ750:250.
GC conditions; column: DB-1, column temperature: 1008C,
injection temperature: 1608C, detector temperature: 1608C,
other conditions were the same as those described in the
general procedure.
H
N
O
Cl
O
N
BrCH2Ph
References
N
N
.
HCl
1. Sugimoto, H.; Tsuchiya, Y.; Sugumi, H.; Higurashi, K.;
Karibe, N.; Iimura, Y.; Sasaki, A.; Kawakami, Y.; Nakamura,
T.; Araki, S.; Yamanishi, Y.; Yamatsu, K. J. Med. Chem.
1990, 33, 1880±1887.
O
N
O
N
H2 / Pd-C
2. Sugimoto, H.; Tsuchiya, Y.; Sugumi, H.; Higurashi, K.;
Karibe, N.; Iimura, Y.; Sasaki, A.; Araki, S.; Yamanishi, Y.;
Yamatsu, K. J. Med. Chem. 1992, 35, 4542±4548.
3. Sugimoto, H.; Iimura, Y.; Yamanishi, Y.; Yamatsu, K. J. Med.
Chem. 1995, 38, 4821±4829.
N
N
Br+
Ph
Ph
5
4. As of 31 July, 2000.